I show up to work early. I set the coffee maker going, but I don’t have time to wait for it. Instead, I make a beeline for the mouse facility.

It’s already gone 8 a.m. and I’m late to feeding my mice. We’re running a time-restricted feeding experiment whereby the experimental group only has access to food between 8 a.m. and 5 p.m.—shown to modulate liver metabolism in a manner that might affect the development of hepatocellular carcinoma (which is what we’re testing). Today it’s my turn to do the cage changes. I don my PPE and get to work: I take the cages out of their light-tight cabinets, put them on the sanitised bench-top next to their food-containing counterparts, then open up each pair of cages and transfer the mice from their fasting cage to their feeding cage. Cages back into the cabinets, wipe down the surface, then back to the main lab.

I head straight to the cell culture room, where I aliquot some medium into falcon tubes and put them in the bead bath to warm up to 37°C.

I briefly return to my desk. Just enough time to clear emails, check my calendar, and savour my coffee (I’m halfway through a bag of lightly roasted washed Ethiopian beans).

I head back to the cell culture room and spend the rest of the morning setting up an assay. Essentially, I’m wearing PPE, moving liquids around (sometimes with cells in them) using tools like the P1000 pipette, and operating specialised equipment like the 5810 R bench-top centrifuge. It’s all manual, physical stuff.

Back to my desk for lunch during an incubation period. I scoff my greek yoghurt and granola, deal with emails again, and read up to figure 2 of a paper that had piqued my interest. I’ll finish it later perhaps.

The afternoon is a bit slower: I’m staining cells with fluorescent antibodies so I can quantify specific proteins in the cells via flow cytometry. I spend the incubation periods analysing data, working on figures for a paper I’m writing, and stopping myself from doomscrolling on X (guilty!).

I make it to the cytometer with my samples by 4 p.m. I sit there for an hour or so setting up the instrument, calibrating it, tweaking the voltages to put my fluorescent cells nicely in the middle of the dynamic range of the detectors, and then acquiring and recording the data. I export the .fcs files then clean and shut down the instrument.

5 p.m. strikes and I’m off to the mouse room again. Same as before. Yoinking the fellas into their fasting cages for the evening now. I clean up. Disposable PPE goes in the bin. Reusable PPE into the autoclave basket. I bring my laptop home so I can finish data analysis after dinner.

§

This is a fairly typical day-in-the-life. I spent maybe an hour and a half at a desk and clocked 14,000 steps while charging around the institute. On any given day, the tasks I do typically fit into one of the following categories:

• animal handling (including surgery/

  procedures)

• moving precise volumes of various liquids into various containers

• operating machines—sophisticated and simple alike

• data analysis

• writing

• reading

Interestingly, only the latter two of these tasks did I have any training for during my far-too-many university degrees. The majority of what I actually do as a life scientist was learned on the job. And a lot of it is physical labour, albeit more taxing of one’s dextrousness than one’s strength or endurance.

Perhaps this is why it has been argued that scientists are, in fact, blue-collar workers merely coded as a white-collar workers. It’s a provocative yet persuasive claim. We wrangle animals like farmers do. Like factory workers, we operate sophisticated machinery and face replacement by robots1. As electricians and plumbers do, we learn vocationally. I suppose we also move liquids around akin to a car mechanic—although our liquids are typically more aqueous.

And yet, despite this, our hiring process is indistinguishable from that of a management consultancy firm. To my mind this is an obvious mismatch and leads to poor hiring decisions which, especially in academia, can be de facto irreversible. The net result of this misaligned hiring process is that we lose great scientists and retain incompetent ones who are good at Zoom interviews but bad at physical science.

§

I recently sat on an interview panel for our lab. We were trying to hire a new research assistant. HR did the primary screening process, and we then were given the resumés of a shortlist of candidates. We interviewed each one via Zoom. Predictable, stale questions:

“Are you comfortable working independently?”

“How much experience do you have handling mice?”

“What steps do you take to mitigate cell culture contamination?”

It was a frustrating process because it was just impossible to get what I wanted from it. What I really wanted to know is how competent they are with a pipette in their hands. How good they are at juggling two protocols simultaneously. How confidently they can perform intra-peritoneal injections. Under our current hiring paradigm I can’t know these things until they’ve already been hired.

The frustration here highlights the need to reform the hiring process for wet-lab science, which, in its current form, is a huge bottleneck to scientific progress and an impasse for meritocracy. I propose trialling a hiring process that emphasises vocational competency. The reformed process would look something like this:

1. Initial screening of candidates is done by HR department (same as before)

2. A shortlist of candidates is invited to Zoom interview. Now, rather than being asked asinine HR-ified questions (“tell me about a time you overcame something difficult”), the candidates are given a single task to do live, while sharing their screen: design an experiment from start to finish, including creating a list of reagents to order. Candidates are asked to think out loud. Tasks here can include plasmid design, flow cytometry panel design, reagent comparison and selection (including cost-benefit consideration), etc.

   The scientific question must be relevant to the lab they wish to join (an experiment that the lab actively wishes to carry out), and should be the same for all candidates for fairness. The experiment should also be doable in one or two days—not a week/months long protracted experiment. The candidates are tested on their ability to figure out how to answer the scientific question.

   As an example, a T cell biologist (my field) can be asked “how would you test whether interferon-γ is secreted at the immune synapse?”. A good candidate with microscopy expertise might find validated IF/ICC antibodies, describe T cell culture, conjugation to target cells or artificial synapses, fixation parameters, staining, microscopy, and microscopy analysis details. (By the way, the answer to this question was always assumed to be “no” until very recently!)

3. The most promising interview candidates are then invited for in-person technical assessment. At this point, their reagent lists have been ordered to the lab and any pre-requisite animal/cell setup has already been performed by lab members. The candidates have also already completed health and safety onboarding (which removes the need for them to complete this later). They are then launched into experimental work—shadowed by a senior lab member (postdoc/staff scientist). They follow their own protocols and experimental plan. They are assessed on their ability to carry out the work faithfully and quickly. The one-day in-person visit concludes with a face-to-face discussion about the experiment and the results (if the experiment can be completed in a single day) with the PI and senior lab members.

4. The candidate who demonstrated the best experimental planning and execution skills is hired for a job that is primarily concerned with experimental planning and execution.

Yes, this requires a bit more investment (of time and resources) up front, but the benefits of picking the best candidate would surely pay back this investment many times over.

It would be fairly straightforward for a research institute or academic department to run a randomised trial to test whether such reforms would enhance the science produced. Labs can be the randomised into reformed vocational hiring or the status quo. Then, over a period of, say, one year, labs hire new staff as needed. The typical performance reviews that new hires undertake can then be used to test which hiring process produces better results. Outcomes can be subjective and objective metrics relating to team cohesion and scientific outcomes (e.g., surveys of lab members, contributions to papers, etc.).

§

When scientists lack technical competence, they procrastinate experiments, they opt for suboptimal but technically easier experiments, and they perform experiments badly. Badly done experiments can generate falsely negative data if technical noise drowns out biological signal, or falsely positive data if technical error produces a bias in one direction. Both of these situations contribute to unreliable, irreproducible data—an enormously concerning problem for science. Even if badly done experiments are discarded and repeated until done well, money and time is wasted.

The degree to which poor physical workmanship limits the sciences is, by my reckoning, massive and under-discussed. Further, it is also a largely invisible problem to PIs, who don’t really spend time in the lab micromanaging the intricacies of the experimental work.

Given that up to 80% of NIH grant money goes towards salaries and benefits for personnel, hiring mistakes are very costly. Further, the human consequences of poor hiring processes are that someone not cut out for the job suffers the pain of being unskilled at something they shouldn’t have been hired for, while talented, skilled scientists were not hired and potentially lost to science forever if they pivot to a white-collar job instead.

In my view, reforming scientific hiring practices is an obvious, tractable, and highly impactful bottleneck to solve, with potential to increase the speed and quality of scientific research while improving job satisfaction and reducing costs. Science is a technical job. We need to hire based on scientific and technical ability.

A common narrative you’ll find in politico-cultural discourse these days is that the internet and social media are contributing to political polarisation and extremism by trapping people in echo chambers. Social media algorithms highlight the most extreme voices on either side of the political aisle and only show people what they want to see, thus driving people further and further to the extremes. Technology created echo chambers; echo chambers created polarisation and extremism, so the theory goes. According to this narrative, algorithms and echo chambers go hand in hand—and both are evil.

To combat the evil of echo chambers, several new media companies have sought to address this problem head-on by forcibly bringing together different sides of politico-cultural debates. Podcasts like pride themselves on their centrism1, level-headedness, and willingness to engage both sides (they recently hosted Ted Cruz and then Mehdi Hasan to discuss the Iran war from opposing standpoints, for example). Jubilee Media goes one further by frequently hosting combative debates—most aggressively in the Surrounded format, whereby a single high-calibre debater is surrounded by 20 amateur disagreeing voices who take turns trying to take down their shared ideological enemy.2 Piers Morgan, the world’s premier example of boomer liberalism, runs a show called Piers Morgan Uncensored where he invites on proponents of opposing views, lets them shout at each other, and then proceeds to censor them by constantly interrupting them. Even the hallowed debating halls of the Oxford and Cambridge Union Societies, with their debates now filmed and uploaded to YouTube, can claim to contribute to the dismantlement of echo chambers. Perhaps the strongest claim to anti-echo chamberism can be made by Ground News, which aggregates news stories from outlets with transparent partisanship labels and puts them all in the same place so that people can identify and overcome media bias.

Sounds great. So perhaps we can be expecting our increasing political polarisation to reverse some time soon, given all of these corrective efforts, right? I’m doubtful.

Despite all of these anti-echo chamber efforts, I really don’t feel like polarisation is getting much better. In fact, it seems to be getting worse. Contrary to their stated intention, I think these echo-chamber-defying platforms are directly contributing to the increased polarisation, and it’s not because they are failing to break people out of their echo chambers—they undoubtedly succeed in that goal. Instead, I conjecture that echo chambers are, in fact, a healthier psychological environment for humans to exist in and probably act to temper political polarisation. But, as I will go on to argue, perhaps this isn’t the correct North Star to aim for.

Echo chambers require gatekeepers

Prior to social media and the internet, the diversity of news sources and opinions available to the public was much more limited, and it was more geographically constrained. This meant that people living in the same area were fed broadly similar information and were thus all in the same echo chamber3. While this made them susceptible to centrally controlled narratives, it also ensured a certain homogeneity of public opinion.

It wasn’t just news, though. All cultural products were disseminated by gatekeepers, rather than being algorithmically optimised for attention. People broadly knew and liked the same music, the same movies, the same ideas.

As Asa Park put it in his brilliant video essay, Why we were told to hate Nickelback:

This meant that on Monday mornings at work or at school, millions of people can talk about the same TV shows they’ve all watched Sunday night; the same songs they’ve all heard on their morning commute. You could say that pop culture was a shared language. Radio DJs decided what songs got played. Blockbuster decided which videos got stocked on their shelves. MTV decided which music videos got into rotation. Essentially, this was gatekeeping. And it sounds restrictive, but behind that gate was our culture. Our shared reference points. Our common ground.

In other words, we lived firmly in echo chambers.

At the level of the individual, this was a wonderful existence. The people around us were predictable and familiar. Small talk with a random stranger would be much more likely to reveal common interests. Encounters with other humans didn’t require a “fight or flight” cognitive response as default.

At the level of the society, though, the information ecosystem was constrained. Overton windows were narrow. Dogmata were firmly embedded. For example, the public opinion of the US invasion of Iraq in 2003 was largely positive. In hindsight, with more information, views on the Iraq invasion are now much less favourable. The key difference here is “with more information”. The point is that, in 2003, information was gate-kept, Americans lived in an echo chamber, and they held opinions that were more or less given to them on a plate, rather than them being empowered to decide for themselves based on all of the available information. There are many such cases of society being led away from the truth under a unified narrative due to gate-kept information (I’m sure the reader can think of a few examples even from the last five years alone).

And yet these unified narratives were beneficial at the psychological level. As the philosopher Joseph James Rogan Jr. recently recounted:

You know, I remember I've talked about this before, but post-9/11, everyone was so connected. Everyone was smiling. People were letting you get on the highway. They're letting you get in their lane. They're waving. Everyone had an American flag on their car. We've been attacked. We were united, you know. And it's just sad that it takes something like that for people to realize, like, this is a gift to be alive in this incredible country at this incredible time in history.

A silver lining to a tragedy, the shared sense of civic compassion Rogan speaks of was only possible due to the echo chambers that Americans lived in at the time. Had podcasters and social media been around at the time, the national narrative of the events of 9/11 would have been fractured and distorted. Otherwise fringe views would have seeped their way into the political conversation and Americans would have been much less likely to have been able to relate to each other on common ground.

Thus, at the level of the individual and from a psychological perspective, echo chambers were pleasant and beneficial, but at the level of a society and from the perspective of seeking the truth, echo chambers were destructive. Ignorance was bliss, and people were blissful. That’s no longer the case now, though.

Breaking down the gates: the algorithm

Now that we live in an era of free flowing information (in large part thanks to large independent social media platforms like X, TikTok, and Instagram/Facebook), information is no longer curated by gatekeepers, but distributed by algorithms.4

Our sources of information have diversified as algorithms have replaced gatekeepers, which is great for disabling censorship and allowing for free speech to take its course. In purely economic terms, decentralising the information market like this will always lead to a more efficient pursuit of the truth. Indeed, worries about algorithms exacerbating misinformation seem to be largely overblown and not supported by any objective data. The reality is that more people know more stuff in 2026 than they did in 2003. More often than not, misinformation alarmism merely reflects the reactive desperation of the gatekeepers trying to cling to their historic power.

While gatekeepers could engineer public opinion on war in the middle east twenty years ago, they are powerless to do so today. The diversified information network managed by algorithms allows for the public to get much more information—and the most interesting and important information inevitably rises to the top. Perhaps that is partly why public sentiment towards the recent and ongoing Iran war is so much less favourable than it was at an equivalent stage during the Iraq war. An algorithmically informed populace is a better informed, more skeptical populace.

Okay so ignoring any potentially negative psychological impacts on the individual, algorithms are good? A welcome change from the gate-kept echo chambers of the past? With an ideal algorithm—yes.

The only slight issue is that the invisible hand which would normally guide the attention economy is not entirely invisible. It is a deliberately constructed algorithm. Tweaking what the algorithm rewards and what it punishes will boost and suppress different kinds of social media posts to different degrees. Maybe information isn’t gate-kept now, but it’s still very much manipulated, squeezed, and pruned. The marketplace of ideas isn’t an entirely free market. The algorithms are imperfect.

Imperfect algorithms

These algorithms typically optimise for time spent on-app. Attention, that is. At a mechanistic level, this means that many of the social media algorithms reward replies to a post much more heavily than likes on a post (think how much more attention typing out a reply requires compared to simply clicking “like”—and how much more attention is garnered from other users by a post with a rich comment section). For X’s algorithm, which is largely open source, if the original poster replies to those replies, then that boost is enhanced even further. This mechanic is designed to foster deeper conversations and commentary.

Now, sometimes this feature of the algorithm works as intended. A post will be genuinely interesting and it sparks a conversation in the replies between the poster and various commenters. Perfect. But I’m sure you can easily see how, especially when posters are cognisant of the intricacies of the algorithm, this system might get gamed.

At the most innocent level, posters who want their post to do well can try to engineer replies/comments by prompting readers with questions (YouTubers absolutely love to end a video with a “do you agree?”, “let me know what you think in the comments down below”, etc). This is fairly organic and probably not that distorting of the information ecosystem.

More Machiavellian posters can’t merely rely on organic comments to boost their posts though. Instead, they’ll rely on an army of algo-cognisant commenters (either human or bot accounts) to spam comments once the post is up—thus boosting the post very quickly and strongly. Notably, there doesn’t need to be active coordination between the poster and the boosters—they can be entirely separate parties with separate ultimate goals but shared proximal goals. Intense boosting of Christian nationalist Nick Fuentes’ posts by accounts largely originating in India, Pakistan, Nigeria, Malaysia, and Indonesia is a case in point.

Perhaps the most pernicious strategy of all algorithm-gaming strategies is a technique known as rage baiting. Due to the conditions our minds evolved in, human psychology has a negativity bias which motivates us to take action (i.e., reply to a post) more greatly when motivated by negative emotions (frustration, hate, anger), than by positive ones (curiosity, kindness, creativity). Rage baiting involves posting deliberately false or frustratingly stupid content designed to grab people by the amygdala and coerce them to correct or berate the rage baiter, thus boosting the visibility of the post.5

Echo chambers create consensus; algorithms create polarisation

Algorithms shattered echo chambers and in so doing have decentralised and increased the quantity of information available. I consider this to be a good thing for getting to the truth—certainly preferable to being fed an incorrect consensus. But the stability of a democracy perhaps doesn’t depend on truth. Perhaps we can blame algorithms (or their creators, at least) for something. As John Burn-Murdoch put it:

For all the attention given to misinformation, a narrow focus on objective falsehoods distracts us from a much more fundamental shift. The emerging democratic risk is not so much that people believe false things — they always have. It is that they no longer believe the same things as one another, false or otherwise.

This is is the political version of what Asa Park was getting at.

But if people believe different things in the age of the algorithm, is that not just another way of saying that algorithms reinforce echo chambers? Are Triggernometry, Ground News, Piers Morgan et al. actually correct?

No. The polarised ecosystem of social media is well-documented, but people tend to misinterpret what’s actually going on.

When people see a graph like the one above, they imagine two groups of people: one shown only far left views and one shown only far right views. But the data points here are views/narratives in posts, not people. Social media users are much more likely to encounter views from both ends of the spectrum. Indeed, as someone on the right myself6, I encounter all manner of far left content on my feeds. Showing me that stuff is more likely to get a rise out of me. Perhaps I’ll comment or share it with a friend so we can ridicule it together. A right wing echo chamber simply wouldn’t grab my attention in the same way—I’d just like, scroll past, and then log off.

In reality, people believe different things not because they are exposed to too little information from the other side, but because they are shown too much information from the other side. Algorithms create polarisation by showing you stuff outside of your would-be echo chamber.

Contra Ground News

A few years ago I believed in the assertion that our psycho-political woes were, at their root, echo chamber-induced anxiety. I came across this new media organisation that promised to find the middle ground on news stories. They present headlines on the same issue from diverse sources to show how media bias shapes opinion. They appeal to well-meaning people from across the political spectrum. They even have a “blindspot” feature, which shows you stories that had been missed by sources that lean in favour of your own bias (we’ve all heard the podcast ad reads …). What a brilliant initiative, I thought. I decided to follow Ground News on Instagram.

Their actual product seems to be really good. The problem is combining their mission with social media algorithms: ~90% of the time I see their posts, I come away enraged. The news item will typically already be known to me (aggregating and then disseminating news is obviously slower than getting it straight from the horse’s mouth). What enrages me is usually the comments. Opinions from ignorant idiots who I don’t know and don’t care about. They’re so ill informed! They draw all the wrong conclusions! What total morons!

I’ll spend more time scrolling through the comments, getting angrier as I do. Sometimes I’ll reply with a long response to a nameless faceless comment that I disagree with. Sometimes it’s an extended back and forth of asinine bickering. Sometimes I’ll catch myself and delete my comment before posting it. Invariably, the net result is that I come away emotionally worse off and with a dimmer view of my fellow humans. Of course, many of the likes and comments I see are probably not human-generated at all.

Beyond personal anecdote, it seems my emotional response here is validated by studies, including a 2025 interventional randomised study that used an LLM-based browser extension to re-rank participants algorithmic feeds in real time. Users were randomised to see an increase or decrease in the prevalence of “antidemocratic attitudes and partisan animosity” (AAPA) content. Corroborating my own experience, users who saw more stuff outside their echo chambers were acutely more angry and sad.

I now realise that the problem isn’t that I’m not fighting hard enough in the comments. It’s not that these people, who hold views I find to be bad and wrong, exist. The problem is that I’m not really supposed to ever encounter such people. Human psychology is not evolved to thrive outside of an echo chamber. We’re wired to resolve political conflicts—by brain or by braun—and then form new, stable echo chambers.

It is this conflict-resolution drive that the likes of Jubilee, Piers Morgan Uncensored, Ground News, etc., capitalise on. Modern day intellectual Roman colossea are a good business model, it seems. Just don’t be fooled that these platforms ever want there to be a resolution to the conflicts. The highs and lows of cognitive tribal warfare are the point, and us consumers should at least be aware of this if we are to continue attending the circus.

And yet, such conflict is essential for the truth to emerge.

If we want to continue to have the most information we can have and understand the world around us as deeply as possible, we must exist in a state of constant exposure to people and things that spike cortisol. We cannot return to the echo chamber. We cannot give up the unrivalled access to information that we, in 2026, have, nor our freedom to decide what to believe in. Humanity’s progress towards the good and the true depends on us being cursed with knowledge—fractured, diverse, and competing knowledge.

In the centre of the garden of Eden, there was a tree which bore fruit that, if eaten, conferred moral awareness—a knowledge of what is good and true and what is bad and wrong. Man contravened its better sense and chose to consume the fruit.

The replacement of gate-kept echo chambers with algorithmically distributed information networks offers Man the same choice once again. Perhaps going down this path represents a second Fall. Perhaps doing so is the inescapable fate of humanity. Ultimately I believe all of us are free to make the choice ourselves. In the same breath that we criticise the algorithms, we must also praise them for what they have given us. Choose: embattled gnosis or controlled, ignorant bliss.

Subscribe now

• “Which diet is best for preventing cancer?”

• “Is climate change killing babies?”

• “How did neanderthal DNA get into Eurasian human DNA?”

But before scientists get to answer their question, they need funding. The first bias to make an appearance is funding bias. You see, certain kinds of questions aren’t really allowed to be asked. They’re just not kosher. Favouring certain types of inquiry over others is a fundamentally political choice, not a scientific one, with the downside being that some interesting areas of research are neglected, while other areas get artificially pushed beyond what raw curiosity would have manifested. At its worst, funding bias directly shapes the outcomes of research by presuming and pushing expected results.

Once they’ve secured a grant, the experimentalists then decide how to go about answering said question. But the method(s) we choose can already predispose a certain outcome. For example, failing to control for obvious confounders in observational data is likely to produce biased results. If we like the direction of this bias, we can do less adjustment for confounders. If we don’t like the direction of this bias, we can correct or over-correct for the confounders.

Even with interventional studies, choice of experimental parameters can introduce biases. For example, let’s say I want to study the effects of a high protein diet on muscle mass. In this hypothetical randomised control trial, will I measure lean body mass via dual-energy X-ray absorptiometry or via bioelectrical impedance? If I wanted to bias my trial in favour of the null hypothesis (i.e., that high protein has no effect on muscle mass), then I’d probably go for the less precise measurement (bioelectrical impedance) because it will create a larger spread of datapoints which would require a much higher sample size to be able to prove that a certain effect is statistically significant.1

Once the data have been collected, we then need to analyse the data statistically. More bias seeps in. Wilcoxon or paired t-test? Benjamini-Hochberg or Bonferroni? Can I assume sphericity for my ANOVA? Some of these statistical choices have no obvious right answer—but they’ll give subtly different results nonetheless, so you can pick the one you prefer the outcome of.2

Then we interpret. This means we piece together the results into a coherent narrative. If this means excluding some of the data we collected because it is inconvenient to the narrative, then so be it. We then craft this interpretation into a story that will persuade editors and reviewers at high-impact journals to publish our paper. Storytelling inevitably involves some creative license (otherwise known as bias). And then if editors or reviewers just don’t like the interpretation of the data (it conflicts with their personal biases, say), they’ll just reject it and the world carries on spinning as if the result doesn’t exist (id est, publication bias).

Once it’s gone through all of this, finally the journalists and science communicators get to have their way with it. Abstracts get skimmed and misinterpreted, nuggets of convenient information get stripped of their context and dishonestly propagated, and sometimes claims are made that aren’t anywhere to be found in the actual paper. It’s not just how the science gets reported though, but which science gets reported on in the first place (aka reporting bias). Important but inconvenient studies just don’t get reported on at all, while poor quality studies that point to a favourable conclusion get way more attention than they deserve.

Once you’re savvy to all of this and you get into the habit of reading papers in full (yes—methods and supplementary data too), then these lenses through which the data are distorted become very obvious and you start to see bias everywhere. It’s exhausting, frankly. And the problem is, once a paper gets reported on inaccurately, it can’t then be unreported on. It’s permanent damage done to the realm of public knowledge. Millions of people who typically pride themselves on being informed end up being worse than uninformed—they become enthusiastically misinformed. This means they go about the place confidently wrong because they read about a study on BBC News.

To illustrate this phenomenon, I’d like to discuss a particularly egregious recent example.

“Vegetarian diet lowers risk of cancer, study finds”

The narrative

This was the title of the BBC article reporting on a recent epidemiological study, published in the British Journal of Cancer, which collated data from prior prospective cohort studies looking at the association between peoples’ answers on dietary surveys and their risk of getting various cancers.

The BBC article goes on to discuss the findings with blatantly causal language:

“A vegetarian diet can slash the risk of five types of cancer by as much as 30%, a new study has found.”

Except there is nothing causal to be inferred here at all. Zero. These are purely observational data. Correlation does not imply causation. Yet this doesn’t stop journalists and scientists pretending otherwise—recall that infamous NEJM study which strongly suggested that higher chocolate consumption causes increased cognitive function.3

The BBC wasn’t alone though. The Independent titled their article “Vegetarian diet can slash the risk of five cancers by as much as 30% – study”. That’s a causal claim, ma’am! Sky News used the same false causal claim for their article. The Guardian was a little better, but went on to include a completely unscientific claim from one of the authors about whether their association is from vegetables being good or meat being bad.

The bias

In case it isn’t obvious, the reason you can’t claim that vegetarian diets “slash” the risk of cancer is because there could be any number of variables that weren’t controlled for (or were over-controlled for) that create this association in the data.4 In the chocolate-Nobel Prize correlation, if you adjust for GDP, the association evaporates (each nation’s wealth is a mediator in this case). For a detailed explanation of why adjusting observational data like this is precarious, I’d recommend this piece:

Cremieux Recueil

You Can't Just "Control" For Things

This was a timed post. The way these work is that if it takes me more than two hours to complete the post, an applet that I made deletes everything I’ve written so far and I abandon the post. You can find my previous timed post here…

Read more

9 months ago · 172 likes · 12 comments · Cremieux

The intricacies of multivariate adjustment present a kind of statistical bias, but of course there were methodological issues in this study too. The dietary info was acquired via a simple survey—it’s probably not that accurate. In fact, we know it’s not that accurate because the researchers had to filter a lot of junk survey responses out of their data, based on assumptions that are subject to the authors’ biases—they filtered out “unreliable dietary data (more than 80% missing), and those with implausible energy intakes”. Would the results be different if they filtered out data that was 70% missing instead? Probably. This is a huge limitation for all nutrition studies and is probably why so much of the field is pure guff.

Another bit of egregious statistical malpractice here is the failure to report FDR-adjusted p-values. This is very important when you are doing lots of different tests on the same data. If, say, you perform 20 tests (i.e., you look at the outcomes for 20 cancers), then one out of those twenty will give you a p-value less than 0.05 (which is “significant”) purely by chance. In this study, they have 20 cancers being tested across 4 pairwise dietary comparisons. With 80 tests, you’d expect ~4 anomalously significant results. Therefore, it is essential to adjust for multiple comparisons. This can be done in a number of ways. In the paper, they perform Benjamini-Hochberg adjustment, which converts your nominal p-value into an adjusted false discovery rate, sometimes called a q-value. The FDR is inevitably a larger number than the raw p-value (meaning less statistically significant after adjustment).

Now, the paper wouldn’t have been publishable if they didn’t do this. The problem is, whenever they describe their results or present them in forest plots and tables, they claim significance based on nominal p-values! They append the descriptor “FDR significant” to comparisons that are actually significant after adjustment, but they don’t report the numerical adjusted p-values anywhere (even the nominal p-values are hidden in supplementary table 3). They also don’t exactly make it easy to get these data if you wanted to find out yourself …

In their discussion, they are forced to admit the pitfalls of basing their narrative on unadjusted p-values:

“Some of the nominally significant associations observed might be due to chance because of the number of tests performed, but we interpreted all the findings cautiously.”

I guess all the excited journalists missed this sentence.

The actual data

The headline result pushed by both the authors and journalists was that the vegetarian diet was associated with a lower risk of five types of cancer. This is true: comparing vegetarians to meat eaters, the vegetarians in the study had (nominally) significantly lower risks of breast, prostate, kidney, pancreatic, and multiple myeloma cancers. Oll korrect.

Were there any other significant results that were pushed later into the news articles and hidden from the headlines? Erm, does the Pope wear a funny hat?

Compared to meat eaters, vegetarians had almost double the risk of oesophageal squamous cell carcinoma (HR = 1.93, 95% CI: 1.30-2.87). Looking at the vegans there was no protective effect for any cancer and they actually had 40% higher risk of colorectal cancer compared to meat eaters. You’ll note the way these inconvenient findings are brushed under the carpet in many of the news articles on the study.

So which of these “significant” results survives adjustment for multiple comparisons according to the authors?

So the only findings that are unlikely to be false discoveries are:

• pescatarians had a slightly lower rate of colorectal cancer than meat eaters

• vegans had higher rate of colorectal cancer than meat eaters

• vegetarians had a drastically higher rate of oesophageal squamous cell carcinoma than meat eaters

Another way to ensure statistical robustness is through something called sensitivity analysis. This basically involves tweaking the setup of the analysis slightly to see if significant results remain significant. If an minor tweak removes significance, its a telltale sign that the result wasn’t significant in the first place, but likely a fluke. Real biology is robust to sensitivity analysis.

The authors of the study carry out sensitivity analysis in two ways: They recalculate the results when

1. excluding the first four years of follow-up, and

2. only including never-smokers

Of the original 11 nominally significant findings, they found only two results remained nominally significant with these alterations:

• higher oesophageal SCC in vegetarians compared to meat eaters (this result actually got way stronger in the sensitivity analysis—the vegetarians who had never smoked got this cancer almost three times as much as meat eaters who had never smoked)

• lower kidney cancer in vegetarians compared to meat eaters.

Every other significant finding disappeared with the sensitivity analysis tweaks.

Taking the sensitivity analysis findings and the multiple comparisons adjustment together, there was only a single truly robust finding from this study: vegetarians had a significantly higher risk of oesophageal squamous cell carcinoma than meat eaters.

This is what the data actually show. You may have noticed that this is pretty much the polar opposite of the narrative pushed in the media. I’m sure you can see my frustration here?!

Going back to the BBC article title, instead of:

“Vegetarian diet lowers risk of cancer, study finds”

it should have read:

“A bunch of vegetarians happened to go on to develop a type of oesophageal cancer more frequently than what would have been expected by chance compared to otherwise similar meat eaters”.

Alas, most journalists and a lot of scientists are statistically illiterate and epistemically inept.

I originally planned on covering three such papers in this article (each relating to one of the scientific questions I posed in the introduction), but I think doing so would be labouring the point. Instead, I’ll leave one of them in the footnotes as an exercise to the reader.5

So what is the average person to do? Well, firstly don’t believe everything you read in the news, but that’s always the case. Then, if you’re interested in science and if you wish to be able to talk about scientific studies to back up your opinions and arguments, being cognizant of all of these biases will in itself immunise you against them to some degree. Learning some stats also probably won’t hurt. Finally, if you want to simply know which Science™ you can trust, I’d recommend finding and following individuals who repeatedly demonstrate competence in statistical methods and scientific interpretation. If in doubt, read the study critically yourself.

Subscribe now

5

Here’s the study: Temperature-related neonatal deaths attributable to climate change in 29 low- and middle-income countries

Here’s some media coverage:

The Print:

The Guardian:

See if you can spot the inconvenient data in the actual study. Hopefully you’ll also get a chuckle out of the x-axis scales of Figure 3c versus d.

For 20-year-old “Clav”, this means administering himself with a concoction of potions including testosterone (his natural production has been destroyed), accutane (a drug for acne), melanotan II (a peptide that enhances melanin production), retatrutide (an experimental next-gen GLP-1 mimetic), Anavar (a synthetic anabolic-androgenic steroid), and heaps of antioxidants to help his liver deal with these drugs, including ungodly quantities of the sleep drug melatonin which also has antioxidant properties at high doses. In addition to the drugs and supplements, he massages his face, mews, and performs bone-smashing—meaning he hits his facial bones with a hammer in zones where he wants to promote bone growth (like the cheek bones). He also claims to want to undergo limb-lengthening surgery soon to increase his height from 6’2” to 6’6”.

Interviewing him last month, the conservative pundit Michael Knowles turned the conversation to Clav’s politics. Clavicular revealed that, in a Vance versus Newsom 2028 presidential race, he backs Newsom “because JD Vance is subhuman and Gavin Newsom mogs”. He goes on to describe all the ways in which JD is ugly and fat, compared with Newsom, who is a “6’3” chad”. The clip is quite funny to be honest:

Clearly, Clavicular is unhealthily obsessed with looks. He’s also a bit of a dork, bad at talking to women, and malcoordinated. But this essay won’t spend time critiquing Clavicular. Plenty of others have done so—none more expertly than comedian gym bro Mike Tornabene. Instead, I want to take his political stance seriously.

Pretty privilege and the halo effect

You see, Clavicular is hyper-aware of something called pretty privilege. If you’re hot, people treat you better: the barista smiles at you, the boss minds less if you’re late to work, and people are more likely to do favours that you ask of them.

Pretty privilege slots into a broader psychological effect known as the halo effect. This describes how people judge one attribute of a thing based on pre-existing judgements of another attributes. For example, a 2020 study published in the Annals of Tourism Research found that the halo effect causes people to rate the location of different hotels in the same location differently based on how good other aspects of the hotels are (i.e., a bad hotel in a certain location gets worse “location ratings” than good hotel in the very same location).

Of course, it applies to peoples’ looks too, though exactly how this manifests is slightly different for men and for women. For example, a 2019 longitudinal, nationally representative American study found that being attractive protects women from being arrested and convicted, but not men. Other studies have found that attractiveness and femaleness also lead to more lenient judgements from jurors and lead to more lenient sentences.

Outside of criminal justice, perceived workplace competency is heavily influenced by looks. A 2024 study using deepfakes to isolate physical attractiveness as the only variable being altered in job applications found that being more attractive led to significantly higher competency ratings and chances of being invited for an interview. Notably, the effect here was stronger in males, with females only experiencing significant pretty privilege for job applications in stereotypically female job roles.

The halo effect can spill into politics too. The wildly different responses we see to the Trump and Obama deportation strategies and mishaps, which are incredibly similar by any objective standard (if not worse under Obama), can be largely attributed to the halo effect (which sits upstream of biased media narratives).

In this case, Obama’s superior rhetoric and stateliness works like a halo—protecting other attributes of his leadership from being criticised as aggressively. Trump, instead, is quite uniquely unpresidential. While this sacrifices the halo effect, it adds an authenticity factor—something I have written about previously:

Obama is also much better looking than Trump. If we like the way someone looks, we might be predisposed to judge their other attributes more positively. Books are always judged first and foremost by their covers, and the same goes for humans. This is the calculus that Clavicular is considering regarding his prediction for the 2028 election.

Polymarket has Vance at 25% and Newsom at 19% at time of writing, but are looks alone powerful enough to switch things up in Newsom’s favour?

Political analysts focus on the wrong things

Clearly there are lots of factors that can go into a presidential election. Vance’s chances will obviously depend on the success of the Trump admin, and the nature of his participation in it over the next few years, while Newsom’s will depend on his ability to memory-hole his gubernatorial failures and continue his fire-with-fire anti-Trump strategy. Momentum will also matter. The passing of prop 50 is a big win for Newsom, and the midterms are Vance’s to lose. Vance is a well-travelled, high IQ, erudite Yale grad and military vet with a proven track record of political campaigning and debating. Meanwhile Newsom is a seasoned politician and great orator with the ability to unite and lead the Biden and the Mumdani factions of the Democrat party. He’s also a mid-wit.1

In a democracy, we get one vote per person, regardless of one’s level of political engagement, education, or intelligence. I propose that this fact renders detailed dispassionate analysis of candidates and policy platforms somewhat irrelevant and overdone. Indeed, a 2020 Science paper argued that political campaigns don’t move the needle very much at all—with background economic factors, incumbency, and candidate characteristics playing more important roles.

Which candidate characteristics are important though? IQ doesn’t matter unless you can use it to get more people to vote for you. Erudition doesn’t matter unless reading Hayek makes you a better campaigner (it doesn’t). Of course, these things matter for the success of a government in power, but not for winning elections—as pointed out by Plato some time ago.2

Since a majority of voting people don’t have the ability nor interest in accurately assessing the best candidate, in reality, the president will be elected on vibes. And one of the most important influencers of vibes is physical appearance.

Being good-looking wins you elections

A landmark 2005 study sought to test the effect of politicians’ looks on election outcomes by using US congressional election results between the year 2000 and 2004. The researchers got participants, who were unfamiliar with the politicians being tested, to judge portraits of pairs of competing candidates based on a number of factors, including attractiveness, honesty, and competence. The participants made their judgements in 1.051 seconds on average, ensuring post-hoc rationalisation didn’t interrupt their instinctual judgements, and if participants recognised a candidate, the data were thrown out.

When competence judgements of portraits between Senate race winners and runners up were plotted against the actual difference in votes between the candidates, a significantly positive correlation emerges.

Incredibly, facial judgements of “competence” predicted the actual outcomes of Senate races in 71.6% of cases (p < 0.001) and House races in 66.8% of cases (p < 0.001). Notably, competence judgements seemed to outstrip all other tested factors, including “attractiveness” judgements.

That judgements of “competence” of faces was much more important than “attractiveness” speaks to an important point about how male looks evolved: Masculine features (low voice, facial hair, prominant brows) seem to have evolved primarily to assist with intrasexual dominance among men, which determined access to mates ancestrally. Yet this selection pressure was/is at odds with facial features that appeal directly to females (intersexual attractiveness)—women actually prefer slightly higher vocal pitches and slightly less masculine features.

In other words, competition against other men drove selection for “competence” looks, while sexual selection for women drove “attractiveness” looks. When we talk about “good looks” relating to men, we inevitably mean a blend of these two facial attributes. Competency of men also overlaps strongly with attractiveness of men in general.

For women, the story is somewhat complicated by female intrasexual competition and significantly lower male attraction to female competency than vice versa (men care more about raw attractiveness of prospective female partners). I think if the Todorov et al. study could be repeated with only female candidates, the results might be markedly different. Luckily, I don’t think AOC or Kamala (remember her?) are particularly strong contenders for president in 2028, so we are fine focusing on male looks only here.

What about presidential elections?

Congressional elections are somewhat lower profile than presidential elections, with the latter receiving much more media coverage and in-depth analysis. So do looks still matter in this instance? I decided to put it to the test.

I looked at all presidential races since 1960—which was when the first presidential debates were televised and looks could really play a role. I took portraits of candidates from the Wikipedia page of each election and plugged them into Grok, asking it to give me a “looks score” from 1-10 based solely on judgements of competency from the portrait.3 I also took height as a factor—again from a Wikipedia page—and calculated whether height, competency-based looks, or the product of both predicted election outcomes better than chance. Crucially, the outcome I care about here is popular vote, not electoral vote.

Amazingly, the looks score predicted the presidential race winner in 81.8% of races that weren’t tied—significantly better than random chance via Chi-squared and one-tailed binomial tests. Height alone was not significantly predictive, but combining looks and height was a stronger predictor than just looks alone.

Incumbency is often thought to be a very important factor in predicting election outcomes, so I included an assessment of the predictive power of incumbency to compare it with looks. I used two definitions of incumbency—one defining an incumbent as a candidate who was elected as president in the previous race, and the other broadening the inclusion criteria to include vice presidents in office who then run for president (as JD Vance would be in 2028). Via these two definitions, incumbency predicted outcomes in 71.4% and 53.3% of races—with neither being statistically better than random chance.

Beyond binary analyses, plotting vote-share difference against looks score differences revealed a (marginally) significantly positive correlation, with looks explaining around 25% of variance in vote-share differences.

I want to ensure that readers are aware of the limitations of my Grok-based looks scoring. I can’t be sure Grok didn’t recognise the candidates and then be biased by the Halo effect, but I couldn’t find free, purpose-built AI tools that would rate facial “competency” over just attractiveness. Feel free to take my results with a pinch of salt in light of this. Human judgements would’ve been better but it’d be almost impossible to find people who don’t know who these candidates are (unlike the congressional races that Todorov et al. analysed).

Caveats aside, it seems Clavicular may be the most perceptive political pundit of our time. And he’s voting Newsom if the election were today. But if looksmaxxing has taught us anything, it’s that our looks are malleable. And if Snapchat filters have taught us anything, it’s that perceptions of our looks are malleable.

Inverse catfishing

There are recent examples of trying to use the halo effect to discredit someone too. In 2021, Joe Rogan got COVID. His doctor and him concocted a “kitchen sink” solution for it, including the drug Ivermectin.4 It’s a very effective anti-parasitic drug that also has a strong antiviral track record in vitro and in mouse models, but which lacks strong evidence in human trials. CNN wanted to denounce this or perhaps their competitor Rogan in general, so they shared a heavily filtered version of his Instagram story in their reporting. Unfortunately for them, Rogan has a much larger audience than them and could simply expose their tactics.

A more recent example is that of Alex Pretti, the anti-ICE protestor who was killed during a scuffle with CBP agents. Hilariously, several media outlets used an AI-enhanced photo of him. The rationale is clear: ugly martyrs aren’t as effective as good-looking ones.5

Interestingly, seeing the real version of both Rogan and Pretti after these failed attempts to warp the halo effect in a particular direction has exactly the opposite of the intended effect. Rogan looks, in reality, markedly healthy for someone with COVID in that video; while Pretti looks quite sickly and pasty compared to his AI-enhanced version.6

The effect is akin to when someone artificially enhances their online dating profile, known as “catfishing”. When they eventually show up to a real date, they get immediately exposed as a catfisher, and the gap between prior expectation and reality makes them seem even uglier to their date.

JD Vance seems to have been, cleverly, leveraging this inverse catfish effect recently. When he became the centre of insulting memes targeting his physical appearance, he laughed along and actually propagated the jokes himself. Not only does this demonstrate great humour and self-awareness, but also means that when people see his regular appearance, he looks comparatively better looking. Consider the meme below— he looks better in reality than most of the meme versions.

And then consider this chad-version of Vance compared to his real portrait—notice how bad he looks when put side-by-side with his AI-looksmaxxed self.

But memes aren’t that important for most of the voting public. His actual physical appearance on TV and at the eventual 2028 debates will be much more important. So how can he avoid being “mogged to death” by “6ft3 chad” Newsom?

Presidential looksmaxxing

To what extent can Vance actually improve his looks? Well the first thing to note is that altering the looks of a politician for political effect is nothing new. Back during WWII, the OSS hatched a plan to lace Hitler’s carrots with oestrogen to feminize him. The goal was to lower support for him as a leader by worsening his physical appearance. Yes, this is real. They really tried to make Hitler trans—before it was cool.

JD can do the opposite of this. At 41 years old, he can probably expect his natural testosterone production to decline over the coming years—especially given his cortisol-inducing job. A safe dose of testosterone replacement therapy wouldn’t be the worst idea for him to maintain muscle mass, boost cognitive energy, and promote leanness.

Indeed, becoming leaner is probably the best thing Vance can do, as this directly affects what one’s face looks like, and he seems to carry a lot of his fat in his face (me too, JD). If he struggles to get time to exercise or doesn’t have the willpower to control his eating habits, he should (seriously) consider going on a GLP-1 mimetic, especially with TrumpRx pricing.

I think Vance knows that he carries fat on his face, and that this is bad for his career as a politician: he uses his beard very well to hide the chubbiness of his face and neck, and he should continue to do this.

Vance also has age on his side. As 58-year-old Newsom’s forehead wrinkles set in, the mogging game is his to lose. The maintenance of Newsom’s hairline should be a top priority for Democrat PACs going forward.

Looks do matter enormously. Clavicular isn’t wrong. But they aren’t all that matter and certainly aren’t even the most important factor for things such as career prospects, wealth, fame, or even dating (outside of Tinder profiles, that is). That said, the halo effect of good looks in politics is clearly strong. As we see more of Newsom and Vance leading up to 2028, it will be interesting to see if and how their looks change. If I were our good friend Clav-“political soothsayer”-vicular, I’d probably stay away from Polymarket for now.

Leave a comment

5

Of course, memes followed:

(As you’ll hear, we originally wanted to tackle the topic of free will too, prompted by an essay I wrote a while ago now, but time got the better of us…)

We begin by discussing how PhDs are actually rather easy unless the student or supervisor makes it otherwise. Universities are financially incentivised to expand the number of students ad infinitum, leading to a lowering of standards at both admission and completion of degrees—including the supposedly elite PhD degree.

Our conversation then turns from skewed financial incentives at universities to skewed financial incentives at academic journals. We discuss what journals get right and wrong, and how they exploit scientists and, ultimately, rip off tax payers.

Finally, we touch on incentives that motivate scientists, and how Substack can leverage these to usurp traditional academic journals. Dayne has been working on a pilot at Substack to index scholarly Substack articles with digital object identifiers and to have them indexed with Europe PMC and, hopefully, Google Scholar. My scholarly review on protein intake is part of this initial pilot.

If you find this discussion interesting and haven’t read my previous posts on these topics, you can find them here:

If you enjoy the conversation, I would appreciate you letting me know by way of liking it or sharing it with friends. This livestream/podcast was Dayne’s idea, but Substack has extremely an frictionless livestreaming interface, so I wouldn’t be averse to recording more interesting conversations in future if there is appetite for it—of course in addition to my written content.

Share

Back in 2015, Ezra Klein (then editor-in-chief at Vox) interviewed Bernie Sanders leading up to his 2016 run for President. At one point, Klein, on the frontier of the woke takeover of the American Left at the time, put to Sanders that, if we are to take global poverty seriously, then we ought to “sharply [raise] the level of immigration we permit, even up to the level of open borders”. Shocked, Sanders responds:

“Open borders? No, that’s a Koch brothers proposal … I mean that’s a right wing proposal which says essentially there is no United States. … It would make everyone in America poorer: you’re doing away with the concept of a nation state, and I don’t think there is any country in the world which believes in that.

…

“What rightwing people in this country would love is an open border policy. Bring in all kinds of people, work for two or three dollars an hour, that would be great for them. I don’t believe in that.

…

“Do you think we should open the borders and let in a lot of low-wage workers? Or do you think maybe we should try and get jobs for [unemployed young Americans]?”

It’s a fascinating exchange between a classical Lefty and a more progressive/woke Lefty. The difference in opinion comes down to a widening of the moral circle that has begun to alter traditional left-wing politics over the past few decades. This phenomenon was captured by a now-famous and often-misinterpreted 2019 study that gauged the effect of political ideology on the expanse of one’s moral circle—finding that conservatives morally prioritise their family, community, and country significantly more than progressives, who instead report to love “all others” more than they love their own family.

In 2025, it seems that Klein’s position has won out for the Left, despite Sanders holding on to his principled position a decade on. Most Lefties, particularly young ones, think strong borders reeks too strongly of nationalism (which equals Hitler), and calling for them demonstrates a lack of compassion for the third world. But is it true that the most compassionate thing to do for the global south is to open the borders to the West?

Some would argue that our compassion shouldn’t be magnetised by people who have nothing to do with us and don’t share our values, culture, or genes. In other words, our compassion should be first-and-foremost allocated to people with which we have personal or familial relations. This is the conservative position, and it is not one that I flippantly dismiss—it is defensible in its own right by invoking evolutionary psychology, game theory, or common-sense moral intuition that is the product of these two things.

Then there is the argument that our compassion should be focused on the working class within our own country. This is Sanders’ position. Sanders’ case is obviously a liberal (in the non-classical, American sense) argument against immigration, but it is one that I won’t be arguing here. His points should be obvious to most, but I find his position to be a halfway house. The logical conclusion of wanting to punish the privileged and uplift the “oppressed” is to make the mission global—which is what Klein was getting at in that interview.

Instead, I wish to lay out three arguments against immigration:

1. Immigration leads to exploitation of immigrants

2. Relatedly, immigration violates the free market

3. Immigration cripples the third world via brain drain

I implore progressives with a wide moral circle to take points 1 and 3 seriously, while the second point should strike a chord with classical Liberals or libertarians.

Colonialism 1.0: stool servants and gondoliers

Most iterations of colonialism—from Roman colonialism in 1st Century Judaea (renamed “Palestine” by the colonialists), to the present day situation in Tibet—involved not only initially conquering another peoples through war, but also subsequently ruling over them via a totalitarian and oppressive regime.

The most familiar iteration of colonialism to the university-educated westerner is that by the Europeans. The British, French, Belgians, Spanish, Portuguese, Dutch, and Italians all sailed to lands anew, set up settlements there, and subjected locals to their laws and customs—in many cases treating the locals as an ethnic underclass whose only purpose was to serve the European overlords.

Taking advantage of their maritime supremacy, the Europeans would also extract and transport resources back to their homelands. These resources included people, who were taken as slaves both to the New World and back to Europe itself.

For example, black Africans bought as slaves were used as “stool servants” by wealthy European families, particularly in Spain, Portugal, and Italy. This is exactly what you might imagine: the servants would accompany their masters to the lavatory and, if requested to, assist with undressing/dressing and post-plop wiping—particularly useful for an elderly or immobile aristocrat.

Bum wiping wasn’t the only useful task that imported slaves would do though. Any task that the local working class wouldn’t do without fair pay could be performed for free by a slave. For example, a plurality of gondoliers in 15th Century Venice were African slaves.

Indeed, this “occupation” offered a path to freedom for African slaves in Venice if the slave outlived their master—how fantastic! You might imagine a gaggle of Venetian women justifying their exploitation of blacks to each other:

“Oh don’t worry, Caterina. My husband treats his slaves very well. Besides, I don’t know about you, but I don’t particularly want to wipe my old husband’s bum and I’m very grateful for the people who do this work. The Paduans and Veronans are far lazier than the Moors—and they demand wages!”

Colonialism 2.0: bum wipers and Uber drivers

If you pay attention to British politics, that imaginary exchange between Venetian aristocrats might sound eerily familiar. That’s because Zack Polanski, the leader of the UK’s Green Party, expressed this exact sentiment when trying to justify mass immigration into the UK on BBC’s Question Time last month:

“One in five care workers are foreign nationals. I don’t know about you, but I don’t particularly want to wipe someone’s bum, and I’m very grateful for the people who do this work.”

And there is it. A momentary mask slippage and all of a sudden the pro-immigration stance of Britain’s hard left reveals itself as not being driven by compassion for foreigners at all, but by the utility of their exploitation. Even the left wing Guardian was shocked—both at the statement itself and the lack of criticism from Polanski’s base, who normally define themselves by their overflowing compassion.

Is this any different from 15th Century Venice?

In most European or American cities nowadays, you’ll find a concentration of foreign nationals and ethnic minorities in jobs that Zack Polanski wouldn’t particularly want to do. Before 2012, London’s Black Cabs were the dominant mode of private hire vehicle in the city. Their drivers were highly skilled (having spent 3+ years studying for the notoriously difficult “The Knowledge of London” examination) and almost all were white British men. Since alternatives like Uber have come about, these taxis are becoming obsolete, and the demographic of the workforce is being replaced by less skilled and cheaper imported labour. In 2023, 94% of Uber drivers in London were ethnic “minorities” or migrants.

Classical Lefties like Bernie Sanders might object to the dispossession of the cockney working class here, and he would have good reason to. The same argument is cropping up among the MAGA Right across the pond, where low-skilled immigrants—often entering the country illegally—are displacing the native working class by taking wages below the legal minimum.1

“So what?”, chime woke Lefties and fiscal conservatives in unison. If the migrants choose to do the job then they aren’t exactly being exploited. It’s not as if they were forcefully imported in the same way Moors were in 15th century Venice. And clearly the low pay is worth it to them. The natives are just too lazy or too spoiled to work for such low wages2. Isn’t this just free market economics giving us the best price for goods and services?

A tempting but flawed argument.

Regarding exploitation, it is obviously true that one can be coerced into choosing an exploited existence if the alternatives are worse. In the case of ethnic Uber drivers, the human toll of sitting in a car for long hours at any and all times of day with little job security or salary guarantee has been well-documented by leftwing sociologists. On the other hand, any degree of labour is, arguably, “exploitative”, so where do we draw the line? Are Elon Musk’s 100+ hour work weeks an example of him being exploited by Tesla share holders?

If the labour market were truly a free market, then I wouldn’t have any qualms with perceptions of exploitation. A free labour market means that workers are free to quit and find another job, and thereby dodge or reduce poor treatment. But, when immigration into western countries is concerned, the labour market is anything but free.

Mass immigration violates the free market

Libertarian economists all the way back to Adam Smith have strongly supported fluid borders as being instrumental to a free labour market. Free labour markets, like free goods markets, produce very efficient economies which raise the material wealth and welfare of everyone. Smith pointed to the open borders of New England (and its massive birth rate) to explain America’s meteoric rise in wealth.3

The “more people = more prosperity” dogma here has clearly been strongly adopted by Europe and, to a lesser extent, America over the past 30 years or so. And while it may hold true that more people raises total GDP, it doesn’t have any relation to material welfare and prosperity of people, as demonstrated by the UK’s completely stagnant GDP per capita since Tony Blaire threw open the borders.

The mistake being made here is equating people with goods—equating immigration with trade, that is. When we talk about “free markets“, what we mean is that individual transactions are controlled by the buyer and the seller (or the sender and the recipient), free from any government intervention. The key here is that transactions are agreed upon by both parties. There is mutual consent and individuals have the freedom to transact on their own terms.

Immigration, on the other hand, does not involve a decentralised, consensual agreement because people can move of their own accord, while goods and services cannot. Thus, there is no “free market” for immigration whereby recipient and sender get to agree on the transaction. This false equivalence of goods and people was highlighted by economist Hans-Hermann Hoppe in his landmark 1998 paper “The Case for Free Trade and Restricted Immigration”. Hoppe points out that the use of the word “free” when referring to trade or immigration mean categorically different things, and that the logically consistent position for a libertarian (or classical Liberal) to take is one for free trade and against “free” immigration:

“[W]hile someone can migrate from one place to another without anyone else wanting him to do so, goods and services cannot be shipped from place to place unless both sender and receiver agree. Trivial as this distinction may appear, it has momentous consequences. For free in conjunction with trade then means trade by invitation of private households and firms only; and restricted trade does not mean protection of households and firms from uninvited goods or services, but invasion and abrogation of the right of private households and firms to extend or deny invitations to their own property. In contrast, free in conjunction with immigration does not mean immigration by invitation of individual households and firms, but unwanted invasion or forced integration; and restricted immigration actually means, or at least can mean, the protection of private households and firms from unwanted invasion and forced integration. Hence, in advocating free trade and restricted immigration, one follows the same principle: requiring an invitation for people as for goods and services.4

In addition to disregarding the consent of citizens to immigration onto their property (it is theirs in the sense that publicly owned goods and property belong to the citizens of the country), the differing laws of different countries also creates a plethora of non-economic incentives for immigration.

For example, if an Iranian wishes to escape theocratic laws, they might immigrate to the US but compete in the US labour market with a much lower requirement for wage or economic incentive—because there is such strong political incentive for their immigration. This means that the Iranian immigrant would put up with considerably lower wages and considerably worse working conditions than a non-immigrant simply because doing so enables them to relish in America’s superior political climate and laws. The net effect of this is to depress wages in the US by mixing in non-economic incentives. The free market becomes heavily skewed by political differences between countries. Throw in a generous welfare state that hands out goodies to immigrants—illegal and legal alike, as many Western governments do—and any semblance of the immigration being “free” in the libertarian sense of a free transaction goes completely out the window.

Employment visas exploit immigrants

As well as pricing out natives, the skewed labour market produced by immigration is highly exploitative of the immigrant worker. Take the H1B visa system.

Employing someone on an H1B visa gives you, as their employer, complete power over their right to be in the country. This creates a skewed hiring incentive: to get away with offering lower wages and to acquire more power over their employees, companies will preferentially hire H1B visa holders. It’s a lose-lose situation for native and immigrant employees—while the companies enjoy power over their employees’ immigration status. It is a fact that H1B holders’ careers stagnate relative to residents’ because they have no leverage with which to ask their employers for promotions, and no means by which to switch employers. It’s modern day corporate serfdom that we should oppose on moral grounds.

Closing the borders would prevent western companies from having a disproportionate grip over the lives of immigrant workers. If immigrants are instead encouraged to work in their own countries, they will find themselves with much greater power as an employee, all else being equal.

Of course, “all else being equal” is easier said than done, but it doesn’t have to be so. Let me put it to you that it is immigration itself that ensures that other countries remain worse places to be than western countries.

Subscribe now

Immigration keeps the third world dependent and poor

The concept of brain drain should be a familiar one to most readers. Since the 20th Century, America has enjoyed its position as the recipient of brain drain from many countries—from Nazi-occupied Europe to Starmer-occupied UK (yes I’m referring to myself here). Attracting the best and brightest from across the globe is obviously a good thing for any country, but rarely do people consider the other side of that equation: it is catastrophic for the country being drained.

India has tremendous potential as a country. Its huge, relatively young population combined with its high quality educational institutions make it a rich source of talent that the US keenly exploits. In 2024, over a quarter of a million H1Bs were awarded to Indians—71% of all H1Bs. Every year, a large portion of the Indian crème de la crème emigrates from the homeland and is put to work for US companies or academic institutions. This is great for the US companies (as I’ve explained) and also seemingly for the individual immigrants, who now get to chase the American dream. But what about the Indians who weren’t so lucky in the H1B lottery, and what about India as a whole?

One supposed benefit of this immigration is the ability to spend US-earned wages with Indian spending power back home. India receives the most remittances of any country in the world—over $137B in 2024 per the World Bank. On the surface this seems great for India—they get to extract wealth from America. Yet this injection of emigrant-earned dollars is massively inflationary, and contributes to the rapid outpacing of domestic wages by asset prices in India. This makes life very difficult for the Indian domestic middle class, who are finding it increasingly difficult to buy homes, for example.

Further, it makes India geopolitically subservient to America. If their economy relies on billions coming back to India from America every year, then it’s hardly surprising that Modi bends over backwards for Trump. Modi’s obsequiousness was, of course, not sufficient to prevent Trump from significantly constraining the H1B program.

Most importantly, though, the brain drain from India to America prevents India from fulfilling its own potential as a country. It lets Indian-American CEOs become points of Indian pride, rather than actual Indian companies in India. Ruchi Gupta put it well last year in an opinion piece for The Indian Express:

“The true cost of this talent exodus extends beyond the immediate brain drain. It perpetuates an outward focus that weakens the urgency for domestic reform, allowing the state to neglect its responsibility to create equitable opportunities and an environment conducive to growth and innovation. The question is no longer whether Indians can succeed globally — they clearly can — but why India struggles to replicate this success domestically and equitably for its masses.”

Thus, the H1B system holds India down by bolstering American hegemony and preventing homegrown talent from investing time and effort in India itself.

Indian migration to the US is a particularly voluminous example, but the same dynamic holds true for other countries. It is simply the case that, whatever the country being emigrated and whatever the reasons for said emigration, the most likely to be able to leave are the most competent. It is as true for Indian H1B holders as it is for illegal immigrants crossing the English Channel on small boats. It is no coincidence that the latter are almost all military-aged men. That’s because this demographic is the most agentic and resourceful among their compatriots. It is this demographic that would otherwise be the most able to find or create solutions in their own countries.

Skimming the upper tail of the talent pool away from third world countries every year is a surefire way to prevent those countries from independently developing and become richer of their own accord.

To conclude

If we care about the rights and welfare of immigrants, we should consider that their presence in western countries is incentivized by those, like Zack Polanski, that wish to exploit what they have to offer us. Arguing for immigration because of the yucky jobs that immigrants will put up with is a disgusting line of argument that is tantamount to colonialism 2.0. If Piers Morgan happens to be reading—yes, arguing for immigration because it means you get to try out diverse cuisines is also rooted in a fundamentally colonial mindset.

If we care about free markets, we should care about ensuring that movement of people involves a free transaction whereby the citizens of a country are free to say “no” to the deal. This is simply a matter of consent and sovereignty. Any libertarian who equates free immigration with free trade is simply not a serious thinker.

Finally, if we care about the sovereignty and prosperity of third world nations, we should care about limiting the emigration of their most competent into the west. Self-sufficiency is the true opposite of colonialism, and is what we should hope for for all countries.

Subscribe now

As a fairly precocious schoolboy, the simplified theology that I was exposed to at Catholic primary school and Anglican secondary school never cut the mustard for me. I recall standing in the Tonbridge School chapel deliberately keeping my hands separate, deliberately gazing upwards at the beautiful stained glass windows, and deliberately keeping my lips firmly pressed together as the rest of the school chanted the Apostles’ Creed. I was like Rosa Parks if Rosa Parks was a young white boy who went to a £60,000 per year British boarding school1. I found the defiance exhilarating.

As I grew older and learned more about philosophy and theology, I became able to reconstruct arguments against the existence of God from first principles. I was no longer just an observer, but a gladiator in the arena. As an undergraduate at Cambridge, the atheistic worldview was assumed. I recall many times sitting in the college bar with a friend bathing in each other’s atheist rhetoric. I knew I needed to challenge my worldview to strengthen it, though.

Fighting a war on two fronts, I would take on woke nonsense at college feminist society meetings, but also flex my logical and rhetorical muscles against Christians when the opportunity arose. On nights out when I didn’t get blackout drunk, I made a habit of debating the volunteers from the Christian Union. They would stand outside Cindies handing out jam sandwiches and bottles of water to drunk, wayward souls. Between 1am and 3am on Thursday mornings I could be found locked into theological debate with these kind Christians. Maybe it was my use of performance-enhancing drugs in the form of caffeine and alcohol from 4 for £10 VKs coursing through my veins, or the exhaustion of my interlocutors who probably actually planned on going to their 9am lectures that morning, but I never lost one of these post-nightclub debates. Atheism remained undefeated.

The problem of evil

My secret weapon in such debates has always been the problem of evil. This is the notion that, if there is an all-powerful, all-loving, all-knowing God, how is it that evil exists in the world? Why does God allow murder? How is it that children through no fault of their own die of cancer? How is it that there are species of parasitic wasps whose life cycles require them to implant their eggs into the larvae of other insects such that the hatched parasitic larvae then eat the other larvae from the inside out?2 Christian apologists have spent centuries—millennia even—trying to answer this question.

One answer is free will: God gifted humans free will, and therefore has to permit them to use that free will for evil purposes. In this version of events, which inspired a couple of fantasy novels out of John Milton, God is an ardent libertarian prioritising human individual liberty over our optimal material wellbeing. To my mind, this is completely valid. Many atheists are deluded into thinking that free will doesn’t exist, but there is even less tangible evidence for determinism than there is for God3. Free will clearly exists and is clearly a moral good. I would expect an all-loving, all-powerful God to permit human evil for this reason, but it does little to explain natural suffering in the world. Human free will has nothing to do with childhood leukaemia, for example4.

Another workaround for theists is to argue that short-term suffering is sometimes necessary for long-term benefit. In other words, evil and suffering are sometimes necessary components of processes in service of good. In this version of events, God is an ardent utilitarian prioritising the greater good over individual wellbeing. While many atheists are themselves ardent utilitarians, this argument rarely persuades them. For me, it works in some cases, but there are always counter-examples for which no great utility for the suffering can be imagined. For example, what utility is there to a random bird getting splattered to death by a North Sea wind turbine blade? Why couldn’t God just give kittiwakes slightly better visuospatial perception?

Another problem, for the Christian, with all of these arguments is that Christians also believe in a perfect realm without suffering or evil. They call it Heaven. If it is possible for Heaven to exist, then why can’t this realm just exist everywhere? Presumably there is free will in Heaven? The Heaven dilemma really does throw a spanner in the works here for Christian theodicy. Of course, you could believe in an omnipotent, omniscient, all-loving God without believing in Heaven, but this is a rare combination.

The problem of evil is a tough one, and has driven some theists to quite heterodox conclusions. Ancient Gnostic Christians didn’t suppose the creator God is all-loving at all, but portrayed him as an ignorant, arrogant, or outright malevolent “Demiurge.” On the other hand, panpsychist philosopher argues that God could be perfectly loving yet not traditionally omnipotent. Mormons take a similar line, seeing God as a benevolent organiser whose omnipotence isn’t absolute but constrained by pre-existing eternal laws and matter. Trimming God’s résumé on either the benevolence or the potency front does allow Him to coexist with evil, but it leaves classical Christians still struggling to rebut polemical drunks outside nightclubs at 2 a.m.

One thing all of these arguments both for and against God have in common is the presupposition that suffering is bad. This seems obvious. An axiomatic truth, almost. Perhaps not?

Leave a comment

To live is to suffer

I recently watched My Oxford Year on Netflix (please skip this paragraph if you don’t want spoilers). The film is about an American woman, Anna, who defers her job offer with Goldman Sachs to spend a year at Oxford studying for a masters in Victorian poetry. While there, she falls for a British PhD student, Jamie, who is also her supervisor5. They fall for each other, but, at Jamie’s suggestion, force themselves to keep things casual and focused on “fun”. Towards the end of her masters, Anna begins to suspect that Jamie is cheating on her. She leaves the Boat Race early, after Jamie didn’t show up, and hurries back to Oxford to try and catch him in the act, only to find him in bed hooked up to a chemotherapy drip. Turns out Jamie has a rare heritable type of cancer that also killed his younger brother, and he had kept this secret from Anna. The couple argue and Jamie insists that she go home to America and forget about him. The ex-girlfriend of Jamie’s dead younger brother also tries to persuade Anna to leave Jamie, insisting that staying with him while he slowly dies of cancer will only bring her suffering. Yet Anna, despite knowing this, actively chooses suffering. She opts in to heartache, grief, pain and loneliness. A short while later Jamie dies and Anna is obviously heartbroken, but she doesn’t regret her decision.

In Man’s Search for Meaning, Viktor E. Frankl recounts his years in the Nazi death camps. His account flips Maslow’s hierarchy of needs on its head. Humans, he argues, can not only survive but psychologically thrive without basic needs and amongst immense suffering as long as they have purpose. Frankl doesn’t just regurgitate Nietzsche, though (“he who has a why to live can bear with almost any how”). Frankl circularises this insight by observing that one of the sources of purpose in the camps was the very act of consciously facing suffering. To quote Gordon W. Allport in the preface to the fourth edition:

“It is here that we encounter the central theme of existentialism: to live is to suffer, to survive is to find meaning in the suffering. If there is a purpose in life at all, there must be a purpose in suffering and in dying.”

I’d take an even harder stance: it’s not merely that purpose allows one to withstand suffering. Suffering is purpose. Suffering is meaning.

Choosing suffering

In the Myth of Sisyphus, Albert Camus declares that Sisyphus—destined to push a boulder up a hill everyday only for it to roll back down at the end of the day—lives a life rich in meaning created by his conscious revolt against his suffering. He concludes:

“The struggle itself towards the heights is enough to fill a man’s heart. One must imagine Sisyphus happy.”

To say that struggling equals happiness may seem strange to the modern human—obsessed with maximising pleasure and comfort and minimising pain and discomfort as we are. Yet we don’t need to imagine anything to see Camus’ point ourselves. We can look to two modern day reincarnations of Sisyphus: David Goggins and Ross Edgley.

For those unfamiliar, Ross Edgley is an extreme endurance athlete, adventurer, sports scientist, and an author. I also consider him to be a philosopher of sorts. Some of his most impressive feats include dragging a car the distance of a marathon, completing an Olympic distance triathlon with a tree on his back, and swimming ~1,800 mi around Great Britain. He recently did Iceland too.

In Ross’s book The Art of Resilience, in which he recounts his swim around Great Britain, he explains his stoic philosophy and pursuit of eudaemonia. A couple of notable anecdotes include him finding chunks of his tongue in his porridge after the chronic salt exposure started to disintegrate it. Another time, he unknowingly swam with a jellyfish tentacle tangled up in his goggles—constantly stinging him in the face for an hour or so until the crew pointed out the source of the pain to him. Both of these events—completely pointless suffering—just made Ross laugh. Of course, every single one of his feats involves immense suffering—suffering that cannot possibly be paid off at the finish line or thereafter. Ross smiles through it all, and it is quite obvious that this man is happy because of his suffering.

David Goggins, on the other hand, rarely smiles. Goggins is an American ultra-endurance athlete, ex-US Navy seal, motivational speaker, and author. I also consider him to be a philosopher of sorts. Goggins lived an incredible life. Growing up in an abusive home, by 24 he had become a fat loser with zero willpower working night shifts as a cockroach exterminator. He turned it around when he decided he “was sick of being haunted by being nobody”, and applied to be a Navy SEAL. In his book Can’t Hurt Me, he describes how he went through three Hell Weeks in a year (failing the first two due to pneumonia and fractured shins) before eventually making it.

He has since completed and won several ultramarathons. In 2005, he ran 101 miles in 19 hours having never run a marathon before. This qualified him to enter one of the toughest ultramarathons in the world, which he ran a few months later with stress fractures and acute kidney failure, placing fifth. Three months after that, he rented a bicycle to place second in a three-day ultra-distance triathlon in Hawaii. He also broke the world record for most pull-ups in 24 hours (4,030 in 17 hours). When he was 35, he found out that he was born with a hole in his heart that limits his cardiac function to 75%.

Now 50, he still runs 70+ miles most weeks for no prize, no audience, no reason except to conquer the voice that wants to quit. His source of purpose is overcoming mental resistance and enduring suffering. You won’t find Goggins smiling. You will find him regularly shouting at the camera telling viewers to “stay hard!”. To most outside observers, this is a miserable and painful existence, and many accuse him of psychopathy or sado-masochism. This is projection, though. Goggins may not smile, but he is happy—his veins course with eudaemonia. He chooses arbitrary, pointless suffering, and by doing so achieves purpose and meaning.

Sisyphean theodicy

They arrived there via two completely different routes, but both Edgley and Goggins have come to the same conclusion—that suffering is not inherently bad but actually something to be sought after. The skeptics who don’t understand Edgley or Goggins don’t understand the human animal. These modern avatars of Sisyphus aren’t psychopaths or masochists. They are human, all too human. To be human is to suffer.

What about non-human animal suffering?6 As I wrote about recently, non-human animals don’t really suffer in any way similar to the way we do. To do so requires complex emotional processing and phenomenal intelligence which other species don’t come close to exhibiting. This is despite our overwhelming tendency to anthropomorphise them—to over-exaggerate their human-like traits and to over-interpret their human-like behaviours. Nonetheless, it is of course true that other creatures—from dogs to dandelions—sense the environment around them and can detect, process, and respond to negative stimuli. To be alive is to suffer.

Many don’t agree with my views on animal suffering, and I don’t want to rehash that argument here, but we don’t have to agree here to appreciate that the suffering-meaning cohesion applies across species. If the intensity of suffering available to humans is proportional to the intensity of meaning derived from it, there is no reason to expect this relationship to breakdown for non-human animals. In other words, if you do happen to believe that animals suffer very intensely, then you must necessarily believe that animals derive intense meaning from that suffering too. Suffering and meaning emerge from the same complex psychological and neurological processes—they are pinned to each other across their spectra of intensities.

Wherever consciousness is complex enough for intense suffering, it is complex enough for intense meaning. The positive moral value of suffering therefore extends from David Goggins to shrimp (or whatever creature is currently in vogue for utilitarians).

This weekend I’ll be running the California International Marathon. It is true that the pain of a marathon, and of the marathon training block, delivers payoff at the finish line, but meaning is derived from the suffering itself, not the payoff. The last time I entered a marathon, I did not reach the finish line due to an injury. No medal, no sense of pride, no personal best. Yet it didn’t become a regretful or pointless experience—it was an incredibly meaningful one. The experience, despite being painful and pointless and not creating any longterm benefit (I lost all the fitness I had gained when recovering from the injury), was valuable to me.7

I wouldn’t want or expect an all-loving God to use His omnipotence to remove the pain and suffering associated with endurance running. Doing so would render it a less worthwhile thing to do. The same is true for involuntary sources of suffering—as attested to by Frankl and fictional Netflix rom-com characters, but also corroborated by any example of unwanted suffering I can think of from my own life.

To live is to suffer. A world without suffering would be a world without life itself. An all-loving, all-powerful God would be remiss to eliminate this moral good from the universe. A cold, dead universe without suffering is morally inferior to a universe rich with life, its suffering, and its meaning. I encourage you, dear reader, to meditate on the moral value that suffering—even pointless, involuntary suffering—adds to your own life.

Subscribe now

Abstract

The recommended daily allowance for protein is 0.8 grams per kilogram of bodyweight per day, or around 50 g per day for the average person. Eating more than the RDA is beneficial for controlling hunger, building muscle mass, and gaining strength—all of which benefit health and healthspan. Yet there is some scepticism towards the notion that “more is better”, with some advocating that excessively high protein intakes, beyond 1.6 g/kg/d for example, yield no further benefit or may be actively harmful. This review synthesises the benefits and risks of a high protein diet. The wealth of observational data in humans comes out in favour of “more is better” except for those with pre-existing chronic kidney disease. To determine the optimal protein intake for maximising accretion of muscle mass—which is predictive of healthspan—seven relevant meta-analyses of randomised controlled trials (RCTs) are dissected in detail. Higher daily protein intakes are beneficial for lean body mass at all ranges of intake. With concomitant resistance training, this dose-response relationship is remarkably linear. Therefore, personalised protein intake recommendations in healthy resistance-training individuals should maximise protein intake within financial and practical constraints, but unconstrained by perceptions of health concerns or diminishing returns.

Introduction

Protein is the only nitrogen-containing macronutrient, and is essential for human survival. Globally, over 180,000 people die per year from protein deficiency—with the vast majority being infants and the elderly (fortunately, this figure is now almost four times smaller than what it was in 1980) [1].

The US Institute of Medicine, based on a 2003 meta-analysis of nitrogen balance studies [2], recommends adults consume 0.8 grams of protein per kilogram of bodyweight per day, or 50 g for the average person eating a 2000 kcal diet. However, this number represents the minimum intake to stave off protein deficiency and muscle wasting, not the optimal amount for human flourishing. The Acceptable Macronutrient Distribution Range (AMDR), on the other hand, offers a recommended range of protein intake as part of an optimal diet [3]. This range is 10-35% of calories, which translates to between 0.8 and 2.8 g/kg/d of protein—the upper limit of which is substantially higher than the Recommended Dietary Allowance (RDA).

Within this large range, it is likely that precise optima vary across different populations. To offer more precise guidance for the elderly, an international panel of experts was brought together by the European Union Geriatric Medicine Society in 2013. The resulting PROT-AGE study group reached the conclusion that, for optimal health, adults over 65 years old should consume 1.0-1.2 g/kg/d of protein as a minimum if they are sedentary [4]. However, they also recommended that such adults not be sedentary—prescribing endurance- and resistance-based exercise at an individually tolerable level—and that the inclusion of such exercise increases the protein recommendation above 1.2 g/kg/d. They also noted that most older adults with acute or chronic diseases, with the exception of untreated severe kidney disease, need an even higher protein intake to optimize health, of around 1.2-1.5 g/kg/d.

Despite this, the notion that increasing protein intake above the 0.8 g/kg/d IoM RDA might be beneficial is still rebuked by academics and journalists alike. In a 2019 perspective, Mittendorfer and colleagues concluded that consuming more than 0.8 g/kg/d is not beneficial for muscle mass outside of weight loss and is not beneficial for overall physical function [5]. More recently, the New York Times argued the same, insisting that 0.8 g/kg/d is “adequate for most people’s basic needs”.

Beyond basic needs, the International Society of Sports Nutrition recommends healthy and exercising individuals consume 1.4-2.0 g/kg/d of protein, with resistance trained individuals requiring intakes at the upper end of this spectrum [6]. A combination of endurance and resistance exercise are already uncontroversially recommended for optimal human health, the benefits of which include increased quality of life and reduced all-cause mortality [7,8]. Therefore, the optimal protein intake must assume endurance and resistance training as a given if we are truly optimizing human health.

Yet even among exercising populations, optimal protein recommendations are disputed. For example, science communicators and journalists have recently argued against protein intakes beyond 1.6 g/kg/d, claiming no benefit beyond this intake and even suggesting that such high intakes may be actively harmful [9,10]. This review will summarise and critique the mechanistic data that have led to such assertions of harm, and look at the broader observational signals in human populations.

Beyond observational data, this review will focus not on preventing deficiencies or recommendations for sedentary individuals, but on the optimal daily protein intake for individuals following other healthspan-optimising recommendations. To this end, meta-analyses of randomised control trials (RCTs) focused on the impact of higher protein intake in combination with exercise will be discussed. Inclusion criteria of large meta-analyses, subgroup analyses, and meta-regression adjustments for covariates will be carefully dissected. For example, RCTs with resistance training versus without resistance training; RCTs in younger versus older populations; RCTs in obese versus healthy individuals; and RCTs of untrained versus trained individuals will be differentiated. The resulting synthesis will converge on a recommended daily protein intake for healthy exercising individuals looking to optimise healthspan.

The importance of muscle mass and strength for healthspan

The primary benefit of a higher protein intake is enhanced accretion of muscle mass and increased strength. While historically a niche outcome to optimise for, enhanced muscle strength and size are increasingly at the forefront of efforts to improve human health and healthspan. For example, muscle mass strongly predicts longevity in older adults, with top quartile for muscle mass having 20% lower risk of all-cause mortality than bottom quartile [11]. Further, a 2020 meta-analysis of observational studies found that sarcopenia—defined as the progressive loss of muscle mass and function—was correlated with a wide range of negative health outcomes, from cognitive impairment and depression to metabolic diseases, hospitalization, and all-cause mortality [12].

In addition to muscle mass, muscle strength as measured by grip strength is also significantly negatively correlated with all-cause mortality risk and cause-specific mortality risk from cancer, cardiovascular disease, and respiratory disease according to a UK Biobank prospective cohort study of over half a million participants [13]. Recent Mendelian randomisation studies have also linked grip strength with adiposity and shown that higher grip strength is causally upstream of increased cardiorespiratory fitness [14,15].

Given the strong correlational and causative evidence that resistance training enhances quality of life and reduces mortality risk [16,17], the potential for optimised protein intake to maximise resistance training-induced benefits on muscle mass and strength, and thereby health and healthspan, is of great importance for public health. Meta-analyses of RCTs specifically focused on this question will be discussed later.

Risks of high protein intake

Arguments against excessive protein intake also deserve consideration. Mechanistically, in vitro and mouse studies suggest that excessive protein intake may drive atherosclerosis by activating mammalian/mechanistic target of rapamycin (mTOR) in macrophages [18,19]. Of course, mTOR activation is also the mechanism by which protein and resistance training drives muscle protein synthesis, which we know to be beneficial [20,21].

In humans, the story is more complex. Mittendorfer and colleagues have argued that a high protein diet without concomitant carbohydrate restriction may present a risk for metabolic disease [5,22]. This is based on the fact that protein ingestion elevates blood insulin and glucagon levels, which, when chronically elevated, are risk factors for type 2 diabetes. However, this assertion ignores the fact that protein-induced elevations in insulin and glucagon are acute and simultaneous, and so do not elevate overall blood glucose concentration either acutely or chronically [23].

High protein diets have also been proposed to place the kidneys under increased strain. Indeed, in adults who already have chronic kidney disease, restricting protein intake was shown to decrease the risk of kidney failure and end-stage renal disease, but not all-cause death events according to a 2018 meta-analysis of RCTs [24]. However, in healthy adults—the focus of this review—RCTs have shown no alteration of kidney function from high vs normal vs low protein diets [25]. Further, a recent meta-analysis of observational studies found an inverse correlation between protein intake and risk of developing chronic kidney disease [26].

Observational studies on protein intake and lifespan or healthspan

To gain more relevant insight into the relationship between protein intake and lifespan, there is a wealth of observational studies. At first glance, the results here seem mixed. One meta-analysis of eleven prospective cohort studies with 350,452 participants found that higher protein intake was associated with increased all-cause mortality, although removing animal protein from the analysis reversed the direction of the association [27]. Yet, a larger meta-analysis of 32 prospective cohort studies with 715,128 participants found a negative correlation between all-cause mortality and higher protein intake across total, animal, and plant protein [28].

It is plausible that certain animal-specific lipids confound these studies in favour of lower (animal) protein intakes, given that a network meta-analysis of prospective cohort studies found that replacing saturated fatty acids (SFA) with ω-3 or ω-6 poly-unsaturated (PUFA) and mono-unsaturated fatty acids (MUFA) was associated with decreased all-cause mortality [29]. Supporting the notion that the correlation between high protein diets and reduced all-cause mortality is confounded by SFAs is The Kawasaki Aging and Wellbeing Project, in which fish (low in SFA, high in PUFA) represented the predominant protein source, which found that the individuals in the top quartile of protein intake had less than half the risk of all-cause mortality than those in the lowest quartile [30].

Degree of food processing is another plausible confounder that can explain why some observational studies find a positive correlation between animal protein intake and all-cause mortality. Prior to commercial expansion, vegan or plant-based diets have historically required an emphasis on whole foods over ultra-processed foods, while many animal-protein-containing foods are highly processed. Ultra-processed foods have been recently shown, in a RCT which equated calories, protein, and fibre between groups, to drive increased fat mass retention relative to a minimally processed diet in the context of caloric restriction [31]. Further, large prospective cohort studies have shown negative outcomes for cardiovascular disease risk, cancer risk, and all-cause mortality risk from processed meat but not unprocessed meat [32,33].

Thus, it is likely that the seeming advantage of plant protein in many large prospective cohort trials is mediated by lower ultra-processed foods, lower SFAs, and higher fibre intakes, the latter also being strongly correlated with reduced all-cause mortality [34]. When compared head-to-head in RCTs, multiple meta-analyses show that isolated animal protein supplementation is superior for muscle mass and strength compared to isolated plant protein supplementation [35,36].

It is worth noting that most prospective cohort studies investigating protein intake and longevity focus on older adults to be able to follow them until death. To address the dearth of insight for younger populations, and to focus on quality of life rather than mortality risk, a 2024 prospective cohort study studied the relationship between midlife protein intake and healthy aging, defined as “being free from 11 major chronic diseases, having good mental health, and not having impairments in either cognitive or physical function” [37]. The authors found a significant correlation between total, animal, dairy, or plant protein intake and healthy aging among 48,762 nurses under 60 years old.

Taken together, prospective cohort studies provide clear support for the notion that high protein diets improve disease and mortality risk, with the caveat that SFAs or ultra-processed foods or low fibre diets may blunt or reverse these benefits at the population level. Speculative mechanisms by which a high protein diet may negatively impact human health, such as macrophage mTOR activation, are therefore seemingly outweighed by the larger positive impacts of a high protein diet. Further, healthy kidney function is unaffected by high protein intakes, although individuals with pre-existing chronic kidney disease might be able to slow their disease progression by restricting protein intake, although their mortality risk is not increased by a higher protein intake. The preponderance of observational data supports the recommendation of a high protein diet for health, healthspan, and longevity, although precise quantitative recommendations are difficult to extract from these studies.

Optimising protein intake for muscle mass and strength

The beneficial effects of a high protein intake on healthspan and lifespan are mediated by improved body composition, enhanced muscle mass, and enhanced muscle function/strength. Meta-analyses of RCTs consistently show that higher protein intakes boost satiety and this can aid fat loss in both healthy and overweight individuals [38,39]. Meta-analyses of RCTs also consistently show that higher protein intakes support muscle mass accretion, strength gains, and the prevention of muscle atrophy during caloric restriction for healthy, overweight, or aged populations—especially in combination with resistance training [40–45]. A single exception here is a 2018 meta-analysis of RCTs by ten Haaff and colleagues which focused on the effects of protein supplementation on muscle size and function in non-frail community-dwelling elderly individuals [46].

To explain this outlier, and to extract precise protein intake recommendations from all of these meta-analyses, the remainder of this review will focus on seven relevant meta-analyses, shown in Table 1, performing re-analysis where appropriate.

Liao et al., 2017

Liao and colleagues performed a meta-analysis of 17 RCTs entitled: “Effects of protein supplementation combined with resistance exercise on body composition and physical function in older adults” [43]. The subjects of the RCTs were elderly and overweight, averaging 73.4 ± 8.4 years old and with a BMI of 29.7 ± 5.5. The RCTs all included resistance training. Higher protein intakes combined with resistance training resulted in significantly greater gains in lean mass and strength compared to just resistance training alone.

The authors also performed valuable subgroup analyses. Firstly, they found that being obese did not change the benefit of increased protein (Figure 1A). Secondly, they found that shorter intervention periods led to greater heterogeneity in results, with improvements from protein plus resistance training becoming non-significant with interventions under 12 weeks in duration (Figure 1B). This is instructive for other meta-analyses, since many include RCTs with very short interventions. Longer interventions create more consistent effects, and these effects are in favour of higher protein.

Interestingly, when separating RCTs by sex of the participants, the researchers found no beneficial effect of increased protein for women (Figure 1C). This has led many to believe that only men stand to gain from high protein diets, yet this interpretation should be made with caution. There were only three RCTs which included women:

• One 8-week trial of resistance training elderly polymyalgia rheumatica patients supplemented the intervention group with 28 g of protein per day and found a significant increase in lean body mass as measured by for Dual-energy X-ray Absorptiometry (DXA)—the gold standard for such measurements [47].

• A second study supplemented resistance training elderly individuals with only 15 g of protein per day and found no significant improvement from protein supplementation on DXA-determined lean body mass [48]. Doses of 20-40 g are typically recommended to sufficiently stimulate muscle protein synthesis [49]. Given that the elderly require even higher doses of protein to stimulate muscle protein synthesis to the same degree as younger populations [50,51], it is unsurprising that this trial of 15 g doses did not produce significantly greater lean mass gains.

• The third trial supplemented elderly women exercising with resistance bands with 20 g of protein per day, which may still be insufficient to stimulate muscle protein synthesis in this population, and measured lean body mass with bioelectrical impedance analysis (BIA) [52]. While BIA is easier to implement and correlates reasonably well with DXA measurements, it has been shown to be less accurate and consistently overestimate fat-free mass (FFM) and underestimate fat mass [53–55].

As such, the null results of the latter two studies can be explained by significant experimental limitations. In this instance, interpreting individual studies might be more informative than a meta-analysis of three RCTs with very heterogenous study designs. At the very least, given the preponderance of evidence that higher protein intakes offer improved outcomes for resistance training in other populations, more studies on elderly women are required before making firm conclusions that higher protein intakes are ineffective in this subgroup.

Beyond this subgroup analysis, Liao and colleagues do not offer dose-specific insights. In elderly and overweight populations, higher protein intakes are therefore recommended with no upper limit.

Morton et al., 2018

While recommendations for the elderly are important, health and healthspan recommendations should target all ages. This is especially true regarding muscle mass and strength, since accumulating more during early and midlife would offset age-related losses independently of interventions late in life.

Morton and colleagues performed a meta-analysis of 49 RCTs on healthy adults of all ages with resistance training. They found that increased protein intake led to significantly greater FFM, strength, and muscle size gains compared to resistance training alone [44].

Subgroup analysis led the authors to make two additional conclusions: the benefit of higher protein decreased with age, and higher protein intakes were less effective when participants were completely untrained prior to the study.

The diminishing effect of age here is based on a univariate analysis, which showed that age was significantly inversely correlated with protein-induced improvements in FFM (β = -0.01, p = 0.02). Yet when multivariate meta-regression was applied, which took into account any mediating effect of protein dose, age no longer significantly affected ΔFFM outcomes (β = -0.01, p = 0.19). Plotting protein dose against age reveals why this is the case: RCTs on older participants tended to use smaller protein doses (Figure 2A, B). Thus, with protein dose removed as a mediator, there is no evidence that older individuals do not benefit from higher protein intakes when undertaking resistance training. Indeed, as previously mentioned, older participants likely need even more protein to elicit the same benefit [50,51], and future RCTs on this demographic subgroup should endeavour to test higher doses of protein supplementation.

The greater benefit of higher protein intakes with concomitant resistance training in habitually trained individuals is a well-recognised phenomenon. This likely comes about since the effect size of the training alone is so large for previously untrained individuals (a phenomenon colloquially referred to as “newbie gains”), that any further benefit induced by diet is relatively small. Morton et al. found that the benefit of high protein intake on ΔFFM in untrained individuals was small in effect size and non-significant (+0.15 kg; 95% CI: −0.02, +0.31; p = 0.08), but large and significant in trained individuals (+1.05 kg; 95% CI: +0.61, +1.50), p<0.0001) [44]. The strong empirical evidence for this phenomenon presented by Morton and colleagues should be instructive for other meta-analyses, which often do not separate untrained from trained subjects. Given that resistance training is recommended for health and healthspan, recommendations for protein intake should be made primarily on the basis of RCTs on trained individuals.

Combining RCTs on both untrained and trained individuals, Morton et al. next analysed the relationship between daily protein intake and ΔFFM. They chose to model their data via segmental regression (“breakpoint analysis”) which fixed a horizontal line from 1.62 g/kg/d and above (Figure 2C). They chose this model because it explained more variance (R² = 0.1907) than a simple linear regression (R² = 0.1170, p = 0.0330) (Figure 2D). This choice of model has led to interpretations that there is no benefit of increasing protein intake beyond 1.6 g/kg/d. However, the same data can be fitted with a one-phase association curve (R² = 0.1443) (Figure 2E), a quadratic curve (R² = 0.1677) (Figure 2F), or a cubic curve (R² = 0.1979) (Figure 2G).

The best fitting model here is the cubic curve, which has a local maximum at 1.852 g/kg/d. Nonetheless, given that these data include both trained and untrained subjects, and the models all fit the heterogenous data quite poorly, caution should be taken before recommending optimal protein intakes from this dataset. It is also worth noting that ΔFFM is one beneficial outcome of a high protein diet. Even if protein intakes above 1.85 g/kg/d do not result in further gains in FFM, there may still be added benefit for fat loss or muscle function/strength.

ten Haaff et al., 2018

A meta-analysis of 36 RCTs of old (>50 years old) non-frail adults both with and without resistance training was performed by ten Haaff and colleagues [46]. They posited that frail elderly adults are more likely than non-frail individuals to benefit from high protein intakes, and so sought to test non-frail populations specifically. They found no significant enhancement of lean body mass or muscle strength/function in this population either with or without resistance training. This meta-analysis stands out as the only one of the seven reviewed here to show a null result for protein supplementation.

The most obvious reason for these null results relates to the phenomenon observed in Morton et al. (2018) [44]: protein interventions in RCTs on older adults tend to be very low. Many of the RCTs in this meta-analysis use doses of 15 g of protein or less, which is insufficient to stimulate muscle protein synthesis in young individuals, let alone old individuals. Of those RCTs that use higher doses (~25 g), many only implement these doses three days per week. Indeed, the authors themselves admit that “the protein supplementation protocol could have been suboptimal in some studies”.

The inferior protein interventions deployed by the RCTs in this meta-analysis make it difficult to form meaningful conclusions. Future RCTs of non-frail over-50s should aim to implement doses more likely to be effective (>25g per day) [56,57].

Tagawa et al., 2020

The largest meta-analysis investigating the effects of protein intake ± resistance training on muscle mass was performed by Tagawa and colleagues, and included 105 RCTs of all non-critically ill adults over a range of dietary settings [42]. By casting such a wide net, the authors sought to impute a dose-response curve, which would inform optimal daily protein intake recommendations. They generated spline curves to model the data and determine the dose-response relationship between daily protein intake and ΔFFM or lean body mass (Figure 3). Adjusting for age, sex, intervention period, and resistance training volume, they found that increasing protein intake always resulted in greater gains in FFM—even beyond 3 g/kg/d. In individuals not engaged in resistance training, diminishing returns were observed beyond 1.3 g/kg/d, but when subjects were concomitantly engaged in resistance training, the dose-response relationship was remarkably linear (Figure 3, middle panel).

This dose-response model has substantially tighter confidence intervals that those generated from Morton et al. (2018), likely due to better adjustment for co-variates and more data (105 RCTs versus 45 RCTs). This meta-analysis offers convincing evidence that lean body mass gains are superior with a higher protein diet at all ranges of protein intake, and that concomitant resistance training maximises the benefit of increases above 1.3 g/kg/d.

Nunes et al., 2022

The notion that higher intakes of protein across a wide range of intakes leads to greater lean body mass gains with concomitant resistance training was corroborated by a meta-analysis of 66 RCTs by Nunes and colleagues on healthy adults of all ages [41]. They found significant benefits of higher protein in resistance training individuals on lean body mass gains (SMD = +0.19; 95% CI: 0.11, 0.28) and also lower body strength gains (SMD = 0.21; 95% CI: 0.08, 0.34).

Subdividing protein intake into three strata revealed no significant benefit of protein intakes below 1.2 g/kg/d on lean body mass (SMD = -0.14; 95% CI: -0.56, 0.27), a modest and significant benefit between 1.2 g/kg/d and 1.6 g/kg/d (SMD = +0.17; 95% CI: 0.06, 0.28), and a large and significant benefit above 1.6 g/kg/d (SMD = +0.30; 95% CI: 0.17, 0.43). This three-level meta-regression directly supports the notion that RCTs where protein doses are too low are unlikely to show an added benefit of protein supplementation with resistance training on lean body mass (as is the case for many RCTs on older subjects—as discussed earlier in the context of the meta-analysis by ten Haaff and colleagues [46]). This same phenomenon was observed for strength gains, with only the highest substratum of protein intake (>1.6 g/kg/d) receiving a significant benefit from their additional protein intake (SMD = 0.40; 95% CI: 0.23, 0.57). Contrary to the assertion that there is little added value of increasing protein intake beyond 1.6 g/kg/d, this meta-analysis offers strong support that intakes above 1.6 g/kg/d are the most beneficial for resistance-training individuals.

Kokura et al., 2024

Focusing on overweight or obese individuals aiming to lose weight, Kokura and colleagues performed a meta-analysis of 28 RCTs investigating the effect of higher protein intakes on muscle mass retention [40]. They found that higher protein intakes were significantly protective against loss of muscle mass during weight loss (SMD = +0.75; 95% CI: 0.41, 1.10). Analysing the dose-response relationship in this context of caloric restriction, they found that protein intakes below 1.0 g/kg/d were detrimental for muscle mass retention, while intakes above 1.3 g/kg/d could not only prevent loss of muscle mass, but increase muscle mass. They found no significant benefit on strength from higher protein intakes, which is unsurprising given only eight of the included RCTs included concomitant resistance training.

This meta-analysis, while focusing on overweight individuals aiming for weight loss and mostly not undergoing resistance training, adds further weight to the recommendation that higher protein intakes (of at least 1.3 g/kg/d in this case) are beneficial for lean body mass.

Refalo et al., 2025

While most of the RCTs in the meta-analysis by Kokura and colleagues did not include resistance training, and all participants were overweight or obese, Refalo and colleagues exclusively focused on the effects of protein intake on FFM in calorically restricted non-obese individuals undertaking resistance training [45]. Further, all participants were experienced in resistance training prior to the study. These inclusion criteria make this meta-analysis of 29 RCTs highly relevant to healthy populations already following resistance training recommendations.

To model the dose-response relationship between protein intake and ΔFFM, the authors deployed a Bayesian approach to compare models. Contrary to previous indications of diminishing returns with higher protein intakes, they found that the dose-response relationship was linear with >97% probability, and that the linear model outperformed other non-linear models via Bayes Factor analysis. The benefit of higher protein for FFM was stronger in longer trials (>4 weeks), if participants had lower baseline bodyfat percentages, and in men. Including random effects, the linear model explained 55% of the variance (R²_conditional = 0.55) [45].

This statistically robust meta-analysis with stringent inclusion criteria corroborates the subgroup analysis on resistance training individuals by Tagawa and colleagues, which also showed a linear dose-response relationship after covariate adjustment (Figure 3) [42], but in the context of caloric restriction. Refalo and colleagues conclude that protein intakes up to 2.5 g/kg/d are associated with greater preservation of FFM during caloric restriction, and that increasing this further up to 4.2 g/kg/d (the highest dose of included RCTs) is “linearly associated with larger FFM gain”.

Summary and conclusions

Governmental recommendations for protein intakes were designed to prevent deficiency, not to optimise health. The updated AMDR recommendation for protein intake provides a very wide range (0.8 – 2.8 g/kg/d), which leaves a lot of room for dispute.

Many worry that excessively high protein intakes might be harmful, or that benefits wane beyond a certain threshold, yet these concerns are not evidence-based. Evidence of harm—either via glucose control, mTOR activation in immune cells, or on kidney function—is not seen in studies of healthy humans. There is evidence that chronic kidney disease patients may be able to slow rate of disease progression by restricting protein intake, but all-cause mortality risk is not improved by doing so [24]. At the population level, vast quantities of observational data show that higher protein intakes increase healthspan and lifespan, particularly when saturated fatty acids and ultra-processed foods are accounted for or avoided.

Higher protein intakes are beneficial for satiety, fat loss, muscle gain/retention, and muscle strength/function—all of which improve quality of life, health, healthspan, and lifespan. By careful review and interpretation of seven relevant meta-analyses, this review converges on the conclusion that higher protein intakes improve muscle mass accretion (or retention in a caloric deficit) at all ranges of protein intakes for young, old, healthy, and obese populations. In non-resistance-training populations, the additional benefit beyond 1.3 g/kg/d lowers—reflecting continued but diminishing returns here [42]. For resistance-training individuals, however, multiple large, well-designed, and statistically robust meta-analyses converge on the notion that the dose-response relationship between protein intake and ΔFFM is linear [41,42,45]. There is strong evidence that higher protein intakes accentuate strength gains induced by resistance training too. In older populations, a large number of RCTs have used suboptimal protein intervention doses, which can skew meta-analyses of these studies. The well-documented necessity for higher protein doses in elderly populations should be incorporated into future RCTs of the elderly.

Since resistance training is already recommended for most people to optimise health, the optimal protein intake for a resistance-training individual should be calculated via cost-benefit analysis that assumes that health benefits scale linearly, and that minimises economic-, convenience-, and preference-based costs primarily—since there is little evidence for any health-based costs of high protein intakes. For many, this number will emerge at the upper end of the AMDR recommendation, upwards of 2.2 g/kg/d.

Methods

Search strategy and selection criteria

Meta-analyses of RCTs investigating the effects of protein on muscle mass were identified by searches of PubMed and GoogleScholar using combinations of the search terms “protein”, “supplementation”, “muscle mass”, “lean body mass”, “fat-free mass”, “resistance training”, “meta-analysis”, and “RCT”. Only meta-analyses published in English between 2015 and 2025 were included.

Data extraction, statistical analysis, and data plotting

Where data was not directly available from the main meta-analysis article or its supplementary data, Plot Digitizer (https://plotdigitizer.com/) was used to extract values from charts. This was only required for re-analysis of Morton et al. (2018) [44]. Re-plotting and modelling of data and all statistical analyses was performed with Prism (Version 10.6.1, GraphPad).

Funding and conflicts of interest

This Review was not supported by any grant or funding source and the author declares no conflicts of interest.

References

1. Deaths from protein-energy malnutrition, by age, World Our world in data. https://ourworldindata.org/grapher/malnutrition-deaths-by-age.

2. Rand, W.M., Pellett, P.L., and Young, V.R. (2003). Meta-analysis of nitrogen balance studies for estimating protein requirements in healthy adults. American Journal of Clinical Nutrition 77, 109–127. https://doi.org/10.1093/ajcn/77.1.109.

3. Wolfe, R.R., Cifelli, A.M., Kostas, G., and Kim, I.Y. (2017). Optimizing Protein Intake in Adults: Interpretation and Application of the Recommended Dietary Allowance Compared with the Acceptable Macronutrient Distribution Range. Advances in Nutrition 8, 266. https://doi.org/10.3945/AN.116.013821.

4. Bauer, J., Biolo, G., Cederholm, T., Cesari, M., Cruz-Jentoft, A.J., Morley, J.E., Phillips, S., Sieber, C., Stehle, P., Teta, D., et al. (2013). Evidence-Based Recommendations for Optimal Dietary Protein Intake in Older People: A Position Paper From the PROT-AGE Study Group. J Am Med Dir Assoc 14, 542–559. https://doi.org/10.1016/J.JAMDA.2013.05.021.

5. Mittendorfer, B., Klein, S., and Fontana, L. (2019). A word of caution against excessive protein intake. Nature Reviews Endocrinology 2019 16:1 16, 59–66. https://doi.org/10.1038/s41574-019-0274-7.

6. Campbell, B., Kreider, R.B., Ziegenfuss, T., La Bounty, P., Roberts, M., Burke, D., Landis, J., Lopez, H., and Antonio, J. (2007). International Society of Sports Nutrition position stand: Protein and exercise. J Int Soc Sports Nutr 4, 8-. https://doi.org/10.1186/1550-2783-4-8

7. Lee, D.H., Rezende, L.F.M., Joh, H.K., Keum, N., Ferrari, G., Rey-Lopez, J.P., Rimm, E.B., Tabung, F.K., and Giovannucci, E.L. (2022). Long-Term Leisure-Time Physical Activity Intensity and All-Cause and Cause-Specific Mortality: A Prospective Cohort of US Adults. Circulation 146, 523–534. https://doi.org/10.1161/CIRCULATIONAHA.121.058162

8. Raafs, B.M., Karssemeijer, E.G.A., van der Horst, L., Aaronson, J.A., Olde Rikkert, M.G.M., and Kessels, R.P.C. (2020). Physical Exercise Training Improves Quality of Life in Healthy Older Adults: A Meta-Analysis. J Aging Phys Act 28, 81–93. https://doi.org/10.1123/JAPA.2018-0436.

9. Lenharo, M. (2025). How much protein do you really need? What the science says. Nature 647, 304–305. https://doi.org/10.1038/D41586-025-03632-1.

10. Topol, E. (2025). Our Preoccupation With Protein Intake. https://erictopol.substack.com/p/our-preoccupation-with-protein-intake.

11. Srikanthan, P., and Karlamangla, A.S. (2014). Muscle Mass Index as a Predictor of Longevity in Older-Adults. Am J Med 127, 547. https://doi.org/10.1016/J.AMJMED.2014.02.007.

12. Xia, L., Zhao, R., Wan, Q., Wu, Y., Zhou, Y., Wang, Y., Cui, Y., Shen, X., and Wu, X. (2020). Sarcopenia and adverse health‐related outcomes: An umbrella review of meta‐analyses of observational studies. Cancer Med 9, 7964. https://doi.org/10.1002/CAM4.3428.

13. Celis-Morales, C.A., Welsh, P., Lyall, D.M., Steell, L., Petermann, F., Anderson, J., Iliodromiti, S., Sillars, A., Graham, N., MacKay, D.F., et al. (2018). Associations of grip strength with cardiovascular, respiratory, and cancer outcomes and all cause mortality: prospective cohort study of half a million UK Biobank participants. BMJ 361. https://doi.org/10.1136/BMJ.K1651.

14. Pinto Pereira, S.M., Garfield, V., Farmaki, A.E., Tomlinson, D.J., Norris, T., Fatemifar, G., Denaxas, S., Finan, C., and Cooper, R. (2022). Adiposity and grip strength: a Mendelian randomisation study in UK Biobank. BMC Med 20, 201. https://doi.org/10.1186/S12916-022-02393-2.

15. Norris, T., Cooper, R., Garfield, V., Hamer, M., and Pinto Pereira, S.M. (2024). Unpicking Causal Relationships Between Grip Strength and Cardiorespiratory Fitness: A Bidirectional Mendelian Randomization Study. Scand J Med Sci Sports 34. https://doi.org/10.1111/SMS.14775.

16. Hart, P.D., and Buck, D.J. (2019). The effect of resistance training on health-related quality of life in older adults: Systematic review and meta-analysis. Health Promot Perspect 9, 1. https://doi.org/10.15171/HPP.2019.01.

17. Shailendra, P., Baldock, K.L., Li, L.S.K., Bennie, J.A., and Boyle, T. (2022). Resistance Training and Mortality Risk: A Systematic Review and Meta-Analysis. Am J Prev Med 63, 277–285. https://doi.org/10.1016/j.amepre.2022.03.020.

18. Zhang, X., Kapoor, D., Jeong, S.J., Fappi, A., Stitham, J., Shabrish, V., Sergin, I., Yousif, E., Rodriguez-Velez, A., Yeh, Y.S., et al. (2024). Identification of a leucine-mediated threshold effect governing macrophage mTOR signalling and cardiovascular risk. Nature Metabolism 2024 6:2 6, 359–377. https://doi.org/10.1038/s42255-024-00984-2.

19. Zhang, X., Sergin, I., Evans, T.D., Jeong, S.J., Rodriguez-Velez, A., Kapoor, D., Chen, S., Song, E., Holloway, K.B., Crowley, J.R., et al. (2020). High-protein diets increase cardiovascular risk by activating macrophage mTOR to suppress mitophagy. Nature Metabolism 2020 2:1 2, 110–125. https://doi.org/10.1038/s42255-019-0162-4.

20. Dickinson, J.M., Fry, C.S., Drummond, M.J., Gundermann, D.M., Walker, D.K., Glynn, E.L., Timmerman, K.L., Dhanani, S., Volpi, E., and Rasmussen, B.B. (2011). Mammalian Target of Rapamycin Complex 1 Activation Is Required for the Stimulation of Human Skeletal Muscle Protein Synthesis by Essential Amino Acids1–3. J Nutr 141, 856–862. https://doi.org/10.3945/JN.111.139485.

21. Song, Z., Moore, D.R., Hodson, N., Ward, C., Dent, J.R., O’Leary, M.F., Shaw, A.M., Hamilton, D.L., Sarkar, S., Gangloff, Y.G., et al. (2017). Resistance exercise initiates mechanistic target of rapamycin (mTOR) translocation and protein complex co-localisation in human skeletal muscle. Scientific Reports 2017 7:1 7, 5028-. https://doi.org/10.1038/s41598-017-05483-x.

22. Fappi, A., and Mittendorfer, B. (2020). Dietary protein intake and obesity-associated cardiometabolic function. Curr Opin Clin Nutr Metab Care 23, 380–386. https://doi.org/10.1097/MCO.0000000000000689.

23. Ang, T., Bruce, C.R., and Kowalski, G.M. (2019). Postprandial Aminogenic Insulin and Glucagon Secretion Can Stimulate Glucose Flux in Humans. Diabetes 68, 939–946. https://doi.org/10.2337/DB18-1138.

24. Yan, B., Su, X., Xu, B., Qiao, X., and Wang, L. (2018). Effect of diet protein restriction on progression of chronic kidney disease: A systematic review and meta-analysis. PLoS One 13. https://doi.org/10.1371/JOURNAL.PONE.0206134.

25. Devries, M.C., Sithamparapillai, A., Brimble, K.S., Banfield, L., Morton, R.W., and Phillips, S.M. (2018). Changes in Kidney Function Do Not Differ between Healthy Adults Consuming Higher- Compared with Lower- or Normal-Protein Diets: A Systematic Review and Meta-Analysis. Journal of Nutrition 148, 1760–1775. https://doi.org/10.1093/jn/nxy197.

26. Cheng, Y., Zheng, G., Song, Z., Zhang, G., Rao, X., and Zeng, T. (2024). Association between dietary protein intake and risk of chronic kidney disease: a systematic review and meta-analysis. Front Nutr 11. https://doi.org/10.3389/FNUT.2024.1408424.

27. Chen, Z., Glisic, M., Song, M., Aliahmad, H.A., Zhang, X., Moumdjian, A.C., Gonzalez-Jaramillo, V., van der Schaft, N., Bramer, W.M., Ikram, M.A., et al. (2020). Dietary protein intake and all-cause and cause-specific mortality: results from the Rotterdam Study and a meta-analysis of prospective cohort studies. Eur J Epidemiol 35, 411–429. https://doi.org/10.1007/S10654-020-00607-6.

28. Naghshi, S., Sadeghi, O., Willett, W.C., and Esmaillzadeh, A. (2020). Dietary intake of total, animal, and plant proteins and risk of all cause, cardiovascular, and cancer mortality: systematic review and dose-response meta-analysis of prospective cohort studies. BMJ 370, 2412. https://doi.org/10.1136/BMJ.M2412.

29. Wallerer, S., Papakonstantinou, T., Morze, J., Stadelmaier, J., Kiesswetter, E., Gorenflo, L., Barbaresko, J., Szczerba, E., Neuenschwander, M., Bell, W., et al. (2024). Association between substituting macronutrients and all-cause mortality: a network meta-analysis of prospective observational studies. EClinicalMedicine 75, 102807. https://doi.org/10.1016/j.eclinm.2024.102807.

30. Kurata, H., Meguro, S., Abe, Y., Sasaki, T., Asakura, K., Arai, Y., and Itoh, H. (2023). Dietary protein intake and all-cause mortality: results from The Kawasaki Aging and Wellbeing Project. BMC Geriatrics 2023 23:1 23, 479-. https://doi.org/10.1186/S12877-023-04173-W.

31. Dicken, S.J., Jassil, F.C., Brown, A., Kalis, M., Stanley, C., Ranson, C., Ruwona, T., Qamar, S., Buck, C., Mallik, R., et al. (2025). Ultraprocessed or minimally processed diets following healthy dietary guidelines on weight and cardiometabolic health: a randomized, crossover trial. Nat Med 31, 3297–3308. https://doi.org/10.1038/S41591-025-03842-0;SUBJMETA.

32. Rohrmann, S., Overvad, K., Bueno-de-Mesquita, H.B., Jakobsen, M.U., Egeberg, R., Tjønneland, A., Nailler, L., Boutron-Ruault, M.C., Clavel-Chapelon, F., Krogh, V., et al. (2013). Meat consumption and mortality - results from the European Prospective Investigation into Cancer and Nutrition. BMC Medicine 2013 11:1 11, 63-. https://doi.org/10.1186/1741-7015-11-63.

33. Kaluza, J., Akesson, A., and Wolk, A. (2014). Processed and unprocessed red meat consumption and risk of heart failure prospective study of men. Circ Heart Fail 7, 552–557. https://doi.org/10.1161/CIRCHEARTFAILURE.113.000921

34. Ramezani, F., Pourghazi, F., Eslami, M., Gholami, M., Mohammadian Khonsari, N., Ejtahed, H.S., Larijani, B., and Qorbani, M. (2024). Dietary fiber intake and all-cause and cause-specific mortality: An updated systematic review and meta-analysis of prospective cohort studies. Clinical Nutrition 43, 65–83. https://doi.org/10.1016/j.clnu.2023.11.005.

35. Liao, C. De, Huang, S.W., Chen, H.C., Huang, M.H., Liou, T.H., and Lin, C.L. (2024). Comparative Efficacy of Different Protein Supplements on Muscle Mass, Strength, and Physical Indices of Sarcopenia among Community-Dwelling, Hospitalized or Institutionalized Older Adults Undergoing Resistance Training: A Network Meta-Analysis of Randomi…. Nutrients 16, 941. https://doi.org/10.3390/NU16070941.

36. Reid-Mccann, R.J., Brennan, S.F., Ward, N.A., Logan, D., McKinley, M.C., and McEvoy, C.T. (2025). Effect of Plant Versus Animal Protein on Muscle Mass, Strength, Physical Performance, and Sarcopenia: A Systematic Review and Meta-analysis of Randomized Controlled Trials. Nutr Rev 83, e1581–e1603. https://doi.org/10.1093/NUTRIT/NUAE200.

37. Ardisson Korat, A. V., Shea, M.K., Jacques, P.F., Sebastiani, P., Wang, M., Eliassen, A.H., Willett, W.C., and Sun, Q. (2024). Dietary protein intake in midlife in relation to healthy aging – results from the prospective Nurses’ Health Study cohort. Am J Clin Nutr 119, 271–282. https://doi.org/10.1016/J.AJCNUT.2023.11.010.

38. Dhillon, J., Craig, B.A., Leidy, H.J., Amankwaah, A.F., Osei-Boadi Anguah, K., Jacobs, A., Jones, B.L., Jones, J.B., Keeler, C.L., Keller, C.E.M., et al. (2016). The Effects of Increased Protein Intake on Fullness: A Meta-Analysis and Its Limitations. J Acad Nutr Diet 116, 968–983. https://doi.org/10.1016/j.jand.2016.01.003.

39. Kohanmoo, A., Faghih, S., and Akhlaghi, M. (2020). Effect of short- and long-term protein consumption on appetite and appetite-regulating gastrointestinal hormones, a systematic review and meta-analysis of randomized controlled trials. Physiol Behav 226. https://doi.org/10.1016/j.physbeh.2020.113123.

40. Kokura, Y., Ueshima, J., Saino, Y., and Keisuke, M. (2024). Enhanced protein intake on maintaining muscle mass, strength, and physical function in adults with overweight/obesity: A systematic review and meta-analysis. Clin Nutr ESPEN 63, 417–426. https://doi.org/10.1016/J.CLNESP.2024.06.030.

41. Nunes, E.A., Colenso-Semple, L., McKellar, S.R., Yau, T., Ali, M.U., Fitzpatrick-Lewis, D., Sherifali, D., Gaudichon, C., Tomé, D., Atherton, P.J., et al. (2022). Systematic review and meta‐analysis of protein intake to support muscle mass and function in healthy adults. J Cachexia Sarcopenia Muscle 13, 795. https://doi.org/10.1002/JCSM.12922.

42. Tagawa, R., Watanabe, D., Ito, K., Ueda, K., Nakayama, K., Sanbongi, C., and Miyachi, M. (2021). Dose–response relationship between protein intake and muscle mass increase: a systematic review and meta-analysis of randomized controlled trials. Nutr Rev 79, 66–75. https://doi.org/10.1093/NUTRIT/NUAA104.

43. Liao, C. De, Tsauo, J.Y., Wu, Y.T., Cheng, C.P., Chen, H.C., Huang, Y.C., Chen, H.C., and Liou, T.H. (2017). Effects of protein supplementation combined with resistance exercise on body composition and physical function in older adults: a systematic review and meta-analysis. Am J Clin Nutr 106, 1078–1091. https://doi.org/10.3945/AJCN.116.143594.

44. Morton, R.W., Murphy, K.T., Mckellar, S.R., Schoenfeld, B.J., Henselmans, M., Helms, E., Aragon, A.A., Devries, M.C., Banfield, L., Krieger, J.W., et al. (2018). A systematic review, meta-analysis and meta-regression of the effect of protein supplementation on resistance training-induced gains in muscle mass and strength in healthy adults. Br J Sports Med 52, 376–384. https://doi.org/10.1136/BJSPORTS-2017-097608.

45. Refalo, M.C., Trexler, E.T., and Helms, E.R. (2025). Effect of Dietary Protein on Fat-Free Mass in Energy Restricted, Resistance-Trained Individuals: An Updated Systematic Review with Meta-Regression. Strength Cond J. https://doi.org/10.1519/SSC.0000000000000888.

46. Ten Haaf, D.S.M., Nuijten, M.A.H., Maessen, M.F.H., Horstman, A.M.H., Eijsvogels, T.M.H., and Hopman, M.T.E. (2018). Effects of protein supplementation on lean body mass, muscle strength, and physical performance in nonfrail community-dwelling older adults: a systematic review and meta-analysis. Am J Clin Nutr 108, 1043–1059. https://doi.org/10.1093/AJCN/NQY192.

47. Björkman, M.P., Pilvi, T.K., Kekkonen, R.A., Korpela, R., and Tilvis, R.S. (2011). Similar effects of leucine rich and regular dairy products on muscle mass and functions of older polymyalgia rheumatica patients: A randomized crossover trial. J Nutr Health Aging 15, 462–467. https://doi.org/10.1007/S12603-010-0276-6.

48. Leenders, M., Verdijk, L.B., Van Der Hoeven, L., Van Kranenburg, J., Nilwik, R., Wodzig, W.K.W.H., Senden, J.M.G., Keizer, H.A., and Van Loon, L.J.C. (2013). Protein Supplementation during Resistance-Type Exercise Training in the Elderly. Med. Sci. Sports Exerc 45, 542–552. https://doi.org/10.1249/MSS.0b013e318272fcdb.

49. Jäger, R., Kerksick, C.M., Campbell, B.I., Cribb, P.J., Wells, S.D., Skwiat, T.M., Purpura, M., Ziegenfuss, T.N., Ferrando, A.A., Arent, S.M., et al. (2017). International Society of Sports Nutrition Position Stand: protein and exercise. J Int Soc Sports Nutr 14, 20. https://doi.org/10.1186/S12970-017-0177-8.

50. Katsanos, C.S., Kobayashi, H., Sheffield-Moore, M., Aarsland, A., and Wolfe, R.R. (2006). A high proportion of leucine is required for optimal stimulation of the rate of muscle protein synthesis by essential amino acids in the elderly. Am J Physiol Endocrinol Metab 291, 381–387. https://doi.org/10.1152/AJPENDO.00488.2005.

51. Moore, D.R., Churchward-Venne, T.A., Witard, O., Breen, L., Burd, N.A., Tipton, K.D., and Phillips, S.M. (2015). Protein Ingestion to Stimulate Myofibrillar Protein Synthesis Requires Greater Relative Protein Intakes in Healthy Older Versus Younger Men. The Journals of Gerontology: Series A 70, 57–62. https://doi.org/10.1093/GERONA/GLU103.

52. Hofmann, M., Schober-Halper, B., Oesen, S., Franzke, B., Tschan, H., Bachl, N., Strasser, E.M., Quittan, M., Wagner, K.H., and Wessner, B. (2016). Effects of elastic band resistance training and nutritional supplementation on muscle quality and circulating muscle growth and degradation factors of institutionalized elderly women: the Vienna Active Ageing Study (VAAS). European Journal of Applied Physiology 2016 116:5 116, 885–897. https://doi.org/10.1007/S00421-016-3344-8.

53. Dzator, S., Weerasekara, I., Shields, M., Haslam, R., and James, D. (2023). Agreement Between Dual-Energy X-ray Absorptiometry and Bioelectric Impedance Analysis for Assessing Body Composition in Athletes: A Systematic Review and Meta-Analysis. Clin J Sport Med 33, 557–568. https://doi.org/10.1097/JSM.0000000000001136.

54. Feng, Q., Bešević, J., Conroy, M., Omiyale, W., Lacey, B., and Allen, N. (2024). Comparison of body composition measures assessed by bioelectrical impedance analysis versus dual-energy X-ray absorptiometry in the United Kingdom Biobank. Clin Nutr ESPEN 63, 214–225. https://doi.org/10.1016/J.CLNESP.2024.06.040.

55. Manole, L.M., Ghiga, G., Iftinchi, O., Boca, L.O., Donos, M.A., Țarcă, E., Ionuț, N., Revenco, N., Margasoiu, I., and Trandafir, L.M. (2025). Bioelectrical Impedance Analysis Versus Dual X-Ray Absorptiometry for Obesity Assessment in Pediatric Populations: A Systematic Review. Diagnostics 15, 1505. https://doi.org/10.3390/DIAGNOSTICS15121505.

56. Paddon-Jones, D., Campbell, W.W., Jacques, P.F., Kritchevsky, S.B., Moore, L.L., Rodriguez, N.R., and Van Loon, L.J.C. (2015). Protein and healthy aging. Am J Clin Nutr 101, 1339S-1345S. https://doi.org/10.3945/AJCN.114.084061.

57. Paddon-Jones, D., and Rasmussen, B.B. (2009). Dietary protein recommendations and the prevention of sarcopenia. Curr Opin Clin Nutr Metab Care 12, 86–90. https://doi.org/10.1097/MCO.0B013E32831CEF8B.

Regardless of where they draw lines in the sand, the important thing to note is that vegans are just better people. Morally superior, I mean. Vegans won’t say this—in fact they make a point of not saying/thinking this because doing so belies said moral superiority—but vegan smugness is difficult to fully conceal, and most can whiff their sententiousness from a mile away.

I quite like smugness and arrogance if it is justified—I find it honest. So the question of importance seems to be whether vegans actually are morally superior to non-vegans. Is their highfalutin worldview deserving of praise and proselytisation, or are the moral fundaments of veganism perhaps less selfless than at first glance?

Here I will break down the psychology and morality of veganism, focusing particularly on “ethical veganism” which is concerned with animal welfare (instead of health or the environment, say). I have three arguments to make:

1. Ethical veganism rests on the presupposition that animals are “conscious” or “sentient” (whatever that means) and that they feel pain. This is false.

2. Ethical veganism requires a utilitarian ethical philosophy, which doesn’t map well onto actual morality as it evolved in humans.

3. When observing ethical veganism meta-ethically or through an Adlerian psychological lens, veganism emerges as a purely selfish exercise.

Pt I: No, animals don’t “feel pain”

EDIT: this claim seems to be unproductively provocative for some readers. The point is to highlight that pain can mean anything from “evasive signal transduction” to “complex emotional distress”. The word was intended to be applied to humans, but we typically use it to describe the lesser pain of animals. Vegans must contend with the fact that plants or bacteria also can sense things and exhibit “pain” if we continue to stretch the definition. In it’s truest form, only humans “feel pain”, and when we see an animal suffering, we imagine that it is experiencing human-like pain, when it is not.

What is pain?

Merriam-Webster gives us two definitions for pain:

“a localized or generalized unpleasant bodily sensation or complex of sensations that causes mild to severe physical discomfort and emotional distress and typically results from bodily disorder (such as injury or disease)”

and

“a basic bodily sensation that is induced by a noxious stimulus, is received by naked nerve endings, is associated with actual or potential tissue damage, is characterized by physical discomfort (such as pricking, throbbing, or aching), and typically leads to evasive action”

These may seem like two ways of describing the same thing, but these definitions couldn’t be more different.

The first definition is a subjective one—focusing on sensation and emotion. This is what most of us associate with pain in our everyday lives. We recall stubbing our toe and think of the negative emotional milieu that accompanied the event. Meanwhile, the second definition is more scientific and objective. It correctly identifies the aetiology of pain as a noxious stimulus and the mechanism of pain involving nociceptors and neural networks. It adds that pain is typically followed by evasive action.

While the second definition is a useful description of how pain works, it isn’t describing what pain actually is. That would be the first definition—an emotionally distressing negative sensation.

This is a key distinction because, while we can say for certain that animals have nociceptors and neural pathways that are somewhat akin to our pain-associated neural pathways, and we can see animals taking evasive action in response to noxious stimuli, we cannot claim at all that animals experience emotional distress or a negative bodily sensation akin to humans. The neural pathways for nociception are evolutionarily ancient, and therefore shared with many animals, but the neural pathways humans use for complex emotional processing, self-awareness, and creating the rich experience of pain evolved much more recently and are completely unique to us.

Correlation-causation fallacy

We know that pain is subjective because humans—being the highly intelligent and conscious beings that we are—can simply choose not to experience pain when receiving a given noxious stimulus. Next time you stub your toe, instead of scrunching up your face, grabbing your foot, or whimpering like an animal, try to meditate on the feeling without making any outward reaction. You’ll notice that the physical sensation is a separate entity to the physical reaction, and these are both separate still from the emotional environment we create in our minds. This ability to parse material stimulus from emotional response is a central tenet of stoicism and buddhism.

Humans can receive noxious stimuli and not experience emotional distress, but we can also create emotional distress from a complete absence of noxious stimuli. Babies often do this once they learn that crying is rewarded by attention from their parents. They’ll simply cry for no apparent aetiological reason at all (although the teleological reason is to garner attention).

Once we acknowledge that the human experience of pain requires two distinct components—nociception and complex emotional processing—we realise that erroneously asserting that animals experience pain (by which we mean and infer “human-like pain”) is falling for the correlation-causation fallacy.

To illustrate this, I’ll borrow a schematic from “The empathy problem”.

“For a human, the x-axis is the input and the y-axis the output. Yet when we sympathise with animals in pain, we use the y-axis as the input and attempt to infer cause from correlation. This is logically fallacious.”

When a vegan sees a fat chicken in a cage, they imagine what it would be like for them to be stuffed and not allowed to roam around, and they conclude that this would be distressing for them and is therefore wrong. But a chicken is not a human. It is incorrect to assert that a chicken considers such a situation to be distressing. Chickens don’t “consider” anything. And “distressing” is a term that only means something for humans. To say that something is distressing for a chicken is just a category error. Chickens don’t think like humans do at all.

This is why René Descartes, of “cogito, ergo sum” fame, also asserted that animals cannot feel pain.

Fake pain science

Humans developed emotive language to describe uniquely human experiences. Yet many now have a habit of aberrantly using these human-specific terms to describe non-human entities—from animals to LLMs. This tendency for anthropomorphism has infected our culture so deeply that even our science is marred by it.

Scientists for decades have been desperately trying to prove that animals feel pain. The problem is, until we can fully explain human consciousness, we are wholly unequipped to assess subjective experiences via objective measurements.

Despite this epistemological hurdle, scientists have pounced onwards looking for physiological or behavioural phenomena that can validate their preconceived conclusion that animals feel pain. As you might imagine with such a biased starting point, the data are extremely dodgy.

Take this 2014 paper, which tried to find evidence that Atlantic cod can feel pain. The authors subjected the lips of wild cod to a range of noxious stimuli and monitored various physiological and behavioural responses. The noxious stimuli were acetic acid injection, capsaicin injection, or fish hook puncture, with saline injection as a control.

When looking at a purely physiological response—the operculum beat rate (fish equivalent of breathing rate)—they found precisely zero effect of their noxious stimuli compared to control. Yet this didn’t stop them writing as if there was an effect!1

They also measured the blood concentration of various metabolites and—again—found zero effect of the noxious stimuli. But don’t worry, they can sweep that under the rug as being “possibly due to time of sampling” while insisting that a statistically non-significant trend in OBR implies fish pain!

The one place they did find significant effects was in behavioural responses. Despite measuring five different responses, they only found an effect of the noxious stimuli in two of the responses. It is worth noting that all of these measurements require substantial human estimation and are therefore ripe for bias.

Interestingly, they found no effect of fish hooks on any fish response, which they argue might be because Atlantic cod have numb lips due to their tough diets (numb to only one stimulus?!).2

They did find an effect of acetic acid and capsaicin injection at the highest concentration. Huzzah! Fish pain! Only one problem. A year earlier, in 2013, researchers characterised the fish TRPV1 receptor, which is responsible for responding to capsaicin, and found that fish actually have mutations in key residues that render their TRPV1 receptors completely unable to respond to capsaicin.3 The authors of the 2014 paper simply missed this.

In other words, it is biochemically impossible for fish to detect capsaicin. This means that the behavioural response to capsaicin shown in the 2014 paper, whose authors include famed fish pain expert Lynne Sneddon, were completely fictitious.

This peer-reviewed paper hasn’t been retracted or anything, don’t be silly. Many other papers in this field suffer similar bias, and yet fish pain supporters don’t skip a beat in citing such papers and telling people how strong the “scientific consensus” is that fish feel pain.

There is no evidence nor any logical reason to believe that animals experience pain in any way similar to how humans experience it.4

A mightily convenient line in the sand

Even accepting that animal “pain” is really not like human pain at all, one might still be tempted to defend an animal’s life regardless of pain. Indeed, many vegans do hold this view. They believe in some “right” to life (although most vegans do not believe in God, so where exactly this right comes from is unclear). The obvious problem with inventing rights here is how do you decide who gets the right to life or not?

It is definitionally true that every extant organism is equally as evolutionarily successful and exhibits an equal “desire” for life as every other extant organism. That the bacterium that causes gonorrhoea fills a different ecological niche to antarctic orcas doesn’t change this fact. Both species have succeeded through the same length of evolutionary time. Both have adapted to their environments equally well (this is a binary—you either go extinct or you do not). Who are we as humans to assign the right to life to only a small subset of these organisms?

Vegans lean on the word “sentience” to fill in for an argument here, but nobody can really define what that means. Merriam-Webster defines sentient as “capable of sensing or feeling”, “aware”, or “finely sensitive in perception or feeling”. If you paid attention in biology class at school, you’ll recall that the S in MRS GREN stands for sensitivity, meaning that every living organism is definitionally sentient.

Indeed, scientists have uncovered many of the impressive ways in which plants display sensitivity, damage sensing, and communication. For example, plants release a wide range of volatile organic compounds in response to being eaten by insects or herbivores. These chemicals can signal to other plants to raise their defences ahead of time, or even alter herbivore eating patterns away from precious flowers and towards leaves that are more dispensable for the plant. Even the manner in which plants signal damage is remarkably similar to the way we do—using waves of membrane depolarisation to transmit electrical signals, which is exactly the mechanism by which neurons fire in us. Below is time-lapse video of plant pain in action from Aratani et al., 2023.

Most strikingly, a couple years ago researchers uncovered that plants literally scream when dehydrated or cut. The scientists suggest that farmers in the future could feasibly use microphones to pick up these gut-wrenching plant yelps and trigger irrigation systems to water the plants when they scream of thirst. Or perhaps pest infestations could be caught early just by listening to the desperate cries of pain from plants being eaten alive.

If you’re going to say a fruit fly can feel pain, why not a plant? Clearly, there is no real line to be drawn between plants and animals here in terms of what “deserves” to live or die. The reason this line in the sand is arbitrarily drawn as such is because otherwise veganism would very quickly go extinct and be obviously outed as the most moronically suicidal cult in human history. It is a necessary pausing of logic to keep their movement alive—literally.

Pt II: Yes, you should kill 10^100 shrimp if it makes you smile

Shrimp simps

A bunch of quite influential effective altruists seem to have recently become obsessed with shrimp welfare. Matthew Adelstein, aka , for example, has been writing for a while now about the moral imperative to donate to charities that implement shrimp stunning. The stunning is supposed to anaesthetize the crustaceans prior to them being killed. Because so many shrimp are killed for human consumption, utilitarians and effective altruists—who think that morality is some kind of arithmetic exercise—insist that paying for these stunners is just about the best thing you can do with your money.

More insistent still was a recent popular essay by titled “Yes, you should save 10^100 shrimp instead of one human”.

I’ve already shown how the presupposition here—that animals “feel pain”—is false. To appreciate the role that anthropomorphisation plays in building such a worldview, check out this emotive description of shrimp death from Adelstein:

“Imagine them struggling, gasping, without enough air, fighting for their lives, but it’s no use.”

The issue here is that susceptible readers of the vegan persuasion won’t imagine a shrimp here at all. They imagine a human struggling, gasping etc. Indeed, the author himself betrays that he is doing this because gasping requires lungs and struggling for air would be the opposite of what a shrimp—currently asphyxiating because it is in air—would require. It is impossible for a human to imagine being a shrimp, so instead we anthropomorphise. (I hope readers appreciate my deliberate attempts to do the same for plants above—two can play that game.)

It is anthropomorphisation that allows vegan utilitarians to assign a numerical moral value to shrimp. They argue that shrimp are at least a bit like humans, so let’s give a shrimp life 0.00001x the value of a human life. Well, if that’s true, then you just need more than a million shrimp until you surpass the moral worth of a human life. So the morally correct thing would be to kill one human if it meant you could save 2 million shrimp. Hah. Logic!

Obviously this isn’t quite right. To understand why, we need to dig a little deeper into morality, ethics, and meta-ethics.

What is morality?

Morality is a complex social sense that aids social decision-making in a way that directly or indirectly benefits ourselves and our close kin with whom we share a substantial amount of genetic material. Morality has only evolved in humans. Ipso facto it evolved for humans. Any application of morality that goes against this is therefore a misfiring and a perversion of morality. As such, contrary to the “rational” claims of the shrimp simps, moral worth isn’t awarded on a gradual scale proportional to how many neurons you have in your brain. Morality is an exclusively human trait and moral worth only applies to moral beings—humans and only humans.

Normative ethics vs morality

Ignoring the fact that the whole vegan argument is debased by the fact that animals don’t really feel pain at all in the sense that we do (despite our penchant for anthropomorphisation) and that the correct moral value for any non-human entity is zero, the utilitarian urge to calculate morality arithmetically is also just not at all how morality works.5

To try to describe how morality works or should work, philosophers have constructed various ethical frameworks. Consequentialism, which is the ethical framework that utilitarianism sits under, judges moral actions entirely by their consequences. You can see why this simple Excel-spreadsheet ethical worldview might be attractive to some—especially those who pride themselves on their rationalism—but it doesn’t map onto the ground truth of actual evolved morality well at all.

This mismatch is easy to demonstrate with a few examples. According to the rational utilitarian, it is not at all morally repugnant for a man to purchase a chicken carcass from his local supermarket to take home and have sex with. Necrophilia is also fine. Incest that doesn’t result in pregnancy? All good. No qualms with child pornography either if the child isn’t harmed. In fact, not only are these things not morally repugnant, but they are actually morally desirable to the utilitarian because nobody is suffering but somebody is gaining pleasure. Pleasure maximisation and suffering minimisation are the only concerns of utilitarianism.

I’d hope that most honest readers can admit that the chicken-fucker or the necrophile are obviously not paragons of morality. This is because morality is a complex emotional sense that evolved under selective pressures to do more than just maximise pleasure and minimise suffering. The single axis of morality that utilitarians focus on (the “care-harm foundation”) is only about one sixth of the story according to moral psychologist . But utilitarians ignore their actual evolved sense of morality to instil their sterile, inhuman, calculated moral philosophy instead.

Utilitarianism also happens to be the ethical framework that has given rise to most of the human atrocities in the last 100 years or so, as I have written about recently.

While utilitarianism focuses on consequences, the other normative ethical frameworks focus on rules (deontology) or character (virtue ethics). I think the latter most closely models how morality actually works—by morally judging actions based on the virtues they reflect in the person, not the material consequences of the actions.

Once you stop massively overestimating the pain of animals, then holding onto a utilitarian ethical worldview leaves you in a sticky situation. If the goal is to maximise pleasure and minimize pain, then all it takes is a human who takes a lot of pleasure in harming relatively distant animals for the calculus to flip in favour of death. The utilitarian must believe that you should kill 10^100 shrimp if it makes the killer smile sufficiently strongly. That’s a little bump in pleasure being added to the universe with little increase in suffering. Or, since vegans round the pain of plants down to zero for practical but illogical reasons, killing 10^100 oak trees is highly moral if the tree killer enjoys doing so. This obviously sounds horrific and immoral, so the utilitarian tries to fix it by pretending that animals do feel a lot of pain, or by minimising the pleasure that humans get from food.6 But the solution here is to throw out utilitarianism altogether since it clearly doesn’t work.

The virtue ethicist, on the other hand, abstains from kicking a dog on the street not because of the interests of the dog, but because of the cruelty it reflects in the human. The human is the only moral entity in this equation, and the corruption of the moral character is what ought to be avoided here. This is why, even though morality didn’t evolve to protect animals, the virtue ethicist does not abuse animals willy nilly.

We can see how human moral character is a better ethical heuristic than consequences of actions by considering that a lion hunting and ripping apart a zebra limb from limb is not considered unethical by most sane people (there are some anti-predation vegans out there … seriously), yet if a human shoots a zebra for fun—a much faster and more humane death—most morally judge the human. From the anthropomorphised zebra’s perspective here, clearly the bullet to the head is preferable. So why do we consider shooting a zebra to be immoral but we don’t judge a lion for hunting for its prey? Because shooting the zebra for fun exhibits moral vice—something the amoral lion cannot exhibit.

But if killing an animal is immoral according to virtue ethics, which I seem to be so keen on, then why isn’t veganism morally virtuous in the same vein?

Pt III: Selfish veganism

Teleological causation

Psychologist Alfred Adler was a contemporary of now-disgraced and debunked, but much more famous, psychologist Sigmund Freud. While Freud pushed the idea of aetiology—explaining behaviours through psychological cause and effect all the way back to childhood—Adler disagreed. Adler, drawing on Aristotelian thought, thought that a teleological framework better encapsulates human psychology, motivations, and behaviours. Teleology involves analysing the purpose of things. It asks what the end goal of an action is and posits this to be its cause. Earlier I mentioned the baby crying not because it is traumatised, but because it desires attention and it knows that crying garners said attention. What does veganism look like through an Adlerian lens?

Who is veganism serving? Well, first and foremost, it serves vegans. Since we are both a moral species and a social species, projecting your own moral goodness to others is a very rewarding thing for humans to do. Not only do we have in-built neurological mechanisms that psychologically reward moral actions, but society has powerful social mechanisms that bolster these psychological rewards further.

Giving to charity feels good. Knowing that someone else knows that you have given to charity feels even better.

The same applies to veganism. The potent psychological reward gained by (1) persuading yourself that you are doing something morally good and (2) letting others know of your moral righteousness (either actively or—even better—passively) are the primary motivations for veganism and utilitarianism more broadly. This is most obviously true when the promised utilitarian benefit doesn’t even accompany the moral posturing, as is the case with most vegan consumer decisions. A small fraction of the population choosing to buy a mushroom burger at the supermarket does nothing to prevent the meat aisle reliably selling all of its stock every week. That no cows are saved here really doesn’t matter though, because subconsciously it is the feeling of having done something good that the vegan is after. Importantly, the fact that the vegan is not cognizant of this emotional reward driving their actions doesn’t affect the veracity of the phenomenon.

Selfish sympathy

Why don’t we like watching Gazans being bombed? It’s actually not directly because of the pain that the Gazans might be feeling. It’s because of the pain that seeing the video footage inflicts on us. This pain exists because of a genius evolutionary chess move that promotes altruism through purely selfish means in humans—sympathy.

Sympathy (not to be confused with empathy!) means that humans literally feel pain when they perceive someone or something else to feel pain too. Sympathy has the effect of tying the emotional fate of ourselves to others, so we work to protect others only because it reduces our own sympathy-induced suffering. It is important to realise that, because of sympathy, the shrimp simp is primarily alleviating their own pain and discomfort when they donate to shrimp charities. Likewise, the vegan abstaining from meat alleviates their own pain and discomfort when choosing not to eat steak. In this way, veganism—but really any utilitarian cause—is motivated purely selfishly. Selfishness dressed up as altruism—because altruism is self-serving at both the social and psychological levels.

Note: I’m not saying that this makes altruism bad. Au contraire. I’m merely pointing out that altruism has mutually beneficial outcomes for the benefactor and the recipient. That’s how it could evolve. Pretending that this is not the case, however, amounts to dishonesty, which actually is immoral.

The issue with sympathy is that it is extremely promiscuous. That we feel sympathy for animals is no indication that animals are actually feeling pain (although most of us assume that they do). We also feel sympathy for inanimate or non-biological objects like lonely bananas in a supermarket or robots being kicked by YouTubers. I have no doubt that in twenty years time we’ll have utilitarians writing essays on how its really important to defend robot “rights” because there is a tiny chance they feel pain. Some people even insist of using “please” and “thank you” when using ChatGPT … just in case. Heck, maybe the video I embedded earlier of plant calcium signalling in response to herbivory-induced volatile compounds is enough to trigger the sympathies of some readers (but what if plants actually do feel pain? We can’t rule it out for sure, right?!).

Virtue or vice?

According to Aristotelian virtue ethics, it is the motivations of the moral being that dictate the morality of an action. I’ve highlighted how the motivations for veganism are at their root selfish. The feeling of moral superiority is a good one after all. That evolution loaded selfish incentives into altruistic and moral behaviours hardly makes these things immoral though—that would be a ridiculous moral paradox. Is there a way to be vegan without being Pecksniffian then? Is sufficiently downplaying one’s moral superiority enough to turn veganism into something genuinely virtuous?

To answer this we can again turn to Aristotle, who noticed something fascinating about virtue: the golden mean.

The golden mean refers to the fact that all virtues sit between two vices at either extreme. Courage sits between recklessness and cowardice. Friendliness sits between being a people-pleaser and being rude. This is an almost frustratingly true observation from Aristotle. Anything done to excess—even if that thing is literally being morally good—becomes a vice.

Crucially for vegans: virtuous compassion is the golden mean that sits between cruelty and excessive sympathy. Most morally calibrated people intrinsically understand that unfounded sympathy for all animals is excessive and is therefore vicious.

Ultimately, the contorted pseudo-rationalism of the utilitarian, which abruptly stops at the plant’s right to life, is inferior to the raw evolved moral sense of the everyday person. Common sense trumps inaccurate ethical philosophy, no matter how deeply carved the philosophy is.

The proof is in the (non-vegan) pudding

While I am sure vegans and utilitarians will disagree with my arguments here, I actually don’t need to persuade them. My philosophical ethical framework is much more closely aligned to ground-truth human morality, while utilitarians mistake the map for the terrain by supplanting actual morality with an inadequate ethical philosophy. Proof of this can perhaps be seen in the success of the veganism movement at large.

Veganism is a heavily bootstrapped movement, meaning big organisations spend a lot of money trying to evangelise for their neo-religion. Around 2% of Brits and 1% of Americans are vegan, and this percentage hasn’t really shifted upwards at all over the past 6 years or so. This is despite ever-growing efforts to convert more omnivores by non-profits like The Vegan Society, which now spends over £4 million per year.

In fact, after a nadir in omnivorous Brits mid-2021 (likely a product of desperate experimentalism during COVID lockdowns), the proportion of the UK that eats meat has risen year-on-year and is now at its highest point since polling began—at 74%.

Even Veganuary—the flagship event for vegan evangelists—saw a ~20% drop in participants between 2021 and 2024.

Clearly, the failure of the movement isn’t due to lack of trying. I think most common sensical people get a hunch for the logical fallacies that lace the vegan worldview even without being able to pinpoint or articulate them precisely. You don’t have to pin down the amino acid sequence of the Atlantic salmon TRPV1 receptor, nor trudge your way through a copy of The Nichomachean Ethics to know that the arguments for veganism are simply not persuasive. If good people aren’t persuaded of your moral worldview, its probably not because they’re stupid or evil. It’s probably because your moral worldview just isn’t any good.

TL;DR

• Academic publishing is broken

• Current “fixes” are partial or open new cans of worms

• I propose integrating key features of non-profit open access journals, preprint servers, and social media platforms to create an optimised publishing system right here on Substack, which I dub Substack Scholar

• I lay out a concrete plan of implementation

In part 1 of “Fixing academic publishing”, I laid out how a scientist might publish their results in an academic journal. Along the way I pointed out many pitfalls of the system. Boiling it down, here are the ten most pressing issues as I see it:

1. slowness

2. lack of motivation or reward for reviewers (with implications for “slowness” and review quality)

3. biased peer review stemming from the power of authors to suggest or reject certain reviewers

4. narrowness of scope of review leading to shallow or poor quality review. Given there are only 2-3 reviewers for most papers, it is unlikely that the editor will have actually found the best people to review a particular paper—especially in smaller journals)

5. politicisation and corruption of science via journal editors

6. increases in article length and incoherence in part due to input of reviewers

7. huge cost of publishing, typically funnelling taxpayer money or philanthropic donations to for-profit journals

8. lack of public access to the research

9. an inability to keep up with the pace and quantity of science

10. the misallocation of papers: bad papers can make it into big journals and be artificially elevated beyond their merit; good papers can be shunned into small journals where they go unread and unappreciated.

I also mentioned in part 1 that there have been some attempts to fix these issues. The problem is, all solutions thus far only solve a subset of issues, or introduce new issues at the same time as solving the original issues. Let’s take a quick look at them before I offer my own—perhaps we can glean something useful from these attempts.

Non-profit open access journals

One of the earliest attempts to offer life scientists new systems for publishing was PLOS (Public Library of Science). Harold Varmus (a Nobel Prize winner and previous NIH director), Patrick O. Brown, and Michael Eisen started the movement in 2000, later launching their first actual journal (PLOS Biology) in 2003. The organisation now runs fourteen different journals and has inspired a swathe of like-minded non-profit journals.

The PLOS model solves several issues with publishing. The main goal was to pioneer open access publishing, which they achieved. The organisation is a non-profit, meaning they are not motivated to erect paywalls wherever they can, and viewing the research in any of their journals is free to everyone.

Several PLOS journals also instruct reviewers to assess papers only on the basis of scientific validity, and not impact or importance. This improves speed and reduces publication bias. This move, though, introduces a new problem in that it now passes the buck to the readers, who will have to browse endlessly to find interesting articles.

Sadly, PLOS, being a big non-profit that employs a large staff, still has to charge fees in the thousands of dollars for publishing (even though there is no surcharge for open access since this is the default). Speed of publishing is still unideal, with PLOS One boasting a 43-day average time until first response and around 4-5 months for the whole process.

Most disqualifying, though, PLOS just doesn’t attract the best research articles. There is very little prestige to be gained by publishing here.

A hypothetical optimal publishing system would have open access to its research, like these non-profit journals, but it would have to offer more prestige to authors and reduce costs to zero.

Preprint servers

Preprint repositories are becoming increasingly popular these days. Arχiv, which was founded in the early 90s, is *the* hub for maths, physics, and computer science papers. More recently, bioRχiv (in 2013) and medRχiv (in 2019) were founded for biological and medical papers respectively.

The emergence of the preprint solves several issues in academic publishing. Most obviously, it speeds it up tremendously. Once you submit an article it is quickly vetted by a moderator and then uploaded, typically within 24 hours. The lack of formal peer review and powerful editors also removes publication bias and allows the public to see the paper exactly how the authors intended it to be.

For example, bioRχiv enabled Colossal Biosciences to publish a preprint exactly when they needed to in order to help defend their claim that their genetically modified wolves can be feasibly categorised as dire wolves during a media onslaught, as I wrote about earlier this year.

Crucially for computer science, these repositories handle *a lot* of volume, with arχiv now receiving more than 20,000 submissions per month. Indeed, the volume of research is all but crippling regular peer review in subfields like machine learning.1

While handling volume is a benefit of arχiv and its biological equivalents, enabled by the fact that they don’t do any peer review, this also means that there is no real quality control mechanism or method for finding the most interesting papers. The websites are labyrinths. Good, high-impact work from a small lab gets buried.

Crucially, while use of preprint servers is growing, they are not replacing regular publishing, but just adding to it. Most arχiv papers still seek peer review at a big conference or journal (the researchers need the boost to their CVs!). That Colossal Biosciences bioRχiv preprint that I mentioned is probably currently wallowing in the mire of peer review at Cell, Science, or Nature if I had to guess.

These preprint servers, like the open-access journals, have failed to overcome the prestige problem, and for that reason they don’t offer a complete alternative to traditional journals. A preprint is great for getting the work out there fast, but the job is not done until a big prestigious journal slaps their name on your work for a few thousand dollars.

A hypothetical optimal publishing system would be rapid, like preprint servers, but it would be the finish line, not a pit stop on the way to a prestigious journal.

Honourable mentions

We have now covered the two main alternatives (or supplements) to traditional academic publishing: preprint servers and non-profit journals. There are a couple other specific examples that deserve an honourable mention.

eLife

eLife, a non-profit biology journal founded in 2012, is currently trying a radical fusion of non-profit OA journals and preprint repositories which they call the “Reviewed Preprint” model.

In 2023, eLife announced that they would publish everything that they send for review as a preprint, and that they would then update the papers post-review, while keeping a version history. This is great for transparency, since you can see exactly how the paper was changed by peer review. Other than that, though, the model leaves much to be desired.

Editors still have ultimate power. In fact, they have enhanced power since they are the sole decision makers over initial publication of the preprint (they reject over 70% of papers that get submitted to them). This means bias and politicisation aren’t eliminated, time-to-publish is longer than ideal, and costs are still significant.

The change in model has also made scientists weary of potential prestige losses in publishing there, and so fewer high impact papers now come out in eLife. Indeed, the journal has recently been stripped of its official impact factor altogether.

The best use-case for eLife now is if scientists wish to get the stamp of approval of peer review without being forced to do extra experiments (say, if the first author has moved to a different lab/career).

A hypothetical optimal publishing system would adopt the capability to edit published papers post-review from eLife.

OpenPsych

The field of psychology is particularly rife with publication bias, politicisation, and data that doesn’t replicate. OpenPsych takes a novel approach inspired by anonymous internet forums. Papers are uploaded (through the filter of an editor) and a sub-forum is created where reviewers, recruited by the editor, can openly discuss the paper with the authors (double-blinded). Authors can re-upload versions as they go.

This is nice because it is a much more streamlined continual process of peer review, and much more like an actual conversation than a judicial decree (which is what traditional peer review feels like). There is also zero cost of publishing because there are basically no running costs and OpenPsych relies heavily on volunteered time. Still, the prestige issue is not solved and editors retain power to inflict their bias if they wanted to. The forum-style of peer review has been instrumental for combatting accusations of pseudoscience in OpenPsych, which stem from the controversial nature of lots of the research there.2

A hypothetical optimal publishing system would adopt the forum-style transparent peer review of OpenPsych.

Subscribe now

A novel solution

Ultimately, the various solutions thus far leave much to be desired. A perfect solution must be fast, free, allow scientists to accumulate prestige (to encourage adoption of the platform), have a filtering mechanism for high impact vs low impact papers, and allow for feedback from experts as a quality control mechanism.

The answer is already, literally, at hand: Substack.

Hear me out.

In part 1, I mentioned how Matt Ridley and Anton van de Merwe were unable to surpass the political filters of editors when trying to submit their scholarly review of the evidence for the origins of COVID-19. Ultimately, Matt published it on Substack. Let’s go through how doing so overcomes many of the issues of academic publishing.

Firstly, it is fast. You can publish on Substack whenever you want and it appears instantly. Further, the inbuilt word processor makes the uploading process even more seamless than preprint servers like bioRχiv.

Secondly, there is no bias or influence from editors or reviewers. Authors can publish what they want as they intended it to be. The politicisation of science and obsequiousness towards editors vanishes instantly.

Thirdly, open access is free. Authors can publish articles with no paywalls. If they want to implement a paywall, it hurts the impact of their research but at least the money goes to the actual scientists (with a small slice going to Substack of course). I would recommend that an academic version of Substack bans paid content altogether though—acting as a non-profit arm of regular Substack (more on that later).

Okay but how is this any different to a preprint server then? What about the prestige issue and peer review? The key difference between a preprint server and Substack Scholar (I’m coining that) is the leveraging of social media mechanisms.

Solving the prestige problem

When you go on Instagram, you aren’t faced with a random selection of all Instagram posts. Instead you are given an algorithmically curated selection of posts from people you manually follow and from accounts that are predicted to maximise your engagement (either negatively or positively). In 2025 we know to decry the social media algorithms as evil and manipulative. Yet the very same type of algorithm could help Substack replace conventional academic publishing. As things currently stand, Substack recommends articles for you to read. This will be no different on Substack Scholar, except all the articles will be academic in nature.

Integrating a recommendation algorithm to academic publishing solves the prestige issue too.

Academics are glorified content creators. Our job is to create content (written articles) that is interesting to other scientists (i.e., has high engagement). Social media, like academic publishing, runs on prestige (or “clout”). On Instagram, the metric for prestige is follower count or likes on a Reel. For academics its h-index and citations on a paper. Charli D’Amelio and Stephen Hawking are truly two peas in a pod, if you think about it.3

Previous attempts to fix academic publishing have tried to ignore or even reject the notion that academics are motivated by prestige. This is a fool’s errand. Why try to deny human nature when you can leverage human nature to create a better system?

Successful academics publishing on Substack Scholar will gain higher subscriber counts. This makes “high impact” an attribute of individual scientists or labs, and not journals (like some kind of cool kids’ club) that you can game your way into or be barred from for unfair reasons.

Beyond subscriber count, individual pieces of research will see the limelight they deserve thanks to algorithmic boosting. If some unknown researcher publishes something truly amazing, scientists *will* see it.

The prospect of gaining prestige via Substack is already an attractor for many academics who dabble in writing (Eric Topol, Steve Stewart-Williams, Erik Hoel, little old me, etc.). If a viable way of merging their regular Substack with Substack Scholar presented itself, I’m sure academics will eagerly take up the opportunity. And if prestige alone isn’t sufficient, academics can have a paywalled regular Substack to accompany their open access Substack Scholar, which they can use to communicate their work in plain English to the general public and garner some revenue.

The need to integrate social media with academic publishing is already apparent to big journals. Social media teams are employed at these journals to ensure maximum reach of their publications. Indeed, I recently published a paper as first author and was asked to make a YouTube Short to promote the paper! (it’s here since you asked—and yes I know I’m a terrible influencer). Preprints, too, rely on the same social media mechanisms for dissemination—either being posted about by the authors or by the automated bioRχiv account itself.

These days I find most science papers to read via my 𝕏 algorithm. Using a thread to explain a paper is now a crucial part of publishing. Indeed, one of the ways in which prestigious journals uphold their prestige nowadays is simply by having larger social media followings. There is no reason why social media accounts of individuals can’t functionally replace those of prestigious journals—in fact they regularly do on 𝕏.

If we published the primary research articles on the very same platform that we shared Notes and SciComm breakdowns of the primary research, the whole process would become seamlessly integrated and we wouldn’t have to deal with link suppression4. Of course, there is no reason why Substack has to take on the mantle here. 𝕏 could equally come out with “𝕏 Research” and try to do the same thing. Substack is already optimised for long-form content, though, its user base is already more intellectually inclined, and it seems to have a first-mover advantage too, as I will come on to.

And peer review?

I’m glad you asked. If you go look at that article I mentioned on the origins of COVID-19 you’ll notice there are quite a few comments!

What you’ll find is (1) a little bit of bickering, (2) some very intelligent debate in good faith, and (3) some links to entire response articles offering the opposing perspective.

A comments section *is* peer review.

Leave a comment

Instead of limiting review to a handful of scientists, every scientist will be capable of chiming in with their positive or negative reviews, or of quickly scribing response articles themselves. This also makes it much more likely that genuine experts can give their opinions on the article, instead of just the few scientists who the editor of a journal thought were appropriate experts with a free enough schedule.

What’s more, timely and salient comments on Substack posts often receive the most likes themselves, and so there is prestige/clout to be gained from engaging in this form of instant peer review. Algorithmic reward for peer review in the form of comments functions similarly to algorithmic reward for good articles in the first place. What’s more, like the review forums of OpenPsych, this format creates a much more conversational and therefore constructive environment.

Substack Scholar: an ideal publishing system

To make Substack Scholar work, there would need to be a few key implementations. Here are my specific recommendations:

Substack Scholar would need to be integrated with, but separate/parallel to regular Substack. I would recommend that only those with institutional emails, validated Google Scholar accounts, ORCiD, or some other measure of scholarliness are able to create Substack Scholar accounts (although excluding too many outsiders might be detrimental). This ensures research is traceable and legitimate (for example, preventing a wave of AI-generated slop from being pumped out by anonymous bad actors).

As I’ve already mentioned, I think insist that all Substack Scholar posts should be free and open access, subsidised by paid content on the regular Substack side of things if needed (although the running cost of Substack Scholar would be minimal anyways). The subsidy will pay for itself in the form of increased overall Substack traffic.

The comment section model of peer review works only if spam is filtered. Algorithmic boosting of the best comments solves this partially. Another simple solution is to have two parallel comment sections under Substack Scholar articles. Those with Substack Scholar accounts can comment in and view both sections, but non-scholars can only view the scholarly comments, not contribute to them (but they can contribute to the normal comment section of course).

Substack articles are currently editable. After publishing, an author can go back and fix typos as they please and it will just update without a trace. This should be changed for Substack Scholar. Amendments should still be possible, but version histories should be kept and made available for viewing always. It should be clear which version comments are referring to (perhaps just create separate fresh comments sections for new versions of articles). This means if authors are persuaded by the arguments of commenters, they can make changes as with regular peer review (or they can just publish a new article altogether).

The referencing/citation UI/UX can be enhanced. As things stand, I typically cite research by linking the article with an in-text hyperlink. Alternatively, I add a footnote and manually type out the citation. Substack Scholar would fix this friction, perhaps with EndNote or Mendeley plug-ins, or its own referencing software add-on. It is essential that researchers can easily cite other work, and also have their Substack Scholar articles themselves cited.

For this to happen, articles need a digital object identifier (DOI), like any other research article or preprint, with an accompanying PDF available for download. A DOI also allows for integration into funding and curation databases. Persuading funding bodies to accept Substack Scholar articles on researchers’ CVs as valid pieces of work is crucial. I recently submitted a bunch of applications for postdoctoral fellowships (mini grants that only cover my salary). All of them bar one allowed me to include bioRχiv preprints in my list of publications. This fact surely plays a large role in the widespread adoption of bioRχiv. There is no reason why Substack Scholar cannot follow suit.

Finally, the algorithm would need constant fine-tuning to ensure it serves scholars well. Whether people choose to get their research articles direct via subscriptions to their favourite scientists or by spending time on the Substack app or site will be up to them. Either way, Substack Scholar should aim to maximise productive scholarly debate and dissemination of high quality and interesting research.

The time is now

Substack Scholar genuinely could work, and I’m not the only one to think so. I was recently chatting with , a product researcher at Substack, who has been eyeing up the ever-increasing number of academics who write on Substack. The idea of assigning DOIs to academic Substack articles is already in the pipeline. Doing so will enable the academic articles selected for the pilot to feature on Google Scholar, which will be a key first step. I hope that this trial goes well and that Substack considers the ideas I have put forth here.

We are at a crucial point with regards to academic publishing. The COVID-19 pandemic and enhanced freedom of information in the social media era has put (scientific) institutions under intense scrutiny—few more so than academic journals.

In part 1, I mentioned how the NIH is instructing all NIH-funded research to be immediately available to the public via PMC. The Howard Hughes Medical Institute, one of the largest private US funders of biomedical sciences, went one step further with their recent announcement proclaiming that they won’t allow their grants to be used for article processing charges at all, and instead insist that all HHMI-funded research is uploaded to bioRχiv from 2026 onwards. Clearly, the appetite for change is there.

It has been said that science progresses one funeral at a time—the idea being that dogmata only die out when the people who uphold them do so first. Fortunately, I think that the adoption of an incentive-laden superior publishing system that leverages the tech of social media will progress one birth at a time—the birth of research labs that is.

Once I run my own research lab—deo volente—the smart move will be to publish via Substack Scholar. The ideal early adopters of Substack Scholar will be end-of-career scientists who have huge personal prestige and nothing to lose by trying something new and postdocs like myself who can be convinced to switch systems ab initio when starting their own labs. Indeed, targeting the birth of new labs may be the best option given the prevalence of closed-mindedness among older academics.

With that in mind, we might expect adoption of Substack Scholar for publishing primary research to be a little slow off the mark. Rome > 24h etc. However, I imagine it will very quickly become the hub for secondary research. Publishing a scholarly review or perspective (which synthesises already-published data, rather than reporting data for the first time) entails much less risk for early adopters, and will likely play a key role in building the credibility of the publishing platform before primary research follows.

In the coming months I plan to write a scholarly commentary summarising what the research tells us about how much protein we should be eating for optimal health outcomes. There have been some strong claims both for and against high-protein diets, and I think a closer look at the evidence is sorely needed. I think I will save myself the the cost of APCs, the strife of dogmatic reviewers, the bias of ideological editors, and the frustration with slow turnaround and just publish it here instead.

The job of an academic scientist is to ideate research, propose doing the research to a funding body (i.e., submit a grant application), do the research, and publish the results of the research in an academic journal. If the journal is sufficiently prestigious, the publication can then be used to persuade future grant committees that your future research proposals should be funded. A side-effect of this cycle of publications and grants is that some scientific discovery occurs, which can then be used by biotech/pharmaceutical companies to actually benefit society.

Research funding not only covers research costs like expensive equipment and reagents, but also the salaries of all scientists involved and institutional overheads. This means that the livelihood of scientists (and the upkeep of universities) depends on them getting their work published in prestigious academic journals. Therefore, publishing being a fair and just system is of existential importance to many scientists. The problem is, the system is markedly flawed.

Financially, the system is backwards. Scholarly research is typically locked behind paywalls, and universities must pay huge sums in subscriptions to the big prestigious for-profit journals in order to access research produced by scientists, including by those from their own institute. If academics wish for their research to be published under an “open access” license, which means that anyone can read the research article for free, they must pay an article processing fee that can be in excess of $10,000. This money, of course, comes from research grants which are either derived from taxes or philanthropic donations.

Beyond the economic exploitation by big journals, whose existence is propped up by the fact that researchers themselves are so hopelessly thirsty for prestige above all else, the quality control mechanisms that decide what does and does not get published are not only ineffective but actively detrimental to research. Those outside of the academe know to revere the authority of a peer-reviewed study, but they may not know what peer review actually entails. The dirty secret is that peer review rarely prevents junk science from being published as long as the editors of the journal in question find the results sufficiently palatable (politically or otherwise) and exciting.

For those who care about this stuff, this is old news. Michael Eisen was writing about how bad the traditional journal model is for science and society decades ago. His 2013 essay is a great primer on the problem. More recently, a scolding critique of peer review published a few years ago by has been doing the rounds again, which I would also recommend as further reading.

Disgruntlement is certainly growing among scientists (in no small part due to the crusade of the Michael Eisens of the world). The issue has given rise to many attempts at new publishing models over the years, including Eisen’s own PLOS, as well as eLife, and OpenRχiv (for biomedical sciences) and arχiv (for other sciences).

Indeed, the NIH—the primary funder for biomedical sciences research in the US—has recently implemented a new public access policy which instructs all scientists who publish NIH-funded research to make it immediately available to the public for free via PubMed Central. We’ll see if this policy is sufficient to get scientists to forgo the prestige of big journals, or whether it will just increase the rate at which scientists dish out NIH dollar for the huge open access article processing fees (I suspect the latter … big journals don’t really seem to be that worried either).

Sadly, none of the solutions so far have been particularly persuasive to scientists—primarily due to the whole prestige thing. In part 2, which I will publish next week, I will offer my own solution which I think overcomes the issues outlined here in part 1 by not only acknowledging prestige as the motivator for academics, but by actually using it. My solution also capitalises on new forms of technology that we already use in other realms to disseminate information to those who want to read it most.

For now, I wish to flesh out the issues in detail here in a way that strikes a cord with academics and non-academics alike. To do this, I will simply recount the process of publishing a paper in an academic journal. By the end of this essay you will hopefully have a grasp of the scale and the importance of the issue.

Let’s imagine a scenario where some scientists have uncovered something very exciting and wish to share the results with other academics.

Submission

The first step is to write a manuscript broadly in the format of a specific target journal. Academics will target a journal whose scope covers their research area, but also whose impact factor is as high as possible.

Impact factor, for those who are unaware, is a metric used to assess journal prestige based on how often the articles published in a particular journal are cited by other subsequent research papers. Since high impact factor journals are more widely read (e.g. Nature, Science, Cell), publishing here maximises the reach of your research and also boosts your prestige as a scientist. Impact factor is a self-reinforcing phenomenon, since high impact journals attract high impact research, and high impact research makes the journal it is published in “high impact”. This self-reinforcement means that it is *very* difficult for a new journal to compete with established journals for prestige.

Once the results are written up, the corresponding author (typically the principle investigator/director of the lab where most of the work was carried out) will write a cover letter to the editor of the journal and submit it with the manuscript. After a week or so, an editor will have skimmed the paper and assessed whether it is good enough to go for peer review or not. If not, then you will have to submit the article to a journal with a lower impact factor (which will be easier to publish in). This can involve reformatting, restructuring, or rewriting the manuscript altogether if the new target journal has different rules or scope.

Since all the power here sits with the editors, there is a strong incentive for scientists to form good relationships with editors. The “best” scientists make a beeline for editors who attend conferences (indeed, “editorial presence” is a key factor in deciding which conferences are worthwhile to attend). The “best” scientists let the children of editors do internships in their labs. The “best” scientists publicly agree with the political views of editors. It is worth nothing that full-time editors are, by definition, not actual scientists themselves.

As an aside, this is also the point at which science can be politicised. Editors can just reject a paper if they don’t like what the results mean politically. This is something that and Anton van de Merwe realised when they were trying to publish their comprehensive review of the evidence that SARS-CoV-2 most likely leaked from a lab (a lab in Wuhan that was actively working on ways to make bat viruses compatible with human cells, by the way). In the end, they published a version in the Telegraph and on Matt’s Substack because editors of big journals simply didn’t like the implication. If you’re interested you can read it here.

That this didn’t end up in a peer-reviewed journal then means that policy makers and pundits can claim that “the scientific consensus supports a zoonotic spillover event and not a lab leak conspiracy”. This is true, but the reason why it’s true isn’t what the public thinks. There are, of course, many such cases. The political preferences of journal editors is one of the main reasons for prolific publication bias in fields like climate science, economics, and psychology, which skews what politicians and the public think is true.

Peer review

If initial submission is successful (perhaps the editor who flicked through your manuscript was in a particularly good mood when they did so), you then wait as your paper is peer reviewed. This means that the editor will pick experts in the relevant field(s) to send the paper to.

While some journals offer double-blind peer review, most of the time the reviewers can see the identity of the authors, which inevitably biases their reading of the paper—you can imagine reading a paper from a Nobel prize winner with slightly less cynicism than one from an unknown researcher from some unknown institution. This immediately lets credentialism seep into the peer review process, where research isn’t assessed purely on merit but on reputation of the researcher/institution.

Another flaw in the peer review system is the suggestion or barring of certain reviewers. Most journals ask scientists at initial submission to suggest or prohibit specific researchers from reviewing their paper. This is supposed to help prevent “professional conflicts of interest” from disrupting fair peer review, but in reality creates a massive loophole that can be exploited to cheat the peer review system. Scientists form clades of friendly peers who tacitly agree to leniently review each others’ papers, while always excluding critical scientists from reviewing their work. If a scientist has previously pointed out flaws in your research or historically disagrees with your interpretations of data, no need to worry—you can just make sure they never review your work by making sure to remember to exclude them when you write your cover letters.

I was given (genuinely) good advice during my PhD to always suggest older reviewers. Young scientists tend to want to prove themselves and will typically read your manuscript with a keener eye for detail and critique your work more thoroughly—which you don’t want of course!

Once the paper has reached its reviewers, they will be given a loose deadline by the editor to return their assessment of the paper. Importantly, this critical peer review step—the poster child of academic publishing—is simply done for free by other scientists, not by the journal itself. Of course, as I’ve already mentioned, scientists are primarily motivated by prestige and kudos, rather than money, but reviewers aren’t rewarded with that either. Some journals now publicise who the reviewers were post-publication, but it’s hardly as lauded as it could be. In fact, the reward for reviewing papers is so low that many principal investigators just ask their students or postdocs to do it for them (it’s called “mentorship” apparently).

Further, since the reviewers who might find the research particularly controversial—and therefore might expeditiously add their critical input—were excluded from review by the authors, the intrinsic interest in reviewing the paper is automatically dampened.

The result of very low motivation to review papers quickly is that the whole process is very drawn out. As a concrete example, I recently waited 99 days to get comments back (meaning a ~500 word response) from two reviewers for a paper from my PhD. I waited 99 days for about 2 hours of work! (NB: I’m not blaming the reviewers here at all—they have far better things to be doing).

To be fair, most journals, like those in the Nature, Science, or CellPress families, boast median turnarounds for first round of peer review of around 60 days.

Wait… “first round”?! Yes, we’re still quite far from the finish line.

Revisions

Once the editor has collated and assessed all of the reviewers’ comments, they then make a decision. In very rare circumstances, reviewers unanimously recommend a paper be published as it is, or with minor textual changes. If a single reviewer (there can be up to four, but at least two) has an issue with anything though, the editor will ask you to amend and resubmit the paper (or present strong arguments against the reviewer).

Peer review is thought, by the lay public, to be a stringent quality control mechanism for the validity of scientific research. If this were genuinely the case, journals would ask other researchers to replicate the experiments in the paper. This would be impossible without much more money and time though, and so experimental validity isn’t actually assessed by peer review, leading to many cases of invalid or fake data getting published in peer reviewed journals.

What the peer reviewers actually do is try to inflict their preferences on the science being done. They do this by insisting on discussion of certain topics, including citing their own work, or by requesting specific experiments be done that they personally would find interesting.

Now, of course more experiments will always make a paper better, so people defend the ability of peer review to “improve science”. But this line of reasoning obfuscates the intended role of peer review. The end result is that peer review makes papers larger, can make them semantically inconsistent (by requesting an experiment that doesn’t particularly make sense in the context of the rest of the paper), and makes the publishing process incredibly drawn out.

Months go by as the scientists work to do extra experiments. The reviewers have no idea of the capacity of the lab or the institute, so they could ask for an experiment that requires equipment that just isn’t available without collaboration (which means even more time and bureaucracy).

Eventually, the scientists submit a revised manuscript. They will have addressed every single point in the reviewers’ comments—either refuting them in a plea to the editor (risky), or addressing them as discussion points, or by adding more data from revision experiments. The editor receives this version and once more sends it out to the same reviewers. Perhaps one reviewer is at a conference, or on annual leave, or is busy submitting their own manuscript. More weeks flit by until the reviewers have time to take a look again.

Most of the time, they will be satisfied with what the scientists have done in the lab at their behest, and might only ask for some amendments to the text at this point. Sometimes, the experiments that the reviewer suggested might turn out to have been bad ones to suggest (oopsies!) and they might suggest new ones now. Many more months can go by before all reviewers and the editor is happy to proceed.

Publication

Finally, after this little song and dance which does nothing to validate scientific integrity, the editor will accept the manuscript in principle. At this point, the journal starts to actually do something, and formats the article properly and goes over typos etc. Weeks later the authors are sent proofs of the final thing. They must approve these and then the article is published online.

It is at this final stage that the authors are also asked which license they wish for the paper to be published under. If they don’t have enough funding to pay for the huge article processing charges for open access, then they have to just face the fact that fewer colleagues will be able to read the paper.

You might have picked up that most of the work here is done by (1) the authors and (2) the reviewers. And yet the journals are the ones raking in the profit. I think Michael Eisen puts it best:

I want you to note just how little the journal actually does here.

They didn’t come up with the idea. They didn’t provide the grant. They didn’t do the research. They didn’t write the paper. They didn’t review it. All they did was provide the infrastructure for peer review, oversee the process, and prepare the paper for publication. This is a tangible, albeit minor, contribution, that pales in comparison to the labors of the scientists involved and the support from the funders and sponsors of the research.

And yet, for this modest at best role in producing the finished work, publishers are rewarded with ownership of – in the form of copyright – and complete control over the finished, published work, which they turn around and lease back to the same institutions and agencies that sponsored the research in the first place. Thus not only has the scientific community provided all the meaningful intellectual effort and labor to the endeavor, they’re also fully funding the process.

In conclusion, mainstream academic publishing as it currently stands is slow, corrupt, frustrating, costly, produces needlessly long and convoluted papers, biases published results through the (political) lens of the editor, and doesn’t even give access to the research to the public.

This obvious farce probably explains why repositories like BioRχiv have surged in popularity. To compare the processes, this is what publishing via BioRχiv looks like:

1. Write your article with whatever formatting you desire and convert it to a PDF

2. Upload said PDF and any accompanying figures or supplementary files

3. Wait less than 24 hours for some straightforward vetting

4. View and share the DOI with whomever you want

Of course, pre-print servers like BioRχiv don’t have the prestige of a Nature paper, and the lack of some kind of critique or peer review means people who cannot independently evaluate research don’t know what to think. This means pre-prints primarily act as a placeholder, and people still hold out for the “actual publications”. Still, these repositories are particularly valuable for increasing the expediency of science and removing politicisation. I can only imagine what kind of hellish peer review process the authors of this Yale study are going through.

It may seem like the entire scientific enterprise is fucked, but I think there is hope. In part 2 of “Fixing academic publishing”, I will discuss further the attempts of others to fix the problem, explain why they too have fallen short, before I give my own novel solution which I think would solve most, if not all, of the issues with academic publishing.

It takes two humans to produce a human. For humanity to carry on existing, therefore, each woman (or man—but tracking maternity is easier than paternity) must produce exactly two humans that also go on to create two humans. Since not all little humans make it to the point where they also create other humans, women must, in fact, produce slightly more children for humanity to persist (around 2.1 for developed countries like the UK; above 3.0 in some African countries with high infant mortality rates). Simply put, falling short of this number for a long enough time equals the extinction of humanity. I am morally opposed to this outcome, since humanity *is* morality as far as we can tell (i.e., without humanity, morality ceases to exist).

In many countries now, total fertility rates (TFR, defined as the average number of children birthed per woman across her lifetime) are significantly below 2.2. In the UK, TFR as of 2024 is 1.41. In the US, despite more religiosity and slightly stricter abortion laws, the situation is not much better, with a TFR of 1.61 in 2023. We have largely avoided negative consequences of this by supplementing our populations with immigration, and due to the population lag effect (it takes a while for demographics to change after birth rates change). But it’s not just a western or caucasian issue only. Thailand, El Salvador, and Iran posted 2023 TFRs of 1.21, 1.78, and 1.70 respectively. Even India is now below replacement, at 1.98. East Asia is particularly bad, with Japan at 1.20, China at 1.00, and South Korea at 0.72.

Let’s do some quick maths to put this into perspective. At replacement level, 100 Koreans now would leave 100 Koreans in 100 years. Perfect. With a TFR of 0.72, assuming an average childbearing age of 33 (so three generations in 100 years), 100 Koreans produces only 4 baby Koreans in 100 years time. Numerically, that’s worse than most genocides in human history.

If TFR remains low in these countries, that’s an obvious catastrophe and tragedy. Worse than just the disappearance of humans over time though, the quality of life for the final generations would be terrible (young people run and fund the economy, while those above retirement age only take from what would be an ever shrinking economy).

Globally, the TFR is still just above replacement (around 2.35)—kept afloat by Africa, where some countries have very high TFRs still (e.g. Niger’s is 6.73). Since this demographic fate seems to pervade all advanced economies, we might expect that, once Africa catches up economically, their fate will probably similar too (although correlation doesn’t necessarily imply causation here, and there are mechanisms of causation in both directions).

For those who’ve been paying attention, I’m stating the obvious here. Yet it still shocks me how many intelligent, educated people I speak to think that we are still facing an overpopulation crisis. Perhaps decades of Malthusian depopulation rhetoric from the likes of Bill Gates is to blame, or perhaps the blindspot stems from wishing to distance oneself politically from the Musk’s of the world, who is fervently talking about this issue. Either way, this brief introduction here should hopefully bring some readers up to speed without flogging a dead horse.

The solutions

While population collapse used to be falsely considered a dog whistle for racists and misogynists, it is now garnering much more mainstream attention—a testament to the reality of the issue and the disingenuousness of the ad hominem-ers. A number of thinkers and writers have proposed solutions to the declining fertility rates issue: from ’s focus on marriage rates or housing conditions, to ’s and ’s focus on financial incentives, to ’s focus on fertility tech for aging mothers. I think they are all great, and these will certainly play a role in solving the problem. But all of these solutions work with a fixed desire to have kids, and try to make it easier for people to achieve these desires. In this way, such solutions only address proximate causes, while leaving the underlying desire to have more kids untouched.

The reason for this is that, as many have pointed out, women in low TFR countries have fewer kids than they say they actually want to have. The problem is, I think this is bullshit. This is for a couple reasons:

Firstly, people change their minds throughout their lives. A woman may have wanted zero kids until age 35 and then all of a sudden decided she actually wants 3 kids. Logistically, she may not get to three by the end of her reproductive years. This scenario gives rise to stats which show that “women actually want more kids than they have in reality”, but these stats are misleading because it’s obviously true that cultural or political nudges towards a desire for natalism earlier in her life would have given her more years post-change of mind to realise her motherhood goals.

Secondly, people lie (including to themselves). Maybe looking back on her life through rose-tinted glasses, a woman might say that more kids would’ve been ideal. Social scientists know, however, that there is quite the chasm between stated preferences and revealed preferences. It is easy to say that you wanted more kids than you got when you aren’t actually faced with the decision to have more kids anymore. In terms of judging peoples’ actual desires, their actions are probably a better metric than their words. When there’s a will, there’s a way, and if the way wasn’t taken, there probably wasn’t the will. Thus, positive changes in the desire to have kids—addressing the ultimate cause of the issue—actually will be hugely impactful.

With that said, I’d like to weigh in here with two solutions. One is policy-based and I offer it with tongue firmly in cheek. I think it would work, but I can’t in good libertarian faith insist on it. The actual solution I think is already underway. It solves the issue bottom-up by invoking that all-powerful but ever-nebulous cause: culture.

Solution I: Have you tried “kill all the pets”?

If I were a benevolent dictator who wanted to boost fertility rates, this is the single law that I would pass:

Mammalian pet ownership is henceforth illegal for childless people below age 40.

Here me out. When I first moved to San Diego I was stunned at how many young people own pets. It wasn’t young families wanting to give their kids the joy of a pet though, it was mainly childless gen Zers and millennials. Pets instead of kids, that is. Lyman Stone has written about how the type of housing and amenities of apartments is anti-natal. He’s definitely got a point here, but developers aren’t demographic engineers—they simply respond to the market. Here in San Diego basically every residential compound I drive past has pet-friendly amenities, from dog parks where you can let them off the leash to pet spa/grooming facilities within the communities themselves. The meme of a “fur baby” isn’t just a meme. In California 47% of dog owners have pushed their dog in a stroller, 36% of Americans give their dogs birthday presents, and pet ownership in America has more than tripled since the 1970s at the same time as birth rates have almost halved.

Breaking down pet ownership by generation is even more revealing. In the US, Gen Z and millennials dominate pet ownership, at 57% (despite only making up ~43% of the population). A stunning 76% of millennials own pets in the US. Further, most of the recent explosion in pet ownership is being driven by Gen Z specifically, with a 43.5% increase in ownership from 2023 to 2024 alone (interestingly, mainly driven by men). Gen Z, despite being at a stage in life that most would consider to be less amenable to pet ownership, exhibit the highest rate of multiple pet ownership out of any generation, at 70%.

It is the attitudes that American pet owners have towards their pets that prove my point. According to an October 2024 Harris poll, 82% of pet owners say their pet is like their own child. When US pet owners were asked to choose between pets and children, 39% said both, 43% chose pets, and only 18% chose children over pets (!).

Many of my Gen Z friends complain that having kids is just too expensive. Disregarding the fact that Gen Z are unprecedentedly rich, this doesn’t seem to stop them from getting pets, which are also very costly. In fact, if you plot the fertility rate versus the percentage of dog owners who would spend $4k to save their dog, by US state (don’t ask me why this data is even a thing), you get a pretty neat negative correlation.

Yes yes there are a gazillion confounders here, I know, and it’s basically GDP vs fertility again. Still, I don’t think it’s wildly implausible to hypothesise that pet owners without kids would spend more on their “fur babies” than pet owners with kids. There seems to be a misdirection of caregiving resources here. Indeed, as has recently pointed out, pet expenditure and birth rates in the US are inversely correlated over time.

When I speak to people about kids, the main reason why they don’t want to have them, besides cost, is just that they can’t be arsed with the hassle. Kids require effort and lifestyle sacrifices. And they’re messy. One friend of mine cited the inevitable months of suboptimal sleep as the main reason why she never wants to have a kid (!).

Pets give you a small dose of this hassle, especially dogs, meaning that pets syphon caregiving energy that could otherwise be spent on a small human. There is a piece of conventional wisdom that having a pet is a good way to “practice” for a couple thinking of having kids. I think this is bad advice. If people are already umm-ing and ahh-ing about having kids with their busy and self-centred lives, throwing a pet into the mix more or less seals the deal against kids for the next few years.

On the other hand, the main short term benefits of kids are a sense of purpose and responsibility that come from having a small thing to look after that depends on you, love/cuteness/adoration etc, and the romantic relationship-enhancing effect of having a shared goal with your partner. All of these things are provided by a pet too. Thus, pets relieve any buildup of parental sentiment while syphoning resources (time/money/energy) that could go towards an actual human.

Recently, I’ve recently seen a few instances of science communicator Andrew Huberman getting dragged on X for being childless at 50 years old, and therefore not being a good masculine role model (I think these accusation are rather harsh). Andrew has spoken extensively of his close relationship with his bulldog, Costello. Huberman was around 35 years old when he got Costello—close to the median age of first-time fatherhood for college educated Americans. I can’t help but wonder if an alternate universe Huberman, deprived of Costello’s affection and full to the brim with caregiving capacity, would have settled down and had kids earlier in his life.

The no pet policy is sound. It focuses on mammalian pets because I think they’re the main culprits for stealing caregiving resources and desire from potential humans. I could flesh it out more, talking about the mutability of the 40yo cutoff, or concessions for the gays and the infertile, but I’ll stop now. No pets would work, but not in a liberal democracy. Let me now offer a realistic solution.

Solution II: Nothing

As I have already alluded to, the desire to have kids is a prerequisite for all enabling factors like fertility tech, tax incentives, or housing availability. For every argument that a lack of X is limiting fertility rates, there are counter-examples where a sufficiently high desire to have kids trumps a lack of X.

For example, it seems to have become an axiomatic truth that young adults won’t settle down and have kids unless they can own a home. But this is obviously nonsense. There are many young families who rent in perpetuity. There are many young families who thrive living in apartment blocks, not houses. Equally, while money can be used to incentivise parenthood, our GDP per capita vs fertility rates analysis clearly shows that wealth isn’t a requirement for having kids. Further, comparing fertility and GDP per capita within one country over time also shows that more money doesn’t equal kids. As I’ve already stated: when there is a will, there is a way. Proximate solutions are great, but the ultimate solution is the desire to have kids. The problem of declining fertility rates needs a bottom-up solution, not top-down ones. But how can we possibly control people’s desires to have kids? You don’t have to be a determinist to already know the answer: culture.

The capacity for culture to outweigh all other birthrate-limiting factors is perhaps best exemplified by Israel. Israel is a rich, progressive country with a GDP per capita greater than the UK and similar home ownership rates as the UK (~65%). Despite this, and their constant threats of war and terrorism, it’s TFR has been stable since the 1990s at around 2.9—higher than peak baby boom UK. Yes, Israel has a swathe of pro-natal policies, but, as I’ve argued before, in democracies, policy is downstream of culture. Israel’s pro-natal policies stem from their pro-natal culture that reaches back to their founding as a country. Politicians follow the desires of the people, and the desires and behaviours of the people en masse is culture.

When we lift the bonnet of our culture over the past several decades, our low fertility rates immediately make sense. It’s no secret that parenthood goes largely uncelebrated, and the desire to have kids has been culturally suffocated. The corporate world is incentivised to keep all adults in the workforce, governments enjoy the double tax intake per family, and the entertainment industry has ridden the wave of sex-positive feminism which rejects the value of the family and upholds the value of singledom and self-prioritisation. The sitcom Friends is probably the archetypical example of the ousting of the nuclear family, but there are many such cases. Moreover, DINKism (Dual Income No Kids) is at its peak right now, as social media and the dopamine economy has worked to glorify what I call the “gap year lifestyle”: no commitments and an obsession with travelling and new experiences. Pleasure-seeking and suffering-avoidance have been the guiding moral principles for gen Z and millennials.

Which forces work on the nebulous behemoth that is “culture”? What’s the secret ingredient here? Is it something that can even be externally steered? Sort of.

Homeostasis describes biological systems that, when pushed in one direction, automatically revert to the baseline. Blood glucose concentration is homeostatic. When we eat a cookie, glucose concentration in our blood rises. The pancreas responds by releasing insulin, which works to bring glucose back down to normal levels, which in turn shuts down insulin production. This course-correcting dynamic exists at every level of biology, from cells all the way up to populations. Indeed, I propose that human fertility rates are strongly homeostatic.

One mechanism for this is basic Darwinism. People tend to inherit culture and beliefs from their parents (twin and genome-sequencing studies have shown political ideology is highly genetic). People who culturally do not value parenthood will have fewer children than those who do. After several generations, people who value parenthood will dominate the gene and meme pools (in the Dawkinsian sense). We are already seeing this in real time with much higher birth rates among conservatives versus progressives, as covered by in the Financial Times.

In terms of cultural cycles, I think millennial DINKism might already be inducing a counter-cultural swing. Slow-mo shots of your girlfriend (of 10 years) skipping through the streets of Santorini were cool in 2016, but in 2025 DINKs are just a bit cringe and sad. Gen Z and alpha, having waded through the swamp of dopamine saturation that was the COVID lockdowns, are now after meaning, not thrills, as indicated by their resurgent religiosity. The full cultural reckoning against DINKism will come when the DINKs enter middle- and old-age. Having the disposable income to fly to the Maldives for a week every year hardly makes up for spending the Christmases of your 40s and 50s alone, or incurring injury or illness in your 60s and 70s and having nobody that will sleep in a stiff chair by your bedside as you recover.

The New York Times published a piece last year titled “The Unspoken Grief of Never Becoming a Grandparent”. It is a depressing read about how would-be grandparents are working with therapists to help reinforce the self-delusion that its fine that their kids aren’t having kids. Their childless children obviously don’t see their anti-natalism as selfish, and even if it was, they’re entitled to put themselves first! The real tragedy will come when they realise, in their 50s perhaps, that the real act of selfishness was on the part of their younger selves against their older selves. The best advert for pro-natalism will be old anti-natalists.

Some cultural shifts are already underway. The aforementioned rising religiosity of gen Z can be expected to create a noticeable bump in fertility rates in the next 5-10 years. DINKs have been met head-on with TradWives, much to the annoyance and confusion of third wave feminists. Donald Trump, regardless of what you think of him, has shifted the Republican Overton window regarding IVF in a pro-science and pro-natal direction. Even left wing British politicians are now ringing the fertility bell.

In media and the arts, the nuclear family is starting to return to centre stage. Apple released a deeply emotional pro-family Christmas ad last year that has now been viewed over 51 million times, while Volvo’s beautiful pro-natal advert received an overwhelmingly positive response. Comedy giant Andrew Schulz sold out arenas across the country on his “Life” tour last year. The show, now available as a Netflix special, is an explicitly pro-natal comedy set about his and his wife’s successful quest for parenthood. He now regularly posts about the joys of fatherhood to his 4.5 million Instagram followers.

While a childless Kamala Harris failed to convince the nation of her leadership skills, the favoured Republican and Democrat candidates for 2028—JD Vance and Gavin Newsom—have seven children between them (I mean … with their wives that is). Perhaps the final nail in the coffin for anti-natalist culture will be when popstar-cum-cult leader Taylor Swift and her politically opaque fiancé announce that they are expecting their first child. I’m quite certain that the attitude towards motherhood of millions of gen Z and millennial women will turn on a dime once Jack Antonoff has penned his first pro-natal Swiftie anthem.

The Economist recently published an opinion piece arguing that we should be optimistic about the fertility crash. Their reason isn’t because they think we can solve the problem, but because a world without any (young) people won’t actually be that bad (because “AI”—seriously). Poppycock, I say. We should be optimistic, but because we are are solving the problem already. Much like the secret ingredient to Mr Ping’s famous secret ingredient soup in the 2008 film Kung Fu Panda, the secret ingredient to solving fertility rates is nothing. Cultural homeostasis is doing its thing and we can expect a course correction soon.

I don’t want to downplay the years of work of pro-natal advocates here. My thesis here is that the cultural tide is already turning. In other words, the pro-natalists have succeeded in their task of highlighting the issue. Pro-natalism is the most human and the most common sense position to have. When suppressive cultural forces are removed, as they are being, the eons-old status quo—that having kids is a good thing—will resume.

I also don’t want to downplay the capacity for top-down solutions to ameliorate our demographic issues. There will undoubtedly be a further dip before we see a rise in birth rates, and economic incentives or biotechnological assistance will be important buffers to smooth out the ride as culture turns pro-natal again.

The bottom line is that having kids is just a good idea—for yourself and for others. When allowed to spread, good ideas are culturally infectious. Once the cultural appetite for pro-natalism is there, policies will follow, our ever-improving fertility tech will gain usage, and the pendulum will begin to accelerate in the other direction. People spoke of a “vibe shift” when Trump won the presidency last year: a shift towards nationalism; a resurgence of masculinity; a return of American exceptionalism. I think pro-natalism is hitching a ride.

Share

Subscribe now

Seeing us struggling, the Afrikaner swaggers over, picking up a small 2.5kg plate on his way. The menacing oke doesn’t have to say anything for us to back off. With no warning, he furiously smacks the weights on the tree with the in-hand 2.5kg plate. He turns to the scrawny teenagers next to him. “Violence. Solves. Everything.”, he blurts in his crunchy accent.

Apparently it was both leg day and arrector pili day for me.

Honeoye Lake, Upstate New York, USA, 2025. My girlfriend, Anna, and I are at an engagement party at a lake house co-owned by the family of the groom-to-be. It’s close to midnight. I’m locked into an engrossing conversation with Anna and about four of her friends from high school. We’re flirting with conspiracy theories. 9/11? Inside job. Moon landing? Fake (van Allen belt and wobbly flag mainly). Epstein? Alive and a CIA asset. Israel comes up (does it ever not?), and I can tell I’m more pro- than the others. Everyone in the circle is to the left of me, but the conversation is interesting, respectful, and fun.

People are speaking honestly now. Someone queries the likelihood that Israeli intelligence didn’t know about Hamas’ October 7th plans in advance, given their phenomenal reconnaissance and infiltration capabilities vis-à-vis the Hezbollah pager shenanigans. A risqué hypothesis. I offer that, even so, such an evil hypothetical turning-of-a-blind-eye by the Israeli deep state depends on the reliable evil of Hamas itself, and in no way diminishes their deliberate acts. Agreement. “Still”, interjects one of my interlocutors—a 23-year-old woman who currently lives in LA, “sometimes violence is literally the only way to change things”.

Laughter plus agreement from the crowd. Laughter plus nervousness from me. I’m genuinely taken aback. Does my generation really think violence is the only real way to get what you want politically? I’m too jet lagged and dehydrated to rebuke. I decide to hold my tongue, but not my pen!

I. THE VIRTUALISATION OF VIOLENCE

Before we answer whether political violence is a justifiable means to certain ends, I think it is important to consider why we are asking the question in the first place.

There is a paradox in our modern age. Violence seems to be everywhere and yet nowhere. The magic light box in our hands shows us devastation in Ukraine, bombs dropping on Gaza, and Druze shot and thrown off buildings in Syria. At home in the west, we’ve see an alarming increase in lone-wolf attacks (on Brian Thompson, Trump, Israeli embassy staffers, Paul Pelosi, Josh Shapiro, anti-Hamas charity workers, a Blackstone executive, et cetera) and mass protests that turned violent (post-Rudakubana riots in UK; BLM riots, J6, anti-Israel campus riots, LA anti-ICE riots in the US). Violence seems to beget violence, as negativity bias and the attention economy drives both classical and new media to push stories of violence wherever they can find them. Indeed, the word “violence” is now contained in written publications more frequently than it ever has been since the invention of the printing press, with the sharpest rise in its frequency seen in the past 40 years or so.

On the other hand, the chances of Americans experiencing violent crime has decreased by at least 50% since the 1990s1. Participation in full-contact American football has dramatically declined over the years, while participation in its non-contact counterpart, flag football, is on the up—indicating a move away from violent sport rather than sport in general. Parents coddle their kids more than ever, and safetyism has metastasised through society: the rate of non-fatal unintentional accidents plummeted by over 50% between 2011 and 2020. Of course, it has also been a very long time since the west has been at war (I’m discounting western involvement in other peoples’ wars here). It is undoubtedly true that the average westerner today actually experiences far less violence than they have at any point in time.

There are a number of things going on here that can explain this gap between the perception and the experience of violence. Not only do news and social media actively or algorithmically push footage of real violence to garner attention (aka, ad revenue), but the ability of films and video games to depict fake violence has been enhanced with technological advancement too. No longer do we see an actor swinging at the bad guy while simultaneously stomping his foot on the ground to mimic the sound of a punch. Instead, our films depict hyper-realistic violence that is indistinguishable from real violence. Indistinguishable except for the fact that it is only seen and heard on an OLED with surround sound, not actually experienced. Further, the psychological association of realistic violence with entertainment has led to a marked aestheticization of violence. As such, our collective psyche is numb to displays of violence. While this hasn’t directly led to an increase in the perpetration of violence, as was previously erroneously thought, it has likely conditioned us to think that violence is inconsequential or, worse, cool.

Beyond shifts in our media environment, academia has long been attempting to broaden the definition of violence. Critical Theory zealots will have you believe that words are equivalent to violence. Far from a fringe hypothetical, restricted to the ivory towers of institutionalised Leftism that we call universities, this nonsense has permeated through the chattering classes, culminating in the invention of the concept of “hate speech” and, in the UK, the spending of taxpayer money on the pursuit of “non-crime hate incidents”. Careful though. If you think staying quiet gives you protection you’d be incorrect. Silence is violence now too.

The shunning of masculinity ties in here too. Men are more physically aggressive than women (this is more true in societies with greater gender equality, by the way). As a consequence, masculine bullying is physical and masculine play is rough-and-tumble. Meanwhile women, who statistically have the same propensity for anger as men, tend to aggress via gossip, reputation attack, and social exclusion. The past couple decades have seen a combination of the deliberate quashing of masculine aggression and the inadvertent enablement of feminine aggression via the digitalisation of human-to-human interaction. This enablement is evidenced by huge rises of cyberbullying over time, with higher rates of its perpetration among teenage girls than boys.

Call me old-fashioned, but I sometimes wonder whether an adult intervening in every scuffle between 8-year-olds has benefitted or harmed society. Allowing kids to settle their own differences teaches agency and resilience, and perhaps some level of mild violence between kids, particularly pre-adolescence, teaches of the seriousness of violence to the boys and girls engaging in and observing it. As an aside—this is also one of the many reasons I am a proponent of contact sports among teenage boys, despite the valid arguments against them (e.g., risk of concussion).

The resultant world is one where men and women are unaware of their own and each others’ capacities for violence. The average woman underestimates the extent to which the average man can physically overpower them, which only serves to increase the likelihood of violence against women. Meanwhile the average man, unsure of their own ability to physically protect themselves or others, feels subconsciously emasculated. This dearth of violence creates a yearning for it by both sexes. It is why men are increasingly watching combat sports, or why they were so fired up by Trump’s actions in Butler, PA. Equally, it is why women are increasingly valorising and fetishising political violence, from Luigi Mangione to Hamas.

Humans are not rational beings by nature. Our emotional reaction normally dictates our opinions and actions, and we rationalise this post-hoc (known as “motivated reasoning”). The virtualisation of violence creates the emotional headroom for why, in 2025, there seems to be such a keenness to justify violence.

II. POLITICAL VIOLENCE IS BAD

They have the right to protest

One argument you’ll hear from those arguing that violence is sometimes necessary to achieve political goals is that “terrorism” is just a label used to discredit one side of a conflict. The problem is, terrorism has a definition: “the unlawful use of violence and intimidation, especially against civilians, in the pursuit of political aims” per Oxford Languages. To be fair, this is a broad definition, but legal definitions are much more precise, as we will come on to.

Some cases are right on the edge. Recently in the UK, members of the radical activist group Palestine Action broke into a Royal Airforce base and vandalised military aircraft. This was illegal, arrests have been made, and the UK’s Labour government have subsequently proscribed Palestine Action as a terrorist organisation. Yet anti-Israel groups from Palestine Solidarity Campaign to Amnesty International UK have expressed outrage that the group be proscribed as terroristic. PSC’s statement included the following:

Palestine Solidarity Campaign (PSC) utterly condemns the grotesque decision to proscribe Palestine Action as terrorists, which threatens all of our freedoms and democratic rights. The government has chosen to redefine the meaning of terrorism in a way that serves to criminalise dissent.

If this ban remains in place, it will be unlike any previous proscription. The vast majority of people rightly understand terrorism as involving acts of violence directed against civilians to achieve political ends. We know that the real terrorists and criminals are those who continue to facilitate Israel’s atrocities against the Palestinian people.

The problem is, the government has not redefined anything. Ironically, PSC have actually sneakily tried to redefine terrorism themselves here—by disingenuously appealing to what the “vast majority of people rightly understand”. In actuality, the Terrorism Act 2000 defines terrorism as any of the following:

• serious violence against a person

• serious damage to property

• endangerment to a person’s life

• serious risk to the health or safety of the public or a section of the public, or

• seriously interfering with or disrupting an electronic system.

If any of these actions are carried out in a manner “designed to influence the government or an international governmental organisation or to intimidate the public or a section of the public”, and “for the purpose of advancing a political, religious, racial or ideological cause”, then it is terrorism.

Whether the £7m in damage caused to military aircraft constitutes “serious damage to property” is therefore the question of importance here, and it will be decided by a judge, not the government.

But this is all semantics, and irrelevant to the point. Palestine Action’s actions were illegal—terroristic or not. Breaking just and moral laws in the name of some perceived higher moral cause is unacceptable, and should be condemned rather than celebrated by these anti-Israel groups. Get on the right side of morality at home, and then people might be more inclined to hear your argument about morality abroad.

For those arguing that the deeds of Palestine Action are simply an extension of freedom of speech: firstly, welcome to libertarianism, nice to have you; secondly, there is no way that these people genuinely can’t see the difference between a peaceful protest (typically involving working with local police) and breaking into a military base and vandalising aircraft in the name of a political cause. If these people wish to stand behind acts of alleged terrorism because they believe, for some reason, so very strongly in the cause du jour, then they should at least have the gall to admit it. That’s fine. You can be like that guy on Jubilee who so gleefully admitted to his support for fascism. I have a lot more respect for someone who owns their position, even if I disagree with that position. That way we could at least have a productive and honest debate (or in Medhi Hasan’s case, refuse to). Pretending that breaking the law is fine, or that terrorism isn’t terrorism, though, is just disingenuous.

Means and ends

When squeezed into a corner, the average Palestine Action supporter (2nd year, Art History, Bristol Uni, probably called Tilly) may rebuke, “yeah fine, it’s technically terrorism and I guess it’s lawful that they got arrested, but it didn’t hurt anyone and the Palestine Action activists are morally justified”. In other words, the ends justify the means.

Herein lies the problem.

The thing is, the capacity for a human to viscerally identify with the importance of a certain cause is boundless. This means that the limit for how far one can justify pushing the means to achieve the ends of the cause is also boundless. Ends justifying means is literally the reason for every single atrocity that has ever stained the history of humanity.

Hitler justified gas chambers with the morally justified goal of a racially pure utopia. Mao justified mass starvation and tens of millions dead with the morally justified goal of a classless communist utopia. Pot justified genocide and forced labour camps with the morally justified goal of a pure agrarian utopia. The Aztecs justified mass child sacrifice with the morally justified goal of prosperous harvests and a bountiful utopia. Yhwh justified the complete genocide of the Amalekites with the morally justified goal of a philosemitic and ethical utopia. I’m justifying this painfully long list with the morally justified goal of encouraging virtue and rationality among my readership.

The point is, the slope that takes you from agreeing with senseless damage of property all the way to the killing of innocents is relatively devoid of friction. For those who are inclined to publicly moralise, they must privately moralise too and reject any ethical deviation along the way to their ends.

If seeing a video clip of an emaciated child on your phone, Reel or not, invigorates your sympathies particularly strongly, virtue ethics might be a good salve to prevent you descending into supporting or committing comparable horrors.

La Rochefoucauld wrote, “We often do good so that we can then do evil with impunity” (Réflexions ou sentences et maximes morales, 1665)

Equally, we often convince ourselves that we will do good so that we can now do evil with impunity.

One man’s terrorist is another man’s freedom fighter

Israel is an apartheid state, so we’re told. This means that when Hamas are accused of terrorism, but our friend Tilly wishes to deflect from their violence against innocents, she invokes Madiba.

“Did you know that they used to label Nelson Mandela as a terrorist?”

Yes, Tilly, they did. That’s because Nelson Mandela was a terrorist. Before he became the great leader that he eventually did, he co-founded and led uMkhonto weSizwe (MK), the military arm of the the African National Congress (ANC). Working with and funded by the Soviets, Iranians, Algerians, and others, they carried out bona fide terrorist attacks, including killing children in bomb attacks, and torturing and executing dissidents and prisoners. They even condoned necklacing, with Mandela’s wife directly encouraging the horrific act. To call the ANC a terrorist organisation at this point was no sleight against Mandela—it was factual. It is no surprise that the “freedom fighters, not terrorists” mantra currently espoused by Hamas supporters was originally promulgated by MK, who were allies and collaborators with Hamas’ predecessors sixty years ago.

The lesson we should learn from Mandela isn’t that bad guys in power always call good guys opposed to them terrorists to silence them. It is that political violence, asides from being unacceptable morally, actually postpones, rather than expedites, achieving one’s political goals. Mandela realised this when on Robben Island. He realised that guerrilla warfare would never break the back of apartheid—terrorist attacks by MK led only to greater and greater apartheid state powers, infringements on liberty, pro-apartheid indoctrination of whites, and deepening racial division in South Africa. Mandela’s philosophical evolution while in prison equipped him with greater empathy for his oppressors (not to be confused with sympathy) and enhanced diplomatic acumen. Whether the track record of sabotage and violence was a necessary backdrop to the diplomacy of the 90s is an unanswerable question, and many might still defend the historical context of MK. Ultimately, contrary to the advice from my Drakensberg gym buddy, violence didn’t solve anything, and it was Mandela’s secret negotiations with the South African government, combined with huge economic pressure due to US-led sanctions, fully enforced by Bush Sr. in 1989, that actually succeeded in ending apartheid.

Once Mandela committed to quashing the violence in his own political movement and dove wholeheartedly into rational diplomacy, the ANC became a serious political movement—one capable of leadership and governance, not just of protest and terrorism. He and F. W. de Klerk were then able to work together to create a better South Africa, not just destroy a bad South Africa.

III. THE ONLY OPTION?

Violence is a skill issue

Mandela’s example highlights how resorting to political violence is, ultimately, a skill issue. As he grew as a person, in wisdom and empathy, his ability to engage in diplomacy became enhanced. This gave him a better option than the unproductive violence of the 70s and 80s. Not just better for the innocents (mainly blacks) whom would then no longer be killed, but better for his own cause.

If you’ve ever been to a pub in a slightly dodgy area (my American readers can imagine a dive bar in BFE), you might have witnessed a bar fight. If you frequent Substack, I’ll pay you the compliment of assuming that you have never been in one of these fights. That’s because, to be blunt, it is typically stupid people who get into fights. That’s not necessarily because smart people don’t get into disagreements in pubs. It’s because smart people are capable of dealing with such disagreements non-violently. If Crick insisted on parallel strands over a pint at the Eagle, Watson, instead of engaging in fisticuffs, could deploy the empathy and the verbal skill required to persuade him of the anti-parallelness of DNA instead.

Those who resort to violence typically lack the capability to resolve issues non-violently. To them, violence is the only option, but with sufficient competence, violence is never the best option, let alone the only option. Empathy—the ability to inhabit the opinions of others—is an intelligence-correlate (and, I’ll stress again, nothing to do with sympathy). Violence occurs when intelligence and empathy are lacking. These same principles apply from the microcosm of a Spoons smoking area all the way up to international geopolitics, because at the heart of all political violence is the psyche of the individual human.

Regarding Israel-Palestine, there was a really good shot at longterm peace in the 90s when Clinton brought Arafat and Rabin together to sign the Oslo Accords. Why did the violence continue afterwards? Because both Palestinian and Israeli leadership were incapable of controlling the extremists on their own side and incapable of empathising with the extremists on the opposite side. This led to an abandoning of the peace process and a continuation of the cycle of violence. Arafat was no Mandela—Rabin no de Klerk. A true end to the Israel-Palestine conflict will surely come from better diplomatic competency, and not continued violence.

Luigi Mangione seems like a smart guy. If he was smarter, better informed, or just wiser, he would not have resorted to violence. He would have realised that UnitedHealth Group—far from the epitome of corporate greed as Mangione would have us believe—has tiny profit margins. UnitedHealth Group typically reports ~6% net profit margins, but in 2024, the year that Mangione allegedly murdered the CEO of their insurance subsidiary, this was 3.6%. For context, the median net profit margins of Fortune 500 companies is around 7.2%. Meanwhile time-vampires like Meta rake in 40% net profit margins (!).

Not only is the United Healthcare business model not particularly exploitative, but their CEO was, by all accounts, much more working-class than bourgeois Mangione. Perhaps worst of all is the complete delusion and overestimation (i.e., lack of skill) displayed by Mangione with regard to the actual positive impact of his actions on the health insurance industry. By my reckoning, his positive impact here has been precisely zero.

How do we get what we want?

So far, I’ve argued that violence is always immoral, per virtue ethics, and that it is largely ineffective, per historical inference. Does that leave us high and dry? Is our only choice to leverage stoicism/buddhism to retreat from our material circumstances and transcend worldly oppression? While it is certainly possible to simply choose to mentally overcome the woes of reality (see: Frankl, Goggins, Mandela, Edgley), we can also change our material reality via better means than political violence.

Inspired by RomneyCare in 2006, the Affordable Care Act has helped to halve the proportion of Americans without health insurance since 2009. Further, the adoption of its expansion by some states and not others in 2014 allowed researchers to infer that ACA expansion saved around 11.3 lives per 100,000 between 2014 and 2018. Yet it is not without its problems. By destroying any semblance of a free market, the Act, while giving the poorest Americans coverage, strained providers by forcing widespread adoption, crushed the middle class with ever-rising insurance premiums, added layers of wasteful bureaucracy to healthcare, and hasn’t actually made much of a dent in rates of medical bankruptcy.

If Mangione were smarter and more virtuous, he might have thought of solutions to these complexities and run for office, campaigning on fixing US healthcare constructively. Alternatively, he could have identified arguably the most exploitative link in the chain of US healthcare—the pharmacy benefit managers—and come up with a way to disrupt their grip on drug prices via competition. If actually creating something of value for society—the remit of entrepreneurs—and the necessary sacrifice that comes with it isn’t for Luigi, he could have simply worked for Mark Cuban, who has done exactly what I have described with his co-founder Alex Oshmyansky. Their company, Cost Plus Drugs, circumvents PBMs, adds transparency, and sells much cheaper drugs to Americans. Billionaire Cuban is doing infinitely more to improve the US healthcare system than Mangione and his 3D-printed pistol.

The ACA and Cost Plus Drugs are cases in point for how best to drive societal change non-violently. The mechanisms at play here are democracy and innovation.

Democracy

It is rare to defend democracy these days. Public satisfaction with democracy and trust in government is tanking, which we can know only because we live in a liberal democracy. Yet whether this reflects an accelerated failure of democracy to bring about positive societal change is a separate question. Perception can, as we have discussed in part I, be quite divorced from reality. A unique feature of democracy versus other forms of government is that it is slow. I suspect that the recent loss in trust in democracy might be due to the coalescence of the slowness of democracy with the shrinking of our collective patience in the modern era.

The dopamine economy has minimised time horizons. Music and film might be instantly streamable now, Lebanese delicacies might appear outside your apartment thirty minutes after you had the thought, the entire corpus of human knowledge might now be constantly at our fingertips, but the pace of politics is the same as it’s always been. This slowness of democracy isn’t a bug, though. It’s a feature designed to instil robustness and stability—to prevent a faddish zeitgeist, even if democratically mandated, from irrevocably harming a country.

In reality, democracy works. You can vote for societal change quite effectively. Most people who insist that this is impossible simply have shitty ideas that a minority of their fellow citizens are persuaded by. Once again, there is a skill issue here, and the skills in question are patience and persuasion.

In 1807, William Wilberforce received a standing ovation in the UK House of Commons after delivering a speech that was followed by a 283-16 vote in favour of the abolition of British participation in the slave trade. Twenty six years later, democracy struck again, this time ratifying the active abolishment of slavery globally—at great financial cost to the British Empire.

Centuries later, the ballot box retains its power. In 2023, Ohioans voted overwhelmingly to legally enshrine abortion rights in their deep-red state. JD Vance took this on the chin, and it persuaded him to slightly soften his stance on the issue. Since he is a likely candidate for #48, this apparently minor state-level democratic vote could bring about a longterm, stable cultural change that depolarises the issue of abortion in America, bringing the cultural milieu here more in line with that in the UK, for example.

National elections too, despite polarisation, are still powerful opportunities for change. In the US, complaints against Donald Trump are mainly centred on his actions, not his inactions. This means that he’s using the power that the people gave him to exert their will (even if half the country disagree). In the UK, threats of a Reform victory in four years time are proving enough to push Keir Starmer to reverse course on issues of immigration and actually respond to the will of working class Brits—Labour’s historical but vanishing voter base. In this case, even the prediction of a vote is tangibly powerful.

There is no denying that, from women’s suffrage to Brexit, the cogs of democracy slowly but surely grind away. Liberal democracies offer stability and freedom, but also a powerful tool for societal change.

Innovation

Perhaps my arguments for democracy are falling on deaf ears (or impatient brains). If pace of change is the primary concern, then innovation is the tool of choice. Innovation has a high skill and effort requirement, making it rare without sufficient incentive. Capitalism solves this incentive issue, offering fair value (definitionally fair, in a free market) for entrepreneurs who solve difficult problems. Mark Cuban’s new drugs company is a neat innovation, and a classic example of how capitalism enhances market efficiency, making things better for everyone. Yet this is small fry. The ceiling for change via innovation, particularly if it is technological, is vast.

In the mid-1800s, over 500,000 barrels of whale oil per year were being produced to light American streets and homes. The innovation of coal gasification and then the discovery of petroleum in Pennsylvania briefly made whale oil redundant, saving around 10,000 whales per year. After WWII, however, whaling resurged as innovation in refinement of and uses for whale products (in cosmetics, textiles, foods, and lubricants) drove demand, while fossil fuel-powered vessels and advanced harpoon tech raised supply. By the 1960s, the whaling industry had brought some species of whale almost to extinction, killing more than 70,000 whales per year.

It turns out that some people care a lot about whales not going extinct. But it wasn’t a whale-hugger assassinating the CEO of Big Whale™, or Just Stop Wh-oil sabotaging whaling vessels that saved the chubby cetaceans. It wasn’t even democracy, the efforts of which—via big international treaties and moratoriums—ultimately proved flaccid in the face of economic incentives. What saved the sperms, minkes, and humpbacks for the second time was yet more technological innovation in the form even better alternatives, including palm oil and modified petroleum products. These allowed companies to abandon whale oil for good, which had become harder to obtain due to whale scarcity. Once innovation allowed for it, the free market was quick to reallocate capital away from whale oil, and whale populations have since recovered.

Nowadays some people care a lot about increases in atmospheric carbon dioxide, responsible for significant greening of the earth in recent decades and causing the dreadful 0.06 °C per decade increase in global average surface temperature we’ve seen in the past couple centuries or so (perhaps).

While some vandalise priceless art to signal to others that they don’t like bad things, others are innovating.

Jonah Messinger and Alex Trembath of The Breakthrough Institute put it well in their essay How Fossil Fuels Saved the Whales, Twice:

“The route to robust environmental conservation, fit to endure industrialization and abundant human flourishing, lies in new and improved technologies that are not merely better with respect to environmental metrics, but genuinely superior on techno-economic grounds.”

Environmentalism is one realm in which technological advancement is poised to stoke change—from de-extinction efforts to small modular nuclear reactors. Yet many other politically charged issues can also be solved with technological progress.

While the motives aren’t fully elucidated, it seems likely that her pro-choice stance on abortion is what motivated the assassination of Minnesota state representative Melissa Hortman and her husband in their home earlier this year. If pro-lifers wish to reduce the number of unborn humans being killed, killing fully grown humans seems like a bad way to go about it. Alternatively, developing the technology to support increasingly premature neonates might be one techno-progressive solution. Since most moderates are only fine with abortion up until the point of viability, leveraging biotech to enable foetal viability at earlier and earlier stages of pregnancy will persuade more and more people against abortions at previously-unviable-but-now-viable stages. Foetal viability is defined technologically, not biologically.

Lab-grown meat is another classic example. Vegans who wish to reduce the number of animals killed for food worldwide might wish to do everything they can to make this as good as it can be—economically, gastronomically, and nutritionally. Steak enjoyers need to be pulled, not pushed. That said, perhaps the goalposts will just shift, and vegans, in search of a moral high ground, will stretch their anthropomorphising tendencies to individual cells after they watch a video of a sad innocent apoptosing fibroblast. I digress.

The point is, technology and innovation solves problems—including moral problems. Yes, sometimes it can create new unforeseen problems along the way, but the engine of progress is perfectly capable of digging itself out of any holes it finds itself in, as the tech-driven oscillations in whale populations over the past two centuries proved. The only reason more people don’t solve problems via innovation is because it is difficult and requires skill.

Next time you see an Extinction Rebellion protestor blocking a road in the middle of your town, remind yourself to judge them not only for their immoral utilitarian worldview, which they have in common with utopian idealism hall-of-famers Mao, Hitler, and Pot, but also for their laziness, stupidity, narcissism, and impetuousness, which withholds them from actually doing anything creative, constructive, and good for the environment.

Tonbridge School, Kent, UK, 2010. It had snowed over the weekend. Day boys didn’t have to come in today but we still had lessons. Obviously we weren’t going to be playing any rugby during games this afternoon. Mr Myslov still took us down to the lower fields, though, and we spent a few hours playing in the snow. A few of us decided to build a massive (mahoosive, actually) snowman. Soon rival snowmen were being constructed. Pubescent boys can get pretty competitive, and we rolled, compacted, stacked, and shaped furiously. Novi didn’t include any structural engineering classes, and one group’s monstrosity was becoming a little top heavy. Suddenly it collapsed. Angered by their own failure and chilblains, the boys proceeded to pelt other erect snowmen with snowballs until they, too, were toppled. A skirmish erupted and soon no men were left, just us boys. Mr Myslov wasn’t pleased. In his thick Russian accent he boomed, “No no no children! We must create! We must not destroy!”. The sun was setting and we trudged back to our boarding houses sulkily and drenched.

Political violence is stupid, always immoral, and it rarely works. Indeed, it can sometimes have the direct opposite of the intended effect, as with Crooks’ failed assassination attempt on Trump, which almost certainly helped orange man get elected. Even when violence achieves its primary goal, the change brought about is rarely precise and human prosperity rarely improves as a result—almost every “successful” coup d’état in history failed to bring about longterm prosperity2.

We should all wish to live in a more prosperous and more moral world. Rejecting the utopian allure of utilitarianism and steadfastly opposing violence, based on Aristotelean virtue, secures a moral future. But rejecting violence on moral grounds doesn’t come at the expense of material wellbeing. Liberal, democratic, capitalist societies offer many greater tools for change—greater in precision, permanence, and scale. Political violence, reached for by those devoid of persuasive skill or innovative creativity, is never necessary, always unethical, and rarely solves anything.

Letter to SoTA

The Society for Technological Advancement (SoTA) invited me to contribute to their journal, SoTA Letters. It is a collection of letters focussed on problems and solutions relating to British technoscientific progress. I really like their mission, and happily obliged. You can read my letter, “Reclaiming the British University”, here.

In general, I try to rail against negativity bias (hopefully all of my writing reflects this). In my letter to SoTA I point out how I think universities have gone astray (particularly elite British ones), but I try to focus on solutions as much as problems (as always). I hope you find the letter to be rousing—there are always reasons to be optimistic.

Scientific article + interview

Some of my PhD work was recently published online in the Journal of Cell Science. The full (open access) research article is here. JCS also wrote a really nice one-paragraph research highlight here. They also published an interview with me, here, which is part of their “First person” initiative to highlight early career researchers.

I distilled the main findings of the paper in an 𝕏 thread, here.

Essays in the works

Much of my writing time is currently being sapped by postdoctoral fellowship applications and other research articles from my PhD. That said—I have some exciting essays in the pipeline at the moment. You can look forward to future articles on political violence, ethical veganism, 16-year-old voting, and population collapse.

Leave a comment

Introduction

Cells are the smallest indivisible units of biology. In my biased view, every biological process and every disease is best understood through the lens of cell biology. Seemingly macroscopic physiological phenomena, like day-night rhythms in our behaviour, ultimately boil down to cells (circadian rhythms are maintained by autonomous molecular “clocks” that run indefinitely in each of our cells). At the other end of the spectrum, atomic or molecular processes are devoid of biological importance until incorporated into a cellular process. For example, Franklin’s famous X-ray crystallograph (“photo 51”) was only useful and interesting to Watson and Crick because they were trying to understand how cells store and read information.

I think understanding how cells work is really important. And really interesting. Here, I want to give an overview of what I consider to be the seven wonders of the cellular world. These are all macromolecular structures inside of cells that each play a key role in sustaining the cell, and, by extension, life. Of course, organelles can be described as wonders of the cellular world (e.g., the Golgi Apparatus or the mitochondrion) but I thought sticking with large molecular complexes keeps the game fairer. Hard to compete with the powerhouse of the cell itself! Factors that come into play for making this list include aesthetic beauty, mysterious aura, remarkable function, and essentiality for life. If you’re not a scientist, this list will hopefully give a broad intro into cell biology, with each of my seven picks taking you on a journey through the cell both spatially and functionally. I say all seven picks, but I mean all six since the last one really is a complete mystery even 50 years on from its discovery. With that in mind, let’s learn some cell biology!

1. The nucleosome

Cells synthesise two main types of complex biopolymers: nucleic acids (like DNA and RNA) and polypeptides (which form proteins). Watson and Crick’s discovery1 of the structure of DNA helped to lay the groundwork for Crick’s later description of the “central dogma of molecular biology”. This key concept basically states that information in DNA can be copied into RNA and, in turn, into protein. But information in proteins cannot go back into nucleic acids (DNA/RNA). Information flows this way because DNA is primarily a stable source of genetic information, and proteins (long chains of amino acids folded into complex 3D structures) are the functional units within cells—they’re the things that actually do useful stuff. Cells transcribe DNA into shorter chunks of RNA, which mostly act as intermediate messengers, and this RNA then gets translated into protein.

The genome is the total DNA code needed to produce every single type of protein that any cell within an organism could need (and much more non-protein stuff, actually). This is a lot of information—for humans it’s about 6.4 billion bits of information in 3.2 billion base pairs, or “letters” of DNA. If the chromosomes were unravelled and joined end-on-end, each human cell would have about 2 metres of DNA. So how the hell does the cell stuff this all into the nucleus, which is about 5 micrometers in diameter at most? Enter: the nucleosome.

Discovered in 1974 by married couple Don and Ada Olins (who sign off letters and emails with “DnA”), the nucleosome is formed by proteins called histones stacked into a cylinder with DNA wrapping around the outside in a coil. If the nucleus of a cell is burst open with pure water and the contents imaged by an electron microscope, you can see nucleosomes as “beads on a string”, as in the image above. The nucleosome is the most fundamental storage unit of DNA in cells and is essential for large genomes and, therefore, complex life.

The insane level of spatial compression needed to cram the two copies of the genome into every one of our cells doesn’t solely rest on the shoulders of the nucleosome though. These coils of DNA themselves form coiled stacks, called chromatin fibres, which in turn are further coiled, compressed, and folded. Chromatin, which describes the overall organisation of DNA in the cell, is not merely static though. The organisation of these nucleosomes is dynamically regulated to control which genes can get turned on or off in a cell. Opening up the chromatin allows greater access to the naked DNA not coiled up in the nucleosome, which allows that DNA to be read and copied into RNA. It’s basically like those archives or libraries with sliding shelves that students try to through before they close. If you want to take out a particular book (in this analogy: a gene), then you open up the channel between the shelves where it is located (in this analogy: chromatin opening up for transcription).

Given their crucial role in affording cells their incredible information density, as well as their simple but beautiful beads-on-a-string arrangement, nucleosomes gain a spot on this list.

2. The nuclear pore complex

We’ve covered how DNA, which is kept in the nucleus, is copied in short chunks into RNA, which can then be used as the template from which to make proteins with useful functions. These proteins don’t get made in the nucleus though—it’s way too crowded. Further, the nucleus is basically a fortress. It’s encapsulated with a thick double membrane so that stuff that can damage or interact with DNA in undesirable ways is kept out, and the DNA itself is kept in. If any DNA were to get out, the cell would think it’s under attack from a virus (which is the only other scenario where you get DNA in the cytosol—the soup of stuff outside of the nucleus), and the cell would probably kill itself. So the RNA gets made inside the nucleus but then must leave and get into the cytoplasm where proteins are made. The gateways between the nucleus and the rest of the cell come in the form of many nuclear pores. Far from just holes in the nuclear membrane though, these are stringently regulated by the nuclear pore complex.

Made up of a bunch of different individual proteins called nucleoporins, the nuclear pore complex is a massive and complicated machine. It sits in the nuclear membrane and forms a pore in its centre which allows small molecules to freely diffuse, but bars anything of size and functional significance from passing. For these molecules to pass through, like big proteins trying to get in or freshly transcribed RNA molecules trying to get out, they have to be tagged by a molecular chaperone that escorts them through and out the other side.

We still don’t know exactly what it looks like at the atomic level, mainly because the very flexible bits on either side are hard to resolve via crystallography or electron microscopy, but some solid efforts have elucidated the beautiful symmetry of the core ring structure, which it turns out is dilatable. Massive geometrically satisfying nightclub bouncers, these things are easily a wonder of the cellular world.

3. The ribosome

Recap: we have DNA in the nucleus, and we know that this gets copied into RNA which can then leave the nucleus through the nuclear pores, and this then forms the template for making proteins. As I’ve mentioned, cells are basically protein machines, spatially organised into lipid-membraned compartments, termed “organelles”. Proteins are biomolecules comprised of chains of amino acids that form unique 3D structures that give them functional properties. Haemoglobin is a protein than can bind to oxygen for its transport around the body. Collagen is a protein that confers tremendous tensile strength that is useful in tendons and skin. Histones and nucleoporins are proteins that you are already an expert on. So how does the cell make all of these proteins? That’s where the ribosome comes into play.

Ribosomes are massive ribonucleoproteins, meaning they themselves are made not just of protein, but also RNA—specifically ribosomal RNA, or rRNA. They assemble around a messenger RNA (mRNA) molecule, produced when the cell reads a gene in your DNA and wants to get the “message” out of the nucleus. Ribosomes loaded onto an mRNA then process along the genetic code, taking in matching amino acids for every codon of genetic message, and stitch them together to form a long polypeptide chain that then folds to become a mature protein. Since the ribosome converts the language of nucleic acids into the language of proteins, using triplets of RNA letters called “codons” to encode one of the 20 available amino acids, this process is called translation.

Since producing proteins is such a key feature of cells, these machines evolved a very very long time ago. This makes ribosomes in different organisms all pretty similar, all things considered. As such, rRNA is the go-to target of sequencing for microbiome research, since every type of bacterium in your gut microbiome will have a recognisable rRNA, making them a good candidate for comparing different bacterial species side by side. rRNA is also a really useful yardstick to help with quality control when sequencing the mRNAs inside a cell, a.k.a. “transcriptomics”—something biologists bloody love doing these days.

Each of our cells has tens of millions of these bulky molecular machines relentlessly chugging along at about 4 amino acids per second. These masterpieces of molecular evolution allowed proteins to first become a thing for life on earth—since the first semblances of life on earth probably started with just RNA. Next time you’re in the gym and notice that your muscles have gotten a little bit bigger, be sure to give a quick thanks to your ribosomes.

4. Microtubules

Once proteins are produced, they have to get to the right place within the cell. The cell has many ways of moving things around to where they need to be. Perhaps the fastest and grandest, though, is via microtubules. The highways of the cell, these are long hollow tubes formed of two protein subunits called alpha-tubulin and beta-tubulin. These dimerise and then stack up in a spiral formation to form tubes with pseudo-helical symmetry comprised of 13 proto-filaments (see image below to make sense of that). Microtubules can polymerise and also undergo “catastrophe” whereby they fall apart from one end. This allows them to grow and shrink, allowing the cell to regulate them dynamically.

Microtubules form “train tracks” that motor proteins like dynein and kinesins use to “walk” along, dragging cargo as they do. If the cell needs to transport something quickly and a long distance, it goes via the microtubules. Long distance high-speed transport is key for many cell types, but perhaps most especially neurons, which have long extensions from the main cell body to the neuronal synapse. These extensions, called axons, contain bundles of parallel microtubules that are crucial for the transport of lipid vesicles, mitochondria, and even mRNA. My favourite cell type, cytotoxic T lymphocytes (a.k.a. “killer T cells”), point one end of all of their microtubules at cancer cells that they engage with, and use the highways to deliver cytotoxic proteins to these cancer cells to kill them.

A quirky feature of the tubulin subunits of microtubules is that they have a lot of a certain type of amino acid in them that allow electrons to become delocalised. Because of the highly ordered arrangement of subunits, the region of delocalisation is very large. This gives microtubules the ability to harness quantum effects at larger scales — effects such as superradiance. This, combined with the fact that microtubules are really abundant and important in the brain, and many anaesthetics seem to work by targeting microtubules specifically, gives rise to the notion that microtubules might be the biological basis for consciousness and free will (a hypothesis that I discuss in greater depth in my essay on why I believe free will exists). Interestingly, one of the key proteins that seems to aggregate in Alzheimer’s disease is a microtubule-interacting protein.

These incredible structures were one of my favourite things to image under the microscope during my PhD. Their structure is beautiful at every scale, they have vital and diverse roles in different cell types, and they are still shrouded in plenty of mystery (what is going on in their hollow cores?!). All of this earns them a solid spot as one of the 7 wonders of the cellular world.

5. The proteasome

We’ve spoken about how the cell needs to churn out a lot of proteins to achieve its functions. But just creating proteins makes for a very messy soup. The cell also needs a way to deliberately get rid of unwanted proteins. Autophagy is one way of doing this, and it allows the cell to recycle lots of proteins to create new amino acids when cells are starved of nutrients. But autophagy is typically a cell-survival mechanism, and can be indiscriminate regarding the proteins that it targets. To target specific proteins for degradation with high precision to keep things running in perfect balance, the cell uses proteasomes.

The proteasome is a very large hollow cylindrical protein complex that targets proteins that have been tagged for destruction via a little molecular red flag, known as ubiquitin. Ubiquitinated proteins are unfolded and fed into the core of the proteasome by the cap structures (magenta in the image above). The core then has a series of proteases (enzymes that chop up proteins) on the inside, using ATP to provide energy to break up the peptide chemical bonds. The hollow structure of the proteasome prevents these proteases from otherwise wreaking havoc on the rest of the cell’s proteins.

Recent work is uncovering how protein sequences, called degrons, lead to certain proteins getting destroyed. In some cancers, these sequences are mutated, meaning proteins that would otherwise get chopped up by the proteasome don’t, and this can lead to cancer. Leveraging degrons to get cells to destroy cancerous or otherwise pathogenic proteins is a promising area of future research. The garbage disposal of the cell. Vital, beautiful, and with great biotechnological promise—proteasomes are a wonder of the cellular world.

6. ATP synthase

To fuel nucleosome assembly, nuclear pore entry/exit, ribosomes, microtubular transport, proteasomes, and most other active cellular processes, the cell uses adenosine triphosphate, or ATP. This little chemical is a great currency for energy, since the third phosphate group can be chemically reacted on and off, either requiring or generating a packet of chemical energy. When proteins use ATP, they cleave that third phosphate and get energy as a byproduct of the reaction. As such, the cell needs to constantly upkeep its ATP stores to make sure there is enough of it diffusing around that proteins can use it all of the time.

This is where ATP synthase comes in. Located in the inner membrane of mitochondria, these molecular turbines are remarkable machines that add the third phosphate to ADP molecules. As you might have guessed, this requires energy. This energy comes in the form of a gradient of hydrogen ions, or protons, which mitochondria maintain across this inner membrane. In fact, the whole point of mitochondria is to establish this gradient of proton concentration. The mitochondria have two membranes. They pump loads of protons into the gap between these membranes, using the chemical energy from our food and oxygen.

The ATP synthase molecule sits in the inner membrane, which is impermeable to ions. There are gaps in between subunits of ATP synthase that allow protons to slide in, attempting to get across the membrane. As they do so, the whole machine rotates like a turbine. This rotating motion is then used elsewhere in the protein complex to move atoms around to whack on the third phosphate group to an ADP to generate ATP.

It’s truly insane. A literal molecular turbine. The analogy of hydroelectric electricity is a good one. Getting the water up there is done by clouds and rain etc, but in the cell its the mitochondria doing that work. The hydroelectric power plant is the ATP synthase. Hydrogen ions flow “downhill” from high to low concentration and spin the turbine, which whizzes away at around 250 revolutions per second (significantly faster than the average jet engine at around 150 revolutions per second).

The dude who figured out the structure and mechanism of this beast won the Nobel prize the year I was born, and he still works in the building I did my PhD in. Despite being as old as me, there is still some mystery surrounding ATP synthase. We don’t exactly know what the outside jacket of the turbine looks like, which is kind of important for knowing exactly how the hydrogen ions generate such phenomenal RPMs.

This marvel just has to be visualised to be fully appreciated, and I’d highly encourage you to watch the embedded video below when you get five spare minutes (perhaps after reading about the 7th wonder below …).

7. The vault

Finally, we come to the most mysterious wonder of them all. In the 1980s, researchers were trying to isolate different kinds of vesicles (lipid-encapsulated little droplets inside cells). They noticed these little specks in their electron microscopy samples. Around triple the size of ribosomes, which held the title for largest protein complex at the time, they assumed these specks were just impurities. Turns out, they were their own thing entirely. Once isolated and imaged properly under an electron microscope, their structure looked a bit like the vaulted ceilings of cathedrals, and thus they were named vaults.

Vaults are huge, beautifully symmetrical, hollow structures. The outside is made of a single protein, called major vault protein (MVP), of which there are 39 copies per half. Mammalian cells have loads of these things—hundreds of thousands up to millions in certain immune cells. But at the same time, plants, fungi, and bacteria lack them. The fact that single celled amoebae have them but things like insects don’t implies that they evolved a long time ago but were since lost by the multicellular organisms that today lack them.

If they’re so ancient and so massive (making them expensive for the cell to make), and so similar across the organisms that have them, then they must be important right? But if you get rid of them in mice, the mice are completely unaffected as far as we can tell, which implies no function at all …

We do have some hints to their function, but no coherent hypothesis satisfies all of the evidence. They tend to cluster around the nuclear pores, so they could be little transport pods for the cell. Their natural cargo seems to be unique RNA species, terms vRNAs, but what these do also isn’t clear exactly. There is some evidence they they help fight bacterial infections in the lung epithelium. Alternatively, given there are many more of them in macrophages, immune cells that migrate a lot, and also single-celled slime moulds which also migrate a lot, perhaps they have a role in cellular locomotion?

The guy who co-wrote the textbook on cell biology thinks vaults “may be part of the massive war with viruses, enabling cells and organisms to survive a type of virus that we don’t even yet know anything about”. While he can’t hide his fascination, he still refuses to put them in his textbook at all—I only found out about these things half way through my PhD!

The enigma endures, and scientists are still trying to solve the mystery. Some super recent evidence from May this year suggests that their movement up and down neurons might be important for learning and memory, but clearly this can’t be a universal function since theyre found in many cells, including single-celled organisms.

Ylvis are kept up at night by Stonehenge. Cell biology PhD students are kept up at night by the vault.

Subscribe now

Trump’s war was declared almost immediately upon his inauguration. It started with a proposed slashing of “indirect costs" rates for National Institutes of Health grants to 15% back in February. This has been recently copied by the National Science Foundation. Harvard has also had its tax-exempt status threatened by Trump himself and billions in funding frozen. The attack on universities isn’t just directed at Harvard, with Columbia University also seeing federal research grants slashed. More broadly, we are told that the NIH itself is “under siege”.

At the same time, Chinese science seems to be stronger than ever. The CCP has formed a secretive Central Science & Technology Commission, swanky new research institutes seem to be popping up all the time there, and the argument can be made that China is now the global scientific superpower in certain fields. Was the David versus Goliath story that we were told about DeepSeek and OpenAI just a foreshadowing of a much grander global shift in scientific hegemony?

I moved to the US to start a postdoc just before Trump took office this year. With the apparent seismic shifts in the scientific landscape, should I be thinking of getting the hell out of here? Should I just go back to Cambridge with my tail between my legs? What about doing some Naturwissenshaften at an MPI in Germany? Or maybe I should abandon the occident altogether and accept the generous CCP relocation package in a shiny new Chinese lab? When I sit back and think about where I would choose to apply for postdoc opportunities if I was finishing my PhD only now, I have to admit that there is only one clear answer: the US of A.

Media headlines telling of the destruction of science are often pre-emptively doomerist, lack key context, or blatantly over-exaggerate the reality of the situation. Here I outline why US science is strong and not really under any kind of existential threat. I make the argument that the US is the best place in the world to do science. Trump isn’t, can’t, and won’t change that, and you should be weary of the incentives behind those telling you otherwise.

Culture > politics

Before I parse fact from fiction regarding the alleged death of US science, I first wish to make a broad foundational point: In a democracy, culture is upstream of politics. The only reason the Trump admin is cutting anything related to science or academia is because that’s what people voted for. There has been growing distrust in institutions, and the institution of Science™ has become politicised, corrupt, and untrustworthy. Peer-review is pretty broken, there is a reproducibility crisis, and the big journals have basically become propaganda outlets for the Democrat party. The goal of dispassionate pursuit of truth in US science has been slowly diluted for decades. Even with these maladies, US science has remained dominant. An attempt to reinvigorate and realign US science shouldn’t be mistaken for an attempt to kill it.

Leftists might be upset that democracy is swinging against them right now, but it is within the very nature of US democracy that, should public opinion of scientific institutions improve in the future, politicians will fulfil the desire of the δῆμος and inflate the NIH/NSF budget once again. The US, with its unfettered free speech, tolerance for peaceful protest, and separated branches of government, is arguably the strongest and most stable democracy in the world. If you want policies that favour science, then first convince the voting populace — in democracies, politicians are followers, not leaders. I would encourage my fellow scientists to direct their anger not at the politicians who are simply acting on behalf of the will of the majority, but towards the institutions that have destroyed public trust in science. This is the root cause of the issue. Trump’s heavy hand is merely a symptom.

Whilst the public reverence for Science™ has decayed, the cultural adoration of American exceptionalism is stronger than ever. Aspiration, meritocracy, and the pursuit of success and excellence are so deeply ingrained in the American psyche that no politician, not even Trump, would dare to threaten that perception. Most Americans care deeply about winning, and winning at science is included in that.

The American left is now full of institutional conservatives sticking up for old establishments like Harvard, whilst MAGA has taken the tech right under its wing. The result is the presence of sizeable factions within both party coalitions that prioritise scientific progress and the funding it requires, meaning pro-science is one stance that has bipartisan support. With this cultural backdrop, it seems unwise to bet against US science.

Signal from noise

It is easy to paint a picture, as I have done in the introduction, of an evisceration of science by the Trump admin. It is less easy to analyse the reality of the situation dispassionately, rationally, and without slipping into attention-grabbing negativity bias.

To level the playing field set up by much of the news media, let me play Devil’s advocate for each of the various “attacks on science” we have been told are unfolding. By reading on, hopefully you can understand each side before forming your own opinion.

1. Indirect costs cap. If a scientist wins an NIH grant for, say, $1M, the NIH also forks over an additional amount to the scientist’s institution itself in order to cover administration and overhead costs. The percentage of the original grant that must be given is individually negotiated by different institutions, meaning places like the Salk Institute (where I work) get $900k extra, whilst University of Florida get just over $500k.

   The secret is, a silent plurality of scientists were actually quite excited at the prospect of the 15% cap. This would mean more grant money is given directly to us scientists, since the total funds are fixed — decided by the Congress, not the White House. Most key core facilities are in fact funded by direct costs, via hourly charges to our grants, meaning most of the cutting would indeed hit wasteful spending and administration. I personally would happily trade better chances of an extra NCI grant to fund some of our lab’s work on liver cancer, for example, in lieu of more indirect funds to go towards the monthly “Salkfest” events at my institute, where taxpayer money gets spent on free food and festivities during work hours (don’t get me wrong, I get free lunch whenever it’s available, but I can see why a regular taxpayer might not want to fund it).

   Regardless of anyone’s stance on this particular matter, in line with my point about the stability of US democracy, both NIH and NSF proposed caps have been blocked by district judges pending further litigation, and US science remains untouched.

2. Visa sponsorship removal. The visa situation at Harvard followed ongoing negotiations for records of students involved in illegal activities during campus protests. Harvard argues that revealing these records would exceed what is legally allowed, whilst the Trump administration argues that Harvard is not complying with reasonable requests. The truth probably lies somewhere in the middle, as always, and time will tell as the court case plays out.

   It is undeniably true that foreign students and postdocs here are simply collateral damage, and this is blatantly unfair. It was nonetheless interesting to me to observe an outpouring on X of offers from lab heads at other institutions to take on these students. Clearly the understanding was that the issue is with Harvard, and not with US science as a whole. It seems likely to me that students potentially scorned by this move would desire first and foremost to relocate within the US, rather than going back home.

   Needless to say, here too a federal judge issued an indefinite restraining order to block the order and foreign students and postdocs will not have to move anywhere whilst the court case ensues. US science remains untouched.

3. Removal of tax-exempt status. The visa issue is the least of Harvard’s worries. It has also had its tax-exempt status threatened, as a means to force them to comply with civil rights laws, which they were violating via their hiring and admittance practices. Let me repeat that point, because this key context seems to be missing from most media articles covering legal battles between the federal government and places like Harvard: multiple prestigious universities in the US were found to be illegally hiring and admitting based on race, the Supreme Court directly ruled that they were breaking Civil Rights laws, and the universities have carried on nonetheless. The judicial branch have tried, is it not the role of the executive branch now to see what it can do?

   Regardless of the motivations for Trump’s actions, Americans may see this tax-exemption change as an outrageous government intervention on a private non-profit educational institution. Interestingly, Brits certainly didn’t seem to think so when the UK’s Labour government implemented this very same change against all private primary and secondary schools last year with very little public resistance (in fact, the move was largely praised because people in the UK see it as immoral to be able to pay for a better education for your kids). Unlike schools in the UK, who are silently adapting to the change, including by raising the school fees such that the schools become even more financially exclusive than before Labour’s move, Harvard hasn’t had to do anything, since the threat was just a mean tweet (or Truth to be precise) — one that made headlines, of course. Tweets, though, don’t constitute a “war on science”, and US science remains untouched.

4. Cutting of scientific funding to specific institutions. Harvard and Columbia University, both hotbeds of anti-American and anti-Jewish cultural rot last year, were also hit hard by aggressive cutting of research funds directly. Because of the grants system I outlined in point 1, scientists’ grants represent a major source of funds for the universities. This means that, to punish the university administrations, the Trump admin must go through scientists. In this way, scientists are collateral damage in a war between the overpaid bureaucrats who (fail to) run the universities and a frustrated government who can’t seem to pull any levers without being kneecapped by district judges (which pushes the executive branch to reach for larger and larger levers).

   Even if Trump was well-intentioned to target the universities themselves, he is doing so via punishing scientists. This is clearly wrong and ineffective, right? We all know that universities, even though they have very large endowments and charge exorbitant fees to students (which is quasi-government funding when you consider forgivable student loans), can’t just spend these funds willy nilly. The endowments can only be appropriated for specific predetermined things. I certainly believed this dogma until I brought myself up to speed with how Columbia are dealing with the cuts. Writing to their research community on March 31st, the senior leadership assured:

   Centrally, we committed to fund those individuals whose salaries and stipends were previously funded with federal support on now-terminated awards, using institutional funds, while we undertake the review.

   Clearly, there are funds to spare and the University can buffer the loss of government funding. On May 6th, Columbia announced that they had reached the end of their review process and were sadly letting go of “about 20% of the individuals who are funded in some manner by the terminated grants”. They also outlined that graduate students and postdocs on terminated grants will all be supported for another year by the university, meaning that the 20% is likely junior research staff or admin staff. Either way, for them to be able to fund 80% in perpetuity is a great outcome and demonstrates the ability of the universities to utilise their own money instead of the taxpayers’. Columbia have also now set up a “Research Stabilization Fund” to aid researchers who are in the throws of uncertain grant applications/terminations.

   How is Columbia able to cough up so much money all of a sudden? In addition to the short-term endowment spending, they are making longterm room in the budget by reducing administration workforce and costs:

   We will continue to make prudent budget decisions that will ensure long-term financial stability across the University, including making significant budget reductions within the University’s central administration. Across the University, we have set parameters to keep most salaries at their current level, without increases for the next fiscal year, with some schools and units providing a modest pool for employees at the lower end of their salary distribution. We have also developed programs to further streamline our workforce through attrition and are preparing to launch a voluntary retirement incentive program, the details of which will be shared next week.

   In other words, they are cutting wasteful bureaucracy, capping the salaries of overpaid admin staff, and incentivising retirement to remove gerontocracy. This is exactly what Trump, and many of his voters, desired. So perhaps his actions directed initially at scientists were well-advised? Rather than killing science, an argument can be made that a rejuvenation and streamlining of Columbia, a scientific powerhouse, will be good for US science in the long run.

5. Cutting of scientific funding for specific areas of research. There have been claims of grant funding being cut or applications rejected due to the presence of specific “woke” terminology in the abstracts or titles of federal grant applications. While many claims, particularly on social media, may just be perceived sleights — used as an excuse for a crappy proposal — there does seem to be a legitimate attempt by the new NIH leadership to shift the focus of what they fund. You could argue this one either way. Either: obviously it’s up to NIH leadership to define what the NIH funds. Or: obviously the NIH shouldn’t be able to shape research objectives through a political lens. Luckily I don’t have to conclude this debate in my head, because earlier this week a federal judge ordered the immediate reinstating of many of these grants. The Trump admin can appeal the order, which could take it up to the Supreme Court. If they don’t, though, then this is another case of US science (even woke science) being untouched.

For every example here except the Columbia cuts (which, as I argued, might turn out to be net positive for science), judges have blocked Trump’s actions. I understand that Trump can be unpredictable and reactive. It’s also important to understand that media outlets, like Nature who dedicate entire collections to Trump’s perceived attack on science, thrive off the clicks and subscriptions that their fear-mongering generates. MAGAs are frustrated with “activist judges”, while TDS sufferers are chomping at the bit for the next headline of “Trump is killing science” to reassure them that their team is the good one. The reality is that American democracy is functioning as it was intended to, and US science is doing just fine.

Brain drain?

In 1933, having removed the power of the Reichstag, Hitler enacted the Law for the Restoration of the Professional Civil Service, which allowed the Nazis to fire any undesirables (Jews, communists etc) from the civil service. In the same year, the Emergency Committee in Aid of Displaced Foreign Scholars was formed in the US, which worked until 1945 to facilitate the immigration of primarily Jewish scientists persecuted by Nazi/fascist rule. The result was a brain drain from Europe, including giants such as Albert Einstein, Enrico Fermi, and Edward Teller. Many of these physicists would go on to join the Manhattan Project, the fruit of which would ultimately play a major role in winning the war via Japan’s surrender.

Earlier this month, the New York Times argued that we are on the precipice of experiencing the reverse of the German exodus, in the form of a Trump-fuelled US brain drain. The notion is corroborated by Nature, who asked their social media followers whether they are considering leaving the US. A whopping 75% of their (anonymous and unverified) respondents said yes, in what I pointed out was a textbook case of stated versus revealed preferences (consider the difference in agency required for waggling your thumbs a few times to stick it to the orange man, versus uprooting your entire life and moving countries).

The Nature publishing group offers a career-finding service. Using this service, they reported “that US scientists submitted 32% more applications for jobs abroad between January and March 2025 than during the same period in 2024”. Interestingly, they only give percentages, and not raw numbers, likely because embarrassingly few people use Nature’s job-finding service (I personally know of zero scientists who have ever used it). Also interestingly, they are too shy to give data from any other years, lest we spot the very likely trend of year-on-year increases in international job applications since COVID.

The key difference between Nazi Germany’s persecution of Jewish scientists and Trump’s funding upheaval is to do with what Richard Hanania calls Elite Human Capital. The total number of scientists that moved from Axis-run Europe to the US in the 1930s and 40s was somewhere around 300. The outsized impact on US science of this exodus was not due to their quantity, but to their quality. Put simply and harshly, scientists that find themselves outcompeted in a more restrictive funding environment in 2025 are not the cream of the crop of US science.

If anything, if uncompetitive researchers are ousted and muster the agency to make their way over to the 36-hour workweek safe haven of France, for example, US science might even be better off for it. I had a conversation recently with a smart young scientist working on a NASA-funded climate project attempting to clean up and analyse huge amounts of citizen-acquired data that is unusable in its current form. He told me how his department is expecting cuts, but he is unsure whether it will kill the entire project, which puts him out of a job (he’ll just get another one, he’s a data scientist), or if the cuts will mean getting rid of the most useless team members, the prospect of which he told me he was secretly very excited about. In this way, the looming “brain drain” may actually be a much needed brain cleanse.

To determine whether a brain drain of the actual top scientists might happen, we mustn’t concern ourselves with polling of the average scientist, but consider incentives for these top scientists. In their article, the New York Times interviewed some of such scientists, including Ardem Patapoutian, an immigrant scientist whose work on the pressure-sensing PIEZO channels in cells earned him the Nobel Prize in 2021. Ardem had posted criticism of the Trump admin regarding science funding on Leftist containment dome Bluesky and was almost instantly approached by the CCP with a relocation offer:

In late February, he posted on Bluesky that such cuts would damage biomedical research and prompt an exodus of talent from the United States. Within hours, he had an email from China, offering to move his lab to “any city, any university I want,” he said, with a guarantee of funding for the next 20 years.

Those outside of academia might not appreciate how unbelievably valuable this offer is. It would take something very extreme indeed to persuade an ambitious scientist against 20 years of guaranteed funding. Ardem’s response?

Dr. Patapoutian declined, because he loves his adopted country.

I rest my case.

China

China’s huge effort to try to grab American scientific talent can perhaps explain some of the media buzz around the death of American science. We know that China has been desperate to topple the US for decades. A bombshell report from last month detailed the vast extent of academic espionage that the CCP undertakes at US universities in attempt to sap knowledge and human capital from the global scientific leader. To think that CCP espionage is limited just to American universities, though, is naive.

New York Times reporter Vivian Wang recently published an extended advert for a new research institute in Hangzhou, China. Vivian notes on her NYT page that reporting on people living in China requires care, since “people can often be punished professionally or legally for criticizing the government”. Given Vivian has now lived in Beijing since 2021, we can take what she says about fantastic new Chinese universities with a pinch of salt.

Nature, too, recently published a very pro-China editorial, which focuses on China dominating the Nature Index top 10 — a measure of how many publications per institution are published in a selection of 145 top journals. This index doesn’t take into account the size of institutions (per capita, Harvard dominates the Chinese Academy of Sciences). It also doesn’t account for the fact that Chinese labs publish a lot of fraudulent science — publishing 17,541 retracted papers between 1996 and 2023, compared to USA’s 3,006. The incentives in China are to game the publication and peer review process, not to produce meaningful science — hence there being just a single Nobel Prize awarded for Chinese science.

Another eyebrow-raiser for me was the “Featured Jobs” section on Nature’s online service, which exclusively contain new postings in China. It may be the case that this is benign — just an algorithmic consequence of China having a great number of well-paid job openings and a dearth of talented scientists to fill them. Still, the terribly convenient alignment of CCP efforts and western media reporting narratives do start to smell slightly marine.

The CCP aren’t the only Machiavellians here though. Like the New York Times, NPR also tried to draw the Manhattan Project analogy, reporting that a number of European universities and politicians have launched efforts to poach US scientific talent. This includes Aix-Marseille Université in France, which has launched a “Safe Place for Science” programme dedicated to attracting and funding American emigrants “in particular those focusing on themes related to climate, the environment, health and the human and social sciences”. Climate science is a heavily politicised field rife with publication and funding bias. I’m not entirely convinced that the fifteen scientists due to fill the spots at Aix-Marseille Université will be particularly missed.

I’ve already alluded to how the media incentive to catastrophize about Trump is self-serving for them. Promulgating that “Trump is killing science” to predominantly Democrat-voting scientists and liberal elites, preying on our evolutionarily hardwired negativity bias, is a powerful way to grab attention, clicks, and subscriptions. The combination of domestic behavioural economics and foreign opportunism more than adequately explain what I would consider to be a large chasm between the reporting on US science and the reality of US science in 2025 so far.

The greenest grass

Ardem Patapoutian and I, despite our different propensities to catastrophize, seem to have reached the same conclusion regarding US science. Even granting that the bombastic criticisms of the Trump admin’s approach to science are true (which, as I have explained, they hardly are), it is worth underscoring why the US is still a superior place for scientists to work.

Based on the latest (2023) data, the US spends $823 billion per year on R&D. With PPP adjustment, the US’s $783.60B is trailed by China’s $723B. The vast majority of China’s funding is government-sourced, with corruption and low trust crippling their philanthropic efforts. Meanwhile, the US enjoys diversified funding with a very strong philanthropic scene. Since Trump’s inauguration, $820 million in grants have been terminated. This $820 million figure doesn’t constitute a real reduction in R&D spending though, since universities and philanthropic organisations are compensating for the Trump cuts. Such a kerfuffle about less than a 0.1% reduction in scientific spending feels a bit like 11-year-old Dudley Dursley complaining about only receiving 36 birthday presents compared to the 37 he was given for his 10th birthday. The fact of the matter is that the US is still the biggest scientific spender by a hefty margin, and this contributes to it being the best place on earth to do science.

Beyond funding opportunities, I believe the US is just a highly desirable place to live and work for driven individuals. In terms of work-life balance, USA represents the Golden Mean between the lazy Europeans and the browbeaten Chinese.

Ambitious scientists often find Europe a frustrating place to work — bogged down by excessive regulations and bureaucracy, and held back by pervasive tall poppy syndrome. Sam Rodriques, founder of computational biotech company FutureHouse, posted about his frustrations running a lab at The Francis Crick Institute, one of the UK’s premier research institutes:

… the primary challenge is cultural, and is both far less visible and far more pernicious. About 6 months in, I was sitting in a meeting with some other faculty and core facility leaders, arguing that we needed to build out an ambitious screening platform similar to those at the Broad. Heads bobbed up and down. And then, as people were filing out, one of the group leaders hung around behind and said, “Sam, you really have a lot of that American energy.” I chuckled. “Don’t worry,” she said, “Stay here for three years, and we’ll beat it out of you.”

Needless to say, the UK killed Sam’s desire to work in academia, and his new company seems to be flourishing in San Fransisco.

I have comparable experiences from my PhD at Cambridge. Most academic scientists, even some high-flying PIs, seemed to start work at around 10:30am. To be fair, they mostly pushed through until 6pm — as long as they’d had their late morning coffee break, extended lunch break, and mid-afternoon coffee break. Any attempt to break the mould there was frowned upon. About 3 years into my PhD, I recall setting up an old whiteboard by my desk to plan and brainstorm with. I wrote down every single remaining objective that I wanted to achieve by the end of my time there. This prompted one senior researcher to come over to my desk, chuckle at my list, and tell me that I should probably focus on just one project and that everything else on the list was too ambitious and unrealistic (she turned out to be incorrect).

At the opposite end of the spectrum is China, with its infamous 9-9-6 workweek. Science requires industriousness, sure, but it also requires creativity and ingenuity. Crushing the individual and removing any space for thoughts, dreams, and ideas harms scientific discovery. More than that, it makes China an unattractive place to work, evidenced by their extremely empty cities and labs. China can throw financial and scientific incentives out as much as they want, but they can’t overcome the fact that their country is a totalitarian sausage fest with a GDP per capita ~6.5x smaller than that of the US. As long as scientists don’t fall for Sinofuturist propaganda, I doubt China will be able to attract much US talent.

America, by comparison, is a rich, liberal, diverse nation that respects individual liberties and rewards individual merit. The US offers unparalleled opportunity — scientific or otherwise. Americans seem relatively unaware of the realness of the American Dream. Arnold Schwarzenegger paid homage to this on his recent appearance on The View:

I am so proud and happy that I was embraced by the American people like that. I mean, imagine, I came over here with the age of 21 with absolutely nothing. And then to create a career like that. I mean, in no other country in the world can you do that. Every single thing, if its my bodybuilding career, if its my acting career, becoming Governor, the beautiful family that I’ve created, all of this is because of America. And so this is why I’m so so happy to see first hand that this is the greatest country in the world and it is the land of opportunity.

It shouldn’t take monologues from immigrants like Arnold for Americans to realise how good they have it, nor should it take essays from immigrants like myself for American scientists to realise that US science won’t be killed anytime soon.

Leave a comment

Subscribe now

Getting the most elite degree from one of the most elite universities in the world should drive some sense of accomplishment, especially for someone raised to be an academic weapon from a young age (although I certainly wasn’t pushed as hard as some). It didn’t. No spark of pride at all really. Reflecting, I think I now know why: getting my PhD from Cambridge was one of the easiest things I have ever done, especially relative to how it is perceived.

This may sound ridiculous, or ridiculously arrogant, but let me clarify my point: I’m not trying to persuade you of my own intellect, I’m arguing that getting a PhD actually doesn’t require particularly high intellect nor does it require particularly hard work—objectively. Yet the societal perception of a PhD is that of a Herculean or even Sisyphean task (almost the opposite is true, as I will explain). There is an incentive for those with PhDs to overemphasise the difficulty of their achievement in order to amplify their perceived expertise and to propagate credentialism, from which they benefit. In reality, doing a PhD is an easy career move (if you can call it that), doesn’t particularly require conscientiousness, and certainly doesn’t make you smarter than the average person.

Subscribe now

The easiest thing I’ve ever done

A caveat to what I am about tell you about my PhD, as a useful anecdote, is that PhD requirements (for both graduating and being admitted in the first place) differ by country, university, department, and even principle investigator (PI=the head of a lab who “supervises” you). This heterogeneity, though, is itself damaging to the prestige of a PhD, since you cannot guarantee that one PhD holder was held to the same standard as another.

So, what does one have to do to get a PhD? The hardest part is being admitted to a PhD programme. For a PhD in the life sciences from University of Cambridge, I had to do the following:

1. Get a 2i or higher at undergraduate level (for those unfamiliar with the British system, this means 60% or higher, with the percentages being normalised such that 60% of the class achieve this standard by definition)

2. Acquire funding and admittance for a PhD by either:

   1. cold-emailing the head of a lab expressing interesting in a PhD and hoping they have money to fund you

   2. applying to a PhD programme which comes with funding. The funding covers fees, a stipend, and maybe even some research costs. Very few PIs will refuse to take you on after you’ve been accepted to a programme, since the financial burden of a PhD student is the main drawback and admission to the programme is seen as enough of a hurdle that busy PIs don’t find it necessary to vet you further.

That’s it in terms of absolute requirements. Having research experience, a masters, and interviewing well all help of course. For me, the research experience and masters were baked into my undergrad programme (Natural Sciences with a “Part III” integrated masters in biochemistry). My pre-PhD degrees being from Cambridge also probably helped, since undergrad admission to Cambridge is genuinely tough and the Natural Sciences tripos is probably much more rigorous than most other STEM courses at other unis. My performance at undergrad wasn’t amazing though. In my final year (the only one that counts at Cambridge) I averaged 64% and ranked 20/27 in my department (this was my best year by the way—I spent second year partying and averaged 52.5%). In fact, if anything, I’d argue that my undergrad record indicated a distinct lack of hard work. Despite this, I was accepted to both PhDs that I applied to (one was directly to a PI, the other to a programme funded by the Wellcome Trust, which I ended up going with because it paid me more).

My experience here was seamless, despite being subpar (at least amongst Cambridge MSci grads). I gather that the US is much more competitive, with severe grade and CV inflation meaning years of research experience is an implicit pre-requisite pre-PhD. To that I would say that the US is probably more competitive across every realm (law firms, consultancy jobs, pharma/industry, tech jobs etc), so my arguments herein are valid relativistically. I also think that British secondary education and undergraduate education is superior to that in the US, meaning more must be done post-graduating to achieve the same level of expertise before beginning a PhD.

Having been admitted to the PhD programme, what did I have to do to actually get the degree?

1. I had to attain adequate termly reports from my supervisor

2. I had to complete a first year report with a viva voce examination

3. I had to present my work once in four years at an internal departmental seminar

4. I had to eventually write my thesis, which doesn’t have a minimum word requirement, and then be examined on it in another viva voce.

Perhaps the hardest parts of this are the thesis and viva (certainly this is what most complain about). In my case, I focused on journal articles instead (scientists might actually read these), meaning I put very little effort or emphasis into the thesis. I wrote it in two weeks, had one round of feedback from my supervisor, and then submitted it last autumn. I had the viva some months later, which was a ~3 hour discussion with an internal examiner (a faculty member who I knew well) and an external examiner (faculty from another university who my supervisor knew well). My thesis was almost embarrassingly small, so I was asked to do “minor corrections” to expand my introduction chapter and add a summary chapter, which I did in a week and resubmitted. That’s it. PhD achieved.

Tramlines

It is almost impossible to fail a PhD. The only time this can really happen is at the final viva exam. Yet, the incentive here is to pass the student. Corrections mean more work for the examiners, and failing a student (almost unheard of) is a huge reputational hit for the supervisor, which is why the supervisor is best off picking lenient examiners.

This isn’t just my hunch. An analysis of over 26,000 PhD candidates across 14 UK universities found that only 3.3% of students failed their viva exam, and only 3% of that 3.3% (less than 0.1% of the total) don’t get any degree, with 97% of the failures still being awarded an MPhil as what is effectively a consolation prize. By contrast, many more (16.2% of the intake) don’t get a PhD simply because they quit. Notably, these data are from between 2006 and 2017. I would confidently wager that the stats are even more egregious nowadays, with lower rates of failure at the viva and higher rates of drop out due to “mental health” (in line with general trends in higher education).

Contrary to the perception of the Sisyphean task promulgated by those who benefit from such a perception, doing a PhD is an anti-Sisyphean task—like rolling a boulder down a hill and trying not to get in your own way. You’d have to do something very weird to stop the boulder getting to the finish line.

If I wanted to, I could have done incredibly little work over four years and still got my PhD. Indeed, many do precisely this. I’m friends with many highly intelligent people who ended up not liking academia (there is a lot to dislike) and completely losing interest in science. Yet the work requirement is so low, the student lifestyle so easy, the job security so valuable, the UK income and council tax exemption so advantageous, the guaranteed CV boost at the end so tempting, that sticking it out is the best thing to do. I absolutely agree with the actions of these friends.

These incentives are actually really important to understanding the sort of person that gets a PhD. At best, a PhD candidate is motivated by a sheer love of science, truth, and knowledge. I’m sure there is a sliver of this in every PhD candidate. If we’re being honest, though, the less virtuous motivators I outlined above are probably more powerful. There is also the giddy delight at the thought of being called “Dr”—the seductive allure of credentialism, the power of which few will admit.

Doing a PhD is also a low agency option. In my case, I didn’t have to move cities, didn’t have to make new friends necessarily, didn’t have to shift my routine to fit that of a real job, and didn’t have to face the turbulence of the job market. For someone conditioned to revere academic success above all else, a PhD can represent the default option which doesn’t require much decision or introspection as to what you actually want from life. It’s much safer to stay on the tramlines even if you aren’t really controlling the direction of the tram. I speak here from experience. If I’m honest—which I am—one of the main reasons I did a PhD was because I thought a city job was for suckers, I wasn’t ready for a nine-to-five, and I wanted to play for the uni rugby team. It was an immature decision that I have since managed to mould into a good one retroactively.

You might be tempted to think that the enormous dropout rate (16.2%) of PhDs implies that they are difficult. Don’t get me wrong—making valuable contributions to one’s field can be difficult. Getting a PhD does not require such contributions though. I would argue that the high dropout rate is instead evidence for the selection for low-agency graduates among PhD applicants. For a PhD project to be successful (measured by actual scientific discovery, not degree completion), failures must be overcome. Science can be unpredictable and involve a degree of chance. Biology, in particular, is complex. Projects can be derailed by an unreliable method, an incorrect assumption or hypothesis, or even being scooped by a competitor. Solving problems and overcoming failures requires both creativity and agency. Lacking these attributes leads to quitting. Negative mental health, which I’m sure many dropouts cite as their reason, is a symptom of low agency.

The alternative

It is a sad fact of skewed incentives that the cream of the undergrad crop tend not to pursue PhDs. In my experience, the smartest undergrads go for the more competitive fields. Most of my undergrad friends who earned first class Cambridge degrees flocked straight to London. Excluding the medics, most work in the classic finance/law/consultancy arena, but also for tech startups and engineering companies. These pathways are much more competitive, much more financially rewarding, and much more difficult. They select for intelligent hard workers with high agency and a sturdier mettle.

To contextualise the 97% PhD viva pass rate, let’s compare it to a postgraduate qualification from the financial sector: the Chartered Financial Analyst (CFA) charter. Similar to a PhD, the CFA charter is acquired after 4-5 years post graduation from university. Applicants come from a similar academic stratum, being predominantly Russel Group graduates with a 2i or higher. The CFA Institute, which runs this global programme, recommends 300 hours of independent study as a minimum for each level, which there are three of. Of course, this is to be done on weekends and evenings whilst working a full-time city job, hence it taking many years—a minimum of 4000 hours of work experience is required to even enroll in the exams. Some firms who sponsor their graduate cohorts may offer some time off close to the exam itself, but many do not at all. What’s more, whereas failing a PhD normally results in you getting an MPhil instead, failing the CFA exam just leaves you thousands in debt. The pass rate? Around 45% for each level, meaning the overall pass rate for all three levels is less than 10%. That’s hard.

The only way to resolve such a vast discrepancy in pass rates is to claim that everyone doing a PhD is just outrageously smart and hard working (which the incentives don’t predict), or to conclude that PhDs are, in fact, easy. What’s more, the global standardisation of the CFA means that, unlike a PhD or even an MBA, the quality of charter holders can be guaranteed.

The lazy genius?

Since PhDs are quite easy, there is a chance they could select for lazy people (2019 Adam is a case in point). But what about actual intelligence? Maybe PhDs fit into the “brilliant but lazy” category.

This may well be true in some cases, but I prefer to draw conclusions based on data rather than anecdote. Studies on the IQ of PhD students or holders are few and far between, typically with small sample sizes. To set the benchmark, we can look at data from 1967 which looked at 148 faculty members from University of Cambridge across various departments. Accounting for the Flynn effect, the average IQ across disciplines was around 122 (±13.7). For context, the distribution of IQ, by definition, has a mean of 100 and a standard deviation of 15. This puts the 1967 faculty, on average, in the 93rd percentile of the general population. Similarly, a 2016 study of 38 PhD students and postdocs from University of Oxford found an average IQ of 125 (±6).

Don’t get me wrong, this is smart, but it is nothing to write home about, and certainly far below the perception. If we take a 125 IQ to be representative of all UK PhD holders (which it definitely isn’t), this places PhDs in the top 5% of IQs. Yet, according to an OECD report, only 2% of the UK population have PhDs. This means the perceived eliteness of a PhD is over twice as high as their actual intellectual eliteness.

Of course, 125 is likely much higher than the real average IQ of PhD holders. For one, the 2016 study is from Oxford, an elite university, likely selecting for smarter people. Secondly, the calibre of PhD graduates has likely dropped since 2016, as admissions have swelled and standards slipped. There isn’t longitudinal data at the PhD level, but a recent meta-analysis looking at the intelligence of university graduates versus the non-university educated found essentially no difference at all, with the mean IQ of university graduates being 102, which the authors describe as “merely average”. Thus, it wouldn’t surprise me if the real average IQ of PhD holders is something like 110.

Make PhDs hard again

So what? IQ is just one measure and it’s not even that important, you may protest. On IQ as a measure for intelligence, you would be wrong—IQ is the single best objective metric that correlates with intelligence. Complaints about the validity of IQ tend to involve politically motivated reasoning. The actual validity, reliability, and predictive power of IQ is unquestionable, despite disgruntlement from biological Marxists. On the importance of intelligence for a PhD, I can offer some wriggle room. Perhaps intelligence isn’t and shouldn’t be important to acquire a PhD. This is certainly the view of those who prioritise inclusivity over excellence. That’s fine, but then the societal reverence for PhDs must decline in concert. My primary objection is to the huge mismatch I see between perceived expertise and actual expertise; between credentialism and intellectualism.

That said, I would rather society did not accept the dumbing down of the PhD as this negates the very purpose of the PhD. We are currently facing an expertise crisis. The average person has been forced to “trust the experts” again and again, all the while the “experts” have been proven consistently wrong on issues ranging from COVID to global warming. Some propose a solution of ignoring the experts and allowing amateur voices to fill the void. I think a better longterm solution is to produce smarter and less ideologically captured experts in the first place. Experts who can debate amateur voices from first principles, rather than by leveraging their credentials. This requires the elevation of more competent experts who pursue critical thinking over ideology. Whilst this is probably best achieved by starting at the level of secondary, or even primary, education, it can be fixed at the level of the PhD by making PhDs actually worth something. PhDs must become more elite.

Some shudder at the suggestion that academia should become more elite. However, this stems from a conflation of intellectual elitism with social elitism. I’m not calling for a rebirth of the Victorian era aristocrat-scientist, who can pour unearned generational wealth into random experiments sewing rats together. I’m calling for academic excellence where it is needed. In fact, intellectual and social elitism are opposed to each other. Siphoning the best talent from only the richest or most socially pedestalized stratum of society is going to produce an inferior cohort of PhDs, versus enabling access to PhDs across financial strata and social class but enforcing stricter intellectual requirements at the point of application and completion.

Making PhDs hard again also has beneficial effects on the structure of science funding. Last year, I attended a panel discussion on “Improving the Structures of Science” at the MRC Laboratory of Molecular Biology in Cambridge. Something all panel speakers converged on was the fact that there seems to be a bottleneck of funding at the stage of scientific independence, that is, after one’s postdoc when trying to acquire funding to start a lab. A longitudinal cohort study of publishing records from over 350,000 scientists from 38 countries found that this is also one of the points at which most people leave academic science, either finding industry jobs or leaving science altogether.

One of the panel members, himself an assistant professor, argued that the application process for early career grants is far too complicated and there isn’t enough funding to go around. This is true. The reason for the exodus is that it is very difficult to get funding after one’s postdoc. But there are two variables affecting this dynamic: the number of scientists applying for funding and the total funding available. Given that academic science is a privilege funded by the tax payer and philanthropy, I don’t think insisting on more funding is a viable solution. To the panel, I offered an alternative hypothesis: there are too many PhDs for the funding that is around. The issue is that there is far too much chaff and not enough wheat competing for limited funding. PhDs are too easy and too common. Some predictable pearl-clutching ensued, but one panel member, a key figure involved in Downing Street’s creation of the UK’s Advanced Research and Invention Agency (ARIA), came up to me afterwards to quietly admit that he completely agrees with what I had said.

Tougher entry into PhD programmes and then a tougher requirement for graduation would remove such ferocious competition and sharp exodus later in scientists’ careers by making the competitiveness of science more ubiquitous across career stages. It is a fact of life that funding for science cannot be infinite. Scientists must compete for finite resources to a certain degree. Making PhDs harder would make the rate of science attrition much more even over time, removing bottlenecks for funding. This is better for everyone. It would reduce wasted funding on PhD students who are unlikely to contribute value to their fields, and it would improve the lives of scientists who would now be less likely to have to change career path at the point in life when most want to start families and seek stability. A good place to start would be a standardised test as part of the PhD application (and no, these aren’t racist).

In defence of the PhD student

That PhD requirements are not challenging does not mean that any particular PhD project cannot be phenomenally difficult. My distinct lack of pride at my graduation or after my viva exam was perhaps compensated for by massive relief and swelling pride when I submitted my work in the form of journal articles to respectable academic journals (although this was and is muted excitement until they actually get published, God willing!).

Whilst PhDs are easy, science itself is actually really really hard. I recall torturous weeks clocking over 80 hours within the walls of my institute; experiments that never saw success despite dozens of attempts, each optimising from the previous; over 20 drafts of my first paper as I both honed my writing ability and jostled with my PI for command of the interpretation of my data; rejection after rejection for postdoc fellowships; constant fear that my discoveries were artefacts or would be scooped etc. And all of this was for some rather unimportant papers in the grand scheme of things. It is very difficult to push the boundaries of human knowledge in meaningful ways, and yet this should be the purpose of a PhD.

Many PhD students do contribute meaningfully to human knowledge, and the degree they receive at the end of it should reflect this achievement, rather than be a footnote—an annoying piece of admin—to the end of 4+ years of arduous research. It should never be the case that a PhD student binds their thesis prior to their viva exam because they are so confident in the triviality of the viva that they know they won’t have to get it rebound with corrections, as was the case for viral smell lady Ally Louks (she’s since quit research by the way, no doubt due to an inability to get funding).

I’m not entirely sure the best way to evaluate PhD students for completion of their degrees. Perhaps theses could be sent for open peer-review online—a democratisation of the viva exam. Perhaps PhD degrees should be withheld until a certain number of citations of one’s PhD work is attained. Perhaps institutions can be given a limited number of PhDs to award per year that is less in number than number of PhD students per cohort, and students can re-compete for their degree the following year if unsuccessful. Certainly preventing choice of favourable examiners with twisted conflicts of interest would be a good start. However we proceed, we must recognise that society can justly pooh-pooh credentialed experts if those credentials are in fact easy to get. Experts must be elite intellectuals. PhDs must be hard.

Should I imitate it? You know my wife gets very upset when I do this. She says, “Darling it’s not presidential”. I say, “Yeah but people like it!”. Should I do it or not?

The crowd cheers “Do it!” and he obliges. In another life, Trump is a wildly successful standup comic.

Since then he posted an AI-generated image of himself as the pope on his personal social media account, which the White House then reposted.

Yes, Trump doesn’t care about being presidential, he cares about being liked. But he is not the only one. Senator John Fetterman dresses like a total slob. It’s not Senatorial, but it draws attention and some people like it. The airline Ryanair posts childish TikToks with bleeped out f-bombs. It’s not professional, but people like it. The Trump White House publicises some of its most bipartisanly laudable accomplishments with a total lack of seriousness. It’s improper, but the other side get rage-baited and people like it. Australian political parties are posting some very strange stuff indeed on social media. It’s not normal, but some people probably like it. Elon Musk plucked the goal of cutting two trillion dollars from the federal budget out of thin air. It’s not realistic and wasn’t the plan, but the crowd loved it.

These are a few particularly blatant examples of unseriousness from public figures, institutions, and companies, but the death of professionalism pervades society. Why can nobody be serious anymore?

Professionalism, to me, means taking yourself and your job seriously, caring about the quality of your outputs, and caring about the impression you give on behalf of your institution/company. Everywhere I look I see evidence of this attitude being eroded not just from the top down, but also the bottom up:

• Three-day in-person work weeks are par for the course since COVID, but even that is too much for the UK civil service, who are planning strikes in response.

• Emails are increasingly “Hey … Cheers…” versus “Dear … Yours sincerely…”.

• Professional dress codes are being slackened across the board. Anecdotally, walking through the City of London makes this abundantly obvious.

• Academic standards and attitudes at elite institutions are crumbling.

• Companies are increasingly giving in to the fact that sometimes their employees just don’t feel like showing up at all, and they shouldn’t challenge it because mental health something something.

• Productivity is declining in certain sectors for mysterious reasons (i.e., people simply aren’t doing their jobs properly).

I detest dogma, pointless rules, posturing, and pretentiousness probably far more than the average person, yet the past decade or so has seen the baby chucked out with the bathwater. I think this is regression, rather than progression. I believe that professionalism, seriousness, and decorum are important facets of society, the revival of which would benefit everybody. There are surely a great number of aetiological factors at play here. I will outline a few before offering my prognosis and paths forward.

Subscribe now

Toxic authenticity

If you wish to compete with other humans in any game that is global and online (any type of content creation—in the broad sense1), conventional wisdom suggests that it is unwise to pit yourself against 8 billion other humans in the same game. The solution is to play a different game—namely being the best and most authentic version of yourself, rather than trying to be a better version of someone else. Being authentic creates a healthier mental framework to act within. It can also be far less stressful than the alternative, since it is more honest and dishonesty takes a toll on the soul and the mind.

The benefits of authenticity aren’t limited only to the internal though. There are several external benefits of authenticity, perhaps explaining why it is all the rage online nowadays. People want to know who it is that they are supposed to be giving their trust and attention to. The more relatable someone seems, the more trustworthy and the more deserved of your attention they are. Thus, authenticity sells better.

The problem is, discerning true authenticity is a difficult task, particularly when given very little information to work with (e.g., a person’s Instagram profile). Given the advantages authenticity offers, and the difficulty of discerning true authenticity, this creates a perverse incentive to engage in faux-authenticity. In the hyper-comparative world of social media and clickbait news media this creates a meta-game that many (particularly those high in trait Machiavellianism) play. Namely, “How can I signal as much authenticity as possible?”.

To state the obvious, attempting to cheat the “authenticity game” in this way is the opposite of authentic. It is manipulative and dishonest. Sadly, it is often successful. Strategies inevitably involve throwing decorum or social standards to the wind, signalling “I simply don’t fit within social norms because I’m such a real person”. Sticking out by rejecting professionalism and its requisite uniformity is the ultimate attention-grab.

In terms of physical appearance, this is exactly what John Fetterman is doing. Cynical take—but this is also what people who insist on wearing their so-called cultural dress at occasions that demand uniformity defined by the host, not the guest, are doing (faux-traditional African dress at English University graduation ceremonies, and faux-traditional tartan kilts at black tie events are two excellent examples of this). The whole notion of “needing to express oneself” is more often than not simply hijacking authenticity to garner attention. You don’t “need” to express anything, you just want the benefits of standing out from the crowd and of signalling “authenticity”.

Crucially, this need to display authenticity isn’t limited to our personal lives, but also to professional settings within corporations or institutions, often transmitted via their social media presence or news media. The textbook example of professional authenticity is Donald Trump, whose unabashed authenticity won him two non-consecutive presidential elections. With the playbook set out by Trump, the same perverse incentives implore others to drop the professionalism and be “real” instead (I recall the “Kamala is brat” phase and shiver with cringe). This is especially true now that everyone—from political parties to airlines—must compete for views online as the first line of engagement with their constituents or customers. In fact, the incentive to seem authentic is arguably even greater at the institutional level than at the individual level.

It is no secret that trust in and respect for previously pedestalized institutions, companies, or positions have waned, exacerbated by a swathe of their shortcomings in recent years and a complete inability to cover things up now we live in the age of free flowing information.

The hardest hit by this fervent distrust have been politicians, large corporations, and the news media. Some first-order effects of this distrust include surprising election results, where incumbency now appears to confer a disadvantage for the first time ever, and a shift in information diet away from classical media and towards podcasters who offer less authority but much more authenticity (maybe Substackers represent the Golden Mean here … 😉)

The examples of Trump’s success and the glaring defeat of classical media to podcasters and new media has told professionals to do one thing: increase their appeals to authenticity. Employ a gen Z social media team and you’re off to the races (refer back to my Ryanair and White House examples at the start of this essay).

In this way, a need to compete via faux-authenticity has led to individuals and institutions killing any professional veneer they might have had in place. Never mind that a professional veneer may have actually served an important purpose. Getting a zoomer to retort “they’re so real for that” seems to be the ultimate goal.

Zoomer brats, millennial cucks

There is a clear generational power dynamic online. Gen Z/α is much more gung ho. Gen Z/α can be brutally honest. Gen Z/α dictates what is cool. Perhaps watching up-close footage from war zones / hardcore pornography / gruesome accidents / police bodycam footage / *insert your favourite psychologically damaging type of internet content* on your phone aged 12 has an effect on the human psyche that makes the zoomer sense of humour much darker and more cynical than that of millennials/gen X. Millennials are also slow to pick up trends (probably because they spend less time online/on their phones) and so risk being out of touch and cringey. This doesn’t stop millennials trying though. I see example after example of millennials desperately appealing to zoomers for approval—adopting zoomer vernacular or imitating zoomer-initiated trends online, for example. Mimicking zoomer behaviour online also seems to be a way in which millennials/Gen X can try to avoid ageing—something they are seemingly terrified of.

The nihilistic and cynical bent adopted by zoomers is reflected in their real-world behaviours. Zoomers just don’t seem to care that much. I am reminded of a discussion with my older brother, who works in the City of London, recounting his experience with a fresh wave of zoomer interns at his office (he himself just sneaks into the bottom end of the millennial generation). None of them wear smart clothes. None of them show up on time. None of them eat before work (that would require waking up early), and therefore go out after the first meeting for breakfast together. But crucially, none of them get reprimanded in any meaningful way. And even if they did, they don’t care, so what would be the point of the reprimand? The funny thing is, Gen Z don’t have to care. In material terms, despite the obsessive focus on the housing market, they are actually better off than every preceding generation (“unprecedentedly rich”, as put by the Economist), and the thing that normally drives a sense of striving or progression in life—dopamine—is easily acquired from the screen in their pocket instead. Worst of all, lax parents, lax teachers, and lax employers have all conditioned gen Z to be demanding and inflexible by letting them get their way all of the time.

Zoomer brattiness and millennial kowtowing decimates professionalism. Nobody feels like they have to do anything anymore. As more zoomers and gen alpha begin to infiltrate the workforce there is a chance this apathy could get much worse, although I have reason to suspect that gen Z will course-correct, as I will get onto later.

Professionalism is right wing. Yuck!

Since this death of professionalism seems to be seen as an example of societal progress, flush with woke epithets centred around self-expression, it is lauded by those on team “progressive”. Meanwhile, those on team “conservative” find themselves reactively anti-self expression, criticising the president of a warring nation for not wearing a suit, for example2.

In general, if a certain behaviour is promulgated by a figure perceived as right-wing, then that behaviour is labelled right-wing and the left recoils from it, and vice versa. In this way, suits are now right-wing coded in part thanks to Jordan Peterson. You could validly argue that wearing a suit is a genuinely conservative thing to do, since it conserves traditional formal attire. But other non-traditional behaviours, like exercising and eating a diet sufficient in protein, have also seen this political lens erroneously applied. There are feedback loops here making the political polarisation of normal things worse—right-wing people more readily gravitate towards these erroneously right-wing coded behaviours and thus exacerbate the stereotype. Equally, people who engage in these behaviours for non-political reasons are labelled as right-wing regardless of their actual political leaning and are then used as examples with which to falsely propagate the stereotype nonetheless. Andrew Huberman as part of the “manosphere” is a case in point. It seems encouraging anyone, but especially men, to pursue excellence in any realm gets smeared as right-wing these days.

The same kind of right-wing coding is at play regarding professionalism, which is the institutional or career-centric manifestation of the pursuit of excellence. Pride in your nine-to-five is seen as an outdated mindset, not relevant in our era of Instagram travel influencers, YouTubers, and crypto traders. Rejecting cynicism and buying into one’s professional duty is therefore seen as yearning for a previous time, which smells awfully conservative. The lens of political polarisation seems to be the nail in the coffin for the death of professionalism.

The future looks (b)right

While I’d love to grab your attention by stressing the negatives of the societal problem I’ve laid out here, I am an anti-doomer and generally find societal homeostasis to be far more powerful than most realise. We may not be at the bottom of the pit of the death of professionalism yet, but I do think we will see it reborn. Regarding the political schism over professionalism, there are two equally plausible paths forward for its resurrection: either the left embraces professionalism, or the professional-embracing right wins out over the left.

For the former, this means the left must acknowledge the value of professionalism and society can move forward agreeing that professional standards, appropriate seriousness, and proper decorum are not to be given up in the name of “progress” or “self-expression” or “authenticity”.

There are whiffs of this already. On the aesthetic front, there is this quite popular left-wing chap who writes and posts about menswear online, often focusing on formal clothing. He mainly leverages his passion for formal attire to dunk on conservatives wearing ill-fitting suits, but regardless of this vindictive motive, it is refreshing to see a die-hard progressive championing decorum and proper attire. The genius of his framing is in highlighting the craftsmanship of the tailors of good suits, contrasting it with fast fashion and the corporate greed of Big Clothing. In this way he moulds professional attire into left-wing ideals.

A much more potent motivator for the left to return to embracing professionalism is their reacting to a lack of professionalism from the right. The more Trump departs from presidentiality, the greater the opportunity for a Democrat candidate to fill this vacuum by appealing to total seriousness and decorum while rejecting cattiness. So far, the left in America have failed to resist the pull towards cattiness, and find themselves hardly better than Trump in terms of professionalism (the benefits of appeals to authenticity are too great for both sides—Trump’s examples just get much more coverage). Nonetheless, I am holding out for the position of the metered and professional centrist to be filled eventually, perhaps taking the lead from the much more professional politicians across the pond (Macron and Badenoch come to mind).

The alternative is that Trump’s presidency marks a global return to conservatism (not that Trump himself is particularly conservative, but his election is a hallmark of a wider trend), and that professionalism piggybacks on this rise of the right in the coming years. The American left seems to be in disarray, and the right is doing well in Europe as the centre-left parties in power still struggle to grapple with the issues people actually care about (chiefly, immigration). We might see a greater cultural resurgence of pride—both national and personal—the latter of which might enable and encourage people to hold themselves to greater professional standards.

Trying to be cool is not cool

Regarding the generational dynamic of the death of professionalism, what millennials and older generations don’t seem to realise is that it isn’t the specific trends that GenZ/α engage in that are inherently cool, it is the fact that these trends are mysterious to older generations and therefore confer exclusivity to young people that makes them cool.

The same goes for Gen Z vernacular. As soon as a previously “cool” phrase comes out of a millennials mouth, it is instantly not cool anymore. What is especially uncool about this is the cringe of trying so hard to be cool by older generations. Younger generations can distinguish faux from real authenticity pretty well when it’s coming from older generations. Since no authentic millennial goes round saying, “no cap fr this burrito is bussin. It’s giving demure. The chef understood the assignment. What a sigma”, it’s obvious that this is pure try-hard and completely fake.

As I was writing this essay, UK’s Labour Party posted on their Instagram page a Keir Starmer edit supposedly demonstrating his “aura”. It might have been funny before the general election, but now he is our prime minister and has a serious job to do. British citizens want a functioning government, not propaganda reels. This type of disregard for professionalism is just not cool anymore. The comments on the video prove my point.

Akin to the political vacuum of professionalism in the wake of Trump, a generational vacuum for professionalism emerges in the wake of uncool and unserious millennials. If millennials are “quietly quitting”, not caring about their jobs, and abandoning wearing a suit, this creates the perfect incentive for young people to embrace professionalism as their own. Again—I detect whiffs of this already with the meme of the “academic weapon” and people like Davis Clarke, an accountant who posts content about his dedication to his job, gaining popularity online.

I hope that we are coming to the end of the era of outrage culture, the coddling of younger generations, and faux appeals to authenticity. I want to live in a society that unironically embraces striving, humanity, and aspiration. This doesn’t mean everyone must don a suit. I certainly won’t be wearing a suit to the lab anytime soon, although I appreciate that my PI does, and there is something that I love about the aesthetic of the 20th century scientist in well-tailored suit with a fag hanging off his bottom lip. Formality isn’t exactly what I’m after, although it is part of it when appropriate. I want real authenticity, pride in quality, and not doing things because, as Trump put it, “people like it”. This is a society that is roused by a Creed halftime show unironically, rather than one in which millennial white women mimic a Kendrick Lamar halftime show to try and seem cool. A society in which people, institutions, politicians and corporations don’t need to compromise their professionalism by making shitty jokes, peddling overused memes and trends, and constantly breaking the fourth wall to rage-bait and subvert expectations for attention. I want us to mature from our current era of societal-level audience capture.

Professionalism means parking features of your private life outside of the work place (e.g., covering tattoos if they would detract from professional efficacy). It means forcing yourself to go to work when you don’t feel like it. It means acting with appropriate decorum even if you would speak or behave differently in the confines of your own home or when spending time with friends. It means aspiring to excellence at what you do and not degrading yourself to garner attention.

I consider a positive attitude towards professionalism to be the epitome of societal progress. This is my plea for a return to this mode over the next decade or so. I am optimistic that we can and will.

Time magazine recently published a cover story celebrating the immense achievement of Colossal Biosciences, who managed to bring back the dire wolf from extinction. Sort of.

What they really did was the following:

• sequence the genomes from two ancient dire wolf specimens with unprecedented genome coverage

• develop a novel non-invasive way of obtaining genetic material from living animals to create reference genomes for a wide range of related canids

• map the phylogeny of the dire wolf, finding that the grey wolf is the extant species with which it shares the most genetic alleles

• determine the function of proteins that are encoded by genes that differ slightly in sequence between dire wolves and grey wolves

• shortlist a subset of these proteins that would give rise to crucial phenotypic traits that separate dire wolves and grey wolves, particularly with regard to their differing ecological niches

• develop a novel multiplex gene editing technique to be able to edit many of the genes encoding these proteins in one go

• make precise edits to the grey wolf genes on this shortlist to recreate the dire wolf gene sequences and thus reproduce dire wolf phenotypes from dire wolf genotypes

This is really really complicated and incredibly impressive science. Yet an alarmingly common response from the scientific community was to criticise: criticisms of the validity of the claim of “de-extinction”; criticisms of the usefulness of the endeavour (as if a majority of academics have ever done anything directly useful for humanity in their lives); criticisms of the savvy marketing of their discovery (I thought public engagement was a key component of what it means to be a scientist?). Compare and contrast the tone and narrative of this New York Times article with this Science news piece. The level of cynicism and contempt spilling out of the Science hit piece-cum-news article is exceeded only by its level of bias and sloppiness. Take this paragraph:

Others have argued that efforts to resurrect lost species divert resources and attention from threatened and endangered species living today. In a comment to the New Zealand Science Media Centre, University of Otago geneticist Nic Rawlence said if it were up to him, companies like Colossal would “develop de-extinction technology but use it to conserve what we have left.”

Yeah nice point except for the fact that leveraging their technologies to conserve and protect extant endangered species is literally exactly what Colossal Biosciences is already doing and what they’ve always said was the motivation for their work. A completely fake criticism that the writer for Science clearly did zero due diligence on. There are other outright falsehoods floating around too, including claims that Colossal chose the wrong starting species and that the black-backed jackal is actually evolutionarily closer to the dire wolf. I think these critics were perhaps thrown off by the order in which the species names appeared in the below figure from Colossal’s pre-print, ignoring what the phylogenetic tree actually shows. As I pointed out to one cynical X user, Colossal wrote the book on dire wolf phylogeny—obviously they wouldn’t deliberately make their job harder than it needs to be.

Many cynics are diving headfirst into pedantry as their primary method of attack—splitting hairs over calling the animals “dire wolves” and insist on minimising the achievement via pedantic and dogmatic (wolfmatic?) definitions of “species”. Yes, of course it’s true that these aren’t 100% genetically identical dire wolves, that the environment these wolves are placed in cannot completely replicate that of the Pleistocene, that they have no dire wolf parents to learn dire wolf behaviours from, that their diet will likely differ from that of Pleistocene dire wolves and this likely further diverges their microbiomes and physiology. To completely adjust for all of these differences would be nigh-impossible. Further, all of these parameters, both genetic and environmental, differ immensely between a Canadian dire wolf in 100,000 BC and a Peruvian dire wolf in 15,000 BC, so which is the “true” dire wolf?

The most valid criticism is the point that these are not 100% genetically pure dire wolves—something that nobody is claiming, least of all Colossal. Might it be possible in the future to completely reconstruct the entire dire wolf genome and create a 21st Century dire wolf that is genetically identical to a Pleistocene dire wolf? Perhaps. What would be the point, though? The critics seem to miss the motivation for why Colossal are doing what they are doing. They are first and foremost a conservation company. They are unique because they develop and leverage synthetic biology to achieve their conservation aims, with hugely beneficial side-effects for fundamental biological research and medical research alike.

In light of this conservationist bent, the species definition that Colossal cares about the most is based on ecological niche. After all, a vacuum in ecological niche is arguably the most impactful negative consequence of species extinction. It’s perfectly valid to define a species by its ecological niche. It’s also valid to define a species by geographical location, genomic sequence, morphological features, or ability to produce fertile offspring with other individuals within the group. All species definitions are somewhat arbitrary human-drawn lines in the sand of the continuous process of evolution, and are therefore mutable. Even if Colossal’s chosen definition of a species is an uncommon one in the post-genomic revolution era, it isn’t “wrong” by any objective measure.

Given the tentative validity of many of the criticisms levied at Colossal, we must ask why some academics have reacted this way. Unnecessary criticism and excessive cynicism serve two main functions: (1) to keep others down, and (2) to inflate one’s own ego. Criticising feels really good. It’s a display to others that you are incredibly smart and have intelligently dissected problems that others did not see on account of your relative talent and knowledge and their relative ignorance. Nowhere does this social dynamic play out more strongly than in academic science. There are structural incentives that create this tendency, as well as a selection effect for people with this bias to become academic scientists in the first place that I will explore in a future essay perhaps.

When failing to be able to pinpoint any specific flaws in the recent pre-print on the dire wolf phylogeny, authored by Colossal’s team, some simply assert, “It’s probably full of errors. I won’t take it seriously until it’s been peer-reviewed.” Peer review is the motte that academics always retreat to when their own attempts at interrogation are insubstantial. We are assured that this mythical process ensures only the most rigorous and best science gets published. Except that’s flagrantly untrue. Loads of totally unreproducible (read: useless) science gets through peer review. If peer review was actually rigorous, it would have caught the blatant data manipulation and fraud published in Nature that led the Alzheimer’s field on a wild goose chase for decades, for example. Peer review has its place, but it is the jewel of credentialism and the ultimate appeal to authority. I have much more to say against peer review. Again—perhaps another time (I say this as if Michael Eisen has left any stones unturned).

Now, to be fair, I don’t know what I don’t know. There could be methodological flaws that I didn’t pick up in my reading of Colossal’s pre-print. Maybe peer review will give them a hard time, but I’m pretty positive it will get published in a top tier journal eventually. To be honest, though, this is irrelevant. Publications are a side quest for Colossal. The great thing about this dire wolf breakthrough is that it is real. The wolves exist. It is a concrete achievement. Publishing a highfalutin Cell paper is great, but doing so rarely brings the metaphorical rubber any closer to the metaphorical road. In two years time when a fully-grown pair of adult male de-extincted dire wolves are stood side-by-side with a grey wolf, peer review or lack thereof becomes completely irrelevant.

I sometimes lament the fact that the endpoint of what I do as an academic scientist is getting a journal article published (or is the ultimate goal to use these publications to get more money? I can never tell). If you’ve learned to frame scientific success through the lens of credentialism, it must hit a nerve that George R. R. Martin and some tech bros without so much as an undergrad degree in biology are authors on a paper that is more impactful than what most of us academics will ever produce in our lifetimes. Creating real progress in the way that Colossal is doing dwarfs the triviality of words on paper (or, more accurately, paywalled PDFs), and some academics hate this fact.

It is no secret that academic scientists are overwhelmingly of a progressive persuasion. Yet the dire wolf saga has revealed this progressivism to be in name only. These “progressives” are in fact conservatives with respect to the authority and dominance of their ivory towers, and their conservatism overpowers any progressivism that they would otherwise exhibit with respect to science, technology, or ecology. There is a bitter irony in how the pearl-clutching conservatism of “progressive” academics is being unmasked by technologically progressive conservationists. This disdain being shown for genuine scientific progress by some academic scientists and adjacent liberal elites is a tragic indictment of academia and the “progressive” establishment.

Subscribe now

Leave a comment

It was recently claimed by Reuters that the CDC, under the stewardship of the newly appointed HHS Secretary Robert F Kennedy Jr, is planning a new study into the potential link between vaccines and autism. The overwhelming response from my scientific colleagues was one of condemnation and disparagement. Some decry the negative impacts of giving legitimacy to “anti-vaxxers”. Others laugh at the juxtaposition of DOGE trying to slash government spending whilst HHS wastes funds on something that is well-settled. There were some pretty funny responses actually.

Whilst it’s very popular to dunk on MAGA these days, I think it is worth giving more sincere consideration to the predicament we find ourselves in with regards to health policy and public trust in science in the post-COVID era.

There is a need for scientists, politicians, and commentators to empathise with the members of the public who distrust the expert class. The thing is, this distrust isn’t random. It was completely deserved. The expert class spent the last five years lying to the public again and again. This included lies about the efficacy of masks, the efficacy of vaccines, the lack of efficacy of alternative COVID treatments, and lies about the origins of COVID. All the while, inflation disproportionately ravaged the poor, the rich got much richer, and Big Pharma basked in their unprecedented legal protections and government-mandated profits in complete absence of real free market forces.

To dismiss skeptical members of the public as moron anti-vaxxers is not only the opposite of helpful, but demonstrates a supreme lack of empathy and surplus of narcissism—driven by the good feeling that one gets when proving the out-group wrong and oneself right, intelligent, and moral. What’s more, much of the credentialed elite are often too arrogant to even attempt to interrogate the claims of people less educated than themselves—dismissing them out of hand because of who the claims are coming from, rather than because of what the claims are. We’re left with an ecosystem where the experts aren’t really experts, but just parrots of dogma. Meanwhile, the distrusting public aren’t being engaged in good faith, but instead reprimanded for questioning the dogma. Resultantly, they are browbeaten downwards into theories of conspiracy.

Truth and honesty is the cure to all of this. So what is the truth when it comes to vaccines and autism? To understand how we got here, we should go back to the origin of the vaccines-cause-autism story. We must go back to the academic fraud of Andrew Wakefield.

The Lancet paper

In 1998, a team of physician-scientists from the Royal Free Hospital in London, UK, published a paper in The Lancet that is now often considered the seed of the vaccine-autism hypothesis. The infamous paper is a rather simple case report that described and attempted to explain the symptoms of twelve children who all seemed to suffer from the same syndrome. Here is the introduction to the report in full:

We saw several children who, after a period of apparent normality, lost acquired skills, including communication. They all had gastrointestinal symptoms, including abdominal pain, diarrhoea, and bloating and, in some cases, food intolerance. We describe the clinical findings, and gastrointestinal features of these children.

Many of the authors, including first author Andrew Wakefield, were gastroenterologists, and gastrointestinal symptoms were the primary focus of the paper. It was the parents of the children who drew the link with the MMR vaccine, with parents of eight of the twelve reporting to the doctors that onset of symptoms quickly followed MMR vaccination (some as soon as 24 hours post-vaccination, others up to two weeks post-jab). One of the other four kids had been infected with measles itself, another had recurrent ear infections, and the parents of the final two didn’t report any coincidences that they had noticed leading up to onset of the syndrome.

The doctors included these parental observations in their study and added a table which included time between MMR vaccination and symptom onset for all twelve children (including the ones whose parents had not made the link themselves). The researchers went on to investigate the gastrointestinal symptoms, performing endoscopies, histology, and urinary analysis. Broadly, they showed signs of chronic inflammation in many of the children’s gut linings, but no signs of bacterial infection. They do no further analysis to explain the neurological symptoms of the children (primarily autism), but connect the dots between their twelve cases and the literature at the time in stating there seems to be a real link between issues in the gut and issues in the brain. The link between gastrointestinal issues and psychiatric disorders has been definitively corroborated by decades of further research, with scientists even proposing modulation of the gut microbiome (via pre/probiotics or fecal microbiota transplantation, for example) during early life as a potential preventative measure against autism.

Whilst the existence of the gut-brain-axis did used to be a fringe area of medical science and was pooh-poohed as pseudoscience for much of the 20th Century, discussing this connection wasn’t what got the paper retracted and Wakefield’s medical license revoked. Most know what I learned in A-level biology class: that it was retracted because of the fake data used to claim that vaccines cause autism. So where and how did they make this claim? I had never read the paper itself until a couple years ago, when I was surprised to find out that, actually, the authors make no such claim.

If your goal is to combat pseudoscience (valid) and publicly dunk on antivaxxers (tempting but ultimately self-serving), being wrong about Wakefield et al, 1998 is not going to serve you well. In actuality, the paper makes no causal claim regarding MMR vaccinations and autism. It doesn’t even claim an association. Indeed, it clarifies as much:

We did not prove an association between measles, mumps, and rubella vaccine and the syndrome described. Virological studies are underway that may help to resolve this issue.

Since the temporal coincidence of MMR inoculation and onset of symptoms for these twelve children was certainly notable, the authors do go on to speculate how investigating such a link should be carried out, and why it might be a valid hypothesis mechanistically, based on the literature available at the time. It’s probably worth noting here that the coincidence of symptom onset post-vaccination, whilst interesting, could simply be just that—a coincidence. Autism happens to appear during early childhood, which is the same period during which we choose to vaccinate humans for maximal efficacy.

One of the arguments people tend to make when questioning the link between vaccines and autism is that an increase in incidence of autism occurred in the years after MMR was introduced. Yet Wakefield et al cite literature pointing to the fact that actually there was no strong evidence of an increase in autism prevalence at the time. Indeed the apparent increase in prevalence to this day may simply be a result of broadened diagnostic criteria, increased surveillance, incentivised diagnosis, and even immigration of predisposed populations. Wakefield et al also cast doubt on a suggestion of a causal link, which might be inferred from their case report, by citing a 1997 epidemiological paper which analysed a large cohort of children and found zero association between measles vaccination and neurological side effects.

Wakefield et al go on to discuss potential genetic factors that have been associated with autism, namely the C4B gene which encodes a component of the complement system—part of the innate immune response. They discuss the perfectly salient point that a genetically impaired innate immune response could increase the likelihood of an adverse reaction to viral infections or attenuated virus inoculations (the basis of many vaccines). Since they found evidence of a vitamin B12 deficiency in urine samples from the children, they also suggest that this could contribute to autism—a hypothesis that is backed up by recent evidence.

I can’t believe I’m writing this, but … the 1998 Lancet paper is actually not bad. Decent even. It doesn’t overstate its claims; it speculates sensibly in the discussion section (this is the point of a discussion section) whilst citing high quality literature; and it pre-emptively caveats any implication that vaccines could cause autism. So why was it retracted?

The cracks begin to show

In the years following the paper’s publication, there was a slew of back and forth publications, each leaning one way or the other. For example, a serological study showed a link between a measles immune response (which could be produced by either vaccine or infection) and autoimmunity in the brain of autistic children. On the other hand, epidemiological evidence, though often underpowered, argued no association between MMR and autism.

Pressure was put on the Lancet, the Royal Free Hospital, and the UK General Medical Council to squash the hypothesis put forward by Wakefield and others, both in light of the disproving epidemiological correlations, and a decrease in vaccine uptake which many blamed on Wakefield.

Notably, MMR uptake was already on the decline before Wakefield’s paper, and insisting Wakefield was the cause here ironically falls foul of the same correlation-causation fallacy that critics accuse Wakefield of regarding vaccines and autism.

Additionally, an increase in uptake of the separate measles, mumps, and rubella vaccines coincided with the dip in uptake of the combined MMR shot—meaning total coverage was probably pretty stable over this period. Perhaps as a result, despite the frenzy, there was no real pubic health consequence post-Wakefield, with no notable increase in measles deaths post-1998. Indeed, there have been only nine deaths occurring in UK between 2000 and 2023 that could have been prevented by MMR vaccine (there was also one measles death in this time period in a vaccinated individual).

Amongst the panic and blaming, an investigation opened up into the validity—both ethical and scientific—of the Wakefield 1998 paper. It transpires that a group of lawyers, representing parents of perceived vaccine-injured children, had commissioned (i.e., were paying) Wakefield to perform a pilot study to gather evidence of a potential relationship between MMR and autism to be presented in their legal case. Some of the subjects in the 1998 Lancet paper were also used in this parallel study and neither this overlap nor the existence of the parallel study were disclosed by Wakefield in the original paper. Herein lied the first conflict of interest.

In 2004, The Lancet editorial board released a summary of the investigation, indicating that ethical approval for the study was valid, but that a conflict of interest should have been disclosed. Wakefield’s own statement argued that there was no actual conflict of interest, since he claimed that the two studies were separate, yet the Lancet editors rebuked that the “perception of conflict of interest” is what mattered, which I would agree with. There is no harm in declaring something that may or may not be an actual conflict of interest.

In this same 2004 issue of The Lancet, the Royal Free Hospital, its NHS Trust, and the UCL Medical School released a statement clearing Wakefield and co-authors of any medical malpractice, but indicating that key conflicts of interest should have been disclosed. Additionally, ten of the twelve authors (excluding Andrew Wakefield and Peter Harvey) published a “retraction of an interpretation”.

We wish to make it clear that in this paper no causal link was established between MMR vaccine and autism as the data were insufficient. However, the possibility of such a link was raised and consequent events have had major implications for public health. In view of this, we consider now is the appropriate time that we should together formally retract the interpretation placed upon these findings in the paper

As already discussed, these “major implications for public health” are fairly nebulous and non-existent. Nonetheless, clarifying the interpretation that these authors placed on their own work is a very useful practice, and surely would have helped prevent over-interpretation of the original paper (even though over-interpretation—in both directions—seems to be rampant to this day). As for Wakefield’s exclusion as an author of this statement, I can only imagine that he genuinely believed that vaccines could cause autism and so refused to take part in this clarification of interpretation.

The whole saga drew to a close in 2010, twelve years after publication, when the UK General Medical Council concluded their investigation and deemed Wakefield unfit to practice. They cited the conflict of interest already discussed, but also some patents Wakefield had filed for a proposed “treatment of an MMR virus mediated disease”. Again, nothing wrong with filing this patent, but failing to disclose it on the 1998 paper is the issue. Beyond these undisclosed financial conflicts of interest, the GMC also did find, contrary to the earlier investigation by the hospital itself, that appropriate ethical approval for all patient interventions was not acquired.

A few days after the GMC ruling, The Lancet issued a full retraction of the original paper, specifically because:

claims in the original paper that children were “consecutively referred” and that investigations were “approved” by the local ethics committee have been proven to be false

This disregard of ethical approval seems to be the real kicker, since failing to disclose conflicts of interests normally doesn’t result in retraction. At least, it didn’t for this 2011 editorial in the British Medical Journal titled “The fraud behind the MMR scare”, which itself failed to declare that the BMJ receives advertisement and sponsorship revenue from MMR vaccine producers Merck and GSK… oh the irony.

Ok science, bad scientist

My takeaway from all of this is that:

• The actual findings of the 1998 paper are valid but minor contributions to the field and are fairly interesting

• The 1998 paper was and is erroneously interpreted (by both sides) as arguing that vaccines cause autism, which it does not. Over-zealous journalists probably contributed to this distortion of interpretation (you don’t hate journalists enough… a topic for another time perhaps)

• Wakefield cut corners and failed to disclose financial conflicts of interest, thus becoming his own worst enemy if he genuinely wanted to move the field to investigate neurological side-effects of vaccines.

So is there a link between vaccines and autism?

The best data we have come in the form of two very large retrospective cohort studies from Denmark—one in 2002, and the other in 2019. The 2002 study, published in the New England Journal of Medicine, looked at over half a million Danish kids, with around 82% of them receiving the MMR vaccine. The results report on the “adjusted relative risk” of autistic disorder or autistic-spectrum disorders.

For readers who are unfamiliar, a relative risk or hazard ratio of 1.00 means that there is neither an increase or decrease in risk associated with the intervention (vaccination here). Below 1 implies negative association (less risk), above 1 implies positive association (more risk). The 95% confidence interval (CI) describes the range that would contain the true RR/HR value in 95% of all possible random samples, if the sampling process were repeated many times. A bit confusing, but basically if the upper bound falls below 1, the negative association is considered to be statistically significant, whilst if the lower bound exceeds 1, a positive association is considered to be statistically significant. Below is the results table from the study:

Whilst there was no statistically significant association between MMR vaccination and autism overall, binning the data by “interval since vaccination”, you can see that there is, in fact, a statistically significant negative association between MMR vaccination and autistic-spectrum disorders at 6-11 months (ARR = 0.31; 95% CI = 0.10 to 0.91) and 18-23 months (ARR = 0.47; 95% CI = 0.26 to 0.86) post-vaccination. This is fascinating. If the analysis is solid and the data are to be believed, this means that MMR vaccination could actually protect children from autism in these time windows! Is there a link between vaccines and autism according to this study? Yes. Yes there is. In the opposite direction.

If true, one explanation could be that the protection against the actual diseases decreases risk of autism. In such a scenario, the viral infection (but not vaccines) might promote autism—something seen in one of the children in the 1998 Wakefield case report who was infected with measles itself prior to autism onset. Alternatively, the vaccine could be suppressing autism via some other neuro-immune mechanism.

Okay, that’s interesting. Well maybe it was a fluke with this particular half million+ kids. Can we find a different or more statistically powerful study to compare?

Enter: the 2019 study. Published in the Annals of Internal Medicine, it was performed by the largely same group of scientists, and was an improvement of the first study. They used a non-overlapping cohort which was larger and they also followed up for longer. Additionally, they included analysis of subgroups at higher risk of autism based on environmental factors (e.g., if the mother smokes during pregnancy) and familial factors (a sibling with autism)—since failing to include these subgroups was a criticism of the first paper.

The headline result: no association (autism hazard ratio = 0.93; 95% CI = 0.85 to 1.02). You’ll note the upper bound of this confidence interval is pretty close to 1 though. When ending the followup at 3-years post-vaccination (the earliest they went this time), the aHR was 0.73 with 95% CI of 0.53 to 1.00—a borderline significant negative association.

When plotting the cumulative incidence of autism, you can see that by age 14, the relative incidence of autism in the vaccinated group is significantly negative.

If the data are of both this and the 2002 cohort study are to be believed, then either we declare that vaccines might prevent autism, or we add caveats to our conclusions from the analysis. Sure—there is no positive association, but the analysis was two-tailed and evidence of a negative association is there. That’s just what the data show.

At the moment it just seems like the scientific and medical establishment choose to ignore this blip in the results, and are happy to conclude no association. I haven’t seen mention of it anywhere, least of all in the papers themselves. My suspicion is that the negative association is not a biologically real phenomenon, even though it is statistically significant within certain time windows. It could perhaps be explained by the adjustments for confounders (or lack thereof). Frustratingly (or not, depending on where you stand on this issue), the creation of a negative association out of thin air by failing to account for unknown confounders would, by the same logic, destroy any biologically real positive association.

This little blip in the data may seem inconsequential to most, or even helpful to the pro-vaxxers amongst us, but to me it is an amber flag. My personal opinion is that this is at least of interest and warrants further investigation. Surely Big Pharma, too, would be interested in a study looking into the prospect that vaccines might prevent autism.

Our best evidence could be stronger

It’s hard to know whether to believe the negative association seen in these studies or not. The solution (as always) is more research. There are three main reasons why we might be unsatisfied with the results of these large cohort studies:

1. Large cohort studies are not appropriate for proving or rejecting causality. Biomedical studies can be designed in different ways, with certain methods of investigation more rigorous than others. Generally speaking, you want a large study where statistics can be applied to rule out happenstance results, and you want to eliminate bias or confounding variables as much as possible. In terms of proving causality, experimental studies are more powerful than observational ones (which can only show a correlation by definition). See the infographic below for the breakdown.

   

   Note that the 1998 Wakefield paper was a case report, the least powerful type of study. Therefore anyone claiming that it proves anything simply doesn’t understand study design. Now these Danish studies, which are typically pointed to as the best evidence we have that MMR vaccination does not cause autism, are nation-wide cohort studies. They are also retrospective, meaning the specifics of analysis were figured out after the data were already out there. A prospective study, where analysis methods and parameters are decided upon before data collection, is slightly less prone to bias since you can’t hypothetically pick and choose stats tests to massage your data into giving your preferred result. Whilst they are pretty powerful (huge sample sizes), they are observational in nature and cannot offer strong conclusions about causality. A large randomised controlled trial would offer more conclusive results.

2. Studies are not always generalisable between nations. Given the well-established genetic inputs to autism, as we briefly covered earlier, comparisons of genetically distinct populations can be expected to produce different results. It is possible (though unlikely) that the Danes have fewer vaccine-triggered autism-associated genetic polymorphisms (if those exist at all…) than the English. The same is true for environmental inputs, including things like food supply, circulating infections, air quality, climate, etc., which all differ between countries. A specific but notable difference between countries is the vaccine schedule. The schedules differ between the UK and Denmark, for example. In the UK, doses are given at 12 months and then again at 3 years and 4 months (this booster was given at 4 years before 2013). Meanwhile, in Denmark doses are given later at 15 months and 4 years of age, with the second dose actually being given at 12 years of age pre-2008, when most data for both studies were collected. Giving doses at different stages of development might be expected to have different impacts. Thus, any positive or negative associations between autism and this vaccine schedule are simply not applicable to other vaccine schedules. Notably, the US vaccine schedule is an absolute Frankenstein’s monster compared to those of the Europeans.

3. Financial conflicts of interest. As we learned from the Wakefield scandal, conflicts of interest, particularly when undisclosed, are problematic. Even if scientists are aware of their own conflicts of interest, they can do little to combat any resultant subconscious bias. The 2019 study laudably disclosed funding from the Novo Nordisk Foundation. It’s pretty tenuous, but a potential issue arises here because the Novo Nordisk Foundation owns a controlling stake in Novo Nordisk (via its subsidiary Novo Holdings A/S). Novo Nordisk (via its subsidiary Novo Nordisk Pharmatech A/S) sells products to vaccine producers for their manufacture. So obviously painting vaccines in a good light would help Novo Nordisk Foundation and they’d be more likely to fund researchers amenable to their bottom line. I don’t actually think this impacted the studies, but it’s an obvious point of criticism for skeptics.

More research, not less

In general, I am a research maximalist. Truth and knowledge are my fundamental values. For this reason, I disagree with the recent decision to pull NIH funding for studies into transgender mice1. Equally, I see nothing inherently wrong with gain-of-function research2. Further, my primary emotional response to Jiankui He, of CRISPR babies fame, is inspiration rather than condemnation.

Hopefully we can acknowledge the lack of perfect certainty regarding any positive or negative association of MMR vaccination and autism based on our best studies, herein described. Hopefully we can also acknowledge the lack of generalisability acr

oss different countries and populations, and a paucity of very convincing evidence in the US specifically. In light of these acknowledgements, I think there are valid scientific reasons to support RFK’s supposed move to fund research into this, regardless of his personal motivations or hypotheses.

Politically, it also makes a lot of sense. If we assume, as many of us do, that vaccines have absolutely no positive or negative association with autism and that the immune system is largely uninvolved in its aetiology, we must acknowledge that there is a non-insignificant population of people in both America and elsewhere that are not as convinced as us enlightened ones. Allowing the most vaccine-skeptic administration in recent history to fund such research is the best possible chance we have to persuade everyone of the truth, whatever it is. An RFK-instigated, NIH-funded, pharma-independent, large, multi-centre, randomised control trial that conclusively shows no role of any vaccination in the development of autism would be the holy grail of silver bullets many of us are looking for to assure the skeptical public of the safety of vaccination. Dr Jay Bhattacharya, newly appointed director of the NIH, has said on the matter:

I don't generally believe that there is a link [between autism and vaccines] ... But what I have seen is that there's tremendous distrust in medicine and science coming out of the pandemic.

I have to say, I agree. I must end this essay the same way I started it—by asking “Do vaccines cause autism?” Let’s get the studies done.

Share

Leave a comment

Subscribe now