Carl Zimmer writes in the New York Times April 15, 2026:
Some researchers hold that evolution hasn’t much altered humans in the past 10,000 years. A new analysis of ancient DNA indicates that natural selection continued to shape hundreds of genes.
Many scientists have contended that humans have evolved very little over the past 10,000 years.
A few hundred generations was just a blink of the evolutionary eye, it seemed. Besides, our cultural evolution — our technology, agriculture and the rest — must have overwhelmed our biological evolution by now.
A vast study, published on Wednesday in the journal Nature, suggests the opposite. Examining DNA from 15,836 ancient human remains, scientists found 479 genetic variants that appeared to have been favored by natural selection in just the past 10,000 years.
The researchers also concluded that thousands of additional genetic variants have probably experienced natural selection. Before the new study, scientists had identified only a few dozen variants.
“There are so many of them that it’s hard to wrap one’s mind around them,” said David Reich, a geneticist at Harvard Medical School and an author of the new study.
He and his colleagues found that a mutation that is a major risk factor for celiac disease, for example, appeared just 4,000 years ago, meaning the condition may be younger than the Egyptian pyramids.
The mutation became ever more common. Today, an estimated 80 million people worldwide have celiac disease, in which the immune system attacks gluten and damages the intestines.
The steady rise of the mutation came about through natural selection, the scientists argue. For some reason, people with the mutation had more descendants than people without it — even though it put them at risk of an autoimmune disorder.
Steve Sailer takes a victory lap. The Akbari et al. paper in Nature is a serious finding. Reich’s lab is the best ancient DNA operation in the world, and the methodology shift Rajagopal describes, modeling time longitudinally rather than comparing populations cross-sectionally, looks like an advance. A twenty-fold increase in detected selection signals is not a marginal result. The core empirical claim, that natural selection continued shaping human genomes through the last 10,000 years at levels much higher than the field assumed, is now mainstream science in the most prestigious venue the field has.
That much is settled. The interesting question is what Sailer is doing with it.
He is running a specific move he has run for years. He finds a mainstream scientific result that has some overlap with something Cochran and Harpending or other heterodox figures argued earlier, and he treats the overlap as vindication of the whole heterodox project. The logic is: they said evolution continued, mainstream science said it did not, now mainstream science says it did, therefore they were right about everything downstream. The downstream claims are where the work happens. Whether selection operated strongly in the last 10,000 years is one question. What it selected for, whether it produced group differences in cognitive or behavioral traits that matter today, and whether those differences explain contemporary outcomes, are separate questions. The Nature paper addresses the first. It does not settle the others.
Rajagopal flags this directly in the passage Sailer quotes. He notes that the polygenic scores used to detect selection on cognitive traits come from GWAS conducted in modern industrialized populations, and how well those scores capture what was being selected in ancient environments is debatable. He also flags that benefit-cost profiles reverse across environments, which cuts against any simple reading of ancient selection as explaining present adaptive superiority. Sailer quotes the caveat and moves past it without engaging. That is the tell.
The 10,000 Year Explosion by Gregory Cochran and Henry Harpending made several claims. The core claim that human evolution accelerated during and after the agricultural transition is looking more right than it did in 2009. Their more specific claims, about Ashkenazi cognitive evolution, about group differences in behaviorally relevant traits, about the timing and magnitude of specific cognitive shifts, are not directly tested by this paper. Sailer conflates the core claim with the specific ones because the conflation serves his coalition.
Turner on tacit knowledge applies here. The mainstream population genetics community and the heterodox HBD community both make claims that rest on tacit background assumptions their opponents do not share. The mainstream position that recent human evolution was minimal rested on methodological choices and priors that the Reich lab is now showing were wrong. Good. But the HBD inference from “evolution happened” to “therefore the specific group differences I want to claim are real and causal” rests on its own tacit moves that this paper does not underwrite. Sailer treats the first correction as authorizing the second inference. It does not.
The Lander detail is worth pausing on. Sailer flagged in 2024 that Eric Lander co-authoring the preprint signaled the result could not be suppressed because Lander is establishment royalty. That prediction turned out right. Nature published it, the Times covered it, Carl Zimmer wrote it up straight. This is a real data point about how suppression works and does not work. Sailer’s coalition narrative about mainstream science hiding inconvenient findings has to account for the fact that mainstream science, in the person of Lander and Nature and Zimmer, is publishing this finding. The suppression model is not wrong in all cases. It is more selective and more contingent than the heterodox coalition treats it.
The closing line, that white scientists doing ancestry research on non-Whites is racist so you may have a while to wait, is Sailer signaling to his coalition that he still believes the suppression thesis even after his own example cuts against it. The Reich lab’s next paper will probably cover non-European populations. Reich has said as much in public. The grievance is rhetorical rather than predictive.
The Akbari paper is an example of a contested empirical claim moving from heterodox to mainstream through better data and better methods. That is the Amy Wax-adjacent question, the one her coalition keeps pressing. Sometimes the suppressed claim turns out to be right and the institutional resistance turns out to be wrong. Sometimes it does not. The Reich paper is evidence that the process can work. It is not evidence that every suppressed claim is correct or that every institutional resistance is cowardice. Sailer treats it as the second. The lesson is closer to the first.
The frame that cuts against both the mainstream dismissal and the Sailer-style victory lap is Stephen Turner’s point that tacit knowledge operates on all sides and that empirical corrections inside a field do not license sweeping coalition claims outside it. Akbari et al. corrects a specific methodological blind spot. It does not authorize the HBD coalition’s full ideological package. Holding that distinction is the honest move.
Turner argues that expert authority in modern institutions rests on knowledge claims practitioners cannot fully articulate or transmit through explicit argument. The expert says: I have looked at the data with proper training, and what I see is X. The layperson or outsider cannot replicate that seeing without the formation. The tacit component, the trained judgment that decides which signals to weight and which to ignore, resists full specification. This is what gives expert guilds their durability and their power. It is also what makes their claims hard to falsify from outside, because any outside challenge can be dismissed as reflecting a failure of formation rather than a disagreement about evidence.
Applied to the population genetics field before Reich, the story Sailer wants to tell is that the mainstream held a tacit consensus that recent human evolution was minimal, and this consensus was ideologically motivated suppression of a truth that outsiders like Cochran and Harpending had already grasped. Turner would partially accept this framing and partially push against it.
Turner’s whole point is that tacit consensus operates in every field, and that the tacit consensus often encodes coalition commitments that the field’s members cannot see because they are inside the frame. The pre-Reich consensus that evolution mostly stopped 10,000 years ago was not purely empirical. It rested on background assumptions about how to model selection, which comparisons to run, what effect sizes to expect, and what would count as a surprising finding. Those assumptions had moral weight because they foreclosed certain downstream inferences about group differences. The field operated within a tacit framework that made the inconvenient finding harder to see. Akbari et al. show that a different methodological choice, modeling time longitudinally, reveals twenty times more selection signals. The methodological choice was always available. The field did not make it.
Cochran and Harpending were not operating from an ideology-free vantage point. They had their own tacit framework, which privileged certain kinds of evidence, weighted certain causal stories over others, and carried downstream commitments about group differences that shaped which questions they asked and which findings they highlighted. Turner would say they were not outside the tacit entirely. They were inside a different tacit framework, one more congenial to heterodox conclusions. When their framework partially matched the Reich lab’s empirical findings, Sailer reads this as vindication of the framework as a whole. Turner would say this is the standard move of every displaced expert coalition: when the mainstream corrects itself in your direction on one point, claim authority over all the points your framework touches.
The tacit knowledge involved in the Reich finding itself deserves attention. The Akbari paper is not raw truth unmediated by expert judgment. It rests on specific methodological choices: which ancient samples to include, how to handle genetic drift, how to model migration, which polygenic scores to use, how to interpret selection on traits whose modern measurements may not map onto ancient environments. Rajagopal flags this last point and Sailer quotes it without engaging. The finding that polygenic scores for cognitive traits show positive selection is not a direct observation. It is a chain of inferences, each link of which depends on tacit judgments about what counts as a valid measurement, what counts as a reasonable model, what counts as a signal versus noise. Turner’s framework insists that we see these as tacit choices, not as neutral descriptions of reality. This does not mean the finding is wrong. It means the finding is expert knowledge of the same kind Reich’s field produced before, embedded in trained judgment that outsiders cannot fully audit.
Sailer’s rhetorical move is to treat the mainstream’s tacit knowledge as ideological distortion and the heterodox coalition’s tacit knowledge as plain seeing. Turner’s framework does not license this asymmetry. Both coalitions operate on tacit foundations. The test is not which one is tacit-free. The test is which one produces better predictions over time and which one survives methodological refinement. On the narrow question of whether recent selection was strong, the heterodox side has scored a real point. That is a reason to take their broader claims seriously enough to examine, not a reason to grant the broader claims on authority.
Sailer uses Lander’s co-authorship as evidence that the mainstream can be forced to publish inconvenient findings when the establishment figure is heavy enough. Turner would say Lander’s involvement is the tacit knowledge system working correctly, not being bypassed. The guild has internal mechanisms for updating when the evidence accumulates past a threshold, and Lander’s willingness to co-author signals that the result had passed the threshold within the guild’s own criteria. This is the expert community running its own update process. The heterodox narrative treats this as suppression being defeated. Turner would say it is suppression never having been the right model for what was happening. The field was operating within a tacit framework that made the finding hard to see, and when a different methodology made it visible within the field’s own standards, the field updated. That is different from ideological enforcement.
There is no view from nowhere available. Sailer presents the heterodox reading as what any honest person would see once the ideological blinkers come off. Turner would say that presentation is itself a tacit knowledge claim, the claim that naive empiricism dissolves the distortions introduced by ideological framework. It does not, because the choice of which empirical findings to treat as load-bearing, which to treat as peripheral, which patterns to read as causal and which as coincidence, all depend on background assumptions that resist full articulation. Sailer’s empiricism is not a neutral method applied from outside the culture wars. It is a tacit framework with its own unverifiable foundations, claiming the authority of objectivity exactly the way the mainstream claims the authority of expertise.
The Reich finding is important and the Sailer response to it is diagnostic. What the finding shows is that a field’s tacit consensus can be wrong in ways that outsiders spotted earlier. What the Sailer response shows is that being right on the narrow point does not confer authority on the broader framework, and that heterodox coalitions run the same tacit-authority plays as the mainstream coalitions they criticize. The honest position holds both insights at once. The mainstream should have caught this sooner and the heterodox coalition should not be allowed to cash the correction for more inferential credit than it earned.
