Astral Codex Ten

schizophrenia

62 min read

schizophrenia

Article

schizophrenia is a recurring concept in the Astral Codex Ten archive, appearing 17 times across 17 issues between January 28, 2021 and March 26, 2026. The archive places it in contexts such as “schizophrenia, which is just a really obvious separate taxon”; “Some, like schizophrenia and ADHD, are very genetic”; “risk genes are shared with other psychiatric disorders, like schizophrenia”. It most often appears alongside US, bipolar, Europe.

Metadata

  • Category: Concepts
  • Mention count: 17
  • Issue count: 17
  • First seen: January 28, 2021
  • Last seen: March 26, 2026

Appears In

Source Context

Recovered passages from the original issue text. When the raw archive preserved outbound links inside the source passage, they are listed directly under the quote.

Ontology Of Psychiatric Conditions: Taxometrics
January 28, 2021 · Original source
In the Brief Taxometrics Primer I’ve been trying to loosely follow in this post, Beauchaine strongly suggests, though doesn’t say outright, that most psychiatric disorders are dimensional, like height and wealth. But a few may be objective distinct categories, especially schizophrenia, narcissistic personality, and endogenous depression (a subtype of depression that happens for no reason, as opposed to the kind of depression you get when something bad happens). This is a completely reasonable set of findings which match my intuition and the intuitions of most other psychiatrists.
Inline links: Brief Taxometrics Primer
And intermittent explosive disorder! This is the fake disorder we made up so that we had something to diagnose angry people with! I don’t think anyone thought this one was real, not even the psychiatrists who invented it and stuck it in the DSM! But here come the taxometricians, saying that schizophrenia and bipolar and whatever are mere dimensional variation, but IED (yes, that’s really the acronym) is an honest-to-goodness crisply-defined category based on objective reality? What gives?
And look what isn’t on here – schizophrenia, which is just a really obvious separate taxon. You know, the condition where sometimes between ages 18 and 25 for men and ages 25 to 35 or so for women, over the course of a few weeks, seemingly normal people start getting extreme hallucinations and eventually devolve into a state where they often can’t live a normal life or even speak meaningful sentences? The taxometricians are saying ah, whatever, it happens to all of us, they’re just the people who it happens to more than average? Is that really what they’re saying? I think they may have bungled this one, maybe by over-focusing on studies of negative symptoms, or of psychosis more generally, but I haven’t looked into it enough to be sure.
Ontology Of Psychiatric Conditions: Tradeoffs And Failures
February 11, 2021 · Original source
If, as Badcock and Crespi claim, schizophrenia is the reverse of autism, it might involve being too imprecise - too willing to declare the identity of unlike data - too quick to pattern-match. If autistics are too quick to mistake signal for noise, schizophrenics are quick to mistake noise for signal. In a well-functioning brain, this makes them creative and socially adept; in a poorly-functioning brain that is constantly getting things wrong, they connect everything to everything else and it's a mess.
Inline links: as Badcock and Crespi claim
Most psychiatric disorders are at least partly genetic. Some, like schizophrenia and ADHD, are very genetic, probably 80% plus. This is strange, because having psychiatric disorders seems bad, so you would expect evolution to have eliminated those genes. Researchers looking into this question argue between two hypotheses.
First, a failure. Evolution is imperfect, so some bad genes manage to slip through. This sounds dismissive, but it's definitely true to some degree. Thousands of different genes contribute to risk for conditions like ADHD and schizophrenia, with each adding only a tiny amount of risk. When a gene is only very slightly bad, it takes evolution millennia to get rid of it, and during those millennia people are getting new very-slightly-bad mutations, so it all balances out at a certain level of bad genes per generation. Those bad genes are sufficient to explain the existing amount of ADHD and schizophrenia; they're just evolution not working as well as we'd hope.
Vitamin D: Much More Than You Wanted To Know
February 16, 2021 · Original source
And also, black people get COVID 1.4x more than white people, and die of it 3x more often. There are lots of potential social causes for health disparities between black and white people. But among potential biological causes, one of the most important is that black people have much less Vitamin D at temperate latitudes - their dark skin blocks the sunlight that would usually help them produce it. This is another pattern that sometimes means Vitamin D could be involved (ask me about schizophrenia rates sometime!), although there are obviously lots of other things that could be going on here.
Inline links: black people get
The loose ends that bother me the most are the seasonal pattern, the latitude data, plus the increased risk of hospitalization and death in Asians. I don't have a great explanation for those. One possibility is that sunlight does help prevent coronavirus, it just isn't Vitamin D mediated. I suspect that the "anything involving sunlight is Vitamin D" assumption a lot of epidemiologists have isn't going to hold up very well - this seems especially true for cancer, where sunlight matters a lot but study after study has shown Vitamin D doesn't help at all. It may also be true for schizophrenia, although I'm going off one really pathetic study there and it could very well turn out to be Vitamin D after all. Some people are doing a little bit of work to clear up what the sunlight-related-Vitamin-D-independent pathways might be; I think nitric oxide has come up a few times.
Inline links: nitric oxide
Welcome Polygenically Screened Babies
July 01, 2021 · Original source
I often have patients ask me something like: "I have a history of schizophrenia in my family. I'm really concerned my kid might get schizophrenia. What can I do to prevent this?" Right now I don't have a lot of answers, besides just staying generally healthy during pregnancy and making sure the kid has a healthy upbringing. But with polygenic screening, you start to get more options. You can IVF lots of embryos, test all of them for genetic schizophrenia risk, and implant whichever one gets the lowest score. How much does that help? LifeView, the pioneering polygenic screening company, has some helpful calculators:
What if you have a family history of multiple conditions, like schizophrenia and breast cancer? Wouldn't that mean decreasing returns on selection? Unless by coincidence the embryo with the lowest schizophrenia risk also has the lowest breast cancer risk, you have to choose to minimize one or the other, right?
Book Review: Crazy Like Us
July 15, 2021 · Original source
The Zanzibaris associate schizophrenia with spirit possession (this seems to be a theme; I assume there is some sort of inspector who comes around and makes sure you attribute mental illness to demons, and if not, they take away your indigenous society license). But they are weirdly blase about this. Their position is that everyone gets possessed by spirits sometimes. If you lose your temper and lash out, that’s a spirit. If you act out of character sometimes, that’s a spirit too. Schizophrenics are possessed by stronger and more dedicated spirits than the rest of us, but it’s a difference in degree, not in kind. When a schizophrenic has a period of lucidity, they interpret it as the spirit having left for a bit, while understanding that it might come back at any time.
Watters focuses on two things he thinks Zanzibaris do right. First, they minimize schizophrenia. Because of the spirit-possession aspect, instead of being marked apart as ill and unusual, they’re treated as on a continuum with everyone else (since we all get possessed by spirits sometimes). Even when they take a more medicalized perspective on schizophrenia, they call it by extremely vague terms that don’t differentiate it from mild illness, eg “an attack of the nerves” (this seems to be a universal developing-world euphemism for schizophrenia, shared by eg Latin Americans). Zanzibari schizophrenics never feel that different from anyone else - everyone gets possessed by spirits sometimes, everyone gets attacks of the nerves sometimes, lots of people never leave their family homes.
Second, Zanzibar is low expressed emotion. Lots of studies have shown that schizophrenics do best in low-expressed-emotion households. Sometimes this is glossed as “people don’t yell at them or criticize them”, but other times it’s taken more broadly, to also include fussing over them and praising them and getting excited about them. At its worst, this line of research sometimes bleeds into the bad old theory that schizophrenia is caused by overly-attached mothers, but some studies suggest it has value - schizophrenics in high-expressed-emotion households seem to have many more relapses, and some studies show that schizophrenics separated from their families do better than those who stay with them (presumably because staff are less emotional).
Your Book Review: Consciousness And The Brain
May 13, 2022 · Original source
A part that really made me think is Dehaene's theory of schizophrenia. I have heard a lot of explanations for schizophrenia, and most of them sound superficially compelling, but collapse pretty quickly when you dig into them. This one... has not collapsed yet, at least not for me. Deahene believes that schizophrenia is what you get if you (partially) lose conscious perceptions. This sounds ridiculous at first, but as always, Dehaene makes it really hard to shrug this off as obvious nonsense. He claims that in order to reach consciousness, a masked signal needs to last "much longer" for schizophrenic patients. I expected that "much longer" means "whatever passes your p-value test", but I looked into the study. It was 90ms versus 59ms. That’s a lot. The maximum of the control group (n=28) was roughly the same as the average of schizophrenic patients (n=28)! And of course, it doesn't stop there. Also for other effects of consciousness, schizophrenics really stick out. (The link is worth clicking!) Dehaene describes how the other neural signatures of consciousness are disturbed or even plainly missing. He describes how schizophrenia can be caused by certain neuro-anatomic damage (in terms of regions and neuron types), which happen to be exactly the type of damage that would impair consciousness. He tells compelling stories on how consciousness would explain the positive and the negative symptoms of schizophrenia. He mentions some weird autoimmune disease which gives you super-strong schizophrenic effects (from hallucinations to paranoia) until you spontaneously lose your consciousness after three weeks and possibly never regain it. I am still cautious about the story, simply because I have a really low prior that we ever find a theory of schizophrenia. But if you hope for such a theory, you should know this candidate.
Inline links: schizophrenics really stick out, positive, negative, weird autoimmune disease
Book Review: San Fransicko
June 23, 2022 · Original source
The San Francisco districts with the highest (left) and lowest (right) homelessness rates. I correlated homelessness rate and population-adjusted density in the same cities I looked at above, but it didn’t add much predictive value to housing prices. Maybe this is restriction of range (all big cities are dense enough to have homelessness, compared to suburbs), or maybe the key feature is relative rather than absolute density (ie the homeless will go to the densest place nearby). Conclusion: No social phenomenon is ever caused by just one thing, but San Francisco’s homelessness rate is around where a housing-cost-based model would predict. San Fransicko briefly touches on this, but overall tries to de-emphasize it in favor of talking about drugs and mental illness. Critiques of patterns of emphasis are necessarily subjective, but the book’s pattern feels misleading to me. Claim 2: Standard Accounts Underemphasize The Role Of Drugs And Mental Illness In Homelessness Having argued homelessness isn’t just about poverty, the book goes on to say we’re neglecting the central role of mental illness and substance abuse: Over the last decades there were many visible signs that homelessness was about much more than poverty and housing. Between 2010 and 2020, the number of calls made to San Francisco’s 311 line complaining of used hypodermic needles on sidewalks, in parks, and elsewhere rose from 224 to 6,275. In 2018, footage of dozens of people slumped over in an entrance to a Bay Area Rapid Transit (BART) station, many with needles in their arm, went viral. “We call it the heroin freeze,” said one local. “They can stay that way for hours.” Said another, “It’s like the land of the living dead.” For decades researchers have documented much higher levels of mental illness and substance abuse among the homeless than in the rest of the population. It’s true that just 8 and 18 percent of homeless people point to mental illness and substance abuse, respectively, as the primary cause of their homelessness, but researchers have long understood that such self-reports are unreliable due to the socially undesirable nature of substance abuse, and the lack of insight that often accompanies mental illness. Using other methods, San Francisco’s Health Department in 2019 estimated that 4,000 of the city’s 8,035 homeless, sheltered and unsheltered, are both mentally ill and suffering from substance abuse. Of those 4,000, about 1,600 frequently used emergency psychiatric services. Shellenberger’s source for 4000 homeless people having these issues is this SF Chronicle article, which seems to based off of this report. The report does estimate 4000 homeless people with mental illness and substance abuse, but it uses a yearly rather than point estimate of homelessness, and finds 18,000 rather than 8,000 people. That means it only finds a 22% rate of these problems, not a 50% rate. Thanks to commenter Sean for hunting down this report and helping explain this. I looked for other statistics to provide context on this number. This 2013 San Francisco Homeless Count found that 29% admitted chronic depression, 15% PTSD, and 22% some other mental illness. About 30% admitted to a substance use disorder, although as far as I can tell this is just the number who admitted it was a disorder, so maybe more used drugs. This article by the Los Angeles Times describes an LA study finding that 25% of homeless people had mental health issues and 14% had drug issues. The Times re-analyzes it in a way that ups the numbers to 34% and 46%, respectively. But they don’t say exactly what choices they made differently, and the few they do give don’t really inspire confidence. Although in some cases they count questions clearly about mental illness which the official definition inexplicably refused to count, in others they decide to count anyone who has ever had mental illness, reversing a government decision to require the mental illness to be long-term (does this mean that if I lost my house tomorrow, the LA Times use me as an example of a “mentally ill homeless person” because I saw a psychiatrist for OCD when I was a kid?) Studies like these don’t show causation. Sure, mental illness can make people homeless. But homelessness can also cause mental illness. One SF study found psych diagnoses among the homeless to be evenly divided among depression, PTSD, and everything else. Homelessness is a depressing and traumatic environment. Just because someone who’s been on the streets for a year has depression or trauma, doesn’t mean that we should attribute their homelessness to mental illness. This study by the California Policy Lab does better. It asks what factors played a role in homeless people losing their homes, and finds that 50% of unsheltered and 17% of sheltered homeless point to mental illness (given SF’s balance, that suggests 37% of SF homeless would point to that problem). But I can’t help but notice that when you add up the percent of people who lost their homes due to physical illness, psych illness, and drug use, it totals 147%. Based on numbers from other studies, it looks like if you added in job loss, eviction, etc, the numbers would total well above 400%. This makes me think people are saying “yes” if the factor played even a minor role in their eventual homelessness, and this shouldn’t be treated as 37% of homeless having mental health issues being their main problem. The same study finds that about 66% of the homeless “have” some mental health problem, but this time they don’t tell us what question they asked or what criteria they use. What about psychosis in particular? This meta-analysis claims that in developed countries (a category to which San Francisco still nominally belongs) about 19% of homeless people qualify for diagnosis with a psychotic disorder, including 9% with schizophrenia in particular. Not all people with psychotic disorders are completely crazy all the time, and some very much are not, but this is at least a specific condition with real criteria. Conclusion: Overall, I’m disappointed in most of the published research on this question, which seems more interested in producing glossy brochures about funding disparities than in informing anybody what any of their numbers mean. But putting it all together and squinting really hard, I think we can tell a story where 10-20% of the homeless are seriously psychotic, and another 20-30% have contributing mental health conditions including depression, PTSD, and others. Somewhere between 25% and 50% of the homeless have substance abuse problems, and this probably mostly overlaps with the 25% - 50% who have psych diagnoses. I think San Fransicko gets this mostly right. Claim 3: “Housing First” Isn’t As Great As People Think, And Might Be Harmful The National Myth About Homelessness is that The Bad People are refusing to give people houses until they’ve “proven” they “deserve” them, thus perpetuating homelessness when they inevitably fail to qualify. The Good People have united under an exciting new banner called “Housing First” to push the revolutionary idea that people should get houses regardless of whether they conform to normal standards of respectability or not. Wherever this is adopted, homelessness rates fall, and the formerly homeless becoming healthier, safer, and more likely to re-integrate into society. Best of all, the program pays for itself in decreased health care and policing costs. The only impediment to solving homelessness everywhere is the Bad People who still insist on not housing the homeless until they’ve “earned” it. In real life, everyone important has been united under Housing First since the Bush administration made it national policy fifteen years ago, and most of the cities with spiraling homelessness crises have been pursuing Housing First policies for decades (eg San Francisco has been trying Housing First since the 1990s). The Obama and Trump administrations both set funding policies that penalized any non-Housing-First welfare programs. Still, everyone is sure that the reason there are still homeless people must be that some Housing First opponent still exists somewhere, ruining everything with their purity-testing ways. But actually these people have already been relegated to the conservative think tanks where moribund ideas go to die. I have looked through a lot of studies and articles to try to see how well Housing First works. I am most sympathetic to the conclusions of Tsai (2020), who basically says that: Homeless people who are given houses are more housed than homeless people who are not given houses. Way, way more housed. You would not believe how strong of an effect giving someone housing has on them being housed. The same is true for other outcome measures like “time spent experiencing homelessness”, “number of days spent in a temporary homeless shelter”, etc. You might think this is obvious, but this is used as the primary outcome in a lot of studies, and “success” on this metric is behind a lot of claims that “studies show Housing First works great!”
Inline links: this SF Chronicle article, this report, 2013 San Francisco Homeless Count, This article, This study, This meta-analysis, some very much are not, made it national policy, since the 1990s, funding policies, Tsai (2020)
The tragic irony is that many of the people who had drawn attention to the poor conditions in mental hospitals had hoped to mobilize the public to increase funding for better care in them, not shut them down entirely […] Idealism and ideology had triumphed over pragmatism and reason. Between 1948 and 1962, the mental health center that reformers had pointed to as the model had not prevented a single case of mental illness or even treated a single individual with schizophrenia or other major psychiatric disorders. As a result, notes a historian, “The majority of lives were little different than they had had while hospitalized . . . and a significant number were considerably worse off.” Some mental health reformers regretted what they had done.
Contra Kirkegaard On Evolutionary Definitions Of Mental Illness
September 07, 2023 · Original source
A 65 year old man who’s only attracted to adult women 40+. Most people in our society would classify 1 (an ephebophile) and 2 (a non-obligate pedophile) as mentally ill or at least worrying edge cases. But I think Emil’s theory rules that only Person 3 (the man attracted to people close to his own age) is mentally ill, since he’s ruled out mating with the vast majority of fertile women. 4: Plato …never had children. “Platonic relationship” jokes aside, I guess he was too busy philosophizing. Great men (and women) who can’t slow down to raise a family seem to be a type. Is an interest in philosophy (or science, or art, or any other worthy endeavor) that reaches the point where it consumes your life a mental illness? Kierkegaard bites the bullet and admits that the priests and monks who took vows of celibacy were mentally ill by his definition. But I think he has many more bullets of this type to bite. Even if we agree that we should classify Plato as mentally ill, this again seems very different from the practical concept of “this person has mental problems and needs help with them”. 5: Chronic Pain, Panic Attacks, Or, If You Insist, Nightmares Is chronic pain a mental illness? It seems pretty bad. But as long as it doesn’t impede your ability to hunt, gather, or have sex, I think Emil would have to say no. Same with panic attacks, anxiety, etc. If it’s hard to imagine a form of chronic pain that doesn’t impede those things, consider nightmares. These surely don’t impede any daytime activity, but chronic nightmare disorders seem very unpleasant! I think Emil has to bite the bullet that conditions which make people miserable and ruin their lives aren’t mental disorders as long as they don’t affect functioning. 6: Severity In his post, Emil includes a few turns of phrase indicating we can talk about severity - ie some mental illnesses are more severe than others. But by his framing, “severe mental illness” would indicate not schizophrenia and bipolar disorder, but homosexuality and asexuality. After all, schizophrenics are more likely to have children than gays. Again, this is pretty different from the way you want to use words when talking about real-world problems around how to help people with mental problems get better. 7: Is Emil’s Definition Of Mental Illness Itself A Mental Illness? Emil’s crusade to reclassify homosexuality as a mental illness doesn’t sound like it would be very popular in his home country of Denmark. Maybe there are even some nice Danish women who would be willing to date Emil otherwise, but are turned off by his un-PC opinions. Willingness to violate taboos couldn’t have been very helpful in the environment of evolutionary adaptedness. I imagine some distant ancestor of Emil’s standing up in front of the tribe and saying “Me think Bear God stupid and ugly! Me piss on Bear Idol!” Might mean fewer Kirkegaards around today. So is contrarianism a mental illness? I would say no, because it’s a matter of personal choice and serves a valuable social function. I’m not sure what Emil’s answer would be. * * * I don’t want to assert any of these too strongly. Maybe Emil knows something I don’t about the EEA, and can prove that actually ADHD would be maladaptive there, or ephebophilia would get you in trouble. If so, I think that would restore some concordance between our intuitive notion of mental disorders and Emil’s version, but that concordance would be coincidental, not necessary. The next day we might learn some different fact about the EEA that would make the two notions discordant again. So to repeat my claim: mental-disorder-(Emil) and mental-disorder-(Scott) both describe useful concepts, but they’re not the same concept. Mental-disorder-(Emil) is useful for talking about evolutionary genetics; mental-disorder-(Scott) is useful for talking about present day mental health problems and what to do about them. We won’t convince people to literally use the terms “mental-disorder-(Emil)” and “mental-disorder-(Scott)”. So who should keep custody of the current term “mental disorder” and who should have to make up a new word for their thing? I think Emil should have to make up the new word, because: There are a few thousand evolutionary psychologists, and a few hundred million normal people who want to talk about mental disorders for normal reasons (like because they have them).
It's Fair To Describe Schizophrenia As Probably Mostly Genetic
February 01, 2024 · Original source
Famous schizophrenia researcher E. Fuller Torrey recently wrote a paper trying to cast doubt on whether schizophrenia is really genetic. His exact argument is complicated, but I feel like it sort of equivocates between “the studies showing that schizophrenia are genetic are wrong” and “the studies are right, but in a philosophical sense we shouldn’t describe it as ‘mostly genetic’”.
Inline links: wrote a paper
Awais Aftab makes a clearer version of the philosophical argument. He’s not especially interested in debating the studies. But he says that even if the studies are right and schizophrenia is 80% heritable, we shouldn’t call it a genetic disease. He says:
Inline links: makes a clearer version of the philosophical argument
Heritability is “biologically vacuous” (Matthews & Turkheimer, 2022), and I think we would be better off if more of us hesitated to assert that schizophrenia is a “genetic disorder” based predominantly on heritability estimates.
Inline links: Matthews & Turkheimer, 2022
ACX Grants Followup Impact Market
March 07, 2024 · Original source
You’ve probably encountered vaticidalprophet in the ACX comments section or the ACX Discord, and heard their spiel about why researching schizotypy is interesting. They want $2K - $50K to research psychosis in velocardiofacial syndrome (DiGeorge/22q11.2 deletion syndrome), which might shed more light on the relationship between schizophrenia, schizotypy, autism, and psychosis. We decided not to fund this because velocardiofacial syndrome is rare, understanding it better wouldn’t directly help many people, and we weren’t convinced this would have knock-on effects for more common psychiatric diseases - but I would love to be proven wrong.
Inline links: research psychosis in velocardiofacial syndrome (DiGeorge/22q11.2 deletion syndrome)
Nobody Can Make You Feel Genetically Inferior Without Your Consent
June 12, 2024 · Original source
Lately we’ve been discussing some of the ethics around genetics and embryo selection. One question that comes up in these debates is - are we claiming that some people are genetically inferior to other people? If we’re trying to select schizophrenia genes out of the population - even setting aside debates about whether this would work and whether we can do it non-coercively - isn’t this still in some sense claiming that schizophrenics are genetically inferior? And do we really want to do this?
Inline links: we’ve been discussing
The problem with claiming that Lance is superior to me isn’t that he isn’t. It’s that it indicates I’m asking the wrong question, in order to make myself miserable. Just as the correct answer to “are schizophrenics genetically inferior?” is “haha, you can’t trick me into using the word that lets you write an article calling me a Nazi”, the correct answer to “am I inferior to Lance?” is “haha, you can’t trick me into using the word that lets you make me depressed.”
I find it clarifying to set aside schizophrenia for a second and look at cystic fibrosis.
P-Zombies Would Report Qualia
June 10, 2025 · Original source
And they wouldn’t answer “IDK, my mouth just moved and formed the syllables ‘this is red’”. Normal humans can easily tell the difference between a voluntary action and an involuntary spasm (eg if your limb jerks because of an electric current or a seizure). In fact, this faculty is so profound that its failures contribute to conditions like schizophrenia; when someone loses the ability to interpret their speech as self-produced, they start formulating hypotheses like “the CIA put a chip in my brain that controls my actions”. Since the p-zombies can do anything humans can (including distinguishing voluntary vs. involuntary actions, and getting schizophrenia) they must be able to report something other than “my mouth moved but I can’t say why”.
Inline links: contribute to conditions like schizophrenia
Contra Skolnick On Schizophrenia Microbes
June 30, 2025 · Original source
Studies find that schizophrenics have very high levels of a gut bacterium called Ruminococcus gnavus.
5: Studies find that schizophrenics have very high levels of a gut bacterium called Ruminococcus gnavus.
All the schizophrenics investigated in VYB were on antipsychotic medication. Previous studies have already shown that antipsychotic medication disrupts the gut microbiome. VYB couldn’t investigate whether the disruptions were caused by medication in their own sample, because they had no unmedicated controls, but they checked whether more medication = more disruption, and it did. They concluded that the most likely cause of the microbiome disruption was the medication:
Inline links: already shown
Highlights From The Comments On Missing Heritability
July 03, 2025 · Original source
I'm going to gently push back against the hereditarian/anti-hereditarian framing (which I understand is probably here as shorthand and scene setting). I am personally interested in accurate estimates that are free of assumptions. I believe twin study estimates are of low quality because the assumptions are untestable, not because they are high. I also think the public fixation on twin studies has created some real and damaging anti-genetics and anti-psychiatry backlash and wrong-headed Blank Slate views. People hear about twin studies, look up the literature and find that peanut allergy (or wearing sunglasses, or reading romance fiction) is estimated to be highly heritable and have minimal shared environment, start thinking that the whole field is built on nonsense, and end up at quack theories about how schizophrenia is actually a non-genetic fungal condition or whatever. I've been very clear that there are direct genetic effects on essentially every trait out there, including behavioral traits and IQ. If someone were to run a large-scale RDR analysis of IQ tomorrow and got a heritability of 0.9 and it replicated and all that, I would say "okay, it looks like the heritability is 0.9 and we need to rethink our evolutionary models". If anything, large heritability estimates would make my actual day job much easier and more lucrative because I could confidently start writing a lot of grants about all the genome sequencing we should be doing.
Inline links: peanut allergy (or wearing sunglasses, or reading romance fiction) is estimated to be highly heritable
This paragraph surprised me, because people coming up with crackpot non-genetic theories about schizophrenia is part of why I think it’s so important to explain that twin studies are usually pretty good!
Schizophrenia has about the same level of missing heritability as IQ, EA, or any other trait (80% heritable in twin studies, ~10% heritable in best polygenic predictors, ~25% heritable according to GREML). I don’t really understand on what grounds you can object to the twin heritability estimates of IQ/EA/etc, but believe the ones for schizophrenia.
Suddenly, Trait-Based Embryo Selection
July 31, 2025 · Original source
Best in what sense? Genomic Prediction claimed the ability to forecast health outcomes from diabetes to schizophrenia. For example, although the average person has a 30% chance of getting type II diabetes, if you genetically test five embryos and select the one with the lowest predicted risk, they’ll only have a 20% chance2. Since you’re taking the healthiest of many embryos, you should expect a child conceived via this method to be significantly healthier than one born naturally. Polygenic selection straddles the line between disease prevention and human enhancement.
Inline links: 2
Herasight’s numbers on how breast cancer risk goes down with number of embryos used in selection. A typical round of IVF produces 1-10 embryos (younger women usually = more). Women with polycystic ovarian syndrome (prevalence: 10%) may get as many as 20. For more, you will probably need to do multiple IVF rounds. Here is a table of different companies’ reported risk reductions, slightly adjusted7 for different reporting conventions but otherwise taking all claims at face value (we’ll talk about how wise that is later). Relative risk reduction for five conditions (gray = no data / disputed data). Here baseline is for embryos neither of whose parents have the condition. GP and Orchid both say their technology has improved since reporting these numbers and they will report better numbers soon. GP numbers are not within-family validated and might be lower if they were. Absolute risk after selection for five conditions (gray = no data / disputed data), ibid. Some people might genuinely want to select on a single condition. For example, people with a strong family history of schizophrenia might want to minimize the chance of their children getting the disease; for these people, reducing schizophrenia risk by 58% (while keeping everything else constant) sounds pretty good. Everyone else probably wants a generically healthy embryo with low risk of all conditions. Exactly how this works depends on the customer’s own values - would they prefer an embryo with lower cancer risk to one who will have fewer heart attacks? - and the exact benefits will depend on how parents make that decision. Genomic Prediction and Herasight try to help by providing semi-objective measures of which embryo is overall healthiest according to different conditions’ effects on longevity and patient-rated quality of life. For Genomic Prediction, that’s the “embryo health score” If you selected the single highest-health-score embryo from a set of five, here’s how they’d do: For Herasight, it’s a “polygenic longevity index”. They don’t give exact risk reduction numbers for each disease, saying that it depends too much on a couple’s specific family history, but say that most people gain 1-4 years of healthy life (when I test it on a set of twenty embryos, the the healthiest gets an extra 1.66 years). How much would you pay to give your children an extra 1-4 years of healthy life? This is no longer a hypothetical question. Here are the costs of the companies in this space: Is it worth it? If: You’re already doing IVF
Inline links: 7, https://substackcdn.com/image/fetch/$s_!0oUh!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe8419603-9239-43bb-8c79-77b078ff0789_548x136.png, https://substackcdn.com/image/fetch/$s_!rpEJ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc539a717-a130-460d-90c9-4ab64619f26d_548x133.png, https://substackcdn.com/image/fetch/$s_!3Kc6!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbda325bb-13fb-4c27-b8c3-24facce5c71a_676x153.png, https://substackcdn.com/image/fetch/$s_!t1Am!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F831ab3c6-4053-4ff9-bc2f-879aee4349cf_673x740.png, https://substackcdn.com/image/fetch/$s_!Q2vE!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fadcf2bb4-8dd1-4a88-9728-6953a820971b_422x575.png
That serves as proof-of-concept that this technology can work, and means other companies’ claims are at least plausible. Scientific Objections: Antagonistic Pleiotropy This is a fancy term for “sometimes genes that are good in one way are bad in other ways”. For example, there is a gene that decreases the risk of lung cancer, but increases the risk of leukemia. If you selected against lung cancer, you might give your child higher leukemia risk. Several of the professional societies raise this concern, and Sasha Gusev gives several examples here, including a correlation between education/IQ and anorexia. When I think about these concerns, I consider the following thought experiment: suppose that I had a natural, unselected child, and that child became high school valedictorian and got into Harvard. Would my first reaction be “Oh no! This slightly raises her risk of anorexia!”? If not, why should this be our reaction to artificially increasing IQ? Genetic selection isn’t doing some different, magical thing. It’s just picking from within the natural IQ/anorexia variation. If you would be happy to have higher IQ (or lower breast cancer risk, or lower schizophrenia risk) naturally, you should be happy to get it through selection too. (Objection one: suppose that the genetic component of IQ is net negative, but the environmental component is net positive to an even greater degree. Then IQ itself might be net positive - so you could still celebrate your valedictorian child - but since the genetic component alone is bad you wouldn’t want to select for it. I have never heard anyone seriously claim this, most studies suggest that genetic components of good things are good in the expected ways, and most critics don’t get this far. I mention it for the sake of completeness only.) (Objection two: is the example above just saying that I value IQ more than non-anorexia? If so, couldn’t I give an alternate example of learning that my child isn’t anorexic, celebrating this seemingly-obviously-good fact, but actually this means they have lower IQ and based on my stated values I should be sad? I don’t think so. There is no claim that the increased anorexia risk from raising IQ is exactly as bad as the IQ increase is good - for example, you could imagine a world where going from moron to supergenius only raises anorexia risk 0.0001%. More generally - although not rigorously - selecting for X should usually increase X more than it increases tangentially-correlated construct Y. So selecting for IQ should be net positive, even though it might slightly increase anorexia risk, and selecting for anorexia should be net negative, even though it might slightly increase IQ. I think this is the intuition that drives parents to be happy both when they learn that their child is smart, and when they learn their child doesn’t have anorexia - not just an intuition that one trait matters more than the other) But also, here’s the table of correlated genetic risks for psychiatric disorders: …where blue means that lowering the risk of one disease also lowers the risk of the other, and red means the opposite (as in the IQ - anorexia example above). Here’s the same table for other conditions, courtesy of Genomic Prediction (except I flipped the colors from the original, to match the one above): Aside from two bright orange squares (gallstones vs. hypertension and hypothyroidism - I don’t know what’s up with this and it doesn’t seem to be a widely-appreciated result) we see that most correlations are zero or positive - that is, selecting against one disease selects against another or at worst does nothing. In this ocean of blue, worrying about those few orange squares feels a bit motivated. Hans Jonas-ism says that no medical intervention may ever cause any harm, no matter how much benefit it produces. By this standard, perhaps slightly raising the risk of gallstones in the process of preventing various cancers and psychoses and other forms of human misery is unacceptable. To anyone with the more normal perspective where something with large benefits and tiny downsides is still pretty good, I don’t think the antagonistic pleiotropy argument carries much weight. Ethical Objection: Cost No way around this one: if these products work, they mean that rich people can have healthier/smarter/taller/prettier kids than poor people. One might object that at least they’re in good company: other products which help rich kids get healthier/smarter/taller/prettier than poor kids include private tutors, gyms, hair salons, health insurance, clothing, books, and food. Is this really the time to declare ourselves against this kind of thing? But maybe we should fight against expanding this already-bloated category. Or maybe there’s something more final about a genetic advantage. Maybe a stronger argument is that rich people get first crack at every new technology, but poor people usually follow close behind. The first cellphone, in 1982, cost $12,000 in today’s dollars. Now you can get something a thousand times better for $50, and Kenyan pastoralists use cell phones to call up the local shaman. The trajectory of genetics has been even more striking: sequencing a single genome cost about $100 million in 2000 and is somewhere around $100 today. Polygenic embryo selection has the potential to follow a similar path. There are two associated costs - sequencing the embryos, and running the analysis. Sequencing costs are decreasing and may eventually be comparable to the sorts of genetic screening (for e.g. Down Syndrome) that most families get anyway. Analysis costs are mostly the one-time expense of inventing the predictor; we might expect these to follow the same pattern as generic medications, where cutting-edge technology is jealously guarded and expensive, but last decade’s technology has made its way off patent and is cheap-to-free. A few groups have already created free open-source predictors; so far these are much worse than the private companies’ versions, but one of last year’s ACX Grantees is working on a better one. Also, it would be crazy for any forward-thinking government not to cover this; it could save hundreds of thousands of dollars in future health care expenses. In countries with public health care, this comes directly out of the government treasury; even in the US, it’s covered by Medicare after age 65. The government should be begging people to select embryos. The most persistent cost barrier is likely to be in vitro fertilization itself, a necessary precursor. In the US, 2-3% of babies are born through IVF. For those kids, this is a no-brainer - even if the cost never comes down, the cheaper products are only a fraction of total IVF expense. What about the other 98%? If those parents feel like they have to get embryo selection (and therefore IVF) to keep up, this could be a significant burden. IVF isn’t fun - it requires pumping a woman full of mind-altering hormones for weeks, extracting eggs in a minor surgery, and then implanting embryos in another minor surgery, all with a decent chance that some step will fail and you’ll have to do it all again. It also costs $15,000 in the US (less in poorer countries), and unlike the genetics, the cost has barely gone down in the past twenty-five years. Some countries, including Israel, offer free IVF for anybody who wants it. And universal basic IVF is surprisingly popular even in the usually government-phobic United States - Donald Trump made it part of his campaign platform. So there’s a plausible path to embryo selection for everyone who wants it. But it’s still going to take a while, it will hit different people at different times, and so far11 there’s no way around the month or two of various miserable medical procedures for women. Ethical Objection: Personhood Is it really correct to say that you have reduced someone’s risk of breast cancer by 46%, if what you’ve really done is closer to replacing them with a different person who is 46% less likely to have breast cancer? I cover this one in more depth here. Ethical Objection: Race This one is awkward: right now the technology works best for white people. Most genetic data available for research/commercial use comes from the UK, US, and Europe - areas which are mostly white. Asian biobanks, and those serving US minority communities, have been more reluctant to share data. So we know a lot about the genetics of white people, and only a limited amount about the genetics of anyone else. Companies are suitably embarrassed about this, and researchers in the field are working hard to wring every ounce of information out of the minority data they have. But for now, white people are the clear winner. Here’s data from Herasight: A European family with five embryos and no family history can cut their diabetes risk by 47%, and an African family 29%, with everyone else in between. As usual, all companies say that they adjust their scores based on the couple’s genetic ancestry. As usual, Herasight challenges them to publicly release data on exactly how they performed the adjustments and how well they work. All companies say they are working as hard as they can to improve cross-ancestry portability, but that progress will remain limited until governments collect/release better genetic data on non-white populations. Ethical Objection: Selection At some point, you’ve got to choose. Genomic Prediction and Herasight offer scores that aggregate overall health risks. Some people will follow them slavishly. Other people will try to second-guess them - would you prefer your child have lower cancer risk, or less chance of heart attacks? And this is the best case scenario! Herasight offers predictors for IQ, height and BMI; Nucleus offers those plus eye color and hair color12. A parent might encounter a situation where the embryo with their favorite eye color also has the highest cancer and schizophrenia risk, and choose to doom their child to cancer and schizophrenia because they really want pretty eyes. On average, even if everyone in the world selected for eye color, it wouldn’t raise cancer and schizophrenia risk. No not-deliberately-perverse polygenic selection choice can make your child worse off in expectation. Still, suppose you got cancer, and your mom admitted that she selected you for pretty eyes and didn’t even check the cancer column of the embryo selection report. How would you feel? And would you feel better or worse than someone whose parents didn’t do embryo selection at all, and spent the money on a Caribbean vacation? What if they selected your brother for everything great, then had you naturally? What if they selected you for IQ, but actually you are very stupid, and you were one of the 20% of cases where a predictor that’s right 80% of the time gets it wrong? Mark my words, one day there will be entire subfields of therapy dedicated to these issues. Going Nuclear Even as outsiders criticize the whole field, Herasight has launched a full-scale attack on competitor Nucleus. Herasight’s white paper compares its own predictors (favorably) to those of Orchid and Genomic Prediction… …but refuses to acknowledge Nucleus at all. In a supplementary note, the authors explain why: they accuse Nucleus of being so bad that it would “not yield a reliable or meaningful addition to our analysis”. They say Nucleus has inflated the accuracy of their scores. This is most dramatic for a few conditions like ADHD, where the leading published polygenic score is based on 2,300,000 variants but explains only ~1% of variance in the condition. Nucleus’ score is based on 12 variants13 and (implicitly) claims to explain 3-6%. This doesn’t make sense. Some of Nucleus’ other scores do use millions of variants. But many of these are 5-10 year old scores downloaded from open-source catalogs, whose accuracy statistics are easily available and far less than Nucleus claims. Here is what Herasight finds when they double-check Nucleus’ numbers: On their Substack, Herasight also criticizes Nucleus’ monogenic screening product. They point out cases where it fails to properly screen for the conditions it claims. For example, the Nucleus website advertises screening for spinal muscular atrophy: But on their gene list… …they don’t screen for SMN, which causes 95% of spinal muscular atrophy cases. They only screen for UBA1, which causes a distinct and much rarer condition called x-linked infantile spinal muscular atrophy. Professional organizations publish guidelines for what genes need to be screened in a screening product, and Nucleus does not appear to be following them. In further discussion, Herasight continued with exhaustive criticism of essentially everything Nucleus had ever done down to the smallest detail. Nucleus reports list the same baseline disease risk regardless of patient ancestry, but different ancestry groups should have different risks14. Nucleus’ physician reports sometimes list lower-than-average risk for patients with positive polygenic scores15. Nucleus’ age-based risk tables don’t distinguish between age and cohort effects (is this bad? see footnote16). My favorite critique is that Nucleus wrote a blog post criticizing competing company Orchid… …which included a section on how Orchid is a polygenic selection company, and polygenic selection companies are inherently “sketchy” and “honestly should be illegal”. But Nucleus is also a polygenic selection company! This is like Marlboro attacking Camel on the grounds that cigarettes are addictive and should be banned! Obviously something went wrong here - my guess is AI - and it’s a really bad look, especially when these scientific issues are so hard to litigate, and so many of us will have to go off gestalt impressions of corporate culture. Nucleus states that they validate their models internally and intend to make their results public soon. A Foothill Of The Future It’s hard not to love this technology. Lots of people (and the aforementioned professional organizations) manage anyway, but it’s hard. If this were a single-use medical treatment, delivered by a doctor after someone got the relevant condition, it would be one of the biggest advances of the decade - imagine a drug that cures 10 - 40%17 of breast cancers with no side effects! But in fact, it works for breast cancer, and schizophrenia, and heart attacks, and approximately everything else. The only things comparable are antibiotics and GLP-1RAs. And then there’s the IQ effects. Even after studying the literature, people have wildly different opinions about the importance of IQ. One of the most important debates is to what degree IQ differences are a cause of poverty, a consequence of poverty, or both. I lean towards both - a country with limited access to schools and medical care will have low average IQ, but as a consequence it probably won’t become the next big semiconductor hub. This technology could close half the IQ gap between poor and middle-income countries, or between middle-income and rich. Or it could give rich countries average IQs that have never been seen before, and let us see what kind of O-ring technologies (and new forms of social cooperation) lie just beyond the frontier. (this is the nice quantifiable argument in favor of IQ enhancement, but I find myself more convinced by fuzzier things - how much is it worth to be able to enjoy great art and literature? To fully comprehend what we know of nature, and be able to fully appreciate the mystery of the rest? To have a sense of why society works the way it does, instead of feeling like you’re being blown back and forth by institutions you don’t really understand? Amateur psychoanalysts like to say that the only people who care about IQ are those looking for an excuse to boast about how high their own is, but my experience is the opposite: I care about IQ because I bang up against the limits of my own a thousand times a day, and I hate it. I fantasize about ways to make my children smarter than I am for the same reason a dog confined in a tiny crate might fantasize about getting her puppies adopted out to a nice house with a big grassy yard.) My biggest qualm is that it might not matter. This is such a tiny foothill, flanking such a vast and foreboding range of mountains, that it might be a mistake to care about it at all. Selecting the best of five or ten embryos is not a very effective way to get the genes you want. There are things in the pipeline that will make this look like Hippocrates draining black bile. By the time the first polygenically selected children are adults, they’ll be old news. And then there’s AI. The average age at diagnosis for Type II diabetes is 45 years. Will there still be people growing gradually older and getting Type II diabetes and taking insulin injections in 2070? If not, what are we even doing here? Many people in the transhumanist community are still bullish on this technology. They think - well, there’s still an outside chance that something comes up and AGI takes another few decades. If we can enhance humans to be smarter, healthier, and more determined by the time it arrives, maybe we’ll have a better chance. Or maybe, if there’s a positive optimistic vision of a human-based high-tech future, people will be more willing to delay AI in the first place. I like this argument, but I also think it’s worth stepping back. What’s the point of anything? Why have kids at all in a world that’s changing this fast? Why save for the future? At some point your answer has to be romantic and aesthetic - it’s never been clear whether anything you do matters in any ultimate sense, but you’ve got to act as if it does and hope for the best. From that perspective, this is the most romantic technology of all. You’re not just giving a better life to your kids. Genes travel from generation to generation; you’re giving a better life your grandkids, your great-grandkids and so on to the point 1.77*log₂(population) generations from now when you are the ancestor of everybody and nobody. Somebody in Macaronesia in 3525 AD will avoid getting breast cancer because of you (if there is still cancer; if there are still breasts). Some combination of reasonable cost-benefit analysis and romantic/aesthetic commitments makes me want to have children despite the uncertainty, and the same combination made me sign up to use this technology despite the same. More later on how that’s going. 1I’m slightly mixing up two different things here - Down Syndrome can be detected with an aneuploidy test, but cystic fibrosis takes a more involved PGT-M test. 2There are two separate questions here. First, how much would diabetes risk decline if you selected the embryo with the lowest risk for diabetes - something you have no reason to do, since you have no reason to privilege diabetes risk over risk of any other disease? Second, how much would diabetes risk go down if you selected the embryo with the lowest health risk overall? Genomic Prediction’s their risk calculator calculator shows, seemingly paradoxically, that you get -38% relative risk by selecting against diabetes alone, but -41% relative risk by selecting against everything at once. Over email, they stand by this surprising result, saying that “for a couple of diseases (type II diabetes and CAD), the EHS actually accomplishes a larger risk reduction than the individual predictors. The explanation is that the EHS takes into account multiple PRS of diseases with high comorbidity”. See eg Figure 3 here: …and the section of the post called “Antagonistic Pleiotropy” for more. However, this paradoxical benefit is only true for a few conditions like diabetes - for everything else, selecting on health index does better than you would naively think, but still does not decrease the risk of a given condition as much as selecting against that condition directly. 3That is, new mutations in that particular baby, as opposed to older mutations already present in the parents. 4Conflicts of interest: I have used Orchid’s and Herasight’s products on my own embryos (not the ones used to conceive my existing kids, but for a potential third child), employees of Genomic Prediction and Herasight have been extremely helpful in contributing expertise to ACX posts on genetics, and I might invest in this field at some point (though haven’t done so yet). This post started as Herasight asking me to write about their white paper, then spiraled out of control. There were some unexpected time pressures and the result is that I didn’t get a chance to run everything in Herasight’s white paper by their competitors as thoroughly as I would like. Although I talked to representatives of all four companies profiled here, I feel like this probably reflects Herasight’s perspective better than other companies’, and that this is a major flaw. If other companies have responses, I’ll publish them. Thanks to all companies involved for their assistance on this article. Finally, I am favorably disposed toward Herasight because of how I learned about them: a professor named Jonathan Anomaly got cancelled from Penn for being too gung-ho about genetic enhancement, and used his newfound freedom to join a very-early-stage Herasight, raise their ambitions, and sell everyone (including me) on the idea. I grew up on a diet of books and movies about mad scientists, and I’m a sucker for a story about a guy named Doctor Anomaly pursuing revenge against the small-minded fools who destroyed his career by creating a race of superbabies. 5The version of the tool I looked at said 5.9 points for five embryos, up to 9 points for twenty embryos. The version of the tool on their current said says 5.3 - 9, so they might have recalculated after I finalized this article. 6Used in quotation marks because these scores were fine for the predictive tasks they were applied for - they just weren’t finding genes that directly caused the outcome of interest. 7Conflict of interest notice: this table was originally unadjusted. A representative of Herasight claimed that this was unfair, because each company used slightly different reporting conventions, and offered to correct for this in a neutral way. I retraced their reasoning, confirmed that the correction did not especially benefit Herasight at the expense of other companies, and accepted the correction. The original unadjusted table is below: Herasight was insufficiently comfortable with Nucleus’ methodology to even be willing to posit a corrected value, so I left their self-reported value in gray. 8Zagorsky (2007) says an extra IQ point means $234-$616/year in higher salary. The midpoint of $425 equals $670 in today’s dollars; assuming a forty-year career, Nucleus’ +1 point estimate is worth $26,800 (vs. $9,249 Nucleus cost) and Herasight’s +6 point estimate is worth $160,800 (vs. $53,250 Herasight cost). 9As part of researching this article, I asked all four major companies about their within-family validation strategies. Here are some details: Genomic Prediction discusses their strategy in this paper. The results are complicated to interpret - the within-family numbers often have such wide error bars that they overlap with both the across-family numbers and with zero - but looking qualitatively it seems like most scores on average lose about 25% of their risk reduction ability (though averages might not be the right way to do this, and some might be much more affected than others). Their website reports unadjusted, not within-family validated numbers; GP says they say this clearly on their site (which is true), Herasight counters that they still present their numbers as applicable to embryo selection (which is also true). To get the most applicable-to-embryo-selection numbers, you might want to adjust GP’s stated numbers down somewhat; it’s hard to say exactly how much, but maybe 20 - 25%?
Inline links: here, here’s, https://substackcdn.com/image/fetch/$s_!G_Lu!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7049550b-4253-4900-9fe1-9f2df009e829_446x432.png, Here’s the same table for other conditions, https://substackcdn.com/image/fetch/$s_!jscV!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9a3ad311-4745-4f16-b82a-7ccdb297c670_1239x1600.png, Hans Jonas-ism, somewhere around $100 today., Donald Trump made it part of his campaign platform, 11, here, https://substackcdn.com/image/fetch/$s_!1Alk!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1af69f39-f353-4aa6-acb9-d8c3b05c7bac_728x895.jpeg, 12, Herasight’s white paper, https://substackcdn.com/image/fetch/$s_!S7lY!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F738260ba-8fe2-4647-8ca2-eeb4d13e0fce_605x341.png, 13, https://substackcdn.com/image/fetch/$s_!u7YE!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F451c286b-c677-47af-8c07-f0d993a14384_612x345.png, their Substack, the Nucleus website, https://substackcdn.com/image/fetch/$s_!XmL4!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff77915b4-3b36-4908-8f9f-032b7cf865ff_562x432.png, https://substackcdn.com/image/fetch/$s_!bxjb!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2df9ffe5-17a3-4448-9a20-e9b27ac9a519_1250x795.png, publish guidelines, 14, 15, 16, a blog post criticizing competing company Orchid, https://substackcdn.com/image/fetch/$s_!MZCB!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2cf9cbf4-825b-4373-9052-80e43c36febf_718x1035.png, 17, GLP-1RAs, O-ring technologies, things in the pipeline, everybody, nobody, 1, 2, their risk calculator, everything, here, https://substackcdn.com/image/fetch/$s_!jtkY!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd645c392-fed1-4f02-9a2e-878b8c7ef7f2_909x878.png, 3, 4, 5, 6, 7, https://substackcdn.com/image/fetch/$s_!Vimq!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fca0f1f15-268d-465a-a70f-b7f1173c6111_566x166.png, https://substackcdn.com/image/fetch/$s_!3B0A!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faccc7a65-b142-4bf6-927d-53eb607d71ef_552x155.png, 8, 9
In Search Of AI Psychosis
August 26, 2025 · Original source
I think now there might be several dozen subreddit moderators who could accurately describe their job as “witch webmaster who runs an online service giving advice to new witches”. And partly it was because there are so many crazy beliefs in the world - spirits, crystal healing, moon landing denial, esoteric Hitlerism, whichever religions you don’t believe in - that psychiatrists have instituted a blanket exemption for any widely held idea. If you think you’re being attacked by demons, you’re delusional, unless you’re from some culture where lots of people get attacked by demons, in which case it’s a religion and you’re fine. This is partly political self-protection - no psychiatrist wants to be the guy who commits an Afro-Caribbean person for believing in voodoo. But it also seems to track something useful about reality. Nietzsche wrote “Madness is something rare in individuals — but in groups, parties, peoples, and ages, it is the rule.” Most people don’t have world-models - they believe what their friends believe, or what has good epistemic vibes. In a large group, weird ideas can ricochet from person to person and get established even in healthy brains. In an Afro-Caribbean culture where all your friends get attacked by demons at voodoo church every Sunday, a belief in demon attacks can co-exist with otherwise being a totally functional individual. So is QAnon a religion? Awkward question, but it’s non-psychotic by definition. Still, it’s interesting, isn’t it? If social media makes a thousand people believe the same crazy thing, it’s not psychotic. If LLMs make a thousand people each believe a different crazy thing, that is psychotic. Is this a meaningful difference, or an accounting convention? Also, what if a thousand people believe something, but it’s you and your 999 ChatGPT instances? III. A Hidden Army Of Crackpots I have a family member who believes that the theory of evolution, as usually understood, cannot possibly work. He has developed an alternative theory called “noctogenesis” which patches Darwinism using ideas from the transactional interpretation of quantum mechanics, and he works on-and-off on various related books and papers. I have told him I suspect he might be a crackpot; he stands by his claims. It’s fine; when I got into the technological singularity and AI safety, lots of people suspected I was a crackpot, and I stood by my claims too. You’ve got to stand by your family members even when they’re slightly crackpottish. This family member is happily married, retired after running a successful business, and generally a normal likeable person. He has no signs of mental illness, and doesn’t talk about quantum evolution unless someone else brings it up first. There must be millions of people like him. Used car dealers with proofs of P = NP, dentists who think they’ve discovered something important about Mary Magdalene, math professors obsessed with destroying the moon. I’m working on evaluating ACX Grants, and these people are out in force. A few propose literal perpetual motion machines. Others have vaguer plans, like some kind of social media app (it’s always a social media app) that will cause world peace. Many of them have decent jobs and seem like upstanding members of society. Their secrets are known only to themselves, their family members, and their would-be grantmaker. …and, increasingly, their chatbots. After years of hiatus (or at least not talking to me about his work) my family member is back on the quantum evolution beat, and LLMs appear to be involved. If I knew him less well, I would think the LLM had caused the quantum evolution theory - but no, it just made it much easier to research and write about. Is this psychosis? The answer has to be no, but it’s once again hard to draw the line. A very small number of crackpots will be vindicated by history. A larger number will be erroneous but sympathetic - the official account of the Kennedy assassination is pretty weird, and reasonable minds can disagree. From there, we get to ones that are maybe not so sympathetic: flat earth, QAnon, the thing where the Queen was an alien lizard. If only one person thought the Queen was an alien lizard, and they never managed to convince anyone else, would that be sufficient evidence for a delusional disorder? I’m not sure. (psychiatry has a diagnosis, schizotypal personality, which sort of involves being a normal person with a few odd ideas, but it’s not a great match for many of these people, and interesting mainly as a genetic curiosity - it travels in the same families as schizophrenia itself) Maybe this is another place where we are forced to admit a spectrum model of psychiatric disorders - there is an unbroken continuum from mildly sad to suicidally depressed, from social drinking to raging alcoholism, and from eccentric to floridly psychotic. People who are eccentric can remain so their whole lives, with the level of expression depending on their social connections and the ease of pursuing their rabbit holes. LLMs, by making it easier to pursue odd theories and serving as a surrogate social connection who always agrees with you, can bring latent crackpottery into the open. IV. Cause And Effect Bipolar disorder has an interesting relationship with sleep. Most manic people sleep very little, or not at all - maybe an hour or two a night. But also, poor sleep can cause bipolar episodes in people prone to them. In a typical case, a bipolar who’s been well-controlled for years will get assigned a big report at work and get poor sleep for a few nights until they finish. At first, this will be just as bad as it sounds, and they’ll be working through a fog of tiredness. Then the tiredness will lift. They’ll feel normal, then better-than-normal, until finally they can’t sleep even if they want to. Then they’ll email the report to their boss and it will be written entirely in Assyrian cuneiform. I increasingly think this isn’t just an incidental feature of bipolar, but part of the reason it exists as a diagnostic category at all. Most people have a compensatory reaction to insomnia - missing one night of sleep makes you more tired the next. A small number of people have the reverse, a spiralling reaction where missing one night of sleep makes you less tired the next. Solve for the equilibrium and you reach a stable attractor point where you never sleep at all. But this does other bad things to your brain - hence the cuneiform. I’m not claiming that bipolar is “just” sleep loss. As Borsboom et al will tell you, psychiatric disorders can be viewed as complex networks of symptoms, each reinforcing the others. In a few pure cases, you can get a ratchet going with sleep alone, and the sleeplessness will spark everything else. More likely, there will be lots of interactions between poor sleep and everything else, and the “everything else” can sink or hypercharge an impending manic episode. Still, I find this a fruitful way to think about bipolar. Sleeplessness is both the cause and the effect. Can delusions also be like this? That is, suppose there’s some personality trait where having one delusion makes you even more delusional. Maybe the delusion makes you excited (who wouldn’t be excited to learn they’re the Messiah?), and you’re more delusional when you’re in an excited state and not thinking clearly. Or maybe it’s a three-symptom cycle - the delusion causes excitement, which makes you unable to sleep, which scrambles your thinking, which makes you more delusional (which makes you even less able to sleep, etc). The point is: delusions are certainly an effect of bipolar disorder. And in the dynamical system model of psychiatric disorders, we should expect that effects are often also causes; that’s how the vicious cycle gets going. This is the best I can do at modeling true LLM psychosis. Someone with a trait where delusions lead inevitably to more delusions starts using an LLM. The LLM accentuates whatever usual tendency towards crackpottery they have and makes them believe something a little crazier than whatever they believed before. Then that crazy belief feeds upon itself and causes other things like excitement and sleep loss, which (if the person is predisposed) precipitates a true psychotic episode. V. Folie A Deux Ex Machina If one person believes a crazy thing, it’s a delusion; if a thousand people believe it, it’s a religion. What if exactly two people believe it? In psychiatry, this is called folie a deux. It fits awkwardly into our nosology and is rarely seen. Still, it happens enough to generate a few case studies. In a typical case, one person has psychosis for some normal reason, like schizophrenia or bipolar, and the second person is a shut-in who lives with them and rarely talks to anyone else. The psychotic person gets some normal psychotic delusion - they’re God, the Feds are after them, etc - and sort of psychically steamrolls over the second person until they believe it too. Usually removing the second person from the first is sufficient for a cure. This slightly challenges the view of psychosis as a biological disorder - but only slightly. Again, think of most people as lacking world-models, but being moored to reality by some vague sense of social consensus. If your social life is limited to one person, and that person themselves becomes unmoored, then sometimes you will follow along. I would expect second-sufferers to believe delusions in a sort of cognitively normal way, the same way people believe true facts, honest mistakes, and conspiracy theories. I would expect them to be less likely (though not zero likely) to have other psychotic features like sleep disturbances, hallucinations, disorganized speech, or a tendency to autonomously generate delusional ideas aside from the one they absorbed from the index case. An introverted person using an LLM has some similarities to folie a deux. If they use the chatbot very often, it might be a large majority of their social interactions. Here the primary vs. secondary distinction breaks down - the most likely scenario is that the human first suggested the crazy idea, the machine reflected it back slightly stronger, and it kept ricocheting back and forth, gaining confidence with each iteration, until both were totally convinced. Compare this to normal social interactions, where if someone expresses a crazy idea that isn’t common in their culture, other people will shoot them down or at the very least nod politely and stop the conversation. So my working theory of LLM psychosis is: Some patients were already psychotic, and LLMs just help them be psychotic more effectively.
Inline links: the transactional interpretation of quantum mechanics, math professors obsessed with destroying the moon, ACX Grants, a spectrum model of psychiatric disorders, As Borsboom et al will tell you, the dynamical system model of psychiatric disorders, folie a deux
Definitely psychotic even before the AI (n=19), if the respondent said the friend had a pre-existing diagnosis of schizophrenia, bipolar, or other psychotic mental illness.
How Natural Tradeoff And Failure Components?
March 26, 2026 · Original source
Michael Halassa: Did John Nash Really Have Schizophrenia? is a good article on the genetics of psychosis. Previous research found that schizophrenia genes decreased IQ but increased educational attainment. Usually IQ and education are correlated, so this was surprising. The new research finds two components to schizophrenia genetic risk. The first component, shared with bipolar, increases educational attainment. The second component, not shared with bipolar, decreases IQ. They average out to the observed full-spectrum genetic signal of constant-to-increased educational attainment paired with constant-to-decreased IQ.
Inline links: Did John Nash Really Have Schizophrenia?
In 2021, I discussed tradeoff vs. failure models of psychiatric conditions, and said that most conditions were probably a mix of both. The new research seems to confirm this: the first genetic component of schizophrenia is a tradeoff: bad insofar as it gives you higher schizophrenia risk, good insofar as it gives you higher educational attainment. Most likely this has something to do with creativity or motivation. The second component is a failure: bad in every way, with no compensating advantage. Most likely this is detrimental mutations in genes for neurogenesis and synaptic pruning.
Inline links: tradeoff vs. failure models of psychiatric conditions
I mostly wasn’t thinking about schizophrenia when I wrote about tradeoffs vs. failures, so I was surprised to see the theory so nicely reflected there. But in retrospect, this is common sense. All multifactorial problems should naturally be combinations of tradeoffs and failures.