The intuitive notion of popularity is that it is a measure of how well one is “liked” by others. A popularity contest therefore discovers the “likedness” of its participants. This seems simple enough. Popularity contest being practiced over meritocracy would just mean people hire who they holistically like the most, instead of who is best at the job.

This pretty cleanly leads to the notion of people having “liking functions.” How much they like someone is some function of that person’s properties, plus a random component. If liking functions equaled merit, then popularity contests would be equivalent to meritocracy.

So we have to ask, what are the actual components of the liking functions? For one, we can see that in wealthy tech and finance spheres, race is a predictor of liking score. Popularity contest is therefore a generalization of discourse on racial discrimination. For example, it would seem that Asians and Indians are disproportionately well liked over white Americans. I would argue that Slavs are also disproportionately favored over white Americans and Western European men, but to a lesser extent than Indians and Chinese are.

This could be causal, or it could be incidental. If it’s causal, it’s direct racial discrimination. If it’s incidental, it’s because these races tend to have better liking scores independent of merit. For example, if the liking score values ability to signal over true merit, and Chinese people are extremely focused on signaling, then they will have liking scores that are higher than their merit, and it will look like racial discrimination against white Americans, who have as much or more merit as Chinese immigrants, but who signal less.

It’s probably a mix of both. Wealthy liberals openly discriminate against white people to some extent, but often in favor of blacks and not Chinese or Slavs. However, Chinese and Slavs are good at maximizing popularity signals, but not at having strictly superior merit to white Americans. This is where the term '“piano-American” comes from. It refers to how they train from childhood to look more “popular” or “cool” by Harvard standards, so they can get in and study engineering. Obviously, the piano has nothing to do with their merit except at a music academy, but it works, so they maximize it, since they lack the hesitance to blow their time playing status games that Western whites have.

Why popularity contests and not meritocracy? Probably the collapse of selection pressures, at all levels of selection. Not just genetic, but cultural and institutional. When competition is weak, people can select for bull-shit “nice to haves” over raw merit. Over long enough time periods under low selection, dumb hiring managers can rise who are popular but who couldn’t even properly measure merit if their life were on the line. This is when they come to believe their their popularity function is merit.

This is likely why popularity contests seem to emerge in large empires at their peak. This is when outside competition is at its minimum. It’s also why schools are known for their terrible popularity contests, both within and between. Within schools, there is no selection whatsoever. It’s pure communism. The dumbest high school students get to remain as students with the worst grades as long as everybody else. Between schools, there is very little competition. The government completely funds most schools, and even largely funds Harvard et al. At a minimum, elite universities have the most selection pressure, but it is relatively small and the consequences of their actions lag.

In other words, I hypothesize the popularity contests emerge when civilizations are in decline. This goes well with Turchin’s “meta-ethnic frontier” theory, which states that inter-population competition is what produces great civilizations, which then rot from the lack of competitive pressure. Popularity contests everywhere are what that looks like from the inside.

What stood out about both articles was how unmeritocratic top AI labs seemed. First, Claude Code. Have you noticed it’s slow? Maybe it uses a lot of memory or battery? That might be because the 7 figure salary geniuses at Anthropic decided to build a rendering image to display text

Brilliant! But I know a software engineer or three that make 1/10 or less of the total compensation of an Anthropic dev who could do better. So how did these guys at Anthropic get hired?

It got me thinking: we live in more of a popularity contest, layered on top of a basic socialism, than in a meritocracy. The popularity contest produces a hierarchy, and is often deceitfully referred to as a meritocracy, but it does not actually sort on merit. As a consequence, you get buggy vibecoded Claude Code, made by engineers making over $1 million per year. Better engineers clearly exist, and can prove they exist, but hirers ignore the signs and keep hiring bad engineers.

What do they hire on? Popularity signals. In tech these are going to a handful of universities, working at a startup funded by a handful of VC firms, working at a handful of big tech companies. Going to these places only correlates with merit or IQ weakly to moderately. I’m guessing overall merit and IQ correlate at around 0.85 if you control for years of experience, while popularity signals correlate with either merit or IQ around 0.50.

As a consequence, we expect people with elite, 3SD popularity signals and positions to have merit or IQs around 1.2 SDs, with a lot of spread. That comes out to about 123 IQ, which is what a lot of people at Google, Stanford, ${Unicorn}, or Anthropic seem to be.

The second article was about AI firms hiring philosophers to write philosophy essays about AI, allegedly paying them up to $10 million total.

Needless to say, they are overpaying. I could find at least a few undergrads at the local university that could write equivalent essays, but they could never “win” this position. Why? Well it’s not because they lack merit.

This article isn’t about AI labs by the way. The same reasoning applies to finance firms and their demand for “pedigree”, top universities and their insane entrance requirements, other tech firms, law firms, consulting positions, corporate executive positions, all of that.

The GLOMP

GLOMP stands for “Girl Law of Moral Panics.” What I noticed is all moral panics in the modern world are centered around girls being allegedly harmed. They’re fueled by girl-tears, or rather people that show others girl tears.

And I say “girls” instead of women because in every example I can think of, the emphasis is on girls and not women. Women might sometimes be the moral entrepreneurs fueling some of the panic, but they talk about girl tears. Girl is defined expansively to mean females younger than about 20 years of age.

The MLOMP

The MLOMP is the Misinformation Law of Moral Panics. Moral panics always thrive on misinformation. Usually there’s a kernel of truth in it, but a large part of the fire is misinformation. The truth-kernel about harms to girls is the tender and the misinformation is the kindling.

The misinformation usually sounds true enough and is hard to debunk. It sounds important because it has to do with girl-tears and people who criticize it can be shut down as assholes who don’t care about girls enough. Basically, if something has a 40% chance of being true for and average person, it will be taken as true because the downsides of a false negative to girls are seen as too great by the activated mob, which can’t parse information at high levels.

The MLOMP causes moral panics to outrun fact audits. It discourages careful people from doing the fact auditing, because they will face harsh criticism for doing so, by many people who care a lot more than they know.

The HLOMP

The HLOMP is the Hysteria Law of Moral Panics. It says that moral panics lead to extremely aggressive, yet often ineffective, actions based on the panic. If the GLOMP is the tender, and the MLOMP is the kindling, then the HLOMP is the logging of the moral panic fire.

Hysteria is defined as “a state of violent mental agitation.” The HLOMP is when panicers start behaving as if they are in a psychiatric episode of hysteria. It leads to violence which, in today’s world, is usually done through regulated channels, since the panic affects enough people to gain a considerable democratic legitimacy.

In modern times the HLOMP is usually expressed as some kind of overly-harsh, unjust law enforcement action, either criminal or civil. It may also be legislative if this fails, meaning that new laws are passed based on the moral panic. It may result in all three.

You can bet civil liberties are abridged when HLOMP time comes and that any decent people in the future will look at the actions like contemporary people look at the Salem Witch Trials.

Examples

The quintessential example of the moral panic is the Salem Witch Trials. While I consider it to have happened somewhat outside of the temporal scope of the “laws” I hypothesize here, it actually fits quite well. Wikipedia says:

The trials began after a few local women in Salem Village were accused of witchcraft by four young girls, Betty Parris (9), Abigail Williams (11), Ann Putnam Jr. (12), and Elizabeth Hubbard (17)

There’s the GLOMP! The MLOMP is believing witchcraft is real and the HLOMP is when they started hanging “witches.” There you go.

The example that inspired this is from Gwern:

NYT and Vaping: How to Lie By Saying Only True Things

This essay looks at a 2022 New York Times article about teenage vaping which juxtaposes synthetic nicotine, flavored vapes, Lizzie Burgess’s hospitalization, and nicotine addiction without ever quite saying that legal flavored nicotine vapes caused her lung injury.

There’s the GLOMP.

Sentence by sentence, the article’s misleading effect comes from technically true phrasing: “vaping”, “vaping THC and nicotine”, “vaping-related lung injury”, and post-hoc narrative order.

The case is a useful example because the ground truth is clear, the article was high-profile and edited, and the manipulation is grammatical and selection rather than simple falsehoods.

There’s the MLOMP. The tender is that vitamin E THC vapes cause lung damage when excessively used. The kindling is the NYT running an article blaming normal nicotine vapes. You’re an asshole who wants to destroy the lungs of girls if you criticize this.

This article appears to have been successful in persuading readers. In the NYT comment section (mirrored below in full), most comments agreed with the article, often going as far as to claim that “addict and kill people—including children”/“here they are exploiting a legal loophole to addict millions more young adults to their lethal products” or “Young people especially become easily addicted to the nicotine…The damage to their lungs resembles the worst pneumonia you can get and still be alive.” or “The purveyors of these drug delivery systems are killing our children with intent. They should all be thrown in jail, and the import of these products should be banned.” (The import of THC vapes contaminated with vitamin E acetate is, of course, already banned or adult ID required etc.) Indeed, only a single NYT commenter (who no one responded to) appeared to get it right in their reply a quarter down the comment section:

Here’s the HLOMP. Criminal procedures usually only happen in one subtype of modern moral panic, and this is not that type, but there are active legislative and regulatory law enforcement actions taking place:

“It’s a whack-a-mole situation”, said Shusterman, whose state banned all flavored e-cigarettes, except in licensed smoking bars. “They’re not following any of the laws.”

Mitch Zeller, director of the FDA’s Center for Tobacco Products, said the agency recognized the problem.

“Disposable e-cigarettes made only with synthetic nicotine pose a particular challenge for the FDA when it comes to our tobacco authorities”, Zeller said in an email. “The FDA is actively investigating this issue and considering how to best address such products.”

[photo of young woman in ICU with face covered by a medical ventilation mask]

All a waste of money of course. Vaping rates aren’t down enough to improve girl health. Taxpayers lose money and freedom due to expensive enforcement from the HLOMP.

Epistemic status

The epistemic status is: statistical study needed, but seems like a good heuristic. I don’t know of any counter examples at the moment, but every moral panic I know seems to fit. There are 2 high profile ones active right now, that I can think of. The ones from the 20th century seem to fit.

If we don’t control for IQ, an English doctor is clearly a larger value add over an average white American. Even if we do control for IQ, childhood is expensive. English doctors don’t consume American school and daycare for themselves, while the babies of white American doctors do. That, of course, does not mean that white physician babies aren’t a value add at the end of their lives. They almost certainly are. It just means that their childhood is expensive and subtracts from the net value add.

Most people being born are not going to be physicians or equivalent types of people. Yet their childhoods are similarly expensive. If, as adults, someone barely produce more than they take, then they can end up a net negative once one counts their childhood.

Wait … are you a Bitcoin millionaire and looking for something to buy without fiat? Are you fed up with Ethereum and looking to dump it? Subscribe to me with BTC or ETH at a reduced price at sociobiology.org!

Read more

The fundamental downside of this approach is that you can often infer things which are not logically implied, and that therefore turn out to be untrue. That’s because verbal thinking is heuristic. Sometimes the heuristic has a positive truth EV, and sometimes it doesn’t. In either case a heuristic is never perfect, so it guarantees nothing. Quantitative reasoning, in contrast, is completely logical. What is inferred from it is known with certainty. This is a necessity for wading into counter-intuitive areas.

The upside to verbal thinking is that it is easy and efficient. The downside to quantitative thinking is that it is difficult, and when applied to the wrong use cases, it provides nothing new over the verbal approach. But I think there are some topics where the quantitative approach is necessary and very beneficial.

Those topics include memetics, social change and behavioral evolution, and social power. To study these topics, I developed some quantitative methods, and I would like to detail the ideas behind explicitly in this article, so that hopefully they can be useful to somebody else.

Individuals as samples from population behavioral distribution

The approach I used to study social change and behavioral evolution started with probability and statistics. It’s too hard to predict individual behavior, but what about mass behavior? It turns out this should be theoretically easy. If individual deviations from the population mean cancel out, then you don’t have to worry about them; you can simply study how the population mean evolves over time. It turns out this is quite quantitatively tractable with evolutionary probability & statistics.

Decomposition of phenotypic pressure/velocity

Now that were are studying population means, we want to explain how they change. We do this with decomposition. It’s analogous to decomposing the forces acting on an accelerating particle in physics.

From phenotypic longitudinal surveys you can measure that the phenotypic mean for leftism changes 0.20 SDs over some decades, for example. Now you break this down into types of forces acting on it. My idea was to do genetics, memetics, and other environment, as something of an analogy to ACE decomposition in twin studies.

So the sum of the forces dG + dM + dE should add up to dP = 0.20 in this example. We immediately see that we only have 2 degrees of freedom; if we can measure two of these terms were get the third. Additionally we can use a simpler equation with two terms, genes and the rest of environment. The extended environment is the sum of “cultural” or “memetic” change and all other types of possible environmental change.

To measure these, we would like them to be as well defined as possible. It turns out dG is already well defined and it’s easy to extend the infinitesimal theory of dG to dM as well. This leads the the next main ideas.

Stripping down phenotypic evolution forces to atomic components

dG at its core is driven by changes in allele frequency, which is mathematically described by the theory of population genetics. dG is a long sum of constituent atomic components that have weights. The theory of regression solves the weights. The weights are the average effect of a change in gene frequency on the phenotype under study.

I realized this model is quite elegant and could also apply to the memetics compoment dM. I read a couple of books on modeling cultural evolution, and they did not really get this; they were filled with intractable differential equations modeling the spread of individual memes. The central idea of these books was that you should model macro memes, like the meme Christianity, and assume that when someone adopts “Christianity”, their entire behavior changes in a saltatory fashion. Thus, you need to track the spread of the macro meme with differential equations, like in viral epidemiology. It is unclear how this should interact with the infinitesimal model of genetics when applying it to the study of polygenic traits.

These models were produced largely in the 1980s and never went anywhere. It seemed that they existed to be pointed to by anti-hereditarians as a kind of road block, “look, here we have cultural evolution, have you dealt with this in your work? go read it, it’s important.” But it was like neoclassical economics in that the models were completely without estimable parameters. They could not be used in real life.

So I decided a better model would be an infinitesimal model of memetics, with macro memes coming as a special case of this, like how Mendelian traits are actually a rare special case of the polygenic model. People don’t adopt macro memes whole sale, they are constantly being blown around by small memes that have microscopic effects on behaviors, and that might add up over time to produce larger effects. Religion is a special case of this model where closed off communities end up with different memetic averages than others due to heterogeneity in the spatial distribution of memes. When that heterogeneity breaks down, you have to use the more general infinitesimal model to get anywhere, because everyone is exposed to dozens of different religions that all vary in a clinal fashion from one another on top of being flopped around by a bunch of other memetic forces that all cancel out to some average.

In theory can you give people large surveys of ideas, asking if they have ever been exposed to this or that bit of information, then you can try to correlate that with their phenotypes. This will produce estimates of the effects of those ideas on them. I was also able to work out some parameters for interactions with genetics. For example, people with different genotypes may tend to produce different types of memes, meaning genetic evolution causes memetic evolution, which can then amplify the genetic evolution. Memetic clustering can be caused by people with different genotypes accepting different types of memes more. Thus the difference between religious groups probably has at its base a genetic origin, although memetic spatial heterogeneity can serve to amplify behavioral differences between such groups.

Applying these ideas to social power

To analyze social power in society, I thought again to use decomposition. That’s the secret formula. What I ended up with was breaking down the structure of society into the elite class and the warrior class. It would seem all social power comes from people with elite positions in the economy, government, or ideological arena, plus people who enforce dictates from such people which is the warrior class.

I then thought about the moral or political substance of laws and determined that, in a perfect democracy, by definition they should follow the median of the population, which under a bell curve is also the mean. And due the the infinitesimal structure of the underlying forces creating that mean, it is in fact a bell curve. So social power really only matters with respect to moral policy if the policy deviates from that mean. Why would it do that? Well, because the elite or the warrior class deviate. So this lends itself to a parameter: elite deviance and warrior deviance. It would seem that warriors today are pretty average, and elites lean a little bit left, but not that far left. It depends on the specific institution. The government does not really lean left very much.

What produces this deviance? Probably population stratification. Likely genetic in origin, resulting from different evolutionary forces acting at different levels of society.

And what is an elite, anyway? In theory we can break it down quantitatively using the techniques I’ve already outlined. If you could measure policy changes in society against people becoming a certain age or dying, you could potentially construct a measure of how influential any given person is. Then you can correlate these scores with their traits. I hypothesize that IQ correlates positively with such scores, as well as some other traits. I expect that it mainly comes down to SES which captures both monetary wealth and some ideas of softer power, though I expect money to matter somewhat more than the latter.

Conclusion

Here I’ve outlined the main techniques I’ve used to quantitatively model society. I view society statistically and I think there’s great potential in users statistical parameters as the measurable observables in models. Mainly these are regression weights, and the regressions come from performing quantitative decompositions and reasoning about the basic parts of whatever aspect of society is being measured. Usually things can be linked all the way back to alleles. Otherwise, one can reason about infinitesimal memes or enviremes.

While I’ve measured some aspects of what I’ve outlined here, such as evolutionary pressures, I haven’t been able to measure memetics or social power parameters due to lack of funding. Maybe that will change in the future.

I’ve slowly come to the conclusion that, like animals, the human economic market is a product of evolution. This is a significantly different viewpoint than that of mainstream economic theory. Mainstream economic theory holds that the economy is the result of a distributed reasoning process. There are many agents in competition, and they all behave in a way which maximizes their utility functions. Essentially, humans are modeled as agents that must use intelligence to find solutions to problems that optimize human scoring functions.

The evolutionary view holds that markets are mainly optimized not by human scoring functions, but by external scoring functions. Humans do blind trial-and-error to find optimal points. The extent to which human preferences shape the market is mainly the extent to which the market is shaped by human biological evolution, because human preferences are primarily evolved and hereditary.

That means that biological evolutionary pressures shape human market behavior. But exogenous pressures on physical processes also shape the market. You can see human intelligence as a way to quickly find optimal solutions when the function changes, but not as utility maximizers.

This sounds extremely abstract, so let me use an example: Tesla. We can think about Tesla like mainstream economists, or we can think about it as evolutionary economists. The mainstream view on Tesla is something like, they entered the market with a product P and a cost C, and they satisfy the market’s collective utility U. Where the product comes from and what the cost is is not really analyzed, and neither is utility. They get an equation with some big unmeasurable letters and a plus or a minus. A little to vague and simple, isn’t it?

The evolutionary view applies the sociobiological analysis of population traits to the market. It says there is some force or pressure on distribution of products and costs of those products. There is also going to be pressure on utility functions, for example, people can evolve, biologically or memetically, to prefer electric cars over gas more, all things held equal, and this will have an impact on the market. The rise of a company is therefore explained by this combination of pressures, like how the motion of a particle in physics is explained by the sum of forces on the particle.

It follows that the rise of Tesla might be explained not by changes in the distribution of utility functions, but by changes in the production and cost spaces. You can think of these functions like probability distribution, or measurable spaces. Humans exert an opening pressure on the space. The production space starts out as 0 nearly everywhere, 10,000 years ago. In other words, there is no technology. More and more parts of it are increased above 0 by humans. Perhaps you could model the value of some part of the space as 1/cost of the production. The cost is infinity when production is impossible. As it becomes possible it may eventually become cheaper. As it becomes cheaper, it becomes more likely to be found in the market. The number can be interpreted as a likelihood of it existing.

Everyone intuits that technology development influences the market. But this is a neat way to formalize that intuition quantitatively, in a way that naturally goes with the rest of evolutionary theory. And this theory lends itself to measurement a lot better than neoclassical economics. In theory, you can measure the forces on the measure spaces and predict how they will change. Their change is the change in the market. Mainstream economic theory is more like a story that uses math than a measurement and prediction engine.

Now let’s apply this idea to trading. Trading models largely abandoned neoclassical theory and use something much closer to what I’m imagining. The value of a traded commodity is modeled as a stochastic process which is the measure of an underlying probability space. Quite simply, we want to think about the variance decomposition of the motion of the stochastic process. Some of it will be utility evolution, some of it will be production evolution, some of it will be exogenous shocks on the “measurable spaces” of those. For example, the Straight of Hormuz. A lot of recent motion in the oil market is explained by this. We could model the war as a pressure that compressed parts of the oil-commodity production space. This is in theory a successor to and generalization of the old notions of supply and demand.

The evolutionary theory also naturally produces neoclassical ideas like elasticity. When the oil area is compressed, it loses “likelihood.” That means other commodities gain relative likelihood. So they will become more “common” in the market as a result. Also, the sum of total costs^-1 decreases, so the sum of total costs increases. Thus, things are generally more expensive as a result. Mainstream economists call that cross-price elasticity.

So, the evolutionary view seems promising when it comes to generalizing disjoint parts of neoclassical economics and explaining them with math that is common to sociobiology. It also mechanistically links market changes to sociobiology through the evolution of U and explains other pressures on the market sometimes in the form of non-economic human behavior, like war. This integrates economics with sociobiology, which is a huge upside. Conceptually, economics should be a subcomponent of sociobiology, since it’s ultimately about the behavior of biologically evolved humans.

It would be interesting to try to formalize this idea, and maybe submit it somewhere like arXiv. There is already a minority field that calls itself evolutionary economics, but it doesn’t like like it’s exactly what I have in mind. Some recentish literature, for example, seems to support the idea that markets have evolving moments, by making “market efficiency” such a moment and showing that it changes, whereas mainstream economics just says it’s static apparently. By efficiency they seem to mean the predictability of future return: perfect efficiency is supposed to mean that the next price cannot be predicted by past prices. Apparently bitcoin has long term swings that come and go, sometimes it moves randomly and sometimes there’s a predictable longterm trend. Well, why is that? It sounds like there’s coming and going evolutionary pressures on bitcoin. Sometimes they persist for a while and then fade. So I suppose this is empirical evidence for my model, although it’s not clear where the focus on efficiency should come from. I suppose my theory might not formally predict market efficiency, rather it would say it’s not that important and it goes up and down, while mainstream economists have a foible of obsession on equilibrium and efficiency. But the underlying theory I’d like to see developed is more or less ignored on the paper, instead the probably irrelevant idea of efficiency is fixated on, with reference to the “Efficient Market Hypothesis” as the alternative hypothesis, become “debunked,” but in a sort of atheoretical, negative way. I much prefer to produce a substantial counter theory which implies this result, and then use the result as support for the theory, instead of acting as a “debunker.”

But it seems like there are some good papers in evolutionary economics nonetheless. The field seems split between people more aligned with a classical evolutionary approach, which is what I’m advocating for, and people who would like to strip it down and seemingly subordinate it to neoclassical economics:

The term “evolutionary economics” might have been first coined by Thorstein Veblen.^[1] Veblen saw the need for taking into account cultural variation in his economic approach; no universal “human nature” could possibly be invoked to explain the variety of norms and behaviours that the new science of anthropology showed to be the rule rather than an exception.^[16] He also argued that social institutions are subject to selection process^[17] and that economic science should embrace the Darwinian theory.^{[18][19][20][1]}

Veblen’s followers quickly abandoned his evolutionary legacy.^[16][21] When they finally returned to the use of the term “evolutionary”, they referred to development and change in general, without its Darwinian meaning.^[1] Further researchers, such as Joseph Schumpeter, studied entrepreneurship and innovation using this term, but not in the Darwinian sense.^[2][22] Another prominent economist, Friedrich von Hayek, also employed the elements of the evolutionary approach, especially criticizing “the fatal conceit“ of socialists who believed they could and should design a new society while disregarding human nature.^[23] However, Hayek seemed to see the Darwin theory not as a revolution itself, but rather as an intermediary step in the line of evolutionary thinking.^[1] There were other notable contributors to the evolutionary approach in economics, such as Armen Alchian, who argued that, faced with uncertainty and incomplete information, firms adapt to the environment instead of pursuing profit maximization

A book from the 80s is credited with being seminal to the field of evolutionary economics. It promotes the use of Markov chains for modeling:

I would say this is a step in the right direction, but it’s not enough. The Markov property is very general and not sufficient for modeling evolution. It says that the next state of a stochastic process is sufficiently predicted by its current state, and no states before that. That’s not enough to break down evolutionary pressures. Evolutionary systems trivially depend on the current state and the pressures on that state, and not past states. I’m also not thrilled with the use of terms like supply, demand, capital stocks, and so on. I would like to replace these terms ,which came out of neoclassical theory, with superseding evolutionary terms, like I outlined above. I’ve always had the intuition that those terms hide a lot of important phenomena and complexity and they don’t really give me what I’m looking for when it comes to economic theory, like heritability does in behavior genetics.

But for now, you have to pay in BTC or ETH. I’ll add other cryptos if there’s enough request.

It also has dark mode:

There’s a comment section too:

I made this website because I don’t like the direction Substack is going. It started out by promising to revive the intellectual blogging scene, then became rife with terrible engagement farming slop newsletters, and is now trying to expand to short form. Its trajectory is negative and generally the ability for a good new writer to grow on Substack has been nil since 2024 at least.

So please check out sociobiology.org as well as its brother site hbdforum.net.

I liked mid-form content internet forums. I think they’re a great space to think and interact with intelligent people. Unlike a blog, a forum is a community, and should have a low effort-barrier to posting, although good forums have a high talent-barrier. That means bright people who don’t want to “build an audience” can post interesting ideas and get them seen by other smart people who are interested in discussion. On Substack, where content is likewise medium form, you mainly pander to an audience, which takes in and does not give back. Without a lot of pandering, nobody sees your writings.

There are always young people out there who are very intelligent, and have interesting things to write, but who don’t want to pander to audiences with slop. If such a young person is interested in HBD, then HBDForum is a great place for them.

I’m going to try to replace my Discord with HBDForum. I have a number of reasons why. The first is that Discord supports medium form content, but the organization promotes shortform, and I don’t like how servers are walled gardens. At the bottom of HBDForum is a live chat and a voice chat feature.

These effectively replace those draws to Discord which classic forums don’t have. Another reason I wanted to do this is that Discord is Orwellian spyware. Discord’s spying practices should be illegal. Sadly, they are popular among boomers and government officials. On Discord, even DMs aren’t private. Discord is always reading your DMs with hired moderators and AI. If a DM is construed to possibly possess illegal content, Discord actively violates your privacy by using its extensive IP logs to forward all of your information to your respective government.

HBD Forum is not intended to be a place for content which is illegal in the United States. However, Europe has stricter “hate speech” laws which could potentially criminalizes some HBD discussions, like race differences in intelligence. HBDForum is hosted in the United States and logs nothing it doesn’t need. I don’t even have it set up to log IP, so proving that someone is posting from the EU is not possible. There are no data retention laws for web forums in the United States, so this is perfectly legal. Logging IP is taken for granted but I’ve never built a website that does it. It’s not default behavior. Any website that does it is essentially malicious spyware.

I hope that younger people realize the goodness of strict, HIPAA like privacy laws that minimize what data internet enterprises are allowed to retain, what type of content they are allowed to let AI or humans read without permission (should only be public content), and what they are allowed to show to any 3rd party, including the government, without any legal order. Boomers conceptualize the internet as a public place that you visit momentarily, and see privacy as like wearing a ski mask to a shopping mall. Younger people use the internet as an extension of their home. A private Discord server is a chat with friends in your kitchen. It’s not a speech on a soapbox on the sidewalk of a busy downtown area. Nobody should ever have a politically incorrect opinion stated to their close friends in their home broadcast to the world.

I made this website in about 24 hours. It was very easy. I used GPT Codex 5.3. It’s not purely vibe-coded slop because I have enough web-dev experience to audit the code and fine-tune it. I’m like a senior engineer and GPT is currently like a really fast junior engineer. I made all of the architectural decisions and an AI skid would have probably made something 200% more sloppier due to taking whatever GPT first gives it But AI is now so good at web dev I’m convinced it’s not a long-term career option any more. There will be people employed in web infrastructure, but they will probably do 10 people’s jobs as of ~2025 for about $100,000 per year in 2026 dollars. I, personally, could probably replace 5 people right now with a good LLM and the firm I work for would have the same level of productivity as of a year or two ago. Productivity is instead going up, which is why that hasn’t happened, but that increase probably won’t absorb all of the LLM boost. That happened for no new technologies like this: cotton gin, calculator, internet, etc. Rather, smart people will have to learn something else besides JavaScript.

I see that as a hugely positive thing. I always thought the idea of a really bright person pooping out web code as ugly. I am glad that it is clear that this will not be the preponderance of my future career. I am glad I can spin up in 24 hours what took a team of millenials 1000 man hours back in 2017. I think smart people will be okay and the proletariat-ressentiment fantasy of smart people digging ditches alongside them or beneath them won’t come to fruition. Smart people are always okay because smart people learn quickly and are better workers. But the question right now is what will smart people do that LLMs can’t do?

One obvious answer is robotics. LLMs are restricted to the digital world. We need to build humanoid bodies for them so they can do embodied tasks we currently rely on functionally illiterate people, who are really egotistical and demand significant control of the political system, for. I think there will be a robotics businesses boom and a lot of web engineers will go there. A lot of the remainder will stay in the web world to “pilot” the LLMs and make the responsible decisions. They will be needed for at least 10 years in the web world because of all of the missing web software; there’s still a ton to be built and LLMs won’t be self-sufficient by definition until the whole economy is automated. There’s a lot of work to be done making sure literally everything informatic is done through a really premium website and producing robots that can automate embodied labor. For one, every government administration in the world needs to be significantly digitized. Every government still uses paper or 00s jank pages for their bureaucracies. My present firm will probably not lay off anyone in tech for at least 5 years because it’s the same there. There’s a ton that needs to be rebuilt from scratch by competent LLMs drivers and a lot that needs to be created for the first time.

I would also like to see more serious science, although I’m not sure if that can happen naturally. Freeing up high IQ people means there are more people available for research, but a lot of research is not profitable. We’ll need a massive expansion of academic research roles, but I’m sure the blessed People will resist this, because they’re narcissists who hate Faustian science and love boosting funding to socialism & People’s safety organizations like old age pensions, schools, secret police, airport security, and so on. Maybe AI will lead to a huge economic boom and there will be money laying around? If it increases the pie it should, once all of the Trump drama is finally over.

The research I mentioned is only important if it’s math-driven. Like physics, quantitative sociobiology (including statistical genetics), machine learning theory, and so on. So the three frontiers I’m seeing are robotics, technical computer infrastructure, and hopefully mathematical science.

Anyway, I would love to discuss this further on hbdforum.net! Please join us.

I thought I would give a summarized table of contents for my articles from the last 3 years. Then maybe I’ll make this a yearly post hereafter. Then at any point in time, the table of content of my writings will just be the sequence of these posts.

Where does leftism come from?

The first thematic sequence, dating back to 2023, addresses the origins of leftism. It includes debates and critiques of non-scientific theories of leftism, as well as my own model building. I ultimately decide that we must use the scientific method of mathematical sciences to satisfactorily answer the question. I create a mathematical model of evolutionary pressure on quantitative traits and treat leftism as a normally distributed psychometric. I provide quantitative evidence fitting the model that leftism’s modern increase was the result of collapsing selection pressures and the resulting accumulation of de novo mutations in the gene pool. This hypothesis was subsequently discussed and critiqued by the broader HBD sphere. I addressed their points and found most of them unsatisfactory; it ended with me concluding that I have about 95% confidence in the theory being true, and 4% confidence in it being closely related to the Flynn effect and increases in height and brain mass (basically, physical, potentially epigenetic and brain-related changes to improved nutrition and other features of modernity, inducing behavioral changes), and amending the theory to include evidence for “Social Epistasis Amplification,” meaning the social change is amplified by memetic effects (which are downstream of evolutionary and non-memetic pressures) by a factor of 25% to 75%.

Here are the articles:

Positive, model building

• Decline in Conservatism Between Generations Correlates Highly with Potential Signals of Mutational Load (2023)

Cross-sectional data indicates inconclusively that de-novo mutations are accumulating, and leftism appears more where more indicators of DNMs are.

• How to compute selection pressures (2023)

• Selection pressures on IQ, leftism, and other traits (2023)

• How to compute mutational pressure (2023)

Deriving some equations for the mathematical evolutionary pressure of quantitative traits model. I now try to avoid explicit equations on Substack. It seemed easy enough to me, but if you are making explicit equations, it should be published in a journal somewhere or folded into a technical white-paper or manuscript. That’s just how the bell curve is.

• Why can’t it be genes? (2023)

• Why can’t leftism be genes? (2023)

• Leftism has happened before (2023)

Accessible discussion of anti-hereditarian bias, fallacies which preclude evolutionary hypotheses to social change, and why you feel like the mutational load hypothesis is wrong, but it’s probably not and you’re just biased against hereditarianism and are engaging in one of the heuristic fallacies I debunk here.

• Evidence for a Paternal Age Effect on Leftism* (2023)

Asterisk indicates scientific article, not a blog article. This article uses a binary leftism metric and finds what I predicted.

• Extinction of the Dark Elves (2024)

Popular application of mutational load theory, addressing population decline since the 19th century and what that means for smart and based people (dark elves) who exist today in lesser numbers.

• Estimating Mutational Pressure on IQ through the Effect of De Novo Mutations in the Exome on Intellectual Disability (2024)

Analyzing direct molecular genetic evidence for mutational load theory.

• A New, Near-Gaussian Measurement of Leftism Shows the Paternal Age Effect Replicates* (2024)

Replicating the previous article with a better, normally distributed continuous measure of leftism.

• What is leftism? (2025)

Discussing why you should treat leftism like a highly heritable, normally distributed, psychometric trait, contra ideas that you need to break it down into “deeper personality factors” and so on (this is just circular political blank slatism).

• Chinese Rice Farms and “Culture” (2025), Evolution explains regional differences in China best (2025)

Fast evolution and Chinese rice farms.

• The Eugenic Origins of Modernity (2024)

Discussing how IQ eugenics created modern science and the industrial revolution.

Criticism of other writers

• Review of “The Origins of Woke” (2023)

Hanania’s theory is what I call the historical legalism theory of Woke. He seems to think laws have a significant random component, and this component shapes society for a long time. In other words, the US Civil Rights act just passed randomly, and laws are powerful, so this caused us to be woke in other contexts. The problem with this is that laws don’t pass randomly, they aren’t enforced randomly, and they don’t evade repeal randomly. All of these things depend on the underlying sentiments of the population. The law is made for man and not man for the law.

• Why Moldbug was wrong about leftism (2024)

Moldbug thinks leftism memetically evolved from Christianity. Basically, the genome and the underlying sentiments of people stayed the same, but modern governments set up an environment where woke people did better than non-woke people due to separation of church and state. Moldbug seems to think “separation of church and state” was an exogenous shock, similar to Hanania on the Civil Rights Act. I just cleaned up his incoherent argument a lot for him, but even this fails all the empirical tests, so it is wrong. You can see this in the person of Moldbug; the man clearly has instincts which weren’t common in the 18th century, like no aversion to homosexuality (quoted in the article). Aversion to homosexuality is instinctual and related to general disgust sensitivity, it’s not learned. Clearly, people changed on a fundamental level and this changed their social norms.

• Why Biology and Math are Key: Commenting on Keith Woods vs Academic Agent, 2024 (2024)

Addressing two YouTubers. One thinks Woke is some kind of conspiracy theory for the elite to profit off of the masses, the other thinks it’s all ideas.

• Birth Order: Mutational Load or Environment? (2024)

Critiquing Cremeiux’s birth order environmentalism. While he showed some good evidence for an environmental component in birth order effects, he downplayed the genetic component too much.

• Werner Zagrebbi and the Origins of Woke (2024)

Moldbug’s intern wrote something fashionable among the bay area crowd going on about “positive feedback loops.” It sounds smart, but he wouldn’t engage with my science. Apparently, I’m not widely read enough among his target audience, and meritocratic discourse is just some BS from 19th century Anglo science. Discourse today must be democratic, which means you must pander to the crowd to get a spot. I believe it was written at the time that my resistance to this is considered “autistic” by the cool people. That offended me at the time, but it’s kind of cool to watch the masses be consistently wrong about something. You just need to monetize your open secret and you’re golden! I am really only constitutionally interested in meritocratic spaces.

• Contra Seb Jensen on Mutational Load (2025)

Responding to a gish-gallop of criticisms that didn’t really hit the theory.

• Re Cofnas: Facts and Logic About Woke (2025)

Responding to Cofnas’s knowledge-model of wokeism. He thinks it’s just people who seriously believe HBD isn’t true.

• Why I am not a Conflict or Mistake Theorist (2025)

Why Scott Alexander’s conflict-mistake dichotomy falls apart under actual logical examination, and how to replace it with my scientific theory.

Who has power?

The next sequence develops my rejection of meaningful “elite theory.” I now believe we live in a peasant democracy and this causes the government to suffer from the moral and intellectual shortcomings of the median person, which are vast. I believe that “elite theory” is possible, and that there is some meaningfully defined conglomerate of a Fighter class and an Elite class which can rule over people, but that Western instances of these classes are genetically democratic, so they give out power to the peasants, thereby failing to rule properly.

• A Theory of Democracy (2024)

Political science usually treats democracy as stemming from a combination of ideas and supposedly exogenous institutions. Basically, books and constitutions. Both are kind of magic and without a book telling people why they should write a democratic constitution, they’ll just be monarchists, and without a well-designed constitution which magically induces everyone to follow it, democracies can fail for trivial, shallow reasons. This is obviously stupid and it would seem that the best theory of democracy is just genes: both high IQ and a certain temperament found among white people are the best predictors of democracy. Books and constitutions don’t matter. The UK has a monarch and no constitution and nobody reads old enlightenment books, but it has the right kind of white people, so it has a stable democracy, for example. Meanwhile, the best constitution hand-crafted by Americans plus the best Arabic translation of Locke airdropped over Iraq will never make it a good democracy.

• America’s elite is mostly a meritocracy (2024)

Shows America’s elite is not very nepotistic and is closely related to the demos.

• Was Moldbug Right about the Cathedral? (2024)

Debunking Moldbug’s theory of academic sovereignty.

• Elites don’t control public opinion (2024)

Quantifying elite theory and testing my mathematical model.

• Elite theory chapter of QSB manuscript* (2024)

Scientific explication of my theory.

What is culture?

Discussing the abuse of the word culture in popular and scientific writing, the problems with mainstream models of cultural evolution, and building my own model.

• The rise and fall of “cultural evolution” (2023)

The problems with cultural evolution as a concept and its roots in academic Marxist activism.

• Memetics theory chapter of QSB manuscript* (2024)

My new model of memetics which posits that ideas which influence behavior are many, and changes in the average behavioral effect of ideas are downstream of non-memetic forces. They therefore have an amplifying effect on non-memetic forces, which precede memetic forces.

• Reply to “Genes, culture, and scientific racism” (2024)

Replying to a new article by one of the very old originators, Feldman, of the cultural evolution idea, exposing how his sources are garbage and offer no empirical support for his premises. Nevertheless, you hear about cultural evolution a lot more than biological evolution of social traits. That’s memetic, but it’s downstream of people like the author existing in greater number than people like me, something which wasn’t always the case.

• Does “Cultural” Transmission Bias Twin Studies? (2024)

Why twin studies aren’t biased by “culture” and how they debunk culture theories.

Eugenics

Why we should have a classical eugenics program. I believe eugenics is the only solution to present social problems, since they were caused by dysgenics.

• Are We Smart Enough for Eugenics? (2025)

Discussing why classical eugenics did not take off (only eugenic people like it an dysgenics overwhelmed us first).

• What’s Wrong with Classical Eugenics? (2025)

Discussing ethical objections to classical eugenics (intelligence is good, being unintelligent sucks, so it’s a moral imperative to do a humane classical eugenics program).

• Can we afford classical eugenics? (2025)

Discussing the cost of a eugenics program.

Fertility and Marriage

• A look at Roman marriage ages (2025)

Original application of an ML method to derive the Roman age at first marriage distribution. Mean around 16 for girls, and 21 for males. The vast majority of girls married in their teen decade between 10 and 19 years of age. The majority of age gaps must have been between 2 and 8 years since most males married between 15 and 25. People frequently deny that antiquity was like this and habitually try to up the actually age at which people married to be closer to modern norms, because for some median people today consider teenage marriage to be extremely offensive, kind of like racism in academia.

• The possible origins of the Modern Western European Marriage Pattern (2025)

Discussing how the ancient pattern faded. It lasted to 1600 AD in Western Europe, contrary to claims that North-western Europeans always married late. Over the next 350 years, the rest of the world caught up, and then the West began a second transition to an even later marriage pattern — instead of marriage in the mid-twenties, it now occurs in the thirties. I allege that the first transition must have been caused by something very similar to the cause of leftism today. Perhaps mutational load, or perhaps some kind of epigenetic change.

• The Fertility Crisis and the Postmodern Marriage Pattern in Germany (2025)

Using Germany as a model white country with very little late 20th century migration, I examine their fertility statistics and show that the transition to 30s marriage is caused by a preceding change in desired number of offspring. In other words, something perturbed the minds of Germans first, and they marriage habits followed to conform to their new desires. This sounds like some kind of psychometric evolution, coinciding with modern leftism. Note also that the MWEMP from the 1600s coincided with early modern leftism: the rejection of medieval monarchy and religion.

Against academic philosophy

• Why “Philosophy” is Invalid (2023)

• The Sleeping Beauty Problem (2025)

What I do could be called philosophy. Any clean division of it from math and science is wrong and done because academic philosophers don’t want to actually put effort into studying anything. Math and science give us much more powerful ways of knowing things than the academic philosopher’s method, which produces nonsense. I see the root cause of their method being bad but still used as lack of aptitude to master math and science. Imagining an infinitely intelligent philosopher, he would be a master of math and science and it would heavily inform his thoughts on all subjects. You can see in the Sleeping Beauty Problem that most academic philosophers struggle with relatively basic probability theory, and this leads to dumb discourses where probability theory is a central topic.

Teenage phone use

• Review of “The Anxious Generation” (2024)

• National Phone Bans in School Are a Bad Idea (2025)

• Some 15 year olds are adults (2025)

• Jonathan Haidt Doesn’t Care about Smart Kids (2025)

Banning me from the internet at 15 would have seriously messed up my life, but there’s a demagogue peddling crap statistics to push privacy violating age verification laws to ban under-16s from most of the internet worldwide. This is extremely obnoxious behavior which panders to the lowest people at the expense of the smartest. Intelligence is a lot more important than age so, while proles intuit that they got better at making decisions between 15 and 30, bright 15 year olds still beat dumb 30 year olds on basically every metric, which should seriously make any decent person question a law that unpersons every 15 year old, instead of, say, unpersoning everyone under 95 IQ in the same way instead.

Economic crowding out of elite whites by Asians

• The severity of Asian Tech Immigration (2023)

Tech is so non-white that if you removed all of them from America, the media legacy American tech worker would make $91,000 more per year given a constant elasticity derivative (as you remove foreign workers, the elasticity stays the same).

• The Narcissism of Asian Immigrants (2026)

Discussing how entitled Asian immigrants act in the USA. Created several talking points on the topic: China is a shit hole and no whites want to move there, Asians claim math contests make them superior to whites despite this fact, their IQs and SATs are inflated by a huge amount by studying, which breaks tests. See How much do intelligence and g correlate?

• Thoughts on Prestige Universities and Alternatives (2026)

Entitled immigrants think that grinding their way into DEI shit hole schools that have trademarks to the names of now defunct 20th century prestige colleges makes them superior to smart whites. I discuss how to build a university that keeps them out by merit and which smart, based whites want to attend.

Future directions

I would like to make the Asian sequence more rigorous since they do seem to clog up all of the job pipelines at the expense of elite young white men. One idea of how to do this is to examine their IQ scores more closely; I would like to show that their fluid intelligence with white levels of training is lower than white levels. I believe their countries suck because they are dumber than whites on average, and they evolved to selfishly grind indicators really hard to exploit ability measures, in their own countries. Now we are dealing with the ruination of our own educated class in white countries by unbridled Asian migration. The other topics seem played out. I would like to start writing about finance and money soon; I’m working through quantitative finance books for a foundation. Stuff like the evolution of money and finance, how HBD fits in, how humans are evolved to do it, group vs individual evolution in the emergence of modern asset markets, theories of memetics and how they operate on pricing.

Read more

Read more

Read more

10. Early marriage is more natural

For all of human history, people married early. See here for a data review on this. Some think that this means that delayed marriage is an improvement; on the contrary, social practices need to be optimized when there is widespread poverty. Robin Hanson has been accessibly blogging on this since at least 2024. My scientific work agrees with his sketches. As an early marriage practitioner, I can confirm my wife and I are very happy with our decision to not delay marriage into our 30s.

9. Early marriage is more romantic

Early marriage means less sexual experience upon marriage, ideally even virginity. This is obviously more romantic than marrying with a large body count. Your spouse becomes your first love, not just the last in a trail of heartbreaks. Most romance stories are also between young people under the age of 25. People under this age might feel romantic love more passionately than older people.

8. Early marriage is more beautiful ; breakups and celibacy are ugly

Early marriage means less breakups and less celibacy throughout life, as well as less sex outside of marriage. Breakups are ugly and immoral, most people are not designed for celibacy and finding it extremely frustrating, and the only alternative is sex outside of marriage, which is risky, unloving, and gross. Thus, early marriage increases the amount of beauty in the world by decreasing the amount of ugliness.

7. Early marriage prevents wasting time dating or being sexually frustrated

Dating is hellish and I feel bad when I see dating discourse. By being early married, I know I will never date again, and that there are no downsides to this. I don’t have to be celibate, and I don’t have to miss out on having a family, because I am married for life. If you had contemporary dating culture, you should support early marriage.

6. Early marriage improves the TFR

Dropping the average age of marriage by 10 years would increase the TFR above replacement, by about 1 extra baby per couple, according to an econometric causal-inference study. The increase is modest due to the presence of family planning, but it’s still positive. If you are pronatalist, there is no downside to advocating for more early marriage. It has a lot more positive spillover effects while being cheaper than the main alternative that pronatalists have to early marriage, which is promoting IVF for very late couples.

5. Early marriage cultivates life-long love better

There are no causal inference studies I’ve found that find the effect of age on divorce risk, although cross-section data usually gives a U shape which shows marriage in your 20s is better than marriage in your 30s. Lyman Stone found that among people where early marriage is normal, the best outcomes were when marrying under the age of 25. Marrying over 30 was associated with poorer marriage quality.

4. Early marriage lets you have kids earlier which makes it easier

Say you have to have a certain number of kids. Do you want to be 45 waking up at night to feed them as babies, or 25? Do you want to attend their high school graduation when you are nearing 60, or 40? Kids take a lot of energy, so being in your 20s and 30s is better for coping with it. You will also see more of your grandchildren, the earlier you have kids. There are biological analogues to this; pregnancy is easier in younger women, and sperm has less mutations in younger men. Babies born to parents in their 20s are slightly healthier than babies born to parents in their 30s and older.

3. Early marriage is a better motive for a strong career

It’s important to not have wicked motives for making money. Single people who don’t want to make the world better who chase money likely have wicked motives for it. This is ugly. If you have a family, however, supporting your family is a good motive for chasing money. If many people are not wickedly greedy, you might even expect that marriage causes men to pursue more money, which boosts the economy in a system where most of the money to be made is through positive-sum interactions. Indeed, Cremieux has showed data indicating this. Even if the games are neutral-sum, it has the side effect of taking money from wicked greedy people, like Sam Bankman-Fried, and giving it to less wicked men who will use it in part to reproduce, which could be consider eugenic or a socially-positive evolutionary pressure.

2. Early marriage prevents STDs

Arguably a huge reason why marriage evolved, and evolved to be early, STDs are a huge problem and usually kill you without modern medicine. Treating STDs is a huge sunk-cost as well. More early marriage means lower healthcare costs and healthier people. Easy win!

1. Early marriage is what God intended ; it improves the Culture

Religion (Christianity but also other major world religions) promotes early marriage for a reason. Thinking in a more secular way, I’ve already listed 9 spillover effects of early marriage. I’m sure there are broader, unexpected effects as well. Radical feminism would probably recede in the public space. The demographic pyramid would be less tilted towards the elderly. This has downstream effects on the community. Things begin to be more family oriented, child-friendly, etc. All of this seems positive to me.

Read more

To answer this question, we have to conceptualize intelligence well. Specifically, we need to know what it’s not, and see if something that is not intelligence causes differences in g-factor scores. One thing that impacts g estimates that is not intelligence is measurement error, but we’ll assume that is minimal, as the reliability of an average of several long g measures (for example multiple g scores extracted from WAIS trials, where each score has a reliability of around 0.90) approaches 1.

Is there anything in a perfectly reliable g measure, like an average of many g scores based on WAIS, that is not intelligence? I think there is. There are many people out there who claim IQs that don’t seem to match an accurate qualitative assessment of their intellectual output. Some of it is exaggeration (self-reported IQs only correlate with measured IQs around 0.30, and online people frequently lie about having taken real tests when they throw out a number), some of it is reliability, but some of it is probably that their g scores are really just inflated relative to their actual intelligence, because they come from backgrounds where people train for the test much more than average.

And there is evidence that training for the test works. One study found that merely taking the WAIS two weeks ago boosts a subsequent FSIQ score by 6 points. For people who train hard and who are at least average, the effect is even bigger.

If 1% of people come from a culture of test training, and are scored on normie norms, their IQs can easily be inflated by 20 points. This won’t show up in reliability corrections; they can take tests again and again, and keep scoring 20 points too high. This would produce people who clearly think at the 125 IQ level, claiming to have scores in the 140s or even higher (but reliability analysis reduces their claims into the 130s or 140s, which is still higher than their real intelligence). I believe there are many such people on X and Substack.

Based on this data, I’m guessing practicing accounts for 20% of g variance with intelligence being the other 80%. That means intelligence and g would correlate at around 0.90. This seems broadly consistent with Arthur Jensen’s correlations between g-loadings and neural measures. It also means the correlation between practicing and g is about .45, or 6.75 IQ points per practice SD. Someone from a culture that is in the top percent of practicing should therefore have 15 to 20 points subtracted from their g score, to estimate their intelligence. This is consistent with the data from the figure above.

Additionally, there’s evidence differences in schooling, motivation, and need for cognition have small effects on g scores, usually around 5% of variance. That would sum to 15% of variance if they are all independent, a good chunk of our hypothesis.

I have also noticed that a lot of sketchy IQ claims come from people with “verbal tilts”. This may be because their training regime is verbal-focused. For example, the schooling and religious catechism of some over-training cultures emphasize verbal abilities while largely neglecting 3D shape rotating skills. When people from these cultures train, it should produce a tilt. Perhaps their spatial IQ scores are therefore better intelligence measures, despite having lower g-loading. They have lower g-loading, because g is in part training, and not loading on training reduces g-loading.

If this estimate is right, it should have application to the intelligence estimates of public figures. Once the g posterior is deduced from observed scores, one must mix in a prior for how much they trained. For many figures there is strong evidence of excessive training; for example, entering the SMPY, especially on math, at age 12, indicates humongous targeted training. Even 160 IQ 12 year olds don’t deduce high school math from first principles; they have to learn it from books, and even at 160 IQ many preteens would rather spend their time doing things that are more fun. Therefore, having gone through the books is a massive training effect, probably over 2 SD, and one could subtract 13.5 IQ points from the estimated IQs of SMPY participants based on this. Importantly, it’s probable that this would have to be done even from a highly reliable adult g-score, since the training effect actually changes their g, but not their intelligence.

Even a neutral prior for how much someone training introduces regression to the mean and uncertainty. 20% of variance in IQ means intelligence will tend to vary about 6.7 points around someone’s g scores, and a 0.90 correlation between intelligence and g means g scores shrink to the intelligence mean by 1.5 points. So, if someone has robust g scores of 140, using a neutral prior for their practice amount leads to an intelligence posterior for that person with a mean of 138.5 and a standard deviation of 6.7. That means there is a 1 in 50 chance their intelligence is really under 125, and a 15% chance it is really under 131. Alternatively, there is a 15% chance it is over 145 and a 2% chance it is over 151. This might help make sense of why there seem to be people who differ a lot in apparent intelligence even at similar g-score levels. For example, Richard Feynman claimed an IQ score of around 125, but his intelligence could have been in the mid to high 130s, even if that score was honestly reported. If it was unreliable then even the 140s are very possible. At the same time, a lot of people claiming 125 will come off a lot dumber than Feynman; their intelligence level can easily be under 119 given robust g scores of 125. This might lead to a 10 or 20 point intelligence gap between Feynman and someone else claiming the same g score, which seems to be the case in reality.

Sapolsky writes:

This chapter’s key fact is that the final brain region to fully mature
(in terms of synapse number, myelination, and metabolism) is the frontal
cortex, not going fully online until the mid-twenties.

For this he cites R. Knickmeyer et al. 2008: “A Structural MRI Study of Human Brain Development from Birth to 2 Years”. Bizarrely, Jay Giedd and friends did something perfectly analogous to this in 2005 in their pamphlet to prevent teen pregnancy (activists!)

In my book, I showed that this claim was false and based on a citation that refers to infants:

The actual data shows that myelination continues into the 40s for most parts of the brain. Below, GCC, Ant CR, and Ant IC are in the frontal lobe area:

Here’s another with multiple myelin measures and scatterplots so you can see how development after the age of 20 really relates to aging after the age of 60 in magnitude:

FL is frontal lobe and as you see there is no line drawn in the mid twenties. From my book, the teenage slope is not noticeably larger than the 20-something slope. The brain is mostly developed at 15 and myelinates into the 40s.

Back to Sapolsky:

This has two screamingly important implications. First, no part of the
adult brain is more shaped by adolescence than the frontal cortex.

The occipital lobe has more development from 20 to 40 than the frontal lobe on the MWF measurement. I guess adolescent vision isn’t working yet? Adolescent eyes, a work in progress?

Second, nothing about adolescence can be understood outside the context of delayed frontocortical maturation

An embarrassingly overconfident thing to say when your basis is nothing of substance.

we’ve just explained why adolescents are so frustrating, great, asinine, impulsive, inspiring, destructive, self-destructive, selfless, selfish, impossible, and world changing.

I wrote my book as a teenager, but now I’m in my mid-twenties, and I can say that both in my teenaged experience, and my present experience, these stereotypes are completely ridiculous.

Think about this—
adolescence and early adulthood are the times when someone is most likely to kill, be killed, leave home forever, invent an art form, help overthrow a
dictator, ethnically cleanse a village, devote themselves to the needy, become addicted, marry outside their group, transform physics, have hideous fashion taste, break their neck recreationally, commit their life to God, mug an old lady, or be convinced that all of history has converged to make this moment the most consequential, the most fraught with peril and promise, the most demanding that they get involved and make a difference. In other words, it’s the time of life of maximal risk taking, novelty seeking, and affiliation with peers. All because of that immature frontal cortex.

I feel like most of these fit my present age group better than teenagers? It seems like he’s just describing things elderly people don’t have (vitality, energy, testosterone), falsely correlating them with the frontal lobe, and the arbitrarily claiming teenagers are defined that way. But this is a better way to describe people aged 21 to 35. In fact, a lot of age-crime curves, basically ones that are not from the USA between 1970 and 1999, show that 20-somethings are actually more violent than teenagers.

People that are not elderly have higher vril, probably due to sex hormones, reproductive instinct, and less physical ageing (hard to be violent when you have an enlarged prostate and need a back surgery and you’re always tired). This applies hopefully to healthy people until as late as the age of 40. I know this is hard to understand with the backwards American population period. Everyone now is old. We’re dominated by old people.

Once you understand that, you can ask about teenagers, (i.e. “adolescents”) specifically. In what ways are they unique when compared to other people who are not elderly? The answer is that they are not that unique, but insofar as they are, they are just more childish. Not violent, or risk-taking, or domineering, or zealous, but more like 9 year olds than they will be in 5 more years. Specifically they seem to like to play more, they are weaker, they are smaller along most horizontal measurements even after they stop growing vertically, and they are more fearful of things like the dark and in general. This makes sense because they are just the age group in between 20-somethings and little children. However, after the age of 14 or 15 they are closer to 20-somethings than children of an equal age-distance because puberty development (ages 11 to 15 usually) involves a sigmoid function.

It’s like walking up a steep hill for 5 miles or on flat terrain for 5 miles. Which is harder?

Back to Sapolsky, he goes on to discuss gray matter. He repeats the misconceptions I debunk in my book, namely that pruning somehow hits a critical point during the teens or 20s. Rather, pruning continues into old age.

In the next section, he writes:

In the foothills of the Sierras are California Caverns, a cave system that
leads, after an initial narrow, twisting 30-foot descent down a hole, to an
abrupt 180-foot drop (now navigable by rappelling). The Park Service has
found skeletons at the bottom dating back centuries, explorers who took one step too far in the gloom. And the skeletons are always those of adolescents.

I thought, hey, everyone gets in accidents, so adolescent overrepresentation is maybe plausible, but saying that all the skeletons were adolescent sounds wrong. Most modern dead cavers are young men. Sure enough, there were 11 people founds, 9 adults and 2 children. 1 was a 12-15 year old girl.

I will stop here for now because it doesn’t seem like he’s even trying to be accurate. If you don’t want to miss a part 2, make sure you subscribe:

• 70: mentally retarded, if wild then savage and incapable of social organization beyond a small clan, can’t pass fractions at school

• 85: doesn’t believe in stories, tactile. Can’t pass algebra 1

• 100: can’t pass calc 1, C tier in algebra 1, sees world through feelings, things they can touch only. believes in stories and myths

• 115: can’t pass real analysis, C tier in calc 1, sees the world through stories. Believes in abstract principles as feelings. Capable of designing rigid protocols for institutions

• 130: sees the world through data. can get at least C/B average in an undergrad math degree. Capable of considering broad unexpected outcomes of feelings based principles.

• 145: sees the world through mathematical models which produce data. capable of contributing at the highest level to knowledge. Can design a state with complex laws that always change based on circumstance and can implement a eugenics program because he understands the necessary math.

We can see from this that it would be nice to know what kind of math ability, ethics, and social organization people of different intelligence levels are capable of. As of yet, IQ only means moderate correlation with some life outcomes like SES. Being able to say “people under 120 IQ can’t do a bachelor’s degree in mathematics” or something similar would be a real game changer. I suspect something like this is true; I see a lot of people who just can’t math at all.

We can tentatively check how calibrated these numbers are. In the Chicago Public School system, 21% failed Algebra 1, but the system was 40% black, so it’s plausible the mean IQ of the students is under 100; our model predicts 21% failure if the mean CPS student IQ is 97 with a standard deviation of 15. So we’re pretty close.

As for 100 IQ people being C-tier in Algebra 1, about 40% of students who take Algebra 1 in 8th grade repeat it, even though only about 20% fail it. And ostensibly, people who take it in 8th grade tend to be a little smarter than average, so 40% makes sense given that people under 100 IQ can only get to a thin understanding of Algebra 1.

For calculus, 35% of people who attempt it in high school in the US fail. But only about 20% attempt to take it in high school. So a decent estimate is that 13% can pass calc 1 in high school, which is close to what I put above (15% of people are above 115 IQ). It seems about 25% of people with university diplomas have passed a calculus class, and about 50% of people get a university diploma, so that’s 12.5% of people. In other words, Calculus filters >85% of people.

Only about 4% of people get a 5 on AP calculus in high school, it’s likely if you can’t do this easily you won’t be good enough at high level mathematics. So that’s roughly people of a minimum IQ of 127. Perhaps at 130 IQ one could get a B average pretty easily in an undergraduate math program.

These numbers may make it looks like less people have pre-requisite intelligence to understand certain math topics than they actually do, since they implicitly assume that all success cleavages in math are due to an intelligence threshold. If work ethic, for instance, effects math performance too, there will be failures who could understand it if they worked harder, and there will be successful people who would not understand it if they worked an average amount. However, these numbers can be interpreted as showing what percent of people are smart enough to grasp a topic, holding their other traits fixed. In some sense, being lazy is an extended stupidity, because lazy people will never think hard with what raw brain power they may have, and therefore will never produce anything really intellectually interesting.

It would be interesting to model thresholds. For example, individual math topics could be interpreted in the IRT framework to have latent discrimination and difficulty. In Item Response Theory (IRT), difficulty is the ability level where someone has a 50% chance of answering correctly (midpoint of the curve), showing how hard an item is, while discrimination (the slope) shows how well the item separates high-ability from low-ability individuals. In other words, high discrimination topics are impassable to people under the steep threshold, while low discrimination topics may vary in difficulty (ie, people over some threshold learn it much faster), but are more amenable to repeat attempts (since, for example, a beneath-threshold person might have a 20% chance of “getting it right”, whereas if the discrimination were very high, they would have a .1% chance).

Discrimination aside, integral solving problems probably have a difficulty somewhere around 115 IQ, quadratic equation problems have a difficulty around 100 IQ, abstract theorem proving problems might have a difficulty near 130 IQ. Sadly it would cost a lot of money to run a large scale study comparing IQ test performance to realized math performance. Maybe an epic based billionaire could fund it?

Where are all the Korean and German 29 year old billionaires? Maybe with something a little more g loaded than a gambling platform? The answer is that the American economy is becoming one big gyp-joint in which the biggest gyp-joint gets all the fake money. Kian Saghedi’s gyp-joint, Nucleus, recently unshockingly ousted for being a scammer, comes to mind.

You could tell Nucleus was gyp-joint, as I did in 2023, because 20 year old college drop outs from money-hungry cultures are not good statistical geneticists. The prior odds were high, knowing the “founder” is a 20 year old Iranian with a high school diploma, that Nucleus would be a get rich quick scam, and not a cutting edge genomics company. That’s because statistical genetics is like physics and actually requires knowledge and intelligence and subject experience to be good at, unlike making shitty gambling platforms.

Herasight is not a gyp-joint and predictably, it’s less funded than the shitty gyp-joint, and it’s actually run by white and Asian 145 IQ quantitative academics. In a serious economy all new companies worth anything would be run by such people, but the story of the 21st century American economy so far is make a shitty gyp-joint app and get a bunch of money you didn’t earn. It takes very little skill so it seems to select for a relatively low intelligence threshold, say 115 or 120 IQ, and maybe energy, specifically energy directed at scamming and not doing good cutting edge research and science. So, in other words, it’s probably worse than a random lottery, shit-head assholes who have high levels of the sin of greed are disproportionately propped up for making a mockery of the US economy.

This happens because the US pöbel are dumb and love a good gyp-joint. They might even see themselves as a future gyp-joint founder. But they know they’ll never be a Wernher von Braun or a Nikola Tesla, so why elevate such people? To elevate such people would require virtues like foresight and humility. But it seems those are in short supply and we’re stuck with mass greed and narcissism.

They also love a good lie. It seems honesty is in short supply too. Nucleus appears to have lied about their product efficacy. Kalshi lies by claiming it’s not a gambling platform. Instead they say something, which sounds super gay: it’s a heckin hivemind for predicting the futurino!!!!!! It’s to incentivize deep research into world events!!! It’s a prediction market!

But the frontpage is all sports gambling and predicting coinflips which are impervious to deep research, such as the winner of the next US presidential election.

I would like to finish by just saying that this is all very ugly and I wish these people would stop turning the US economy into one giant scam. I think them getting free money is bad and these “founders” ought to work real jobs or not be in the West. The US is not your “free money glitch.” I’d also like to say I’m disappointed in the pöbel who vote for this and consume this crap. Honestly you could probably justify investigating Nucleus for fraud and Kalshi for racketeering under the current letter of the law, but that doesn’t mean it will happen. Everyone is too busy worrying about a dead guy getting some massages from teenagers 25 years ago to go after companies that might be causing large-scale economic damage. That is, Nucleus predatorily takes from IVF couples to deliver an inferior product, by way of stealing R&D and advertising resources from better firms, through scammy and unethical competitive practices like backdoor networking and lying about product performance. Kalshi is a casino and it takes energy but never gives back. I don’t believe the world will improve because of “prediction markets” anymore than it improves by opening a new casino. So it’s an enrichment scheme for its founders who are not exactly wise, virtuous leaders deserving of billions of dollars. They’ll be wasting and harmfully spending all of that money just like George Soros.

Subscribe now

Read more

I ended up finding an alternative formula, but it yields the same results as Frisell 2012, which per my simulation actually produced a correct formula. Sadly, my formula is less efficient, i.e. it appears to produce noisier estimates. It is also only asymptotically unbiased in number of points per cluster, while theirs is unbiased. For now I’m just putting it up on github here.

Plugging in the data from the Herasight paper for the UK Biobank, I found a Frisell corection of 1.33. It looks like they used about 1.28. Using 1.33 would give .472 instead of .456. The confidence intervals overlap enough that this doesn’t matter.

I’m quite confident their correction factor is not flawed in any practical way.