Thoughts on everything I have written

Human Science

2023 — Against full scale IQ fetishism : I consider myself an “IQ lukewarmist”. I think IQ tests are generally decent measurements of intelligence. Said measurements tend to correspond well with financial success (explains ~25% of variance in permanent income and productivity) but not much else; correlations with other variables of interest like divorce risk, happiness, ideological beliefs, and taste tend to be null or disappointingly weak.

Despite the high correlation between IQ and productivity, I don’t think selecting employees for IQ works well because the tests employers use don’t tend to be accurate (such as the raven’s matrices test) and the existing job selection process already filters out people based on IQ, demonstrated by the lower variance in intelligence within a job position in comparison to the applicants. Those who doubt this are free to select people based on tests all they want — literally, it’s legal. Griggs vs Duke Power only discouraged using them.

Even if it were the case that IQ tests did work as selection mechanisms, I doubt people would want to use them anyway. They’re nerdy, boring, and make judgements too explicit and legible, which is something people don’t like. Having a bunch of numbers like IQ scores decide filters makes it more difficult to make looser decisions based on charisma and style, which are valued but might not correlate that highly with productivity. As Robin Hanson notes:

Years ago a NYC based software firm ran some prediction markets, hoping in part to find & promote “diamonds in the rough” employees who predict especially well. They did find such, but then said “Oh, not them”; such folks didn’t have the polish & style they wanted.

Anecdotally, I’ve noticed that the most successful and competent artists and creatives don’t seem to be as gifted in terms of IQ as the top tech titans (e.g. Thiel, Musk, Zuckerberg, Gates) or the greatest scientists (e.g. Einstein, von Neumann). I can even see it in their social backgrounds; scitech titans often attended top universities or came from wealthy families, while creative titans usually do not attend top universities and are only more likely to come from a wealthy family.

Perhaps this is due to the fact that IQ is capable of measuring the cognitive talents associated with science and tech well, but not the talents that are associated with writing narratives, making music videos, or painting pictures. Or maybe these kinds of creative pursuits depend on a range of non-cognitive traits like volition, taste, or openness that are not captured by intelligence. Either way I don’t think IQ is a great signal of artistic ability the way it is for scitech ability.

2023 — Signalling theory of hobbies : I endorse the theory, but I advise against adopting hobbies or activities that are high status because they are high status. I personally choose hobbies based on how pleasurable I find them (e.g. video games) or how talented I am at them (e.g. drawing). Productive hobbies are better than unproductive hobbies, but I feel like deliberately choosing productive hobbies takes the fun out of doing them.

2024 — Taking mutational load theory seriously : a comprehensive refutation of the theory that mutational accumulation due to lowering infant mortality is making us more genetically unfit.

In summary: the biggest theoretical issue I have with mutational load theory is that the more generalized mutations are, the more likely it is that mutations that cause early mortality will be selected out through miscarriages, survivalpregnancy failures, or reproduction. The less generalized mutations are, the less problematic it is that mutations associated with early mortality accumulate in the population.

2024 — Narcissism and selfishness are the human default : I still believe humans are selfish in a fundamental way. Not because of any scientific evidence, but because it’s not actually possible. I also don’t think this is a conclusion that people can arrive at through scientific thinking, it is something that must be personally discovered.

I no longer define narcissism as self-enhancement, but as an identity disorder that involves difficulties with identity formation and self-conception. I’m still outlining my thoughts on self-conception and identity, so I don’t have any concrete thoughts on the matter, but I did note in the article that distorted self-perception is the rule and not the exception.

2024 — The Role of Cultural Drift in Obesity : I will double down on the fact that the social class does not prevent obesity, but I’m not sure if it is cultural either. I think it might be a product of cross-trait assortative mating, where higher SES individuals exhibit a stronger preference for thinner mates, and so over time the correlation between SES and obesity increases.

2025 — Economic inequality increased ever since the dawn of civilisation : my argument is that the Gini coefficient is bad because it measures inequality in terms of proportion and not actual resources. I defend the post insomuch that I think economist’s insistence on using the Gini coefficient to measure inequality is extremely irrational. Otherwise I think the case that economic inequality increased since the 50s (as I claimed in the post) is admittedly thin. Numerically it did, but mass production has made it so resources like computers, food, good clothing, spices, furniture, transportation, housing, and healthcare are now available to most people in developed countries regardless of income. Inequality discourse should ideally include measures of the actual standard of living, but I lack the historical knowledge and life experience to evaluate the problem at that level.

2025 — the answer to the missing heritability problem : although I am right to dismiss the missing heritability problem and “molecular genetics studies” as red herrings, I still think that heritability research is valuable. Public understanding of genetics is surprisingly decent from what I’ve gathered, but behavioural genetics increases the dimensionality and rigor of this knowledge.

Academia

2022 — Top 5 problems with sociology : the problems were overreliance on observational data, genetic confounding, lack of real world testability, unintelligent students can still get good grades, and left wing bias. Not a bad list, but an awkwardly written article (as is the case for most of my early posts).

2023 — Why Ivies should exist — I think I should have made a macro-post on selection and Ivy league universities rather than a few loosely connected ones. My view is that Ivies are not necessary, but work fine as standins for public SAT/ACT scores, and that criticism of their methods of selection are misguided — many claim they’re not focusing enough on non-academic ratings, when in reality those “personality ratings” and “extracurriciulars” hardly predict anything in the real world and they would be better if they just took the top 20k students + some athletes.

2023 — University quality vs Student quality — I statistically prove that statistical measures of “university quality” all correlate highly (r > .9) with the academic quality of the students they admit. Some univeristies are obviously better than others in terms of aesthetics, features, and location, but I don’t think that those things are subject-independent or easy to measure. I’m also not sure how well they would scale with the academic quality of the students.

2024 — Academia’s prestige will nosedive in 50 years — I reaffirm the prediction, but restate it in the form “it will have nosediven in 50 years”.

Philosophy

2021 — Moral Unism: a Case for Objective Morality : here I attempt to construct a theory of objective morality based on the fact that preferences for certain states of existence are subjective but the collective preference for which state of existence maximises people’s preferences is objective and constitutes a valid goal to optimise for.

I see three large problems with this theory. The first is that there is no cosmic justification for why the collective preference should be optimised for at all — one could argue that it is irrational for somebody to value their preferences but not the preferences of others, but this is not the case if they value selfishness.

The second is that the theory doesn’t adequately or explicitly deal with cases where people’s preferences, even if revealed, are bad. Some people may pursue status in the first half of their lives but later discover they don’t actually care about it, while others may do the opposite.

The theory also does not wrestle with the fact that people have different higher values (egalitarianism, libertarianism) beyond different preferences for individual states. As such, they will have different ways of aggregating the value equation. One could argue that these “higher values” devalue themsevles and there is no reason for them to be taken into consideration, but if one acknowledges that is the case then the attempt to create a grand theory of morality is not exactly “the play”.

2021 — In Defense of the Ad Hominem : I argue that the Ad Hominem, commonly considered a fallacy, is indeed a valid argument because a person’s arguments are downstream from their selves. I do detail how genes and motivated reasoning play a role in the construction of beliefs, but I feel like the one line explanation of the idea is sufficient.

2022 — Absolute Monarchy is not a Panacea : around this time I got interested in the neoreactionary movement, particularly Curtis Yarvin’s work. I (sloppily) explained his philosophy and the idea behind his solution, neocameralism: having the government be a corporation governed by a CEO accountable to a board of directors.

The first problem I raised with his solution is that the difference between a corporation and a government is meaningful — a government can control its own currency and run a deficit indefinitely, while a corporation cannot (this point can be easily refuted by noting that the East India Company issued coins).

The other issue is the board of directors. If the board is appointed by the king, then the king is in control (so neocameralism is a dictatorship). If something else (e.g. lottery, public) appoints the board, then the board is in control (oligarchy).

Yarvin’s rebuttal wrt this is that, in monarchies, the public is politically apathetic because they believe they have no influence on politics. I would argue that this should be true about oligarchies (commonly known as democracies) if having no influence leads to no interest, but that’s somewhat beside the point. I’m also not sure if a politically apathetic public can exist in the current media and cultural environment even in a different political system.

I still largely agree with this piece, with the added point that leader-led political systems are unstable and are predicated on the competence of their leader; oligarchy might be stale and opressive, but at least it works consistently.

2022 — the Accelerationist Manifesto: The Eternal Whitepill : In the post, I concede to the opposition that there are many ways technological development has made the world a much more unfree place. Agricultural improvements have led to increases in population sizes, which have encouraged the creation of police (restricted freedom). Industrialisation has led to specialised jobs, which has led to the education system (restricted freedom). Technology facilitates taxation which encourages governments to build massive spending programs that accomplish very little of value (restricted freedom). Markets and a combination of other factors have encouraged the formation of corporations and modern jobs, which involve a lot of forced socialisation that some people aren’t going to like.

One might think that, perhaps, we could rise up and destroy all of this technology because it is bad for us. But nothing ever happens, so that’s not an option.

Rather than mope at this fact, my argument is that accelerating technology further is actually likely to make us more, not less free. If I look at the big technologies of the current day that have yet to run their course, they are:

• Internet

• Embryo selection

• AI

The interent encourages remote work which reduces the amount of forced socialisation we must endure. It also allows people to distribute knowledge more freely, so much that groups and advertisers have to shackle big tech into censoring their users. One could make the case that the internet has led to the creation of a more digitally exposed culture, but I think this is somewhat overplayed and not that much of a concern in everyday life.

Embryo selection encourages higher levels of intelligence and health which should lead to higher human capital, which is a massive boon to society: less criminals, less educational strain, higher productivity, and more desirable areas for housing.

I didn’t specify what the effects of AI would be in the article, as I only started getting into AI in mid 2022. It’s hard to predict what will happen because we don’t even agree on what version of AI we are getting: singularity, strong AI, or just more marginal improvements in currently existing technology.

Personally, I think the idea of iminent AGI is capeshit, because modern AI do not possess a lot of the features you would want in an autonomous, generally intelligent machine: innate values and processing that converts sensory/visual/audio data into information (aka, g). I don’t see how upgrading these machinese gives them those abilities. As long as these features are not striven for, current models will not yield AGI no matter how much training is done.

I think it’s likely that AI will, in the long term, reduce the barrier of entry of most white collar jobs. I’ve seen some make the case that AI will break down the education system the way the printing press broke down monasteries. I really, really, really, really, really, really, really want to believe this but the pessimist in my thinks that professors and educators will find ways to adapt, like with in person testing or oral exams.

In the original article I argued that population decline would also contribute to freedom by reducing population density, pollution, and littering. I still largely agree with this view, but though I’m not that interested in convincing anybody else of it.

2022 — Problems with libertarianism — in summary, I argue that libertarianism is not tenable because it is either framed through deontology (which makes no sense as a theory of morality, because rules must be founded on something or risk being arbitrary) or consequentalism (so liberty isn’t sacred anymore). I also brought up the criticisms raised against unregulated markets (externalities, limited information) which can result in the markets being a net negative for the group. I am not very interested in economics (and never pretended to be, even 3 years ago I admitted as much), but these arguments seem credible enough to me.

With regard to libertarians who are anarchists, I argue that anarchism is untenable because of the problem of national defense — an army that collects income from citizens is subject to a free rider problem where people who don’t pay the military benefit from their services, so there is no incentive for the anarchists to pay their own army. If the army decides to loot their citizens, they are arguably a government at that point. If anarchism is global, then all that will happen is that the political order will devolve into polycentric law, where everybody buys into law systems that benefit themselves, which is obviously unsustainable and eventually leads to one law system winning out.

The strongest case I could make for libertarianism is that we have far too many taxes, far too many regulations, far too many laws; liberty is underrated but not a self-asserting high value. Libertarianism is an emotional reaction to this massive state that should stay as a political movement, but not as an ideology.

2022 — Debate is mostly useless : The “debate mindset” assumes people believe things because they are true. They don’t. They believe in things because they feel good. Changing your opinion on a descriptive belief can change a prescriptive one, but usually the former exists to protect the latter. Some people do care about the truth: those people don’t watch debates, they research at their own speed or consult their peers. They also possess volition that others do not and cannot have.

A lot of times, when people disagree, it’s due to underlying beliefs or values which cannot really be debated on the fly. Good luck kidnapping a progressive and trying to sort out whether “authority” is better than “equality” (meaningless abstract words that don’t mean anything anymore). The specific debates are proxies for the general debate.

If you have to “debate” somebody, something is going wrong. Should I debate my mom on whether we should have pizza or cheese casserole for lunch? Should I debate my father on whether his driving is too slow?

2024 — Statistics vs Words vs Power? — not a very good article, but a decent attempt to counterbalance two irrational forces: the force which discards all verbal argument as “wordcelery” and the force that discards all data as “lying statistics”.

Politics and current events

2022 — is American Education Uniquely Bad? : I make the case that it is not — empirically, American students are in the middle of the developed world when it comes to scholastic performance, and on top of it if one controls for race. I also highlight statistics suggesting that policies that increase school choice and funding for schools are also unlikely to yield better grades.

Not exactly an uncommon opinion on this part of the internet anymore, but it was a solid attempt at answering the question.

2022 — COVID19 Theory Breakdown : I outline different points of contention regarding the COVID pandemic (e.g. lab leak vs zoonic, efficacy of the vaccine) and then map them to whether they are more supportive of the pandemic being planned or unintentional. I argue that a lot of the observations as of now are more consistent with the hypothesis that the pandemic was planned, but I leaned against the plandemic hypothesis anyway because I didn’t see a good candidate for a culprit.

I still dislike the plandemic hypothesis, but my reason for this has more to do with priors for an intentional pandemic being very, very low.

2024 — Does the Right Have a Stupidity Problem : No. There is like a 2-4 IQ difference between Republicans and Democrats in the modern day. The right’s loss of cultural power occurred far before it started getting dumber in the mid 00s.

2024 — the alt-right lost : I argue that the alt-right lost because it didn’t affect public opinion with regard to race or immigration, a perspective which I then debated with Walt Bismarck. Now I feel entirely different: the alt right temporarily lost because of the Charlottesville incident, but then inevitably grew afterwards into what we have now.

It was effective in offering an alternative to people who were skeptical of “the taboos” and wanted a more intense cultural experience, but it couldn’t deliver anything en masse besides memes that were judged by the collective consciousness to be inoffensive enough.

2025 — what happened to the dissident right? : somebody told me in a DM that this post could have been much improved from detailing things out or providing examples, which was definitely true, but I didn’t want to single out sane people who left and erased their tracks or deranged but talented people who insisted on staying.

Although I think the DR is in decline at the moment, I think there will always be a demand for something like it. I do think that the specific brand of alt-right culture that emerged between 2010 and 2020 that was polymathic (interested in philosophy/game/polsci/HBD/history…) and overtly edgy is on its way out due to decreases in literacy, more censorship, and lower levels of intellectual interests in the recent generations.

2025 — left wingers who are not awful — independent of political disagreements, left wing political pundits are almost always boring. They might doubt morality on philosophical grounds but their revealed beliefs betray their moralism. Most of them don’t seem to be very deep thinkers.

Left wing philosophy and punditry all converges towards the same policies and opinions due to a shared culture and value system. Even if individual pundits disagree on individual policies, they are all disagreeing on how to best optimise for those values. The same is not true of the right, where there is a wide variety of political camps with different value systems and outlooks (accelerationist Landians, street safetyist GDP-maxxing LKYites, catholic traditionalists, Fascist bodybuilders).

That said, I refuse to disavow Zizek. He’s charming. Funny. Interesting. I also find that left-right preference often does not track quality when it comes to domain-specialists.

Statistics

2022 — Is there a g-factor in political views? : More or less yes — it’s left wing vs right wing. It doesn’t load on every political question in existence but it loads on 96% of them.

In intelligence research, the “g factor” is a word used to label a hypothesised statistical variable that causes variance in all cognitive abilities (in other words, intelligene). People have also subject political surveys to factor analyses, and the vast majority of questions load onto the first factor: left vs right. It’s not immediately clear why this is: it could be social media echo chambers, it could be genes, or it could be due to how ideological beliefs are constructed. Personally, I lean towards the latter two.

2022 — Data damage in an AI study predicting professional LoL : I proved that a study predicting wins in League of Legends matches suffered from data damage, where the red side player winrates and blue side support winrates are no longer consistent with any of the other starting around row 1860. I tried contacting the authors and predictably got no response.

2022 — Reaction time and gaming skill : I find a low correlation (r = .2) between reaction speed and League of Legends rank based on survey data. I considered this evidence that reaction speed has a weak effect on actual performance, which is dictated more by the ability to anticipate events or react to different possible stimuli.

2023 — Database of IQ Estimations : using a (relatively) uniform methodology, I attempted to estimate the IQs of ~50 different public figures. I think I went a little overboard in terms of effort, but I endorse the implicit message I was trying to send: highly successful people are less intelligent than commonly thought and the >150 IQ figures that people tend to float around are made up.

2024 — Ageism is justified : I reafirrm the idea that people past the age of 30 or so are unlikely to dramatically shift in terms of their competence or social class, but I think I was too ready to dismiss the idea that luck matters.

The heritability of intelligence and big 5 personality traits are all about 75%, presumably the heritability of competence would be similar, but the heritability of income is only 45% (this means only about 60% of income in a given year can be explained by competence). The heritability of long-run income is higher (about 60%), which implies that 80% of long-run income is caused by differences in competence. At least according to my model…

(Note: competence and merit are synonyms, but not perfect ones).

2025 — The art of data analysis : a little navel-gazey and pretentious, but I disavow nothing I wrote here. Nothing can systematise the process of information processing.

Culture

2022 — How to estimate the quality of an anime in 5 minutes : a failed attempt to communicate the true intuition, which is that people can subconsciously predict with decent accuracy whether they are going to like something before they watch it.

2024 — Why do anime bloggers have such bad taste in anime : my argument is that watching things to evaluate them produces a systemtically different experience from watching things for pleasure, which leads to a distorted value system.

Posts I have no opinion on or tacitly endorse

Everything else.

Posts that were mistakes

2022 — Wordcels vs Shape Rotators

2022 — Midwit red flags

2022 — Analysis for TLAU

2023 — Why are Great Intellectuals So Rare : the triple event logic, that multiple traits are necessary to promote good and tasteful ideas, checks out in my head, but I think I ignored the far greater contributor: people with the talent to be good intellectuals also probably have skillsets that let them engage in other endeavours that are more fruitful.

2024 — What’s in a nerd : basically what I said is that stereotypes of intelligent people (unaesthetic, physically weak, socially stunted) have no basis in reality, but that the type of the “nerd” emerged anyway because it embodied a set of traits that distinguish people who perform well in school/work and people who perform better in normal sociological hierarchies.

I have no particular reason for disliking it, just wish I addressed the topic from a better angle and with more detail. There is also definitely a distinct “nerd” type that emerged in modernity that did not use to exist, but that the type itself is not reflective of correlations that exist in the real world (e.g. smart people aren’t more likely to be unathletic even if there is a distinct type of person who is smart and unathletic).

2024 — Does Saturated Fat Intake Cause Mortality? : I argued that the hypothesis that saturated fat causes mortality (particularly CVD) was empirically unsupported, which was actually true, but I’m starting to think that the actual theory itself is true, just to a very limited extent where saturated fat slightly elevates LDL which has an impact on CVD death risk, but the effect itself gets attenuated through the multiplication of effect sizes.

2024 — Midwits as impostors : I think I failed to explain the problem with the term midwit, which is that it unsuccessfully combines Gaussian (linear) and Jungian (typological) ways of categorising people. I also don’t think people with above average but not superlative levels of intelligence define themselves by having… Above average but not superlative levels of intelligence.

2025 — The Case for Utilitarianism : utilitarianism is just morality, but consequentialist and conflating value and utility. As I think value and utility are meaningfully different, I no longer endorse utilitarianism.

Ending thoughts

I’m mostly proud of what I have written. The biggest issue is that I invested too much energy into combatting ideas that were not worth fighting against, usually regarding people’s incorrect ideas on hereditarianism. In fact, I decided to compile a list of all ideas I don’t like that don’t deserve their own blogpost to save my energy for more constructive efforts. I also did the thing where you get interested in things you don’t actually care about that much but you see other people care about so you care about them too (in other words, I am mimetic like all other humans).

This was the case when it came to my interest in health. By nature I am healthy and tend to recover from illnesses, even bad ones, in just a few days. I’ve had my range of health issues (hyperflexibility causing all sorts of issues, varicocele) which I fixed through my own effort and knowledge, but otherwise I feel like a lot of my interest in diet and lifestyle came from seeing other people talking about it.

In terms of ideas, I think that most of my errors have stemmed from being too optimistic, despite being one of the most blackpilled people on the internet. I don’t think I am pessimistic, just asocial — somebody who isn’t interested in keeping up morale or thinking that other people and groups are headed in good directions. I also have a tendency to forget my own opinions due to having so many of them, which leads to my thoughts being less interconnected than they should be.

Other stuff

2024 — Politics of Hereditarianism : I reaffirm my thesis that talking about hereditarianism will not change anybody’s minds. Alex Young’s recent admixture study is also another case of researchers getting a lot of attention in niche internet circles that translates to absolutely zero media attention or real world impact.

The only way the discussion is going to reemerge in public discourse is if a high status and prominent person decides to argue in favour of the hereditarian hypothesis, which could go well (increase in visibility and dialectical advancement of hereditarianism at the cost of station) or badly (loss of station with no increase in visibility or dialectical advancement). I lean towards the latter.

2024 — How common is incest in the west? — I stand by the estimate that 30% of people have engaged in sexual activity with a relative; to those who argue about muh false self-reports, the high figure is partially downstream from the wide definition which includes playing doctor. The fraction who engaged in fondling or sex were a minority of that 30%.

2025 — Contra Lyman Stone on national IQs — I stand by everything I have written here, and it was fun to watch Lyman Stone write a reply that I never read, but my time is worth orders of magnitude more than his, so this post was ultimately a net negative for the world.

2025 — Childhood stories — believe it or not, everything I wrote here was the universe experienced through my eyes.