This is true, but there is strong evidence from the second cohort that low number of events --> higher social isolation --> lower productivity. Should be positive on net.
koreindian
Karma: 280
If more Inkhavens were to happen they would be in off-seasons for Lighthaven, so the next one is most likely April/May. A Summer session seems unlikely for 2027, (but it’s possible).
Quick Thoughts:
Daily pace is excellent. I get a lot of looks at what I want my writing to look like. I had some very concrete ideas coming into Inkhaven about how I expected to improve, and I’ve already overshot those goals. I expect Weekhaven to be weaker growth stimulus overall.
There should be one day off per week. This would allow people to rest, plan more ambitious & fun events, go on hikes and actually decompress, etc. It would also give residents doing technical work the chance to focus on research rather than writing.
More mandatory, scary feedback from peers and residents sounds good. I’m not making myself get feedback. This is low-agency behavior on my part. At the end of the day, it’s up to the residents themselves to get the most out of the experience. That being said, yes limiting the action space of the residents a bit here would probably help.
Even though I have heavily underutilized them, I’ve gotten a lot of value out of the contributing writers. I overheard a snippet of feedback Andy Matuschak was giving someone else that I’ve been thinking about for days now.
I’ve read 70% of the submissions, and median writing quality is much higher than I expected coming in. There are definitely posts that seem both low insight and low effort however.
I feel pretty well socialized. I don’t expect to form lasting friendships, but many people have already left very deep impressions on me.
Maybe scheduling two, two-week Inkhavens a year will reduce the opportunity cost, and allow you to iterate & filter participants more.
Maybe this is inappropriate, but is there a path to convincing you to stay? I disagree with some of the details of what you’re saying, but much seems directionally correct and important. It would be a shame if those with your values were to vacate the commons.
As David Duvenaud said, a surprisingly high number of researchers and engineers at leading capabilities labs believe in transformative AI, but have underdeveloped or overtly incoherent models of the future. I suppose one model of how said people could be this way is that they only value money and power, or are only interested in securing their slice of the lightcone. I personally think most are persuadable (though probably not persuadable in time). Maybe what is needed in the community is more courage, as per Nate Soares?
Maybe this is because I avoid the Bay Area, but the median rationalist and effective altruist I know (and particularly those who are my friends) demonstrably care about more things than maximizing wealth and living forever. In fact, I honestly have a hard time persuading those around me that they should care about living forever.
And separately, even if that were all in the comment, don’t prescribe dating preferences to other people. People are allowed to be (not) attracted to anyone they damn well please.
This is confusing. Prescription != proscription. I prescribe that people not be fat and sedentary. I don’t thereby think that people are “not allowed” to be fat and sedentary.
Saying that “I’m not racist” means “I did nothing wrong” is actually making up a definition. That is not a criterion used in talking about racism pretty much anywhere in academia. “Sexual racism” is an apt term as all the behaviors I can think to attach to that label squarely fall within the ordinary language definitions I see here, here, here, etc. It is implausible that “racism” is only a valid label to apply to things like employment choices, but not to sexual choices.
Now it may be the case that racial sexual preferences are a non-central instance of generic racism, hence the fallacy. My point in bringing up Is Sexual Racism Really Racism? Distinguishing Attitudes Toward Sexual Racism and Generic Racism Among Gay and Bisexual Men is to pose the argument and evidence that these things are actually much more centrally clustered to generic racism than you might have thought.
I don’t think we should be held hostage by the conversational standards of social media. In the common discourse “eugenics” is synonymous with evil. I am still going to use the term “eugenics”, and not going to shuffle around to try to find adjacent pointers. I do taboo “racism”—I have an addendum talking about what my mental model for that label is. You don’t like that you have to scroll down to read it? I don’t think that’s a big issue, sorry.
I agree that one should be concerned about bad incentives in the cosmetic surgery outcomes research. Most of the “researchers” are surgeons. Outcomes research in medicine in general suffers from this problem. I am still comfortable with saying that cosmetic surgery outcomes research does provide net evidence for surgery results being good, even after correcting for conflict of interest concerns.
I’ll write a follow-up post focusing on this question (on my blog, not LW).
You are overconfident about this.
I posted some literature on limb lengthening outcomes in this comment.
There is also more natural variation in limb proportions than you are allowing for. Just ask your bouldering friends for anecdata about ape index variation. I have taught many of my friends how to squat and deadlift. There is enormous natural variation in shin/femur ratios. It does not at all seem ridiculous to me that many men could add two inches to their femurs without going outside natural variation.
When I buy gis for jiu-jitsu, I have to buy separate sizes for tops and bottoms and hem the bottoms down. This is because compared to Caucasians my height, my hips start three inches lower. People do not realize that I’m disproportionate in this way by looking at me, and it took me decades to realize this about myself.
I said this to some friends recently:
...people on social media can tell [that shin-lengthening surgery looks weird] because these guys usually go viral after posting before/after photos, and so yeah you can convince yourself that the proportions look weird in that context, but if you saw these guys post surgery on the street you wouldn’t be like “hey everybody get that long-shinned freak!”
One other interesting sort of anecdata I’ve heard is the trouble with Asian American males trying to date Native Asian females (like a tech or finance professional working overseas in Singapore, China, SK, etc.). There isn’t an attractiveness gap, but there is a cultural gap where Asian women are more forthright about asking things like “do you have a car?”, “do you have a house? why not?”, etc. There’s like an odd maneuver we do in Western dating were women are not allowed to explicitly ask about financial matters, but the men are expected to make it legible.
I’m interested in this idea you’re raising about neuroticism and race. I find Gold et al. (2008):
Yeah it looks like you’re correct, Asians are more neurotic due to a deficit in even-temperedness, though the other subcomponents of neuroticism are favorable to Asians. But there do seem to be large effects favorable w.r.t. attractiveness for Asians for agreeableness and conscientiousness.
Interestingly it seems like Blacks have better 5 factor than Asians in pretty much every respect:
Hm, I don’t know. This seems pretty subject to what Aella was talking about:
Plastic surgery is more common than you think (people don’t tell you) and has better results than you think (the only surgery you notice is bad surgery; all the great surgery is invisible). It’s also a great mental health boost, if I supported public funding I might for that
Maybe there are some non-subjective local factors going into this as well? I live in Los Angeles, I see a lot of pretty good plastic surgery.
It’s possible that the plastic surgery literature is compromised and is just a part of the sales department of the plastic surgery industry, but I don’t think so? The literature is quite large and diverse, and it doesn’t seem to me that they are making big enough methodological mistakes to line up with your data defiance.
Some defense of the Michael Thai methodology, their Study 3 does look at non hard and fast rules (e.g. “Asians preferred”), and they still find penalties against perceived racism and dateability for those profiles even within the “preferences != racist” cohort.
To the essence of what you’re saying, I hear you and am sympathetic to the point you’re putting forth. I guess stepping back a bit, it just feels memetically unstable to live in a world where all implicit racial preferences are basically fine and unquestionable but it is a grave offense to actually articulate that preference to another person. Like let’s say Person A judged 100% of race [x] to be unattractive, but wasn’t even conscious about this and is careful not to humiliate people by expressing this preference in real life (but does act on it). Let’s say Person B makes similar judgments, but is aware about this and will tell people frankly that he has this preference. I think what’s weird (I don’t want to say ‘offensive’) about both of these situations is the 100% unattractiveness judgment, not that Person B has the self-access and confidence to articulate their tastes. Removing my own moral judgments from the scenario, functionally I think what’s going on is that we are socially punishing Person Bs (a minority) in order to exonerate Person As (the majority). To be clear, I’m not in favor of witch-hunting the Person As of the world, nor even concocting elaborate social schemes to prevent them from acting on their attractiveness judgments. Rather I think it’s fair game to point out “it is weird that you have such a strong implicit judgment, and it seems like there is some confusing socially-subjective mechanism that is putting those judgments into you”.
Right yes thanks for this insightful comment Jenn. This rings very true when I reflect on my own experiences as well. I think part of my middle-school and high-school attraction to middle-class white women involved factors very similar to the ones you’ve mentioned: a lightness of mannerism, more social confidence and situatedness, their family dinner tables were lively and fun, more likely to read books for fun, their parents actually had friends, etc...
It makes sense that these deep middle class manners take a generation or more to settle in. Similar to this “Room of One’s Own” analysis you’re giving on women writers, I think about Proust’s analysis of Jews in France. Swann, the Verdurins, the narrator himself, they are all the grandchildren of Jewish stockbrokers or traders, and it takes them more than a lifetime to bury this fact and successfully join the social circles of, absorb the mores of, and ultimately supplant the older aristocratic French families like the Guermantes. Cultivating the appropriate lack of baggage and grievance takes generational time.
Thanks for the great questions.
Technological changes make this difficult to track. A lot of the research on explicit racial preferences in online dating looks at the last generation of online dating platforms (Match, Yahoo Personals, etc.) because when you specify racial preferences it is visible on your profile. The current generation of platforms do not do this generally, and instead race preferences are specified on the matching side. I’m assuming then that stated preferences have dropped very low, in part because they are functionally unnecessary. Grindr had but recently removed algorithmic race filtering, causing men to put racial preference in their profiles again, but it seems like there is community consensus that this is bad manners. Another anecdote, there is a man in Nathan Fielder’s The Rehearsal Season 2 that claimed to have been kicked off of every dating platform due to writing “no t-girls” in his profiles. This man seemed an unreliable narrator, and in general seemed socially clueless, but it does seem plausible that you could be on the receiving end of a moderator action for explicitly stating preferences.
So more recent experiments on race preference are typically surveys done in-person or on Qualtrics. Gendered Black Exclusion (2014) could be taken as representative of this type of research, which finds:
Which are comparable exclusion rates to what Robnett & Feliciano found with the 2005 Yahoo Personals data. Meanwhile, explicit races preferences have probably dropped quite low. This 2022 article is also useful data.
This doesn’t quite get to the core of your question, which is about how people’s actual behavior around having racial preferences can diverge from their ability to articulate to others (and themselves!) that they have such a preference. Maybe to get some estimates we can take Robnett & Feliciano’s data as representative of the latter, and compare that to the analysis of reply rates by race in the OkTrends data as representative of the former.
So the reply rate data shows the same directional trends as the Robnett & Feliciano data, but there are clearly a lot of interpretative subtleties here that I leave to you.
Comparing the reply rate data to the match % data, this suggests to me the possibility that people say and think they have strong race preferences even when they do not (maybe these preferences soften when you put yourself in an actual position to not follow them). The difficulty is that this data is conditional on the reply rates, which in turn are conditional on sending rates, which in turn are conditional on stated racial preferences (you probably won’t send messages to people who say they are excluding your race), so it could be that the match % data merely shows that people actually have super accurate explicit race preferences.
2. The above OkTrends data gives more sub-Asian granularity (here’s another blogpost). Outside of this I don’t recall seeing many studies with high granularity. This one looking at Chinese, Japanese, and Korean international students might prove of interest.
Yes I was hoping for this sort of reaction.
But also to add some color on recommendations here. The two people that I shot down on steroid usage IRL were about to take trenbolone, which is a truly terrible idea. For attractiveness purposes, cycling a medical dosage of anavar (i.e. 2.5 mg daily) is more than sufficient. For context, that’s a dose that is anywhere from a fourth to a fiftieth of what athletes and bodybuilders take. This is not going to destroy your fertility, and is not going to add significant cardiovascular risk as far as I can tell.
And what, in general, should such a reaction be?
If you are low-race status in the West and you want to be comparably desirable as your peers, you need to be prepared to compensate for your low-race status (on top, of course, of whatever other factors are reducing your general desirability). Your white peers ceteris paribus do not need to do this extra bit of racial compensation. Both Black men and Black women do have to do this. Asian men have to do this. Due to various asymmetries in desirability and endophilia, Asian women do not need to do this and Black women have to do more of this more than Black men. The means to increase this desirability could be pushing harder on the generic answers (working harder on your physical health, collecting more money, more status, etc.) or the less generic, exogenous answers that I’m encouraging readers to take more seriously (surgeries, & hormones). Alternatively, you could accept having lower desirability and tinker with low desirability strategies.
If you are not low-race status, I’m not telling you to do anything really, except perhaps, following Amia Srinivasan, I am encouraging people to reflect more on what affects what they see as attractive.
If the driver is perceived attractiveness, then how is this different from any other aesthetic preference, in terms of how either party should react to the preference’s existence? … If the driver is stereotyped perceptions of personality and culture, well, that can be very different. Then it becomes a question of whether the stereotypes are completely wrong, which they often are, or whether they are derived from an actually-racist-in-the-evil-sense distortion of some underlying-but-real trend, which they also often are.
I think there a case to be made that some of the important inputs to sexual attraction are culturally-scoped. I argue for this by elimination in much of the post (culturally-invariant theories don’t seem sufficiently explanatory w.r.t. particular asymmetries in endophilia across genders within a race).
Here’s an additional argument to suggest that important inputs to sexual attraction is culturally-scoped. Asian American men are seen by women of all races to be significantly less attractive due to being seen as under-masculinized. From the perspective of evolutionary psychology, the perception of masculinity is functionally meant to indicate good genetic health in men. The principle thing that should be approximated is free testosterone levels. If Asian American men had lower testosterone levels than the average American man, it would make sense that the lowered perceived masculinity levels are actually tracking an important, culturally-invariant measurement of genetic health. However, Asian male free testosterone levels do not differ from White males (link 1, link 2, link 3). This causes me to update towards believing that the racial masculinity gap has a culturally-subjective input.
Ah okay this is helpful. It seems that many commenters do not share my priors on the high returns of the exogenous procedures I recommended, which I did not expect.
I think the evidence here is really strong that these procedures are great, and underappreciated by men outside of the entertainment industry.
Starting anecdotally: Aella tweet on plastic surgery, other Aella tweet on cosmetic surgery. She estimates a 5%-10% increase in attractiveness due to two facial cosmetic surgeries.
Literature on Limb Lengthening:
Marwin et al.’s review article Cosmetic Stature Lengthening (2020) finds:
Cosmetic stature lengthening provides favourable height gain, patient satisfaction, and functional outcomes, with low rate of major complications. Clear indications, contraindications, and guidelines for cosmetic stature lengthening are needed.
Mean height gained is 6.7 cm. The most common problem is limited ankle mobility, but again the rate is low.
Similarly Bulut et al.’s Meta-analysis of complications and functional outcomes in cosmetic limb lengthening (2025):
Seven studies were included, with a total of 489 patients undergoing EF or IMN for cosmetic limb lengthening. Procedure-related mortality was zero across all studies. EF was associated with higher rates of Paley problems (56 %), obstacles (48 %), and complications (10 %) compared to IMN, which had 37 % problems and 42 % obstacles, with no significant complications reported. Patient satisfaction was generally high, with EF studies reporting satisfaction rates of up to 95 %.
...
EF and IMN are both viable options for cosmetic limb lengthening. While EF may present more complications, it maintains high patient satisfaction, whereas IMN offers a safer profile with fewer complications. A collaborative approach, combining surgical options with psychosocial and psychiatric support, can help individuals achieve improved physical and emotional well-being.
Literature on Plastic Surgery:
Rankin (1998) tracks large mental well-being benefits:
...Mean scores for depression [determined by using the Center for Epidemiologic Studies Depression Scale (CESD)], improved from 11.2 preoperatively to 6.5 at 1 month, and to 6.3 (f= 79.3, p = < .0001) at 6 months after surgery. Surgical intervention produced no significant differences between preoperative and postoperative ways of coping and social support scores. Cosmetic surgery produces positive psychological benefits by significantly improving quality-of-life outcomes that persist long term, without adversely affecting social support and ways of coping.
Parsa (2019) shows for men there are large increases in obviously attractiveness, but also masculinity, youthfulness, likeability, extroversion, and trustworthiness:
Score increases were significant for perceived attractiveness (0.29; 95% CI, 0.13-0.46), likeability (0.41; 95% CI, 0.24-0.57), social skills (0.25; 95% CI, 0.08-0.40), and trustworthiness (0.27, 95% CI, 0.11-0.44) when evaluating all facial cosmetic procedures together (upper blepharoplasty, lower blepharoplasty, face-lift, brow-lift, neck-lift, rhinoplasty, and/or chin implant). Upper blepharoplasty was associated with positive changes in perceived likeability (0.72; 95% CI, 0.06-1.50) and trustworthiness (0.74; 95% CI, 0.22-1.25). Lower blepharoplasty was associated with decreased perception of risk seeking (−0.78; 95% CI, −1.45 to −0.10). Face-lift was associated with increased perception of likeability (0.69; 95% CI, 0.08-1.30) and trustworthiness (0.66; 95% CI, 0.05-1.27). Neck-lift was associated with increased perception of extroversion (0.60; 95% CI, 0.10-1.09) and masculinity (0.70; 95% CI, 0.21-1.19). Patients who underwent rhinoplasty had improvements in perceived attractiveness (0.51; 95% CI, 0.03-1.00) and likeability (0.40; 95% CI, 0.03-1.00). Chin augmentation did not show any significant improvements.
And then obviously there are many results showing how many facial plastic surgery procedures make you more attractive, with the procedures having low regret rates.
Ah okay yes. What I’m saying there is not well phrased.
What I meant is that it is common to blame irrelevant factors in the racial losers or to blame relevant factors disproportionately to the intensity of the racial dating preference.
So for instance if someone said “I don’t date race [x] because they are short” this is both that person taking ownership of the preference and putting “blame” on the racial losers (“blame” wouldn’t even be the appropriate word here).
I was thinking more of cases like an Asian American woman saying Asian American men are too patriarchal to date. There is a moral failing being attributed to Asian men, but it is not plausible that avoiding patriarchal partners explains the Asian woman’s racial preference, due to reasons I describe in the section “Common Arguments Attempting to Explain Asian Women’s Endophobia”. It seems like there is an attempt here to draw attention away from having a racial preference as such. The Steffi Cao article on the “Oxford Study” that I discuss in one of the later sections also contains examples of this.
Can’t be bothered to read the rest of it since it is extremely long.
Downvoting your comment because this seems poor.
I still feel somewhat wary of pushing people to take steroids out of a desire to be perceived as more masculine: it can go really badly. In general, I am wary of recommending extreme interventions to people without having a lot more context of their situation. It is hard to steer complex systems to outcomes you desire.
Yes this is very good. I think due to doing martial arts I’m more comfortable with steroid usage than I should be. I am also realizing that the two times I heard people around me say they were considering doing steroids for what I thought were insufficiently serious reasons, I told them very strongly not to do that.
I guess my mental model is something like this: some people get negative desirability points for things like their race. Those race demerits still exist legibly even when you do plausible amounts of compensation by ordinary means. So consider exogenous interventions (hence steroids, among other things). But those interventions should be narrowly tailored to compensate for the relevant invariant, and as you’re very correctly pointing out people are going to do that incorrectly by default.
Yeah I guess my intuition is that marginal returns on attractiveness are pretty significant basically everywhere. It definitely has high returns when you’re trying to attract men, and is softer when trying to attract women. But I don’t think there is a “cusp” or discontinuity.
On the first point, yeah I’m not sure the compliments model is a good one. Race and attractiveness matter a lot in online apps for reply rate, which in turn matters a lot for matches, resulting in effects that are a lot larger than like 2x.
Yes you can come to these of course.