On reflection, no. Thanks for pointing that out. Swap out “perfectly” for “adequately”, and I stand by the rest of the comment.
re: relevance of the Royal Society
...why do you doubt the details are a crux?
Well, it’s not a double crux. Maybe it’s a crux for you.
The reason it’s not a crux for me is because it just seems so remote from the matter at hand. I already have years of direct personal experience productively collaborating with Achmiz on the kind of intellectual work that I want to do. His value to me is not in doubt. It’s just really hard to see what I could possibly learn about 17th-century England that would make me change my mind about that!
It seems like the argument would have to go through Achmiz’s mere presence on the website discouraging so much contribution—despite the existence of the user-level ban feature!—that somehow I’m better off privately benefitting from Achmiz’s counsel via email while he’s officially disgraced and banned from the website which is the central gathering place for the kind of work I do.
I suppose it’s not logically impossible, but it would be super weird for the empirics to shake out that way, and empirical evidence from 17th century England just seems too distant in time and culture to move the needle. The Royal Society isn’t totally irrelevant, of course: human nature hasn’t changed; they were our honorable ancestors doing the natural philosophy thing, and we’re doing a natural philosophy thing.
But our founding texts (which we agree, per the footnote on your 30 May comment, that the current project is running off the fumes of) are very explicit about aiming to do better thanour ancestors. The Sequences are exhorting the reader to follow the timeless ideal of Bayesian reasoning wherever it leads, not imitate how Robert Boyle won friends and influenced people in 17th-century England. (And my discourse-normsposts are likewise trying to appeal to the timeless ideal.)
re: Less Wrong 1.0
Do you have an explanation for the decline of Less Wrong 1.0?
I do think this is at least more relevant than the Royal Society, just because there’s less of a generalization gap to cross between “running a website in 2016” to “running the same website in 2026.”
What I think I can say (and I think you’d agree with the chatbot on these) is that the decline of the original website was not monocausal, and that the differences with LessWrong 2.0 are multidimensional. Things like LessWrong 2.0 having dedicated staff (instead of charity hours from Tricycle) and a less notoriously cursed codebase (in which to, e.g., implement moderation tools that can deal with the Eugene_Nier problem) are factors in LessWrong 2.0′s success.
As a result, when choosing interventions to make sure LessWrong 2.0 doesn’t die, it’s not a binary choice of “intervene or don’t”; there’s a high dimensional space of potential things to push on. I don’t find Habryka’s claim that “the specific way Said has been commenting on the site had a non-trivial chance of basically just killing the site” to be credible—particularly in light of my investigation of alleged author complaints in §III.2. To the extent that authors being discouraged by criticism was a problem, there were other available levers to pull that would save the website with less damage to the mission, such as encouraging use of the user-ban functionality, as I argue for in §IV.1 and suggested in July 2025. When I see Habryka reaching for the lever of purging someone he clearly personally dislikes when less intrusive measures were available and untried, I think I’m justified in judging this as a betrayal of the mission rather than a sincere disagreement about how to pursue the mission (forced by the need to avoid repeating the fate of the original site). If the claim isn’t that the Achmiz ban was necessary to literally save the site, but merely to optimize levels of user engagement, that’s not really substantively disagreeing with my model: I’m saying that you’ve betrayed the mission for the sake of popularity.
my read of you is something like “well we didn’t try to Correct Plan hard enough”. Like, if Yudkowsky had more personal virtue, then he wouldn’t have stopped posting on LW 1.0 and there wouldn’t have been the decline.
I wouldn’t go to “there wouldn’t have been the decline” (because of the high-dimensional multi-causuality), but yes, I do think that factor would have helped, and that Yudkowsky has more generally relinquished his Art and lost his powers. Granted, it’s not obvious how to intervene on a leader’s personal virtue (as contrasted to followers, who can often be bullied as Vassar did to me in 2008), but I think giving up on the idea of personal virtue is much worse!
re: people also being similar
But I think people are also similar, and there’s some value in us peeling back the layers and comparing how we’re put together
Right. That’s why I agree with 2013-Vaniver that “responding appropriately to criticism” is a trainable skill. I don’t think I’m an “emotionally tall” mutant preaching an impossible standard. I think other people could learn it, too, if the culture weren’t saturated with anti-epistemology claiming that they shouldn’t need to.
conflict vs. mistake theories of the present discussion
I am trying to understand what is going on
Are you? In your 30 May comment, you mocked me for spending too many words explaining why prioritizing hedonic and reputational effects trades off against error-correction and therefore correctness, “as the alert reader might have anticipated immediately”. But if it’s supposed to be so obvious that the mod team isn’t maximizing correctness, why would I believe you when you claim to be trying to understand what is going on, if I think that understanding what is going on would almost certainly have negative hedonic and reputational effects (because it would be a weird coincidence if the story that was true also made all the local power players look good)? Maybe you’re trying to understand in the privacy of your own mind, while carefully avoiding any speech acts that would deal unacceptable reputational damage? But the difference isn’t decision-relevant to me; I don’t get to interact with your private intentions in their purity; all I get to see is your speech acts, which strike me as evasive.
(Sorry, I know that was a super rude thing to say, but given what you told me about not putting correctness above all other concerns, you should have expected that I would entertain the hypothesis that you were telling the truth about that and consider the implications.)
I think Ben Hoffman’s theory about “a policy of unprincipled and unbounded submission to threats by the people (or personas) whose feelings are supposed to matter” seems like a better fit to the behavior I’m seeing from you and Anna Salamon. You report that you didn’t have “a personal problem with Said or his comments”, that you “generally found them easy to read and easy to respond to.” Salamon reports that her “personal experiences of Said on LW were clearly and substantially net-positive.” And yet both of you are arguing in favor of the ban to me—not because either of you are willing to testify in your own voice that you think Said’s comments are so bad that you’re not willing to share the website with him, but because other people—largely other unspecified people, given my investigation in §III.2!—aren’t willing to share the website with him. I think that’s weird! Naïvely, if I personally think something is fine, I should go on the record opposing banning it (for the record, even if I’m not the boss and it’s not my decision).
I need something like Hoffman’s theory in order to explain the behavior I’m seeing in you. Importantly, it’s a conflict theory, not a mistake theory: a story about different agents with conflicting goals, rather than everyone and the God-Empress sharing the same utility function. What you consider reputational and hedonic “costs” (to be minimized by the God-Empress for the good of all), Hoffman and I are construing as “threats” (in which an agent claims that they’ll deal disutility to everyone else if their demands aren’t met).
I think if you were actually trying to understand what was going on rather than trying to rationalize a policy handed to you by upstream social forces, you would show more evidence of having thought about the impacts on the site’s mission if the moderators took the other side of the conflict and stood up to the threats by telling complainants to downvote and move on with their lives.
In previous discussion, you’ve told me that the mod team isn’t going to tell users to “get gud” and that trying to shame you into stopping isn’t going to work. I understand that my complaints aren’t going to work. The ban isn’t going to be reversed; the function of this discussion is to get a shared accounting of what happened, where I’m not currently accepting the disvalue attributed to Achmiz as a legitimate cost on the shared ledger. You’ve told me that you felt more optimistic about communication of the form, “Okay, what is the thing that matters here, and how do we get it?” and not “the mods care about Something that is not Logic.” But the problem with that communication prescription is that it presupposes a mistake theory which you haven’t given me reason to believe in!
Do you seriously think that the mods telling complaintants “downvote on move on with your lives” would kill the site? (To be absolutely clear, I’m asking for a conditional prediction, not a policy change; I understand that the mods aren’t going to do that; I’m asking what you think would happen if you did that.)
I would expect some marginal users to bounce, but I don’t think the effect would be quantitatively large, and I think the resulting culture would be much more aligned to the mission. I think most people would grow, the way Michael Vassar helped me grow in 2008.
I could imagine you disagreeing with me on this as an empirical matter of human psychology, but there’s a missing mood in your behavior given that you’ve already agreed that a pro-criticism, pro-being-questioned emotional orientation is good for individuals and good for Society. If it turns out that I’m empirically wrong to think that it’s possible for people to learn the thing we both agree is prosocial, then my reaction is not, “I guess my approach was impractical, therefore wrong, and the new culture is a better way”, but rather, “oh God, we’re screwed, apparently almost no humans are capable of rationality; we were already dead; we were never going to survive.”
Previously, you’ve told me that you think Less Wrong has changed and that Habryka is the person most trying to surf those changes. I think the difference between you and me is that when civilization has fallen to a zombie apocalypse, I think it’s more dignified to go down fighting rather than joining the zombies because they’re the winning side (because surfing the changes means siding with the winners).
Again, I realize that that’s not a nice thing for me to say about you, but I think you’ve been commendably clear about your position, and I think it makes sense for me to be clear, too.
re: encouraging sportsmanship
I think there are obviously means for encouraging other people to do things besides arguing for it at length. For example, one could reward people for doing it, or one can model it in one’s own behavior.
Thanks. That answers my question of how else is one supposed to do it; now that you point it out, I agree that rewarding and modeling could also work.
Conspicuously, that doesn’t answer my question about what textual evidence supports Achmiz doing more to push people away from the right sort of sportsmanship, such as would justify your “not true on net” judgement. I think that after having conceded, as you have, that the pro-criticism orientation is prosocial, the “obvious” conclusion is that the Achmiz ban was unjustified; in that context, the “not true on net” judgement, stated without evidence, looks suspiciously like an ad hoc goalpost-moving excuse. (It’s an odd notion of sportmanship that demands that players not only exhibit good sportsmanship themselves, but also do so in a way that causes others to as well. Traditionally, it’s “how you play the game”, not “how you influence others to play the game.”) I don’t think “he wasn’t modeling it” is credible. Is your claim then that “he wasn’t rewarding it enough”? What would rewarding it look like?
Thanks for your continued patience.
On reflection, no. Thanks for pointing that out. Swap out “perfectly” for “adequately”, and I stand by the rest of the comment.
re: relevance of the Royal Society
Well, it’s not a double crux. Maybe it’s a crux for you.
The reason it’s not a crux for me is because it just seems so remote from the matter at hand. I already have years of direct personal experience productively collaborating with Achmiz on the kind of intellectual work that I want to do. His value to me is not in doubt. It’s just really hard to see what I could possibly learn about 17th-century England that would make me change my mind about that!
It seems like the argument would have to go through Achmiz’s mere presence on the website discouraging so much contribution—despite the existence of the user-level ban feature!—that somehow I’m better off privately benefitting from Achmiz’s counsel via email while he’s officially disgraced and banned from the website which is the central gathering place for the kind of work I do.
I suppose it’s not logically impossible, but it would be super weird for the empirics to shake out that way, and empirical evidence from 17th century England just seems too distant in time and culture to move the needle. The Royal Society isn’t totally irrelevant, of course: human nature hasn’t changed; they were our honorable ancestors doing the natural philosophy thing, and we’re doing a natural philosophy thing.
But our founding texts (which we agree, per the footnote on your 30 May comment, that the current project is running off the fumes of) are very explicit about aiming to do better than our ancestors. The Sequences are exhorting the reader to follow the timeless ideal of Bayesian reasoning wherever it leads, not imitate how Robert Boyle won friends and influenced people in 17th-century England. (And my discourse-norms posts are likewise trying to appeal to the timeless ideal.)
re: Less Wrong 1.0
I do think this is at least more relevant than the Royal Society, just because there’s less of a generalization gap to cross between “running a website in 2016” to “running the same website in 2026.”
I think my honest answer here has to be No, I don’t have an explanation—in the sense that I don’t think I can outperform the replacement-level chatbot answer to this question.
What I think I can say (and I think you’d agree with the chatbot on these) is that the decline of the original website was not monocausal, and that the differences with LessWrong 2.0 are multidimensional. Things like LessWrong 2.0 having dedicated staff (instead of charity hours from Tricycle) and a less notoriously cursed codebase (in which to, e.g., implement moderation tools that can deal with the Eugene_Nier problem) are factors in LessWrong 2.0′s success.
As a result, when choosing interventions to make sure LessWrong 2.0 doesn’t die, it’s not a binary choice of “intervene or don’t”; there’s a high dimensional space of potential things to push on. I don’t find Habryka’s claim that “the specific way Said has been commenting on the site had a non-trivial chance of basically just killing the site” to be credible—particularly in light of my investigation of alleged author complaints in §III.2. To the extent that authors being discouraged by criticism was a problem, there were other available levers to pull that would save the website with less damage to the mission, such as encouraging use of the user-ban functionality, as I argue for in §IV.1 and suggested in July 2025. When I see Habryka reaching for the lever of purging someone he clearly personally dislikes when less intrusive measures were available and untried, I think I’m justified in judging this as a betrayal of the mission rather than a sincere disagreement about how to pursue the mission (forced by the need to avoid repeating the fate of the original site). If the claim isn’t that the Achmiz ban was necessary to literally save the site, but merely to optimize levels of user engagement, that’s not really substantively disagreeing with my model: I’m saying that you’ve betrayed the mission for the sake of popularity.
I wouldn’t go to “there wouldn’t have been the decline” (because of the high-dimensional multi-causuality), but yes, I do think that factor would have helped, and that Yudkowsky has more generally relinquished his Art and lost his powers. Granted, it’s not obvious how to intervene on a leader’s personal virtue (as contrasted to followers, who can often be bullied as Vassar did to me in 2008), but I think giving up on the idea of personal virtue is much worse!
re: people also being similar
Right. That’s why I agree with 2013-Vaniver that “responding appropriately to criticism” is a trainable skill. I don’t think I’m an “emotionally tall” mutant preaching an impossible standard. I think other people could learn it, too, if the culture weren’t saturated with anti-epistemology claiming that they shouldn’t need to.
conflict vs. mistake theories of the present discussion
Are you? In your 30 May comment, you mocked me for spending too many words explaining why prioritizing hedonic and reputational effects trades off against error-correction and therefore correctness, “as the alert reader might have anticipated immediately”. But if it’s supposed to be so obvious that the mod team isn’t maximizing correctness, why would I believe you when you claim to be trying to understand what is going on, if I think that understanding what is going on would almost certainly have negative hedonic and reputational effects (because it would be a weird coincidence if the story that was true also made all the local power players look good)? Maybe you’re trying to understand in the privacy of your own mind, while carefully avoiding any speech acts that would deal unacceptable reputational damage? But the difference isn’t decision-relevant to me; I don’t get to interact with your private intentions in their purity; all I get to see is your speech acts, which strike me as evasive.
(Sorry, I know that was a super rude thing to say, but given what you told me about not putting correctness above all other concerns, you should have expected that I would entertain the hypothesis that you were telling the truth about that and consider the implications.)
I think Ben Hoffman’s theory about “a policy of unprincipled and unbounded submission to threats by the people (or personas) whose feelings are supposed to matter” seems like a better fit to the behavior I’m seeing from you and Anna Salamon. You report that you didn’t have “a personal problem with Said or his comments”, that you “generally found them easy to read and easy to respond to.” Salamon reports that her “personal experiences of Said on LW were clearly and substantially net-positive.” And yet both of you are arguing in favor of the ban to me—not because either of you are willing to testify in your own voice that you think Said’s comments are so bad that you’re not willing to share the website with him, but because other people—largely other unspecified people, given my investigation in §III.2!—aren’t willing to share the website with him. I think that’s weird! Naïvely, if I personally think something is fine, I should go on the record opposing banning it (for the record, even if I’m not the boss and it’s not my decision).
I need something like Hoffman’s theory in order to explain the behavior I’m seeing in you. Importantly, it’s a conflict theory, not a mistake theory: a story about different agents with conflicting goals, rather than everyone and the God-Empress sharing the same utility function. What you consider reputational and hedonic “costs” (to be minimized by the God-Empress for the good of all), Hoffman and I are construing as “threats” (in which an agent claims that they’ll deal disutility to everyone else if their demands aren’t met).
I think if you were actually trying to understand what was going on rather than trying to rationalize a policy handed to you by upstream social forces, you would show more evidence of having thought about the impacts on the site’s mission if the moderators took the other side of the conflict and stood up to the threats by telling complainants to downvote and move on with their lives.
In previous discussion, you’ve told me that the mod team isn’t going to tell users to “get gud” and that trying to shame you into stopping isn’t going to work. I understand that my complaints aren’t going to work. The ban isn’t going to be reversed; the function of this discussion is to get a shared accounting of what happened, where I’m not currently accepting the disvalue attributed to Achmiz as a legitimate cost on the shared ledger. You’ve told me that you felt more optimistic about communication of the form, “Okay, what is the thing that matters here, and how do we get it?” and not “the mods care about Something that is not Logic.” But the problem with that communication prescription is that it presupposes a mistake theory which you haven’t given me reason to believe in!
Do you seriously think that the mods telling complaintants “downvote on move on with your lives” would kill the site? (To be absolutely clear, I’m asking for a conditional prediction, not a policy change; I understand that the mods aren’t going to do that; I’m asking what you think would happen if you did that.)
I would expect some marginal users to bounce, but I don’t think the effect would be quantitatively large, and I think the resulting culture would be much more aligned to the mission. I think most people would grow, the way Michael Vassar helped me grow in 2008.
I could imagine you disagreeing with me on this as an empirical matter of human psychology, but there’s a missing mood in your behavior given that you’ve already agreed that a pro-criticism, pro-being-questioned emotional orientation is good for individuals and good for Society. If it turns out that I’m empirically wrong to think that it’s possible for people to learn the thing we both agree is prosocial, then my reaction is not, “I guess my approach was impractical, therefore wrong, and the new culture is a better way”, but rather, “oh God, we’re screwed, apparently almost no humans are capable of rationality; we were already dead; we were never going to survive.”
Previously, you’ve told me that you think Less Wrong has changed and that Habryka is the person most trying to surf those changes. I think the difference between you and me is that when civilization has fallen to a zombie apocalypse, I think it’s more dignified to go down fighting rather than joining the zombies because they’re the winning side (because surfing the changes means siding with the winners).
Again, I realize that that’s not a nice thing for me to say about you, but I think you’ve been commendably clear about your position, and I think it makes sense for me to be clear, too.
re: encouraging sportsmanship
Thanks. That answers my question of how else is one supposed to do it; now that you point it out, I agree that rewarding and modeling could also work.
Conspicuously, that doesn’t answer my question about what textual evidence supports Achmiz doing more to push people away from the right sort of sportsmanship, such as would justify your “not true on net” judgement. I think that after having conceded, as you have, that the pro-criticism orientation is prosocial, the “obvious” conclusion is that the Achmiz ban was unjustified; in that context, the “not true on net” judgement, stated without evidence, looks suspiciously like an ad hoc goalpost-moving excuse. (It’s an odd notion of sportmanship that demands that players not only exhibit good sportsmanship themselves, but also do so in a way that causes others to as well. Traditionally, it’s “how you play the game”, not “how you influence others to play the game.”) I don’t think “he wasn’t modeling it” is credible. Is your claim then that “he wasn’t rewarding it enough”? What would rewarding it look like?