I mean, I really tried to explain a lot of my models for what I think the underlying generators of this are. That’s why the post is 15,000 words long.
You spend a lot of words trying very hard to explain a thing that is not the same thing that I wanted to know. Perhaps lots of other people wanted to know it? I can only speak for myself.
To be clear, LessWrong is not a democracy
Yeah, I know. I provided an example of a way you could have chosen to openly run it as an oligarchy (“anyone the entire mod team is sick of for any reason is banned because we have no one to moderate them”) and that I would have respected. Let us call this proposed oligarchical model something fun like “Modularity”—once no mod(ule)s are compatible with somebody, that somebody can’t be on the site. You are doing a thing other than that with more moving parts.
what I think is going wrong in conversations with Said
These moving parts in particular. If you were implementing Modularity, it wouldn’t matter what was going wrong. Your bare word is more than sufficient to convince me that something is going wrong, and “something is going wrong” would be enough for you to refuse responsibility for further dealing with Said where Things Go Wrong All The Time for whatever reason. No troubleshooting burden would exist, no explanatory burden would be called for. You could just not like his face, and get the same result. But you’re not doing Modularity. You’re doing something where you write thousands of words about why you think his face is objectively bad.
Thank you for the numbers!
I did really try to explain
I know! I can tell you tried. You did not communicate all the things I wanted to know, though you have here ameliorated that somewhat; I am agnostic about whether this is because you were originally trying to communicate some other thing (perhaps to some other audience) or because communication is just hard and even trying to communicate the correct thing sometimes does not work.
Maybe that’s where we disagree and you think I am doing something that is actively bad by trying to elaborate on my models here, instead of just owning up to this being something closer to a personal preference.
No, not particularly. You’re not implementing Modularity, you’re doing this complicated model-backed thing and you want to explain your complicated model-backed thing. I think that if you were doing Modularity that would be a) respectable and b) require very little digital ink-spilling, but you are not in fact doing either the policy I made up nor any policy closely related to it, so you might as well explain what you are doing.
Your OP comment to me reads as clearly upset and implying that I’ve done something worthy of harsh social judgement
Yeah, you reading additional content into text as though it were clear is a theme here. I said in words that I was disappointed and dismayed. If you are ascribing more emotions to me than those ones, kindly cut it out. I wouldn’t have bothered commenting at all if you hadn’t expressly announced you wanted to hear from the peanut gallery on this one, I’m not making any claims about what social judgment you deserve.
It doesn’t sound to me like you are actually saying that I don’t have to answer these questions.
You literally do not have to; the power of Alicorn might completely fail to compel you and then the consequences of this would be nil. Is there some passphrase which communicates a question and also acknowledges that nothing bad happens to you if you don’t answer them or are we operating under a guess culture so extreme that this is impossible? I mean, if you ignored me I might be sorta irked. Maybe I would make sarcastic remarks about it with my pals. I’m not gonna try to get you fired or anything.
maybe … you think the bad thing that happened is me making bad arguments for banning Said, which is much worse than no arguments for banning Said
No; given that you are doing an argument-driven decisionmaking process, providing the arguments is the right call. I just brought up a process that would be respectable yet not argument-driven.
Let me know if so, if not, I don’t think your top-level comment feels compatible with your assertion here that I don’t actually have a burden of proof here
I’m not sure what you mean by this but perhaps some other thing I said will happen to clarify something usefully.
Cool, I think this clarified a bunch. Summarizing roughly where I think you are at:
In moderation space, there is one way to run things that feels pretty straightforward to you, which you here for convenience called “modularity”, where you treat moderation as a pragmatic thing for which “I don’t have the resources to deal with this kind of person” without much explanation or elaboration is par for the course. You are both confused, and at least somewhat concerned about what I am trying to do in the OP, which is clearly not that thing.
There are at least two dimensions on which you feel concerned/confused about what is going on:
What is the thing I am trying to do here? Am I trying to make some universally compelling argument for badness? Am I trying to rally up a mob to hate on Said so that I can maintain legitimacy? Am I trying to do some complicated legal system thing with precedent and laws?
What do I actually think is going wrong in conversations with Said? Like, where are the supposed terrible comments of his, or things that feel like they are supposed to be compelling to someone like you by whatever standard by system of moderation is trying to achieve? There are a lot of words in the OP, but they feel like they aren’t really addressing this (in part because it’s not super clear to you what standard the explanations are aiming for establishing).
Probably I got various pieces of affect wrong, but this is currently my best guess of where you are at in this thread.
I think this is a reasonable perspective, so I’ll respond to it, assuming that it’s reasonably accurate, though let me know if I got something important wrong.
What is the thing I am trying to do here?
I agree with you that something like the “modularity” approach is a reasonable baseline. And indeed, I do think a huge fraction of this announcement should be seen as “look, it’s been too long, I’ve spent a lot of energy on this, I am not dealing with this anymore, let’s part ways”, and I would, in many circumstances consider that a sufficient thing for a moderator of a forum like LessWrong to say.
But I think there are a few reasons why I am not satisfied with that in this circumstance:
LessWrong is less straightforwardly my personal fiefdom to run than most online forums. I did not found this place, Eliezer did. I have inherited what I consider a really important cultural institution, and in that context, I feel a responsibility to justify what I do with it with something more like good universal principles. Realistically as moderators I will need me and my team to have some freedom to occasionally just say “look, I can’t deal with this person”, but IMO at a larger scale, I think the right to run LessWrong should be earned.
Beyond that, the thing I am trying to do with LessWrong is aiming bigger, and aiming to do something somewhat more robust than to “just” build a functional community. The Rationality community is one of the biggest online communities in the world, and has been surprisingly impactful on the trajectory of civilization[1]. I think something closer to courts and debates on the nature of justice and principles and checks and balances is appropriate for something that I want society to be able to put weight on (for example, LessWrong is by far the most active discussion forum for AI Alignment, and within those confines occupies an important social role).
And then there is also a third thing, which makes me want to be particularly detailed and concrete and give lots of arguments and models here, which is more relevant for Said in-particular. And that thing is the feeling of being insane, of at the same time really feeling like something hurtful and harmful is happening, while tons of people around you are denying that any such thing is going on, is at the core of a lot of the complaints I have received over the years about Said, and is also at the core of my own bad experiences with Said. More specifically the thing where I read a top-level that I cannot help but read in a sneering voice, dripping with judgement, pointing a finger at me or the author in a way that summons judgement and punishment, but which as soon as its called out disappears, denies it ever existed, or keeps slipping away, redefining itself in endless circles.
And I don’t know, maybe any attempt to disentangle the things that are going on and to try to make some kind of compelling demonstration that I and others are not insane is doomed. But I currently think it’s a valuable service, and something I owe to a lot of the people who had bad experiences on LessWrong over the years with Said, and something for which I myself would benefit from recognition and understanding. I also think it’s the cause of the death of many many institutions in the world at large, and while I think LessWrong could survive for a while on the moderators just appealing to “look, I just can’t deal with this guy”, my best guess is we would lose some social legitimacy each time, and this would substantially limit what LessWrong can achieve over the long run.
What do I actually think is going wrong in conversations with Said?
So now maybe let’s get a bit more into the object-level. I do think I have pointed to the biggest component of what I think is going wrong in the previous two paragraphs.
My model of you, based both on your last reply, and other things I know about you, is currently not very compelled, and probably thinks I am chasing shadows or something like that. That no, Said is indeed really actually not intending to sneer at people all the time, is not intending to summon up lots of social judgement, does not carry emnity in his heart, and is just writing comments, usually with disagreement, sometimes with approval. Maybe he feels some things while writing, sometimes those feelings are negative, but it’s not that there is some overarching complicated optimization process that tries to get most authors on LessWrong to leave and stop doing whatever they are doing. They are just comments, trying to point out flaws and correct the record. For reasons of local validity.
But man, I do just think that’s false, after almost a decade of thinking about it on and off, being in dozens of comment threads with Said, and spending many hundreds of hours on it. I link to many of the relevant posts where I think something more like “Said is trying to unilaterally enforce social norms of his own choosing while denying any such thing is going on, by making lots of comments that imply the people he doesn’t like are idiots, or are making elementary mistakes, or are being deceptive, or lack common sense, but shift and twist when pushed on” is going on.
I write about the pieces of this a lot in the post above. It IMO comes through quite a lot in comments like this:
Of course people have such preferences! Indeed, it’s not shocking at all! People prefer not to have their bad ideas challenged, they prefer not to have obvious gaps in their reasoning pointed out, they prefer that people treat all of their utterances as deserving of nothing less than “curious”, “kind”, “collaborative” replies (rather than pointed questions, direct and un-veiled criticism, and a general “trial by fire”, “explore it by trying to break it” approach)?! Well… yeah. Duh. Humans are human. No one is shocked.
or this:
This is only because most people don’t bother to ask (what I take to be) such obvious, and necessary, clarifying questions. (Incidentally, I take this fact to be a quite damning indictment of the epistemic norms of most of Less Wrong’s participants.) When I ask such questions, it is because no one else is doing it. I would be happy to see others do it in my stead.
Or:
I am, of course, ambivalent about harshly criticizing a post which is so laudatory toward me.[1] Nevertheless, I must say that, judging by the standards according to which LessWrong posts are (or, at any rate, ought to be) judged, this post is not a very good one.
Like, I think you can see the disdain in these comments. The disdain for almost everyone on LessWrong. Definitely disdain for me, and many others. And I think I could handle disdain fine, if it was carried openly, and could be argued with. But I don’t know how to argue with it. I don’t know how to take it as an object. Whenever I try to point at it, I get comments like this:[2]
I don’t make claims about my interlocutors being stupid or evil or any such thing (heck, I generally don’t even call people’s ideas “stupid”, or anything similar). And you also know that I’ve said quite explicitly that I don’t “hate” anyone here, or really have any strong feelings about any particular person on Less Wrong. So why read such negative valence into my comments? I don’t see any good reason to do so…
And I at least don’t know how to deal with it. I find the experience of having this kind of dynamic present on LessWrong extremely costly. I know many authors have felt similarly. Maybe you wouldn’t find it costly. I currently don’t believe that you wouldn’t, and instead believe that you too would choose to ban Said from forums that you ran, after you ran into it a few times yourself, probably much faster than I did.
But maybe I am wrong. While I feel some temptation to try to restore some of my sanity by providing compelling demonstrations of what has felt like gaslighting to me, it is not the primary thing I am hoping to do with this post. The key thing is to have any clear moderation announcement at all, and to make it easy for other people to form their own judgement of how good of a job we are doing with moderation, and to explain some of the principles that will guide future moderation decisions, for the reasons I listed in the second part of this comment.
IDK, that’s a lot more words. Maybe they help communicate something. Maybe they again fail. In general thank you for many of your contributions to LessWrong and the community over the years.
Please forgive me not linking to all of these. I don’t have all the links handy, and searching for the text of the comment should be relatively fast if someone wants to look up the underlying comment.
I think your model of me as represented in this comment is pretty good and not worth further refining in detail.
I read something into those comments—I might even possibly call it “disdain”, but—“disdain (neutral)”, not “disdain (derogatory)”. It just… doesn’t bother me, that he writes in a way that communicates that feeling. It certainly bothers me less than when (for example) Eliezer Yudkowsky communicates disdain, purely as a stylistic matter. If I thought Said would want to be on my Discord server I would invite him and expect this to be fine. (Eliezer is on my Discord server, which is also usually fine.)
It bothers you. I’m not trying to argue you out of being bothered. I’m not trying to argue the complainants out of being bothered. It bothering you would, under the Modularity regime, be sufficient.
But you’re not doing that. You’re trying to make the case that you are objectively right to feel that way, that you have succeeded at a Sense Motive check to detect a pattern of emotions and intentions that are really there. I don’t agree with you, about that.
But I don’t have to. I don’t have your job. (I wouldn’t want it.)
But you’re not doing that. You’re trying to make the case that you are objectively right to feel that way, that you have succeeded at a Sense Motive check to detect a pattern of emotions and intentions that are really there. I don’t agree with you, about that.
I think the claim I’d make is not necessarily that Oli’s Sense Motive check has succeeded, but that Oli’s Sense Motive check correlates much better with other people’s Sense Motive checks than yours does, and that ultimately that’s what ends up mattering for the effects on discourse.
Like, in the sense that someone’s motives approximately only affect LessWrong by affecting the words that they write. So when we know the words they write, knowing their motives doesn’t give us any more information about how they’re going to affect LessWrong. For some people, there’s something like… “okay, if this person actually felt disdain then the words they write in future are likely to be _, and if not they’re likely to be _ instead; and we can probably even shift the distribution if we ask them hey we detect disdain from your comment, is that intended?”. But we don’t really have that uncertainty with Said. We know how he’s going to write, whether he feels disdain or not.
I am somewhat interested in his True Motives, but I don’t think they should be relevant to LW moderation.
(This is not intended to say “Said’s comments are just fine except that people detect disdain”.)
But you’re not doing that. You’re trying to make the case that you are objectively right to feel that way, that you have succeeded at a Sense Motive check to detect a pattern of emotions and intentions that are really there. I don’t agree with you, about that.
Makes sense. I think I probably could, with many more hours of examples and walking you through things, convince you of that. Maybe that’s worth it. Or maybe I’ll be a better writer in a few years and can get it across more easily. (Of course, you disagree, for if you did agree with that, you would probably agree with me now, conservation of expected evidence and all that)
Not planning to give it another try for now, though if you want me to try, I would do it. Just doesn’t seem, on the margin, the best use of either of our time.
You spend a lot of words trying very hard to explain a thing that is not the same thing that I wanted to know. Perhaps lots of other people wanted to know it? I can only speak for myself.
Yeah, I know. I provided an example of a way you could have chosen to openly run it as an oligarchy (“anyone the entire mod team is sick of for any reason is banned because we have no one to moderate them”) and that I would have respected. Let us call this proposed oligarchical model something fun like “Modularity”—once no mod(ule)s are compatible with somebody, that somebody can’t be on the site. You are doing a thing other than that with more moving parts.
These moving parts in particular. If you were implementing Modularity, it wouldn’t matter what was going wrong. Your bare word is more than sufficient to convince me that something is going wrong, and “something is going wrong” would be enough for you to refuse responsibility for further dealing with Said where Things Go Wrong All The Time for whatever reason. No troubleshooting burden would exist, no explanatory burden would be called for. You could just not like his face, and get the same result. But you’re not doing Modularity. You’re doing something where you write thousands of words about why you think his face is objectively bad.
Thank you for the numbers!
I know! I can tell you tried. You did not communicate all the things I wanted to know, though you have here ameliorated that somewhat; I am agnostic about whether this is because you were originally trying to communicate some other thing (perhaps to some other audience) or because communication is just hard and even trying to communicate the correct thing sometimes does not work.
No, not particularly. You’re not implementing Modularity, you’re doing this complicated model-backed thing and you want to explain your complicated model-backed thing. I think that if you were doing Modularity that would be a) respectable and b) require very little digital ink-spilling, but you are not in fact doing either the policy I made up nor any policy closely related to it, so you might as well explain what you are doing.
Yeah, you reading additional content into text as though it were clear is a theme here. I said in words that I was disappointed and dismayed. If you are ascribing more emotions to me than those ones, kindly cut it out. I wouldn’t have bothered commenting at all if you hadn’t expressly announced you wanted to hear from the peanut gallery on this one, I’m not making any claims about what social judgment you deserve.
You literally do not have to; the power of Alicorn might completely fail to compel you and then the consequences of this would be nil. Is there some passphrase which communicates a question and also acknowledges that nothing bad happens to you if you don’t answer them or are we operating under a guess culture so extreme that this is impossible? I mean, if you ignored me I might be sorta irked. Maybe I would make sarcastic remarks about it with my pals. I’m not gonna try to get you fired or anything.
No; given that you are doing an argument-driven decisionmaking process, providing the arguments is the right call. I just brought up a process that would be respectable yet not argument-driven.
I’m not sure what you mean by this but perhaps some other thing I said will happen to clarify something usefully.
Cool, I think this clarified a bunch. Summarizing roughly where I think you are at:
In moderation space, there is one way to run things that feels pretty straightforward to you, which you here for convenience called “modularity”, where you treat moderation as a pragmatic thing for which “I don’t have the resources to deal with this kind of person” without much explanation or elaboration is par for the course. You are both confused, and at least somewhat concerned about what I am trying to do in the OP, which is clearly not that thing.
There are at least two dimensions on which you feel concerned/confused about what is going on:
What is the thing I am trying to do here? Am I trying to make some universally compelling argument for badness? Am I trying to rally up a mob to hate on Said so that I can maintain legitimacy? Am I trying to do some complicated legal system thing with precedent and laws?
What do I actually think is going wrong in conversations with Said? Like, where are the supposed terrible comments of his, or things that feel like they are supposed to be compelling to someone like you by whatever standard by system of moderation is trying to achieve? There are a lot of words in the OP, but they feel like they aren’t really addressing this (in part because it’s not super clear to you what standard the explanations are aiming for establishing).
Probably I got various pieces of affect wrong, but this is currently my best guess of where you are at in this thread.
I think this is a reasonable perspective, so I’ll respond to it, assuming that it’s reasonably accurate, though let me know if I got something important wrong.
What is the thing I am trying to do here?
I agree with you that something like the “modularity” approach is a reasonable baseline. And indeed, I do think a huge fraction of this announcement should be seen as “look, it’s been too long, I’ve spent a lot of energy on this, I am not dealing with this anymore, let’s part ways”, and I would, in many circumstances consider that a sufficient thing for a moderator of a forum like LessWrong to say.
But I think there are a few reasons why I am not satisfied with that in this circumstance:
LessWrong is less straightforwardly my personal fiefdom to run than most online forums. I did not found this place, Eliezer did. I have inherited what I consider a really important cultural institution, and in that context, I feel a responsibility to justify what I do with it with something more like good universal principles. Realistically as moderators I will need me and my team to have some freedom to occasionally just say “look, I can’t deal with this person”, but IMO at a larger scale, I think the right to run LessWrong should be earned.
Beyond that, the thing I am trying to do with LessWrong is aiming bigger, and aiming to do something somewhat more robust than to “just” build a functional community. The Rationality community is one of the biggest online communities in the world, and has been surprisingly impactful on the trajectory of civilization[1]. I think something closer to courts and debates on the nature of justice and principles and checks and balances is appropriate for something that I want society to be able to put weight on (for example, LessWrong is by far the most active discussion forum for AI Alignment, and within those confines occupies an important social role).
And then there is also a third thing, which makes me want to be particularly detailed and concrete and give lots of arguments and models here, which is more relevant for Said in-particular. And that thing is the feeling of being insane, of at the same time really feeling like something hurtful and harmful is happening, while tons of people around you are denying that any such thing is going on, is at the core of a lot of the complaints I have received over the years about Said, and is also at the core of my own bad experiences with Said. More specifically the thing where I read a top-level that I cannot help but read in a sneering voice, dripping with judgement, pointing a finger at me or the author in a way that summons judgement and punishment, but which as soon as its called out disappears, denies it ever existed, or keeps slipping away, redefining itself in endless circles.
And I don’t know, maybe any attempt to disentangle the things that are going on and to try to make some kind of compelling demonstration that I and others are not insane is doomed. But I currently think it’s a valuable service, and something I owe to a lot of the people who had bad experiences on LessWrong over the years with Said, and something for which I myself would benefit from recognition and understanding. I also think it’s the cause of the death of many many institutions in the world at large, and while I think LessWrong could survive for a while on the moderators just appealing to “look, I just can’t deal with this guy”, my best guess is we would lose some social legitimacy each time, and this would substantially limit what LessWrong can achieve over the long run.
What do I actually think is going wrong in conversations with Said?
So now maybe let’s get a bit more into the object-level. I do think I have pointed to the biggest component of what I think is going wrong in the previous two paragraphs.
My model of you, based both on your last reply, and other things I know about you, is currently not very compelled, and probably thinks I am chasing shadows or something like that. That no, Said is indeed really actually not intending to sneer at people all the time, is not intending to summon up lots of social judgement, does not carry emnity in his heart, and is just writing comments, usually with disagreement, sometimes with approval. Maybe he feels some things while writing, sometimes those feelings are negative, but it’s not that there is some overarching complicated optimization process that tries to get most authors on LessWrong to leave and stop doing whatever they are doing. They are just comments, trying to point out flaws and correct the record. For reasons of local validity.
But man, I do just think that’s false, after almost a decade of thinking about it on and off, being in dozens of comment threads with Said, and spending many hundreds of hours on it. I link to many of the relevant posts where I think something more like “Said is trying to unilaterally enforce social norms of his own choosing while denying any such thing is going on, by making lots of comments that imply the people he doesn’t like are idiots, or are making elementary mistakes, or are being deceptive, or lack common sense, but shift and twist when pushed on” is going on.
I write about the pieces of this a lot in the post above. It IMO comes through quite a lot in comments like this:
or this:
Or:
Like, I think you can see the disdain in these comments. The disdain for almost everyone on LessWrong. Definitely disdain for me, and many others. And I think I could handle disdain fine, if it was carried openly, and could be argued with. But I don’t know how to argue with it. I don’t know how to take it as an object. Whenever I try to point at it, I get comments like this:[2]
And I at least don’t know how to deal with it. I find the experience of having this kind of dynamic present on LessWrong extremely costly. I know many authors have felt similarly. Maybe you wouldn’t find it costly. I currently don’t believe that you wouldn’t, and instead believe that you too would choose to ban Said from forums that you ran, after you ran into it a few times yourself, probably much faster than I did.
But maybe I am wrong. While I feel some temptation to try to restore some of my sanity by providing compelling demonstrations of what has felt like gaslighting to me, it is not the primary thing I am hoping to do with this post. The key thing is to have any clear moderation announcement at all, and to make it easy for other people to form their own judgement of how good of a job we are doing with moderation, and to explain some of the principles that will guide future moderation decisions, for the reasons I listed in the second part of this comment.
IDK, that’s a lot more words. Maybe they help communicate something. Maybe they again fail. In general thank you for many of your contributions to LessWrong and the community over the years.
With unclear sign, unfortunately, but like, we IMO are having a large effect about how the whole AGI thing is going in some way
Please forgive me not linking to all of these. I don’t have all the links handy, and searching for the text of the comment should be relatively fast if someone wants to look up the underlying comment.
I think your model of me as represented in this comment is pretty good and not worth further refining in detail.
I read something into those comments—I might even possibly call it “disdain”, but—“disdain (neutral)”, not “disdain (derogatory)”. It just… doesn’t bother me, that he writes in a way that communicates that feeling. It certainly bothers me less than when (for example) Eliezer Yudkowsky communicates disdain, purely as a stylistic matter. If I thought Said would want to be on my Discord server I would invite him and expect this to be fine. (Eliezer is on my Discord server, which is also usually fine.)
It bothers you. I’m not trying to argue you out of being bothered. I’m not trying to argue the complainants out of being bothered. It bothering you would, under the Modularity regime, be sufficient.
But you’re not doing that. You’re trying to make the case that you are objectively right to feel that way, that you have succeeded at a Sense Motive check to detect a pattern of emotions and intentions that are really there. I don’t agree with you, about that.
But I don’t have to. I don’t have your job. (I wouldn’t want it.)
I think the claim I’d make is not necessarily that Oli’s Sense Motive check has succeeded, but that Oli’s Sense Motive check correlates much better with other people’s Sense Motive checks than yours does, and that ultimately that’s what ends up mattering for the effects on discourse.
Like, in the sense that someone’s motives approximately only affect LessWrong by affecting the words that they write. So when we know the words they write, knowing their motives doesn’t give us any more information about how they’re going to affect LessWrong. For some people, there’s something like… “okay, if this person actually felt disdain then the words they write in future are likely to be _, and if not they’re likely to be _ instead; and we can probably even shift the distribution if we ask them hey we detect disdain from your comment, is that intended?”. But we don’t really have that uncertainty with Said. We know how he’s going to write, whether he feels disdain or not.
I am somewhat interested in his True Motives, but I don’t think they should be relevant to LW moderation.
(This is not intended to say “Said’s comments are just fine except that people detect disdain”.)
Makes sense. I think I probably could, with many more hours of examples and walking you through things, convince you of that. Maybe that’s worth it. Or maybe I’ll be a better writer in a few years and can get it across more easily. (Of course, you disagree, for if you did agree with that, you would probably agree with me now, conservation of expected evidence and all that)
Not planning to give it another try for now, though if you want me to try, I would do it. Just doesn’t seem, on the margin, the best use of either of our time.