16 Jan: Ben to me/Anna on old thread: esotericism is less of a problem if flagged
18 Jan: gender refugees, moral patiency, and BBL
29 Jan: me to Kay Brown: have you noticed LGBTQ/queer culture getting worse?
+7 Feb: Messenger with "Wilhelm" about Rust All-Hands
8 Feb: me to Katie about Rust All-Hands
8 Feb: me to Ben and Michael about social competence
20 Feb: impulsive email to Scott (subject: a howl of despair which is nameless)
3 Mar: more draft, more thanks for social proof
5 Mar: Scott gets back to me
5 Mar: 12 short stories about language (and, expressing regret that I only email Scott when I need something from him)
+6 Mar: to "Wilhelm" about anti-trans complaints against Anna
6 Mar: SJWs without guns doing this than you are about mostly white men with money, gavels and guns doing it? ... SJW agenda is to play by the same rules as the legal system plays by, and they do it with less support from tradition and from ritual, so it's more obvious,
6 Mar: much better draft to EY (defending against alt-right)
10 Mar: send revised draft to EY
15 Apr: me—concern about bad faith nitpicking, Jessica says I'm not
15 Apr: Jessica's comment, the lesson of Funk-tunul
15 Apr: trying to recruit Steven Kaas
+15 Apr: Signal conversation with Ben—
+I'm having some ideas on how I could write my post about disillusionment with the rationalists well still not being too hostile and not leaking info from private conversations
+
+> Yeah, I think the right target to keep in mind is people like you a few years ago who haven't yet had the experiences you've had
+> So they don't actually have to freak out to the point of being imprisoned and then demand help from community leaders and not get it, they can just learn from you
+> Like, the actual sympathetic-but-naive people
+> Not the people fucking with you
+
+[...]
+
+people also deserve credit for the help they did give me (care resources from Alicorn et al.) even if I don't think there intellectual performance was good enough to deserve the Mandate of Heaven implied by the "rationalist" brand name
+
+> Yes, but we need to keep separate accounts for _personal_ gratitude, and institutional fraud
+> Like, I think that if somehow Alicorn gets hurt by the Rationalists dissolving, and we're in a position to help her, we should try to help in any way that _doesn't_ work via continuing to validate the collective narcissistic fantasy that the Rationalists are a thing
+> Insofar as that makes things hard, we have to think of it as a hostage situations
+> But, it's not obvious that this imposes undue costs on the individuals that have helped you a lot, except maybe Anna
+[...]
+> And, _not_ doing this is imposing costs on people like Michael (and all of humanity) instead, which is a lot less fair than harming Anna's interests incidentally by breaking up a scam she's drawing power from
+
+[...]
+should I just reread you posts about EA, or is there more?
+
+> People are systematically conflating corruption, accumulation of dominance, and theft, with getting things done.
+[...]
+
+> These both seem like evidence that CFAR and MIRI see themselves as building an army, not a truth-seeking intellectual community.
+[...]
+
+> This isn't about direct benefit vs harm
+> This is about what, substantively, they're doing
+> And they claimed to be cultivating an epistemically rational community while in fact building an army of loyalists
+
+[...]
+
+last night I was up until 2 a.m. and wrote Anna two paper postscards asking for her to engage with my "... Boundaries?" post when she ahs time and to give me an estimate of when she might have time [...] I figured that spamming people with hysterical and somewhat demanding physical postcards is more polite (and funnier) than my usual habit of spamming people with hysterical and somewhat demanding emails
+
16 Apr: I feel sad and gaslighted
16 Apr: Jessica sparks engagement by appealing to anti-lying
Ben— I think a commitment to lying is always wrong just straightforwardly leads to deciding that anything power demands you say is true.
-16 Apr: If it were just a matter of my own pain, I wouldn't bother making a fuss about this. But there's a principle here that's way bigger than any of us: that's why Michael and Ben and Sarah and Zvi and Jessica are here with me.
+16 Apr: If it were just a matter of my own pain, I wouldn't bother making a fuss about this. But there's a principle here that's way bigger than any of us: that's why Michael and Ben and Sarah and [...] and Jessica are here with me.
we're in favor of Jessica's existence
17 Apr: plantiff/pawn/puppydog
17 Apr: Jessica— This person had an emotional reaction described as a sense that "Zack should have known that wouldn't work" (because of the politics involved, not because Zack wasn't right). Those who are savvy in high-corruption equilibria maintain the delusion that high corruption is common knowledge, to justify expropriating those who naively don't play along, by narratizing them as already knowing and therefore intentionally attacking people, rather than being lied to and confused.
18 Jul: my accusation of mis-citing Ozy was wrong
20 Jul: me to Anna and Steven about LW mods colluding to protect feelings; "basically uninterested in the mounting evidence that your entire life's work is a critical failure?"
20 Jul: Michael—Court language is the language that we have for "you don't have the ethical option of non-engagement with the complaints that are being made"
+> We *also* need to develop skill in the use of SJW style blamey language, like the blamey language about feelings that was being used *on* us harder and to a great extent first, while we were acting under mistake theory assumptions.
23 Jul: "epistemic defense" meeting
24-25 Jul: Michael Vassar and the theory of optimal gossip
Kelsey thinks "threatening schism is that it's fundamentally about Vassar threatening to stir shit up until people stop socially excluding him for his bad behavior"
15 Sep: Yudkowsky will say, "Whoops, I said the wrong thing" re his interaction with Glen Weyl, which seems more proportional to Weyl's power rather than the magnitude of the offense https://twitter.com/ESYudkowsky/status/1164275185622507520
16 Sep: SK: "I think there's a gradient where sanity is mostly flowing from core people outward toward non-core people."
me—That's what I would have thought, too! I consider this falsified [...] twenty-thousand-word, five-month extraordinary effort [...] The gradient is running in the other direction!
+
+19/20 Sep: Signal conversation with Ben—
+> Van Nostrand mentioned that Merlin has asked you why you're a boy, an dwe're curious what you're answer was now
+... I stonewalled him because of my political incentives :'( (semi-consciously mishead the question as "Are you ab boy?" and said "Yes" after a long pause)
+I would have preferred a straightforward honest answer like, "Because I have a penis, like you", but that answer felt inappropriately "political", speech I want to avoid in that house (which is an insane situation to be in, but)
+
21 Sep: Being widely taken seriously is important if the plan is to lobby Congress to pass Safe AI regulation. Being able to add 6 + 7 is important if the plan is to make unprecedented scientific breakthroughs. Remind me what the plan is?
22 Sep: There are also forces pulling in the direction of not selling out your principles for power and popularity, if there exist valuable people who care about principles. [..] If my actions (implausibly) represent a PR risk to someone else's Singularity strategy, then they're welcome to try to persuade or negotiate with me.
24 Sep: "Heads I Win" was really successful; reasons why you might think this is good besides measuring idea quality democratically
30 Sep: the third time that someone has responded to my "help help, everyone is being stupid about the philosophy of language for transparently political reasons, and Michael Vassar's gang are the only people backing me up on this; what the fuck is going on?!" sob story with (paraphrasing), "Your philosophy hobbyhorse is OK, but Michael's gang is crazy." [...] Jessica's assessment from earlier: "Another classic political tactic: praise one member of an alliance while blaming another, to split cracks in the alliance, turning former allies against each other." / Where the three incidents seemed more coherent on the "praise Zack, diss his new friends" aspect, than on the specific content of the disses, whereas in the worlds where Michael's gang is just crazy, I would expect the content craziness allegations to correlate more
Oct: model sync with Jessica/Alyssa/Lex/Sarah
+19 Oct: https://www.lesswrong.com/posts/bSmgPNS6MTJsunTzS/maybe-lying-doesn-t-exist
3 Nov: actually good criticism from Abram at MIRI!!!
Isn't the problem with bad (shortsighted, motivated) appeal-to-consequences, rather than appeal-to-consequences in general?
example: predator-avoidance
16 Jan: do you think you could explain this in a way Said would find satisfying?
18 Jan: Yudkowsky on Caplan: https://www.econlib.org/scott-alexander-on-mental-illness-a-belated-reply/#comment-237783
21 Jan: still have this "Vassar Group vs. Corrupt Mainstream MIRI/CfAR/LW/EA rationalist civil war" framing in my head that I need to unwind, because it's the wrong frame, and it's ironically the wrong frame by our ("our") own standards. I already made the mistake of expecting a community once, but by now I should know that no Lord hath the champion, and no safe defense. There is no group; there is no community; there are just grown-ups who think for themselves (or ought to) and sometimes coordinate with their friends on those specific plans that actually require coordination.
-
-
-
-
-
-
-
+I indicate I might enter the Caplan-Alexander debate (I don't actually)
+25 Jan: Said thinks Jessica's political thinking is vague and under-researched
+27 Jan: Anna has to ("has to") let both me and Ziz down because big-tent coalitions require incoherence (e.g., progressives prefer not to admit out loud that trans people and Muslims have conflicting interests).
+27 Jan: Vassar's trans reversal—lame
+1 Feb: sleep disruption due to M.M.B. situation
+1 Feb: I reply to Scott on "AGP is common", and M.M.B. situation
+4 Feb: you exist in the physical sense. To whatever extent your behavior is controlled by the balance of "not wanting to fight pro-trans people" and "not wanting to fight Zack", then you don't exist decision-theoretically.
+6 Feb: me to Anna on fog of war and MMB
+(the conversation with Somni is relevant)
+12 Feb: "I'm a girl and I'm a vegetarian"
+23 Feb: connection with Iceman
+23 Feb: discussion on meaning of "zero-sum"
+25 Feb: I'm still talking about finishing writing a memoir
+28 Feb: "belief check" conversation with Jessica on origin of trans ideology
+3 Mar: Scott's new COVID-19 post is great on the "shared maps" issue! https://slatestarcodex.com/2020/03/02/coronavirus-links-speculation-open-thread/ in contrast to how last month he was acting like he didn't understand the problem https://slatestarcodex.com/2020/01/30/book-review-human-compatible/#comment-847143 https://www.greaterwrong.com/posts/yaCwW8nPQeJknbCgf/free-speech-and-triskaidekaphobic-calculators-a-reply-to/comment/mHrHTvzg8MGNH2CwB
+4 Mar: niche skill of being able to translate between high-church Yudkowskyism and Vassarian moonspeak
+9-10 Mar: talk to Mom/Nicole about COVID prep
+22 Mar: ping Jack for social call ("You seem a lot more vicious than the sixteen-or-whatever-year-old I remember meeting at a Less Wrong meetup back in '13-or-whatever, but it looks adaptive")
+31 Mar: I'm skeptical of Eli as a mediator between me and Anna
+"Human Diversity" review published
+I've discovered that you can just stop being scared of things and then nothing bad happens. All swans are white! There are no other swans!
+24 May: Jessica on high-precision claims
+24 May: Anna seems to be regaining power of speech (Facebook post on U.S. decline)
+2 Jun: I send an email to Cade Metz, who DMed me on Twitter
+25 Jul: rubber-duck philosophy for "Unnatural Categories"!!
+4 Sep: misguided by the hideousness of our weapons?! or, theory of universal algorithmic bad faith
+runaround argument about free speech that never gets anywhere, possibly because of a bug on my end
+in principle, "What if telling the truth about X has bad consequences?" is an empirical question that could be asked in good faith and, having been asked in good faith, deserves a good-faith answer
+in principle, I don't believe people are asking it in good faith; I think it's a weapon of psychological warfare
+our most recent Less Wrong exchange. I characterized you as being on "Team Seek Power For The Greater Good". You explained why that's not a fair characterization. I explained why selectively omitting evidence is a problem. You explained that you only advocate omitting things that will make the audience respond irrationally, not omitting everything that would be inconvenient to the case.
+I don't think that thread was a good showing for me: if we apply the principle of charity to my behavior, we would say I was a bit slow on the uptake, forcing a four-comment thread to restate things that had already been covered previously.
+I'm better modeled as trying to humiliate him or waste his time
+... I still don't understand how to engage with it
+"I'm curious what evidence changed your mind" is still a counter-weapon
+13 Sep: out of patience
+13 Sep: me to Anna—focusing on my email to Yudkowsky is more important than entering the Facebook comments
+Anna asks me to be calm, high in denotation
+14 Sep: the end of the Category War
+suspect Ben will be sad that I'm "bought off", but I think not being at war makes it easier for me to do clarity-creation, in the spirit of "Here's what the thing is actually doing" rather than "flames on the side of my face"
+21 Sep: me to Don Fallis—the argument over whether merely withholding information is deception reminds me of arguments over distortion: it's not surprising that the answer takes the form "it depends on the zero point" rather than "Extoriton: Yes or No"
+25 Sep: anticipating that people are going to (not entirely inaccurately?! :'( ) round off everything I say as being backchained from my gender war, which makes the fact that Ziz totally agrees with us (about the Bay eating our culture and our people) very reassuring https://sinceriously.fyi/cached-answers/#comment-897
+still this urge to prevent the old beacon from eating people who don't know what it's doing
+maybe my razzing him about the Human Compatible review (https://slatestarcodex.com/2020/01/30/book-review-human-compatible/#comment-847143) was causal in his COVID shared maps thing—maybe this entire line of thinking is still poison: thinking about my effects on Scott in particular (as if he were special, which he's not)
+3 Oct: Do the smart-leftists feel contempt for all the retarded children, or do they just not-notice them as being worth paying attention to (the way I seamlessly don't-notice ordinary Trump voters)?
+1 Nov: knowledge boostrapping with EvN
+4 Nov: destructive creature Olympics
+25 Nov: me to Michael—I think your reputation (in the jungle growing around the ruins of what we once called the "rationalist community") has some amount of instrumental value
+25 Nov: "No reputation in the jungle" meaning that people are just going to praise or throw you to the wolves depending on the expediency of the moment's shifting alliances, rather than there being any way to gain or retain standing for being Actually Good?
+25 Nov: notes from conversation with Michael
+in lawful environments, your reputation is by default true
+in lawless environments, bad rep just means you're important enough to have enemies, which is good
+It's easy to move the conversation from what is going on to whose side are you on.
+There's a game of trying to build the largest coalition: previously, it was the European coalition, now the not-European coalition is ascendant. [ZMD's note: this reminds me of both NRx's https://spandrell.com/2017/11/14/biological-leninism/ and Ben's http://benjaminrosshoffman.com/hierarchy-wings/] We want to carve out a space for reason within the winning coalition.
+Reason is the motte and pragmatism is the bailey
+the officers vs. the enlisted; the enlisted claim to recommend the common people, the officers claim to represent the rich; both sides totally fake ideas
+Ziz is that least plausibly a girl of all trans girls, but by declaring a protected identity and having wealth, can just barely survive.
+my objection: but, but as part of the whole "Truth instead of expedient lies" part of rationality, don't we at least want Ziz to know that?!
+ I can operate for short periods in territory in which I only have the option of being a (closeted) trans woman—as an energy-minimizing play, to inhabit someone else's stereotype—but I can't live there, because that story makes no fucking sense
+Trauma is expecting coordination for evil to win, and causes you to be part of that coordination in a way that's minimally ego-dystonic.
+Surprise minimization means doing what you think that that they think that ... a free-energy constaint. If you're given a lot of "responsibility", you have to behave in a way that fulfills expectations, and eventually the only thing left is being bad in a stereotyped way that everyone knows how to expect—the type of conspiracy you can only get out of physics.
+By defying people's stereotypes, Shawanna imposes enormously large costs on cynical do-gooders, so they equalize the energy differential by imposing pragmatic costs on her.
+surpised that Ziz can't get a lawyer despite having money
+14 Dec: destructive creature Olympic, cont'd: It's great to have common knowledge of the mutual ability to interpret conditional statements literally! (Not something you can always depend on these days.)
+if Kelsey wrote her own 10,600 draft Document [...] that would be really interesting! [...] no one else seemed interested in having a theory, as opposed to leaping to institute a social convention that, when challenged, is claimed to have no particular consequences and no particular objective truth conditions, even though it's not clear why there would be moral urgency to implement this convention if it weren't for its consequences.
+I don't think the 19th-century Chinese who bound their daughters' feet, or the 18th-century Italians who castrated little boys to perserve their singing voice were innately less empathetic than us [...] How will history analyze the moral culture of 21st-century Californians?
+17 Dec: But parts of it felt like concern trolling as a cult-recruitment tactic? In caricature: "Your writing and research doesn't matter because there's no living audience out there; you can only talk to us." You can see why Dagny Taggart would perceive that as hostile?
+--- psych care disaster
+21 Dec: me to Anna—you either didn't hide it, or I noticed anyway. But if other people in the care coordination group had seized on that to turn against you and paint you (to me) as dangerous, that would have been bad for me (independently of whether or not it would have been fair to you), because the fact that you were there was vastly more helpful, than the fact that you didn't want to be there was unhelpful
+2 Jan 21: super-weird unsolicited mental/physical health email about blood sugar and sleep
+4 Jan: sleep strategy notes
+7 Jan: I've been distracted by the news (I told Anna that Wednesday was a scary news day—referring, of course, to OpenAI DALL-E
+Feb: New York Times hit piece on Scott
+18 Feb: my "incredibly shallow and transparently self-serving" reaction to Yudkowsky on Scott email leak
+22 Feb: Technically, he started it this time! I totally would have left him alone if he didn't kick the shitpile again!
+dude can't really expect to get away with pulling out the pompous-register "feelings don't get to control everybody's language protocol" in that context; that's _my_ line
+27 Feb: me to Marcus—Everyone who's read Scott's work deeply is completely unsurprised by the content of the leaked email
+28 Feb: "Are you trying to convince him, yourself or those watching?" Those watching, plus testing my acquired immunity to pompous-register Yudsplaining (which I used to psychologically vulnerable to, because in 2009, it was always right)
+28 Mar: block warning
+9 May 21: trapped priors—at home!
+ I think the effort asymmetry here is kind of hilarious, where it's not that hyperbolic to say I spent three years of my life (early 2018–early 2021) trying undo the cultural brain damage from that one post.
+when I talk to AI researchers like Abram Demski or John Wentworth, they get it instantly ...
+7 Jun: dolphin war
+https://twitter.com/zackmdavis/status/1402067999977132035
+9 Jun: I want to use the phrase "all-out nuclear war with MIRI", but that's just another expression of anger
+23 Jun: I am introduced to Slay the Spire
+28 Jun:
+> I hope you find a way to let go of your war.
+You know, this is awfully condescending! You wouldn't tell a transgender person who was transitioning, "I hope you find a way to let go of this."
+But when I, who no one even seems to doubt has the same underlying psychological condition (long story—I recently finished a 16,000-word essay explaining my thing), put in years of effort to counter what I'm claiming is disinformation about the counterintuitive true nature of underlying psychological condition, I get told to drop it. Why the double standard? If [...] feelings count, why don't mine?
+4 Jul: Independence Day party
+feeling scared about the Singularity
+6 Sep: final Twitter showdown
+2 Nov: It's a black mark, but let's be quantitative: the guy helped me a huge amount in 2017-2019 (long, separate story) for reasons that made sense to him, and he also hurt me a decently large amount in December 2020 for reasons that made sense to him. These things can both be true at the same time!
+13 Jan 22: blessing to speak freely, and privacy norms?
+---END
> When I look at the world, it looks like [Scott](http://slatestarcodex.com/2014/11/21/the-categories-were-made-for-man-not-man-for-the-categories/) and [Eliezer](https://twitter.com/ESYudkowsky/status/1067183500216811521) and [Kelsey](https://theunitofcaring.tumblr.com/post/171986501376/your-post-on-definition-of-gender-and-woman-and) and [Robby Bensinger](https://www.facebook.com/robbensinger/posts/10158073223040447?comment_id=10158073685825447&reply_comment_id=10158074093570447&comment_tracking=%7B%22tn%22%3A%22R2%22%7D) seem to think that some variation on ["I can define a word any way I want"]() is sufficient to end debates on transgender identity.
But have you fucking tried trying to talk about biological sex with Berkeley psuedo-rationalists? They won't fucking acknowledge it.
-I said, "I need language that asymmetrically distinguishes between the original thing that already exists without having to try, and the artificial thing that's trying to imitate it to the limits of available technology."
+I said,
+
-Kelsey said, "[T]he people getting surgery to have bodies that do 'women' more the way they want are mostly cis women [...] I don't think 'people who'd get surgery to have the ideal female body' cuts anything at the joints."
Elena said, "'the original thing that already exists without having to try' sounds fake to me" ... and gets 4 "+1" emoji.
-Eliezer, I've spent my entire adult life in the subculture you created. You created me. Everything I'm doing now, I learned from you. I've put in a lot of effort—along with Ben and Michael and Sarah and Zvi and now Jessica—trying to explain that there's a collective sanity problem that we could really use your help with. (You should see the strategy email threads we've been having behind your back!) I don't understand how repeating that people live in different mental universes is a relevant response to our thousands of words of careful arguments?
+Eliezer, I've spent my entire adult life in the subculture you created. You created me. Everything I'm doing now, I learned from you. I've put in a lot of effort—along with Ben and Michael and Sarah and [...] and now Jessica—trying to explain that there's a collective sanity problem that we could really use your help with. (You should see the strategy email threads we've been having behind your back!) I don't understand how repeating that people live in different mental universes is a relevant response to our thousands of words of careful arguments?
I'm pretty sure everyone here is strongly in favor of Jessica's existence. But maybe, if someone thinks really carefully about it, there could be some way to protect Jessica's interests without destroying the faculty of language?! (Especially given that Jessica herself may have interests in being able to talk.)
Michael—
-> My guess is that at this point Zack and Ben should both sign off actually, and maybe Sarah as well, and that this should become a conversation between Eliezer and Jessica, Zvi, myself, and other people who have or have historically had strong MIRI affiliations and have paid serious attention to logical decision theories. Maybe we should bring in Wei Dai, since along with Eliezer, he created logical decision theory, so in so far as there's a natural heir to the Caliphate other than Eliezer it's him. At this point, I think that if this doesn't resolve with Eliezer backing down I am pretty much obligated to involve everyone who I believe has historically taken seriously Eliezer's claims of strong principles against lying. Among other things, I think that it's clear that without standing on such a moral bedrock, Eliezer is no longer claiming to be the spiritual leader of the Caliphate (intuitively, that title passes to me?), and that the secular leader of the AI Risk Caliphate has been Demis for many years now.
+> My guess is that at this point Zack and Ben should both sign off actually, and maybe Sarah as well, and that this should become a conversation between Eliezer and Jessica, [...] myself, and other people who have or have historically had strong MIRI affiliations and have paid serious attention to logical decision theories. Maybe we should bring in Wei Dai, since along with Eliezer, he created logical decision theory, so in so far as there's a natural heir to the Caliphate other than Eliezer it's him. At this point, I think that if this doesn't resolve with Eliezer backing down I am pretty much obligated to involve everyone who I believe has historically taken seriously Eliezer's claims of strong principles against lying. Among other things, I think that it's clear that without standing on such a moral bedrock, Eliezer is no longer claiming to be the spiritual leader of the Caliphate (intuitively, that title passes to me?), and that the secular leader of the AI Risk Caliphate has been Demis for many years now.
> Sarah, don't worry, this definitely doesn't imply involving shit-heads who want to carpet-bomb us. It might involve explicitly discussing the pros and cons of involving leftist shit-heads as part of motivating Eliezer (and/or Scott) and I think that we should try to become psychologically immune to their carpet-bombing in any event, but there may be good reasons for Sarah to remain economically coupled to their good-will, since of all of us she's the only one who may stand to receive resources from the Corporate Center-Left Coalition or to direct their resources. It does seem to me that technically, in so far as their attacks are harmless to those not seeking privilege, leftist shit-heads are Guided by the Goodness of their Weapons, and working out whether this is in fact true may be related to the logical decision theory that will enable us to integrate political threats with compassionate rational engagement and bring Eliezer and Scott back from the team of deception and privilege.
Michael—
> I don't think that we should be permanently giving up on people, but I do think that we should be regarding them as temporarily without the appeal to rights in relationship to ourselves. Varelse in Orson Scott Card speak.
-> I think that it was extremely worth our time to engage at this level of detail, and at an even greater level of detail in order to achieve common knowledge, as I don't think that we have achieved common knowledge*yet* if Zvi still thinks that we shouldn't confidently act on the assumption that Eliezer is not doing useful things unobserved.
+> I think that it was extremely worth our time to engage at this level of detail, and at an even greater level of detail in order to achieve common knowledge, as I don't think that we have achieved common knowledge*yet* if [...] still thinks that we shouldn't confidently act on the assumption that Eliezer is not doing useful things unobserved.
> I would defend engagement with Eliezer for as long as we did in the past and in several future cases if we had several future relationships that seemed comparably valuable to the one we were hoping for, e.g. with a reformed Eliezer Yudkowsky, but I think that it's important that we be clear in our own minds that as of now there should be no presumptive loyalty to Eliezer over, for instance, Demis who is the Schelling Point for leader for an AI Risk coalition not Guided by the Goodness of it's Weapons.
(zero-sum actors, e.g. neo nazis, are going to be pushing in various ways that make it easier to push along in their direction than it would if it were just you).
-I think this is how the political factions end up being "protect minorities and be incoherent" and "harm minorities and be locally coherent". Different respectability/violence tradeoffs (to get a mix of zero sum and non zero sum energy). And these define the Overton window, whereas many other Overton windows could be drawn instead.
+> I think this is how the political factions end up being "protect minorities and be incoherent" and "harm minorities and be locally coherent". Different respectability/violence tradeoffs (to get a mix of zero sum and non zero sum energy). And these define the Overton window, whereas many other Overton windows could be drawn instead.
+
+In the possible world where the parameters of male sexual psychology are such that autogynephilia doesn't exist "but everything else is the same", then I don't think you get the kind and scope of trans-activism movement that Saotome-Westlake's and Adams-Miller's writing is a reaction to. In this world, you still get Judith Butler (AFAB, Gender Trouble published 1990) and Anne Fausto-Sterling (AFAB, Sexing the Body published 2000), but I think you don't get pronoun stickers at conferences in 2018 and you don't get a "three-year-old [...] is a girl now; she verbally confirmed it!" social-reality enforcement campaign in 2020.
+
+To your second question, it might actually depend on how you operationalize "influence"! Notably, the campaign to trans MMB (whose parents met on lesswrong.com) is being run predominantly by AFAB people; they're just combatants in the service of an ideology that I don't think would have been so memetically fit if it weren't such a convenient accommodation for AGP. (At least, that's my theory under the doctrine of "algorithmic intent"; the people involved don't think of themselves as combatants in the service of anything.)
+
+
+I think your reputation (in the jungle growing around the ruins of what we once called the "rationalist community") has some amount of instrumental value: most of why you were so useful to me during the Category War when everyone I trusted was fucking with me, is that I was leaning on you for internal social proof ("Michael thinks it makes sense that I'm murderously angry about this, and that makes me feel brave enough to keep harrassing Scott and Eliezer instead of quietly committing mind-suicide"), but that only worked because I had a cached reputation of you as one of the Old-Time Rationalist Elders. It seems unfortunate that you don't have the asset of that reputation anymore among today's jungle-dwellers.
+
+
+-----
+
+7 Feb conversation with "Wilhelm" about Rust All-Hands
+
+M. showed people this funny Tweet on his phone: https://twitter.com/reduct_rs/status/1093599017303457793 and my uncontrolled facial expression was probably one of annoyance rather than mirth
+ReductRs on Twitter
+
+https://twitter.com/reduct_rs/status/1093599017303457793
+> Man Who Finds ‘They’ Pronouns Confusing Has No Problem Calling His Computer "She"
+
+You sent
+I don't know if he noticed; it's not like he was testing me for facecrime; showing ppl funny Tweets on your phone is just what everyone does. But maybe the little moments add up?
+You sent
+I lamented that the karoke machine didn't have the background lyrics for the part of "Prince Ali" where the women are fawing over Aladdin, and wokebro said that now that he thinks of it, the song is problematic, and I said, jovially, "It was 1992; it was a different time"
+You sent
+Obviously, no one is going to throw a fit over me casually defending the Disney films of my youth; that's not purity test material. But maybe the little moments add up, as people gradually get a read of who's sending what signals
+
+oh, I guess there was also a moment where I told the wife, "You're our only female vocalist; you should make sure to sing the 'Hope he doesn't see right through me' line in 'Be a Man' from Mulan", which I thought was an obviously innocuous/relevant comment (it was in fact the case that no one else had the voice for that part), I could imagine someone having a negative micro-reaction on the grounds that I was being exclusionary by calling attention to her sex, or because of counterfactual transphobia (I wouldn't have said that if there was also a trans woman in the room, which maybe makes it a bad thing to say even if there wasn't??)
+You sent
+that's sheer speculation; I don't know how the 2019 algorithm works; I haven't been downloading patches since 2009
+
+except, oops, I'm unironically guilty of sexism by referring to the woman as one of the dev's wife, when she's a Site Engineer at GitHub in her own right and therefore deserves one-of-us Hacker status even if she hasn't been working with Rust specifically
+You sent
+you, at least, will forgive me
+
+-----
+
+to Wilhelm 6 Mar 2019
+
+so much stupid drama among the "rationalists"
+You sent
+two non-passing MtFs are kicking up a fuss about my friend Anna allegedly discriminating against trans women, and people are taking this seriously
+You sent
+in contrast to my view (speaking only for myself; Anna is totally innocent) that statistically "discriminating" (i.e., using Bayesian reasoning) against trans women when hiring for a rationality org is just straightforwardly CORRECT
+
+
+at least men's and women's bodies will make similar numbers of paperclips
+You sent
+where "similar" means Cohen's d around 0.9
+
+8 Mar:
+
+watching television is different now that I have anti-prog antibodies
+
+on The Good Place (2016—), the neurotic philosophy professor character is black, and the idiot comic relief guy is Asian
+You sent
+I like Superstore (2015—) as a moderately-funny workplace comedy, but sometimes it veers into prog signaling in a way that just—isn't even well-done? (the most egregious one was when the disabled mulatto main-cast character is attracted to a woman at an anti-trans bathrooms protest, and is faced with the dilemma of, "She's hot, but her views are beyond the pale")
+
+Lauren Ash's character on Superstore is written in a behaviorally-masculine way that makes me swoon (in a world with magical body-swapping, AGPs would be this kind of woman), but Lauren Ash probably doesn't actually behave that way in real life
+
+[11 April:] Star Trek: Deep Space Nine (1992–1999) got preachy ... like, three episodes in a seven year run
+
+
+30 March
+> The cynicism may not be misplaced? The language of allyship is not a credible signal (in the language of https://en.wikipedia.org/wiki/Signalling_theory) of actually being an ally if non-allies can translate their agenda into that language and are incentivized to do so.
+
+he Liked it; as if you can get away with explaining the game theory of Havel's greengrocer as long as you frame it as "some greengrocers are actually enemies of the people"
+You sent
+"even if they have the sign up"
+
+
+13 April
+
+> Another problem with the view that "how to draw category boundaries is based on subjective priorities" perspective is that it is solipsistic. Your priorities re the territory aren't something you can always dictate; sometimes the territory is dictating priorities to you, and you need to figure out the right categories to address this pressure.
+> For instance, if you are a fisherman, you may not care about whether a dolphin is a fish: you can catch both in nets. But if you are a dophin breeder, then dolphins being mammals matters a lot more.
+> In the case of dolphins, most people won't have to breed dolphins or house them, so they lack skin-in-the-game to care about the exact categorization of dolphins
+
+one possible failure mode of this marketing campaign is that people who saw the object-level culture war version "... To Make Predictions" will have an Absolute Denial Macro reaction, but people who see the meta-level version about dolphins will just be like "This is obvious; everyone already knows this, not worth upvoting or discussion"
+
+there's no way to win
+
+> Perhaps because they don't think they have skin-in-the-game
+
+EVERYONE has skin in the game of Bayesian epistemology!!!
+
+------
+
+Discord conversation 23 Jan 2019
+
+but I need the phrase "actual women" in my expressive vocabulary to talk about the phenomenon where, if transition technology were to improve, then the people we call "trans women" would want to make use of that technology; I need language that _asymmetrically_ distinguishes between the original thing that already exists without having to try, and the artificial thing that's trying to imitate it to the limits of available technology
+
+Kelsey—
+what's wrong with 'cis women'?
+
+me—
+it's _pointlessly obfuscatory_
+in this particular context
+
+Kelsey—
+also, cis women get boob jobs all the time
+the people getting surgery to have bodies that do 'women' more the way they want
+are mostly cis women
+cis women get labia reconstruction sometimes
+cis women get tummy tucks
+I don't think 'people who'd get surgery to have the ideal female body' cuts anything at the joints
+
+Elena—
+"the original thing that already exists without having to try" sounds fake to me
+
+
+I agree that the "SneerClub et al. hates us because they're evil bullies" hypothesis has a grain of truth to it, but stopping the analysis there seems ... _incredibly shallow and transparently self-serving_?
+
+
+------
+
+https://www.facebook.com/yudkowsky/posts/pfbid0WA95Anng7UZZEjqYv2aWC4LxZJU7KPvcRnxkTdmNJpH4PoQQgEFtqszPbCiCnqfil?comment_id=10159410429909228
+
+If you listen to why _they_ say they hate us, it's because we're racist, sexist, transphobic fascists. The party-line response to seems to trend towards, "That's obviously false (Scott voted for Warren, look at all the social democrats on the LW/SSC serveys, &c.); they're just using that as a convenient smear because they like bullying nerds." (Fair paraphrase?)
+
+But ... the smears have a grain of truth to them, right? If "sexism" means "it's an empirical question whether innate statistical psychological sex differences of some magnitude exist, it empirically looks like they do, and this has implications about our social world", the "SSC et al. are crypto-sexists" charge is ABSOLUTELY CORRECT (e.g. https://slatestarcodex.com/.../contra-grant-on.../). (Crypto-racist, crypto-fascist, &c. are left as an exercise to the reader.)
+
+You could plead, "That's a bad definition of sexism", but that's only convincing if you've _already_ been trained in the "use empiricism and open discussion to discover policies with utilitarian-desirable outcomes" tradition; the people with a California-public-school-social-studies-plus-Tumblr education don't already _know_ that. (Source: I didn't know this at age 18 back in 'aught-six, and we didn't have Tumblr then.)
+
+In that light ... can you see why someone might find "blow the whistle on people who are claiming to be innocent but are actually guilty (of thinking bad thoughts)" to be a more compelling ethical consideration than "respect confidentiality requests"? The "debate ideas, not people" thing is a specific meta-ideological innovation, not baseline human morality!
+
+If our _actual_ problem is "Genuinely consistent rationalism is realistically always going to be an enemy of the state, because the map that fully reflects the territory is going to include facts that powerful coalitions would prefer to censor, no matter what specific ideology happens to be on top in a particular place and time (https://www.lesswrong.com/.../heads-i-win-tails-never...)", but we _think_ our problem is "We need to figure out how to exclude evil bullies", then we're in trouble!!
+
+
+> We also have an inevitable Kolmogorov Option issue but that should not be confused with the inevitable Evil Bullies issue, even if bullies attack through Kolmogorov Option issues.
+
+Being transparent about the game theory I see: intuitively, it seems like I have a selfish incentive to "support" the bullies (by publicly pointing out that they have a point, as above) insofar as I'm directly personally harmed by my social network following a Kolmogorov Option strategy rather than an open-dissidence Free Speech for Shared Maps strategy, and more bullying might cause the network to switch strategies on "may as well be hung for a sheep as a lamb" grounds? Maybe I should explain this so people have a chance to talk me out of it?
+... hm, acutally, when I try to formalize this with the simplest possible toy model, it doesn't work (the "may as well be hung ..." effect doesn't happen given the modeling assumptions I just made up). I was going to say: our team chooses a self-censorship parameter c from 0 to 10, and faces a bullying level b from 0 to 10. b is actually b(c, p), a function of self-censorship and publicity p (also from 0 to 10). The team leaders' utility function is U(c, b) := -(c + b) (bullying and self-censorship are both bad). Suppose the bullying level is b := 10 - c + p (self-censorship decreases bullying, and publicity increases it).
+My thought was: a disgruntled team-member might want to increase p in order to induce the leaders to choose a smaller value of c. But when I do the algebra, -(c + b) = -(c + (10 - c + p)) = -c - 10 + c - p = -10 - p. (Which doesn't depend on c, seemingly implying that more publicity is just bad for the leaders without changing their choice of c? But I should really be doing my dayjob now instead of figuring out if I made a mistake in this Facebook comment.)
+
+> Eliezer is not a private person - he's a public figure. He set in motion a machine that continues to raise funds and demand work from people for below-market rates based on moral authority claims centered around his ability to be almost uniquely sane and therefore benevolent. (In some cases indirectly through his ability to cause others to be the same.) "Work for me or the world ends badly," basically.
+
+> If this is TRUE (and also not a threat to destroy the world), then it's important to say, and to actually extract that work. But if not, then it's abuse! (Even if we want to be cautious about using emotionally loaded terms like that in public.)
+
+> We've falsified to our satisfaction the hypothesis that Eliezer is currently sane in the relevant way (which is an extremely high standard, and not a special flaw of Eliezer in the current environment). This should also falsify the hypothesis that the sanity-maintenance mechanisms Eliezer set up work as advertised.
+
+> The machine he built to extract money, attention, and labor is still working, though, and claiming to be sane in part based on his prior advertisements, which it continues to promote. If Eliezer can't be bothered to withdraw his validation, then we get to talk about what we think is going on, clearly, in ways that aren't considerate of his feelings. He doesn't get to draw a boundary that prevents us from telling other people things about MIRI and him that we rationally and sincerely believe to be true.
+
+> The fact that we magnanimously offered to settle this via private discussions with Eliezer doesn't give him an extra right to draw boundaries afterwards. We didn't agree to that. Attempting to settle doesn't forfeit the right to sue. Attempting to work out your differences with someone 1:1 doesn't forfeit your right to complain later if you were unable to arrive at a satisfactory deal (so long as you didn't pretend to do so).
+
+--------
+
+7 May—
+> I'm still pretty frustrated with the way you seem to dismiss my desire for a community that can get the basics right as delusional! Yes, I remember how you said in 2009 that you're not going to say the Things You Can't Say, and you keep repeating that you tried to tell me that public reason doesn't work, but that seems really fundamentally unresponsive to how I keep repeating that I only expect consensus on the basic philosophy stuff (not my object-level special interest). Why is it so unrealistic to imagine that the actually-smart people could enforce standards in our own tiny little bubble of the world? (Read the "worthless cowards" political/rhetorical move as an attempt to enforce the standard that the actually-smart people should enforce standards.)
+
+> I'm also still pretty angry about how your response to my "I believed our own propaganda" complaint is (my possibly-unfair paraphrase) "what you call 'propaganda' was all in your head; we were never actually going to do the unrestricted truthseeking thing when it was politically inconvenient." But ... no! I didn't just make up the propaganda! The hyperlinks still work! I didn't imagine them! They were real! You can still click on them: ["A Sense That More Is Possible"](https://www.lesswrong.com/posts/Nu3wa6npK4Ry66vFp/a-sense-that-more-is-possible), ["Raising the Sanity Waterline"](https://www.lesswrong.com/posts/XqmjdBKa4ZaXJtNmf/raising-the-sanity-waterline)
+
+> Can you please acknowledge that I didn't just make this up? Happy to pay you $200 for a reply to this email within the next 72 hours
+
+Or see ["A Fable of Science and Politics"](https://www.lesswrong.com/posts/6hfGNLf4Hg5DXqJCF/a-fable-of-science-and-politics), where the editorial tone is pretty clear that we're supposed to be like Daria or Ferris, not Charles.
+
+> But, it's kind of bad that I'm thirty-one years old and haven't figured out how to be less emotionally needy/demanding; feeling a little bit less frame-locked now; let's talk in a few months (but offer in email-before-last is still open because rescinding it would be dishonorable)
+
+
+> I liked the conversation and like being on good terms and do not wish to be negative in what I am about to say. That said, if it's okay with you, I in general don't want to receive offers of money anymore (nor actual money either); I think I used to regard them as good-faith proposals for mutual benefit between consenting adults a la libertarianism, and at the moment I am afraid that if I ever accept money (or you ever give me money for an email or something) it'll appear in some Ben Hoffman essay as "using you" or some similar such thing, and receiving such offers (or taking them), given that I am not in fact in much need of money at the present margin, looks to me like not a good way to further my own goals as a consenting adult. (I appreciate the email you then sent afterward; I'd planned to write this before you sent that email, and it still seemed worth saying, but I appreciate also your own efforts to create a low-drama context.)
+
+> Re delusions, perhaps "ideals, which are useful but also partly intentionally simplified/inaccurate so as to be easier to unite around, and which for simplicity one coordinates as though others share" might be a better way to describe them, since sometimes simplified models have uses. Also, of course, the desire part can't be delusional; desires don't have truth-values; only predictions can be false-and-protecting-themselves-from-evidence. You can desire whatever you want, and can work toward it!
+
+> Regarding whether those ideals were ever the thing that a sensible but unbiased seeker-for-uniting-ideals, observing the 2008-2010 community, would've extrapolated from our community's speech and writing:
+
+> I agree such a person might've gotten your ideal from "a sense that more is possible", "raising the sanity waterline", and for that matter the facebook post about local validity semantics. Also from Michael Vassar's speech at the time about how he just went ahead and discussed e.g. race and gender and IQ and didn't seem to get ill effects from this and thought other people should do so too. too.
+> I think there are also pieces of speech one can point to from 2008-2010 that point toward it being unsurprising if people avoid controversial issues, e.g. "politics is the mind-killer" and its reception/quotation"; my own discussion of the essay "things you can't say" and my getting of Benton House to read this essay; various other things that others in good regard (not Eliezer) said to me privately at the time.
+> My own extrapolated ideal from all of this at the time was something like: "Ah, our ideal is 'let's try to figure out how to actually think, and practice it in concrete cases, while meanwhile dodging examples in public that risk getting us into trouble; practicing on less politics-inducing examples should work fine while developing the art, and is a net better idea than heading toward places where our nascent community may get torn apart by politics'".
+
+> (Looking at that last bullet point myself today, that claim seems far from obviously true, though also far from obviously false. It is therefore not a sentence I would say today (to avoid saying things that aren't likely enough to be true). It also seems to me today that we have done less to develop the core art of epistemic rationality than Eliezer hoped for in those posts, partly for lack of diligence and partly because the project itself proved harder and more likely to stir up hard-to-avoid psychological and social cans of instability than I initially expected. I guess my current guess is something like: "politics: can't rationality with it, can't rationality without it", although "it's hard to figure out how to" is maybe more accurate than simply "can't".)
+
+
+
+> When forming the original let's-be-apolitical vision in 2008, we did not anticipate that whether or not I should cut my dick off would become a political issue. That's new evidence about whether the original vision was wise! I'm not trying to do politics with my idiosyncratic special interest; I'm trying to think seriously about the most important thing in my life and only do the minimum amount of politics necessary to protect my ability to think. If 2019-era "rationalists" are going to commit a trivial epistemology mistake that interferes with my ability to think seriously about the most important thing in my life, but can't correct the mistake (because that would be politically inconvenient), then the 2019-era "rationalists" are worse than useless to me personally. This probably doesn't matter causally (I'm not an AI researcher, therefore I don't matter), but it might matter timelessly (if I'm part of a reference class that includes AI researchers).
+
+> Fundamentally, I just don't think you can do consisently high-grade reasoning as a group without committing heresy, because of the "Entangled Truths, Contagious Lies"/"Dark Side Epistemology" mechanism. You, Anna, in particular, are unusually good at thinking things without saying them; I think most people facing similar speech restrictions just get worse at thinking (plausibly including Eliezer), and the problem gets worse as the group effort scales. (It's easier to recommend "What You Can't Say" to your housemates than to put it on a canonical reading list, for obvious reasons.) You can't optimize your group's culture for not-talking-about-atheism without also optimizing against understanding Occam's razor; you can't optimize for not questioning gender self-identity without also optimizing against understanding "A Human's Guide to Words."
+
+-----
+
+
+there's this thing where some people are way more productive than others and everyone knows it, but no one wants to make it common knowledge which is really awkward for the people who are simultaneously (a) known but not commonly-known to be underperforming (such that the culture of common-knowledge-prevention is to my self-interest because I get to collect the status and money rents of being a $150K/yr software engineer without actually performing at that level, and my coworkers and even managers don't want to call attention to it because that would be mean—and it helps that they know that I already feel guilty about it) but also (b) tempermentally unsuited and ideologically opposed to subsisting on unjustly-acquired rents rather than value creation
+
+(where I'm fond of the Ayn Rand æsthetic of "I earn my keep, and if the market were to decide that I don't deserve to live anymore, I guess it would be right and I should accept my fate with dignity" and I think the æsthetic is serving a useful function in my psychology even though it's also important to model how I would change my tune if the market actually decided that I don't deserve to live)
+
+> but the "Everyone knows that Zack feels guilty about underperforming, so they don't punish him, because he's already doing enough internalized-domination to punish himself" dynamic is unsustainable if it evolves (evolves is definitely the right word here) into
+
+a loop of "feeling gulit in exchange for not doing work" rather than the intended function of "feeling guilt in order to successfully incentivize work"
+
+You've got to be strong to survive in the [O-ring sector](https://en.wikipedia.org/wiki/O-ring_theory_of_economic_development)
+
+(I can actually see the multiplicative "tasks are intertwined and have to all happen at high reliability in order to create value" thing playing out in the form of "if I had fixed this bug earlier, then I would have less manual cleanup work", in contrast to the "making a bad latte with not enough foam, that doesn't ruin any of the other customer's lattes" from my Safeway-Starbucks-kiosk days)
+
+------
+
+Discord messaging with Scott in October 2021—
+
+However subjectively sincere you are, I kind of have to read your comments as hostile action in the rationalist civil war. (Your claim that "it’s fair for the community to try to defend itself" seems to suggest you agree that this is a somewhat adversarial conversation, even if you think Jessica shot first.) A defense lawyer has an easier job than a rationalist—if the prosecution makes a terrible case, you can just destroy it, without it being your job to worry about whether your client is separately guilty of vaguely similar crimes (that the incompetent prosecution can't prove).
+
+[context for "it's fair for the community to defend itself"—
+https://www.lesswrong.com/posts/MnFqyPLqbiKL8nSR7/my-experience-at-and-around-miri-and-cfar-inspired-by-zoe?commentId=qsEMmdo6DKscvBvDr
+
+> I more or less Outside View agree with you on this, which is why I don't go around making call-out threads or demanding people ban Michael from the community or anything like that (I'm only talking about it now because I feel like it's fair for the community to try to defend itself after Jessica attributed all of this to the wider community instead of Vassar specifically) "This guy makes people psychotic by talking to them" is a silly accusation to go around making, and I hate that I have to do it!
+]
+
+
+[note: comment this is quoting is at 92 karma in 40 votes]
+> hey just responded with their "it's correct to be freaking about learning your entire society is corrupt and gaslighting" shtick.
+
+I will absolutely bite this bullet. You once wrote a parable (which I keep citing) about [a Society in which it becomes politically fashionable to claim that thunder comes before lightning](https://slatestarcodex.com/2017/10/23/kolmogorov-complicity-and-the-parable-of-lightning/). Imagine living in that Society in the brief period where the taboo is being established. Imagine taking epistemic rationality seriously, thinking that your friends and community leaders take it seriously, and trying to make sense of all of them, seemingly in lockstep, suddenly asserting that thunder comes before lightning, and acting like this is perfectly normal. When you point out that no one believed this ten years ago and ask what new evidence came in, they act like they don't remember. When you pull out videos and textbooks to argue that actually, lightning comes before thunder, they dodge with, "Well, it's depends on what you mean by the word 'before.'" (Technically, true!)
+Eventually, you would get used to it, but at first, I think this would be legitimately pretty upsetting! If you were already an emotionally fragile person, it might even escalate to a psychiatric emergency through the specific mechanism "everyone I trust is inexplicably lying about lightning → stress → sleep deprivation → temporary psychosis". (That is, it's not that Society being corrupt directly causes mental ilness—that would be silly—but confronting a corrupt Society is very stressful, and that can [snowball into](https://lorienpsych.com/2020/11/11/ontology-of-psychiatric-conditions-dynamic-systems/) things like lost sleep, and sleep is [actually really](https://www.jneurosci.org/content/34/27/9134.short) [biologically important](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6048360/).)
+
+This is a pretty bad situation to be in—to be faced with the question, "Am I crazy, or is everyone else crazy?" But one thing that would make it slightly less bad is if you had a few allies, or even just an ally—someone to confirm that the obvious answer, "It's not you," is, in fact, obvious. But in a world where everyone who's anyone agrees that thunder comes before lightning—including all the savvy consequentialists who realize that being someone who's anyone is an instrumentally convergent strategy for acquiring influence—anyone who would be so imprudent to take your everyone-is-lying-about-lightning concerns seriously, would have to be someone with ... a nonstandard relationship to social reality. Someone meta-savvy to the process of people wanting to be someone who's anyone. Someone who, honestly, is probably some kind of major asshole. Someone like—Michael Vassar!
+
+
+From the perspective of an outside observer playing a Kolmogorov-complicity strategy, your plight might look like "innocent person suffering from mental illness in need of treatment/management", and your ally as "bad influence who is egging the innocent person on for their own unknown but probably nefarious reasons". If that outside observer chooses to draw the category boundaries of "mental illness" appropriately, that story might even be true. So why not quit making such a fuss, and accept treatment? Why fight, if fighting comes at a personal cost? Why not submit?
+
+I have my answer. But I'm not sure you'd understand.
+
+
+> I disagree with your assessment that your joining the Vassarites wasn't harmful to you
+
+As compared to what? In the counterfactual where Michael vanished from the world in 2016, I think I would have been just as upset about the same things for the same reasons, but with fewer allies and fewer ideas to make sense of what was going on in my social environment.
+
+I've got to say, it's really obnoxious when people have tried to use my association with Michael to try to discredit the content of what I was saying—interpreting me as Michael's pawn.
+Gwen, one of the "Zizians", in a blog post about her grievances against CfAR, has [a section on "Attempting to erase the agency of everyone who agrees with our position"](https://everythingtosaveit.how/case-study-cfar/#attempting-to-erase-the-agency-of-everyone-who-agrees-with-our-position), complaining about how people try to cast her and Somni and Emma as Ziz's minions, rather than acknowledging that they're separate people with their own ideas who had good reasons to work together. I empathized a lot with this. My thing, and separately Ben Hoffman's [thing about Effective Altruism](http://benjaminrosshoffman.com/drowning-children-rare/), and separately Jessica's thing in the OP, don't really have a whole lot to do with each other, except as symptoms of "the so-called 'rationalist' community is not doing what it says on the tin" (which itself isn't a very specific diagnosis). But insofar as our separate problems did have a hypothesized common root cause, it made sense for us to talk to each other and to Michael about them.
+Was Michael using me, at various times? I mean, probably. But just as much, _I was using him_: from my perspective, I was leaning on Michael for social proof for high-variance actions that I wouldn't have been socially brave enough to take unilaterally—but because Michael had my back, and because my brain still had him tagged as a "rationalist" Authority figure form back in the Overcoming Bias days, that made it okay. Particularly with the November 2018–April 2019 thing (where I, with help from Michael and Ben and Sarah and later Jessica, kept repeatedly pestering you and Eliezer to clarify that categories aren't arbitrary—that started because I flipped the fuck out when I saw [Eliezer's "hill of meaning in defense of validity" Twitter performance](https://twitter.com/ESYudkowsky/status/1067185907843756032). That was the "Vassarites" doing an _enormous_ favor for _me_ and _my_ agenda. (If Michael and crew hadn't been there for me, I wouldn't have been anti-social enough to keep escalating.) And you're really going to try to get away with claiming that they were making my situation worse? That's absurd. Have you no shame?
+
+
+
+I just wish you'd notice that it is a choice, and that there is a conflict: the lightning-first dissidents in your parable absolutely should not be applying [mistake theory](https://slatestarcodex.com/2018/01/24/conflict-vs-mistake/) or the principle of charity to smart thunder-first interlocutors like Kolmogorov, because you've specified in the story that Kolmogorov isn't being intellectually honest! The theory that Kolmogorov is making mere "honest mistakes" makes bad predictions, because actually-honest cognitive mistakes would be random, rather than systematically favoring the orthodoxy.
+
+ Eliezer was offering a vision of a new mental martial art of systematically correct reasoning: "rationalist" in the sense used in the Sequences meant someone who studies rationality (like [Robin Dawes](https://www.amazon.com/gp/product/0155752154/) or E.T. Jaynes), the way "physicist" means someone who studies physics.
+
+Contrast that to how, in the Alicorner conversation the other day, Kelsey expresses anger at me over ["the thought that there are people out there who wanted to be rationalists and then their experience of the rationalist community was relentlessly being told that trans women are actually men [...] and are traumatized by the entire experience of interacting with us"](https://discord.com/channels/401181628015050773/458419017602826260/899705780402552852). I think this sense of "wanting to be a rationalist" (wanting to belong to the social cluster) and making sure that "we" don't alienate potential should be seen as obviously non-normative. I don't care about the movement except insofar as it succeeds in advancing the ideology.
+
+See Eli (with many emoji-agreements) in the Alicorner conversation arguing that I shouldn't have taken "DMs open for questions" literally on the presumption that once someone has declared that they "are trans", theory is unwelcome. I appreciate that this crowd is much more intellectually tolerant than people who are actually woke. But insane religious fantatics who "merely" want heretics to know their place (as opposed to wanting to hurt or exile them) are still insane religious fanatics. How do you live like this?!
+
+But those are cases of different conceptual boundaries (which represent probabilistic inferences about the real world) being conventionally attached to the same word/symbol, such that you should disambiguate which concept is meant when it's not clear from context. If you do the disambiguation, you're not accepting an unexpected X deep inside the conceptual boundaries of what would normally be considered Y; you just have two different conceptual boundaries that both use the symbol "Y".
+
+And the persistent lack of clarity on this point is very obviously motivated. When people like Quillette editor Colin Wright say, "Look, I'm not questioning your 'gender', whatever that is; I'm just saying that this female {locker room/prison/sports league/&c.} is on the basis of sex, not gender", then the people who praise you for writing "... Not Man for the Categories" are not satisfied.
+
+the reason the bitterness is there is because of this situation where our marketplace of ideas is rigged, the marketplace owners know it's rigged, and will admit to it in private, and even in public (as in "Kolmogorov Complicity") if you ask the right way—but the "rationalist" and "EA" brand names and marketing machines are still out there sucking up attention and resources based on claims to excellence that everyone in the know actually _agrees_ are false! (And people are still citing "... Not Man for the Categories" without noticing that the edit-note's admission completely undermines sections IV.–VI.) I spent four years of my life arguing that you were wrong about something in increasing technical detail, and Yudkowsky still says you're ["always right"](https://twitter.com/ESYudkowsky/status/1434906470248636419)! Obviously, I get that it's not your fault that you write faster than everyone else and Yudkowsky is an arrogant prick. But it's still not fair, and it still stings, and I don't think it's Michael fault that I feel this way, even if I used specific framings (marketing machines, &c.) that came from conversations involving him (would it be better if I felt the same way, but were less articulate about why?), and he was my political ally for a couple years. (I haven't even talked to the guy in ten months!! My last experience with him was actually very bad!)
+
+And when Jessica writes a post to talk about her experiences, it seems messed up that you barge in to make it all about Michael, get the top-voted comment (341 points!), and Yudkowsky boosts you as a reliable narrator ("leaving out the information from Scott's comment about a MIRI-opposed person who is [...] causing psychotic breaks in people"), and the voting on the post turns around (it was [at 143 karm in 97 votes](https://archive.md/MJHom), and now it's down to 50 karma in 162 votes). This looks like raw factional conflict to me: Jessica had some negative-valence things to say about mainstream "rationalist" institutions, so the mainstream rationalist leaders move in to discredit her by association.
+
+You know, I also had a sleep-deprivation-induced-psychotic-break-with-hospitalization in February 2013, and shortly thereafter, I remember Anna remarking that I was sounding a lot like Michael. But I hadn't been talking to Michael at all beforehand! (My previous email conversation with him had been in 2010.) So what could Anna's brain have been picking up on, when she said that? My guess: there's some underlying dimension of psychological variation (psychoticism? bipolar?—you tell me; this is supposed to be your professional specialty) where Michael and I were already weird/crazy in similar ways, and sufficiently bad stressors can push me further along that dimension (enough for Anna to notice). Are you also going to blame Yudkowsky for making people autistic?
+
+"A lesson is learned but the damage is irreversible." https://www.alessonislearned.com/
+
+Take all what, specifically?! I briefly told you about a specific incident when Michael and Jessica and Jack were super-mean to me, in response to my perceived negligence during an unusual crisis situation. I do think this should be a negative update about Michael's and Jessica's and Jack's character and judgement. (Even if being super-mean to me would have been justified if it had helped Sasha, I don't think it actually helped or could have been expected to help, and the timescale-of-a-month effects of being super-mean to me turned out to be pretty bad for me because of my preexisting psych vulnerabilities.) That doesn't mean you can assume Sarah is being harmed just by having any social connection to them! (Ask Sarah, if you must!)
+
+I messaged Sarah the next day to talk about what happened. When I mentioned that I was rethinking my earlier vague intention to move to New York to be closer to the Vassarites and to her, "because they (everyone, but Ben being relatively more merciful) are super-quick to escalate the moment their ideology says that you defected first", she replied, "I like em but I could see how being their roommate would be a pain."
+
+[...]
+
+But you know, I also think there are reasons that getting too close to the Valinorians could be bad for someone (in subtler, lower-variance ways). Michael will yell at you and give you terrible drug advice, but at least I can Actually Talk to him; Kelsey will slowly destroy your sanity with superficial "niceness" carefully engineered to block all substantive communication. (Perhaps the best illustration of what I mean is that after typing the previous sentence, I felt an impulse to add "Don't tell Kelsey I said that," whereas Michael you can tell anything. For that, I can forgive a lot. I don't necessarily expect you to understand, but do try to have some empathy that being super-mean over the phone during a crisis is a lesser crime on this side of the trenches than on yours.) In both cases, I want to remain on good terms, without confusing the relationship I actually have with the one I wished I had.
+
+
+Insofar as I'm skeptical of Jessica's narrative, I suspect she's not putting nearly enough emphasis on the acid + weed (!) as a direct causal factor. But I don't think that invalidates everything she has to say about the culture of you guys who aren't us (anymore) and who we don't like (as much anymore). "MIRI's paranoid internal culture did a number on Jessica psychologically, to which there are structural parallels to Zoe's report about Leverage" and "subsequently, trying acid made things much worse" and "subsequently, Michael's recommending weed made things even worse" could all be true at the same time.
+
+I agree that this is a real concern. (I was so enamored with Yudkowsky's philosophy-of-science writing that there was no chance of me bouncing on account of the sexism that I perceived, but I wasn't the marginal case.) There are definitely good reasons to tread carefully when trying to add sensitive-in-our-culture content to the shared map. But I don't think treading carefully should take precedence over getting the goddamned right answer.
+
+For an example of what I think treading-carefully-but-getting-the-goddamned-right-answer looks like, I'm really proud of my review of Charles Murray's 2020 Human Diversity: http://unremediatedgender.space/2020/Apr/book-review-human-diversity/
+I'm definitely not saying, Emil Kirkegaard-style, "the black/white IQ gap is genetic, anyone who denies this is a mind-killed idiot, deal with it." Rather, first I review the Science in the book, and then I talk about the politics surrounding Murray's reputation and the technical reasons for believing that the gap is real and partly genetic, and then I go meta on the problem and explain why it makes sense that political forces make this hard to talk about. I think this is how you go about mapping the territory without being a moral monster with respect to your pre-Dark Enlightenment morality. (And Emil was satisfied, too: https://twitter.com/KirkegaardEmil/status/1425334398484983813) ["Probably the best book review ever of an HBD book" https://archive.ph/2cSZh]
+
+I think this should just be normal, and I'm willing to fight for a world in which this is normal. When something is hard to talk about, you carefully flag the reasons that it's hard to talk about, and then you say the thing. If you don't get to the part where you say the thing, then your shared map doesn't reflect the territory!!
+
+It's ahistorical to talk about "the rationalist community" in 2007. There was a blog where Eliezer Yudkowsky and Robin Hanson wrote about rationality as a subject matter. No one thought the subject matter of reasoning itself was the exclusive property of any particular social group that could have a "caliph." That would be crazy!
+
+I confess, in my frustration at this, sometimes I indulge in the vice of flashy rhetoric that sounds like "Everyone is lying!" (which is surely false), rather than reaching for something wordier like, "This is an astoundingly hostile information environment!" (which is surely true). The reason it's tempting is because I think the outcome of a sufficiently hostile information environment looks a lot like "Everyone is lying", even if the psychological mechanisms are different.
+
+-----
+
+Discord with Scott December 2019
+
+> Don't answer if it would be too unpleasant, but - I'm interested in asking some questions to test the autogynephilia hypothesis on the next SSC survey. Do you have any suggestions beyond the obvious? Also, do you know any intelligent and friendly opponent of the hypothesis who I could ask the same question to to make sure I'm getting advice from both sides?
+
+20 December
+Hi, Scott! You shouldn't worry about answering being unpleasant for me—speech is thought, and thinking is good! I am actively at war with the socio-psychological forces that make people erroneously think that talking is painful!
+
+-----
+
+zackmdavis — 12/24/2019 2150
+okay, maybe speech is sometimes painful
+the _Less Wrong_ comment I just left you is really mean
+and you know it's not because I don't like you
+you know it's because I'm genuinely at my wit's end
+after I posted it, I was like, "Wait, if I'm going to be this mean to Scott, maybe Christmas Eve isn't the best time?"
+it's like the elephant in my brain is gambling that by being socially aggressive, it can force you to actually process information about philosophy which you otherwise would not have an incentive to
+I hope you have a merry Christmas
+zackmdavis — 12/24/2019
+oh, I guess we're Jewish
+that attenuates the "is a hugely inappropriately socially-aggressive blog comment going to ruin someone's Christmas" fear somewhat
+
+---------
+
+ScottAlexander — 12/25/2019
+Eh, I know we've been over this a thousand times before, but I still feel like we disagree and can't pinpoint the exact place where we diverge and so I'm reduced to vomiting every step of my reasoning process out in the hopes that you say "That step! That's the step I don't like!". Most of the reasoning process will be stuff you know and agree with and I'm sorry.
+
+My confusion is something like - I think of "not trying to replicate God's dictionary" as equivalent to "attempts to design language are a pragmatic process rather than a 100% truth-seeking process". But if you agree that designing language is a partly pragmatic process, then you can't object to appeals to consequence in it, since appealing to consequence is what you're supposed to do in a pragmatic process. I'm sorry if I made that point in a way that insulted your intelligence, but I don't understand how, even though we agree on all the premises, we keep getting opposite conclusions, and I don't have a great solution other than vomiting more words.
+
+Probably you are overestimating my intelligence and ability to understand things, I'm sorry. I continue to regret that something like this has come between us, and I really like you, and if there were something I could do that would prevent us from constantly being at loggerheads and prevent me from constantly offending you and making you feel like I am insulting and misunderstanding you I would, but I just really think I'm right about this, and so far haven't been able to figure out why I might not be, and don't like pretending not to believe something when I do believe it.
+ScottAlexander — 12/25/2019
+And I feel like thanks to you getting involved in the Group It Would Be Paranoid Of Me To Blame All Bad Things In The Rationalist Community On, whenever I try to sort this out with you you accuse me of playing dumb and mentally destabilizing you and ruining your day and whatever, and so I don't do it. I don't know how to convince you that I actually think I'm right and am trying to do the right thing as a rationalist and this isn't all part of an attempt to destroy you.
+zackmdavis — 12/25/2019
+Thanks and merry Christmas 💖 🎅 ; I think I see some productive things I can write that haven't already been written yet that do a better job of pinpointing the step
+ScottAlexander — 12/25/2019
+I reread your Boundary post and I honestly think we disagree on the f**king dolphins.
+I just 100% literally believe that a fish group including dolphins is exactly as good as one that doesn't, whereas it seems like you are taking it as a preliminary obvious step that this is worse.
+Am I understanding you right?
+zackmdavis — 12/25/2019
+oh, that's interesting
+my reply is that the fish group including dolphins is going to be a "thinner" (lower-dimensional) subspace of configuration space
+at least, compared to some alternatives
+ScottAlexander — 12/25/2019
+Imagine a superdolphin that has had 100 million more years to evolve to live in the sea and converge to sea-optimal, so that it is identical to fish in every way (eg scales, cold-blooded, etc) but still has tiny bits of mammalian DNA that indicate it is evolutionary-relationship a mammal. Do you agree that this could reasonably be classified as a fish?
+zackmdavis — 12/25/2019
+yes, but real dolphins are predictably not superdolphins. Again, I use the configuation space visual metaphor for everything: if the superdolphin cluster overlaps with the fish cluster along all dimensions except the "these-and-such tiny regions of the genome" axis, then I want to put superdolphins and fish in the same category (unless I have some special reason to care about those-and-such tiny regions of the genome), but the reason evolutionary relatedness is probably going to be a thicker/more-robust subspace because DNA is the "root" of the causal graph that produces morphology &c.
+ScottAlexander — 12/25/2019
+Huh, it seems plausible to me that the number of characteristics dolphins share with fish (fluked tails, smooth bodies, fins, same approximate shape, live in water) is larger than the number they share with mammals (tiny irrelevant hairs, bears live young but I think some fish do too, warm-blooded but so are some fish)
+Does it seem to you that somewhere between dolphins (which you classify as obviously mammals) and superdolphins (which you classify as plausibly fish) there's a broad zone where reasonable people could disagree about whether the animal should be classified as a mammal or fish?
+zackmdavis — 12/25/2019
+in principle, yes, but the tendency I'm at war with is people saying, "Whelp, categories can't be wrong, so there's nothing stopping me from using pragmatic considerations like politeness as a tiebreaker", which seems like a memetic-superweapon for ignoring any category difference you don't like as long as you can find some superficial similarity in some subspace (sorry for the linear algebra jargon; this is actually how I think about it)
+"Categories can't be wrong; language is pragmatic" is an appeal-to-arbitrariness conversation-halter (https://www.lesswrong.com/posts/wqmmv6NraYv4Xoeyj/conversation-halters), when the correct move is to figure out which subspace is more relevant to your goals
+ScottAlexander — 12/25/2019
+Forget what tendencies we're at war with for now, I want to figure out where we actually differ.
+zackmdavis — 12/25/2019
+thanks, that's right
+ScottAlexander — 12/25/2019
+I'm thinking one possible crux is something like "you believe thingspace is a literal space with one correct answer, even though puny human intelligences cannot perfectly map it, whereas I believe thingspace is a vague metaphor that at best we can agree on a few things that would be obviously true about it". Is that actually a crux of ours?
+zackmdavis — 12/25/2019
+it's getting close
+it's not a literal space, but
+ScottAlexander — 12/25/2019
+Or, like, if dolphins were at the exact spot between real dolphins and superdolphins where reasonable people (including all the smartest people) disagreed whether they were mammals or fish, do you feel like there would still be one correct answer that they just weren't smart enough to converge on?
+zackmdavis — 12/25/2019
+no
+ScottAlexander — 12/25/2019
+Huh.
+zackmdavis — 12/25/2019
+maybe the intuition-generating difference (not a crux, but the difference in intellectual backgrounds that generates cruxes) is that I'm not trying to think of "reasonable people", I'm doing AI theory
+ScottAlexander — 12/25/2019
+Okay, then well-designed AIs who could do other things right and hadn't been programmed to specifically grind an axe on this one point.
+zackmdavis — 12/25/2019
+different AIs would use different category systems depending on their goals, depending on which variables they cared about being able to predict/control
+ScottAlexander — 12/25/2019
+Huh, then I think maybe that is our crux. I feel like there are a bunch of easy ways to solve this for AIs, like "use two different categories to represent sea-adaptedness and evolutionary-descent as soon as you realize there's going to be an issue". Or "figure out why you personally care about dolphins and only use the category that reflects why you care about them". I feel like it's only when we live in a society of hard-to-coordinate humans who already have a word and various connotations around it and who are all pursuing different goals that this becomes a problem.
+zackmdavis — 12/25/2019
+Right! And like (sorry if this is stepping back into the memetic-warfare thing for a moment which is unhealthy, because we actually trust each other enough to do philosophy), my first reaction to "Against Lie Inflation" was agreeing with the post, and feeling rage that you obviously wouldn't let Jessica get away with saying, "An alternative categorization system is not an error, and borders are not objectively true or false, therefore you can't object to me defining lying this way"
+And then only after I read it more closely and zeroed in on the paragraph about "This definition would make people angrier", was I less angry at you, because that meant we actually had different views on how linguistic pragmatism worked and you weren't just making an unprincipled exception for trans people
+ScottAlexander — 12/25/2019
+Just to make sure I understand this - you were annoyed I wouldn't let Jessica get away with unprincipled pragmatism because you thought that was something I had supported before and I was being a hypocrite, not because you agree with unprincipled pragmatism.
+zackmdavis — 12/25/2019
+right
+I agreed with the post, but was angry at the hypocrisy
+(percieved hypocrisy)
+you could argue that I'm being incredibly unfair for having an axe to grind over something you wrote five years ago (I've written lots of things five years ago that were wrong, and it would be pretty annoying to get angry mail from people wanting to argue about every single one)
+It must suck being famous 😢
+ScottAlexander — 12/25/2019
+No, I agree if I'm wrong about this or hypocritical I want to know.
+I guess my claim is something like "language should be used pragmatically to achieve goals, but there is actually a best way to use language pragmatically to achieve goals in any given context".
+I think I also am just much more likely than you to think any given clusters in thingspace are in the vague "is a red egg a rube or a blegg" category where thingspace has nothing further to tell us and we have to solve it practically. I was really surprised you thought there was a right answer to the dolphin problem.
+zackmdavis — 12/25/2019
+the reason I accuse you of being motivatedly dumb is because I know you know about strategic equivocation, motte-and-bailey, the worst argument in the world, because you taught everyone about it
+and when I say, "Hey, doesn't this also apply to 'trans women are women'", you act like you don't get it
+and that just seems implausible
+
+ScottAlexander — 12/25/2019
+Can you give a specific example of that? What's coming to mind is someone saying "Women have uteruses," I say "Sounds right", someone else saying "Caitlyn Jenner is a woman", I say "I'm committed to agreeing with that", and them saying "Therefore Caitlyn Jenner has a uterus", and me saying "Well, wow". Obviously that doesn't work, can you give an example of strategic equivocation in this space that does?
+(I'm not saying there isn't anything, just that it's not immediately coming to mind which may be a flaw in my own generators)
+zackmdavis — 12/25/2019
+(yes, one moment to dig up a link from my notes)
+things like ... The Nation, a nationally prominent progressive magazine writes, "There is another argument against allowing trans athletes to compete with cis-gender athletes that suggests that their presence hurts cis-women and cis-girls. But this line of thought doesn’t acknowledge that trans women are in fact women." https://www.thenation.com/article/trans-runner-daily-caller-terry-miller-andraya-yearwood-martina-navratilova/
+ScottAlexander — 12/25/2019
+I agree that this is stupid and wrong and a natural consequence of letting people use language the way I am suggesting.
+zackmdavis — 12/25/2019
+this is where the memetic warfare thing comes in; I don't think it's fair to ordinary people to go as deep into the philosophy-of-language weeds as I can before they're allowed to object to this
+ScottAlexander — 12/25/2019
+I think my argument would be something like "the damage from this is less than the potential damage of trans people feeling lots more gender dysphoria". I think the part of your Boundaries post that I marked as a likely crux for that is "Everything we identify as a joint is a joint not 'because we care about it', but because it helps us think about the things we care about" but I didn't really follow - if you agree this is our crux, could you explain it in more detail?
+zackmdavis — 12/25/2019
+I think "pragmatic" reasons to not just use the natural clustering that you would get by impartially running the clustering algorithm on the subspace of configuration space relevant to your goals, basically amount to "wireheading" and "war" (cont'd)
+If we actually had magical sex change technology of the kind described in https://www.lesswrong.com/posts/QZs4vkC7cbyjL9XA9/changing-emotions, no one would even consider clever philosophy arguments about how to redefine words: people who wanted to change sex would just do it, and everyone else would use the corresponding language, not as a favor, but because it straightforwardly described reality
+ScottAlexander — 12/25/2019
+(agreed so far, will let you continue)
+zackmdavis — 12/25/2019
+and similarly, I don't want to call Jessica a man, but that's because her transition actually worked; the ordinary means by which my brain notices people's secondary sex characteristics and assigns "man"/"woman"/"not sure" and decides which pronoun to use has been successfully "fooled". If it looks like a duck, and quacks like a duck, and you can model it as a duck without making any grievious prediction errors, then it makes sense to call it a "duck" within the range of circumstances in which that model continues to perform well, even if someone considering different circumstances (looking at a "thicker" subspace of configuration space) would insist that they need to model it as a duck-oid robot or that that species is actually a goose
+ScottAlexander — 12/25/2019
+(are you going to continue, or is that the end?)
+zackmdavis — 12/25/2019
+um, I have thousands of words like this in my notes
+ScottAlexander — 12/25/2019
+I guess I asked because I don't understand how you supported the "either wireheading or war" claim (which I just realized you might have thought I agreed with, I only meant I agreed with the instant sex change thing), and I don't understand how this answered my question.
+zackmdavis — 12/25/2019
+right, "wireheading" insofar as if I were to transition today and I didn't pass as well as Jessica, and everyone felt obligated to call me a woman, they would be wireheading/benevolently-gaslighting me
+making me think my transition were successful, even though it actually wasn't
+that's ... not actually a nice thing to do to a rationalist
+ScottAlexander — 12/25/2019
+Like, my claim is that transgender people have some weird thing going on in their head that gives them an electric shock if you refer to them as their birth gender. The thing isn't an agent and so anti-extortion rules don't apply. So you consider the harms of electric shocking someone a bunch of times, vs. the harms of remembering to say a different word than the one that automatically comes to mind (which is the same problem I face everytime I don't tell a stupid person FUCK YOU YOU ARE AN IDIOT, and isn't a huge cognitive burden), and accept The Nation making specious arguments which it would probably do about something else anyway, and overall you come out ahead.
+I think negative connotations of wireheading are doing a lot of the work there? If I change the metaphor to "give a depressed person a good antidepressant that works for them", I don't think it changes the situation at all but it definitely changes the valence.
+zackmdavis — 12/25/2019
+so, I claim to know a lot of things about the etiology of trans about why I don't think the electric shock is inevitable, but I don't want the conversation to go there if it doesn't have to, because I don't have to ragequit the so-called rationalist community over a complicated emprical thing; I'm only required to ragequit over bad philosophy
+ScottAlexander — 12/25/2019
+Fair.
+zackmdavis — 12/25/2019
+I think the philosophy crux is likely to be something like: in order to do utilitarianism, you need to have accurate maps/models, so that you can compute what the best thing to do is. If you let utilitarian considerations corrupt our models of the world, then you don't actually maximize utility; you're just gaslighting/wireheading yourself (I know, negative connotations of those words)
+You might think you can change language without changing models, but "37 Words Ways Can Be Wrong" explicitly and at length explains why you can't
+ScottAlexander — 12/25/2019
+I agree with that, but I feel like I'm proposing something that may be a completely-delimited 0.0001% decrease in world model understanding for a rather substantial increase in utility. Do you disagree with my numbers, or with the claim that it's ever acceptable to make that kind of a tradeoff at all?
+zackmdavis — 12/25/2019
+Where did you get that 0.0001% from?!?!!
+It's not 0.0001% for me (I've been in constant psychological pain for three years over this), but maybe there aren't enough copies of me for the utilitarian calculus to care
+ScottAlexander — 12/25/2019
+I currently call trans people by their self-identified gender and don't feel like my world-model has changed much.
+(are you saying it hasn't been a 0.0001% utility decrease for you, or a 0.0001% world-model-clarity decrease?)
+Hm, I only want to debate this if it were our actual crux, do you feel like it is?
+zackmdavis — 12/25/2019
+empirical magnitude of trans is not a crux
+ScottAlexander — 12/25/2019
+So getting back to my point that I feel like I'm making a very lucrative utility vs world-model-clarity tradeoff, do you think you should never do that regardless of how good the numbers are?
+(I actually think I would agree with you on this if I felt like it was even a noticeable world-model-clarity decrease, or if I thought it had the potential to snowball in terms of my Parable of Lightning, it just doesn't even really register as a clarity decrease to me)
+zackmdavis — 12/25/2019
+... uh, what sex people are is pretty relevant to human social life
+much more so than whether lightning comes before thunder
+ScottAlexander — 12/25/2019
+Yeah, but anyone who cares about it routes around it.
+zackmdavis — 12/25/2019
+so, the problem with the fictional "thunder comes before lightning" regime is that they didn't choose the utilitarian-optimal truth to declare heresy???
+ScottAlexander — 12/25/2019
+Like I'm straight and I don't date transwomen. If for some reason I got "tricked" into dating transwomen by linguistic philosophy (not by lack of knowledge of who was biologically female or not) then I wouldn't regard this as a failure, I would regard this as linguistic philosophy changing my sexual orientation.
+zackmdavis — 12/25/2019
+http://unremediatedgender.space/2019/Dec/more-schelling/
+ScottAlexander — 12/25/2019
+I think their problem was that they actually made people ignorant of things. I don't feel like anyone is being made ignorant of anything by the transgender thing, their thoughts are just following a slightly longer path.
+zackmdavis — 12/25/2019
+I have a paragraph in my notes about this, one moment
+The "national borders" metaphor is particularly galling if—[unlike](https://slatestarcodex.com/2015/01/31/the-parable-of-the-talents/) [the](https://slatestarcodex.com/2013/06/30/the-lottery-of-fascinations/) Popular Author—you actually know the math.
+Slate Star Codex
+Scott Alexander
+The Parable Of The Talents
+[Content note: scrupulosity and self-esteem triggers, IQ, brief discussion of weight and dieting. Not good for growth mindset.] I. I sometimes blog about research into IQ and human intelligence. I …
+Image
+Slate Star Codex
+Scott Alexander
+The Lottery of Fascinations
+I. Suppose I were to come out tomorrow as gay. I have amazing and wonderful friends, and I certainly wouldn’t expect them to hate me forever or tell me to burn in Hell or anything like that. …
+Image
+If I have a "blegg" concept for blue egg-shaped objects—uh, this is [our](https://www.lesswrong.com/posts/4FcxgdvdQP45D6Skg/disguised-queries) [standard](https://www.lesswrong.com/posts/yFDKvfN6D87Tf5J9f/neural-categories) [example](https://www.lesswrong.com/posts/yA4gF5KrboK2m2Xu7/how-an-algorithm-feels-from-inside), just [roll with it](http://unremediatedgender.space/2018/Feb/blegg-mode/)—what that means is that (at some appropriate level of abstraction) there's a little [Bayesian network](https://www.lesswrong.com/posts/hzuSDMx7pd2uxFc5w/causal-diagrams-and-causal-models) in my head with "blueness" and "eggness" observation nodes hooked up to a central "blegg" category-membership node, such that if I see a black-and-white photograph of an egg-shaped object, I can use the observation of its shape to update my beliefs about its blegg-category-membership, and then use my beliefs about category-membership to update my beliefs about its blueness. This cognitive algorithm is useful if we live in a world where objects that have the appropriate statistical structure—if the joint distribution P(blegg, blueness, eggness) approximately factorizes as P(blegg)·P(blueness|blegg)·P(eggness|blegg).
+Disguised Queries - LessWrong 2.0
+Imagine that you have a peculiar job in a peculiar factory: Your task is to take
+objects from a mysterious conveyor belt, and sort the objects into two bins.
+When you first arrive, Susan the Senior Sorter explains to you that blue
+egg-shaped objects are called "bleggs" and go...
+Image
+Neural Categories - LessWrong 2.0
+In Disguised Queries [/lw/nm/disguised_queries/], I talked about a
+classification task of "bleggs" and "rubes". The typical blegg is blue,
+egg-shaped, furred, flexible, opaque, glows in the dark, and contains vanadium.
+The typical rube is red, cube-shaped, smooth, hard, trans...
+Image
+How An Algorithm Feels From Inside - LessWrong 2.0
+"If a tree falls in the forest, and no one hears it, does it make a sound?" I
+remember seeing an actual argument get started on this subject—a fully naive
+argument that went nowhere near Berkeleyan subjectivism. Just:
+
+"It makes a sound, just like any other falling tree!"
+"...
+Image
+Causal Diagrams and Causal Models - LessWrong 2.0
+Suppose a general-population survey shows that people who exercise less, weigh
+more. You don't have any known direction of time in the data - you don't know
+which came first, the increased weight or the diminished exercise. And you
+didn't randomly assign half the population t...
+Image
+ScottAlexander — 12/25/2019
+I'm still interested in whether you think, if I were correct about the extent of the minimal inconvenience and high utility gain, I would be correct to have the position I do.
+zackmdavis — 12/25/2019
+"Category boundaries" are just a visual metaphor for the math: the set of things I'll classify as a blegg with probability greater than p is conveniently visualized as an area with a boundary in blueness–eggness space. If you don't understand the relevant math and philosophy—or are pretending not to understand only and exactly when it's politically convenient—you might think you can redraw the boundary any way you want, but you can't, because the "boundary" visualization is derived from a statistical model which corresponds to empirically testable predictions about the real world. Fucking with category boundaries corresponds to fucking with the model, which corresponds to fucking with your ability to interpret sensory data. The only two reasons you could possibly want to do this would be to wirehead yourself (corrupt your map to make the territory look nicer than it really is, making yourself feel happier at the cost of sabotaging your ability to navigate the real world) or as information warfare (corrupt shared maps to sabotage other agents' ability to navigate the real world, in a way such that you benefit from their confusion).
+ScottAlexander — 12/25/2019
+Now we're getting back to the stupid dolphin thing.
+I don't want to accuse you of thinking God has a dictionary, but it sure sounds like you're saying that God has pretty strong opinions on which categories are more natural than others.
+...actually, that's unfair, God does have some opinions, just not as deterministic as you think.
+zackmdavis — 12/25/2019
+I mean, yes? Thingspace is really high-dimensional; the multivariate distributions for females and males are actually different distributions, even if there's a lot of overlap along many of them; your brain is going to want to use this information and want to use language to communicate this information; although maybe an alien surveying earth to decide which rocks to harvest wouldn't bother noticing the difference, because it's only looking at rocks and not the differences between apes https://www.lesswrong.com/posts/cu7YY7WdgJBs3DpmJ/the-univariate-fallacy
+ScottAlexander — 12/25/2019
+Also, wait a second, you wouldn't want Jessica competing in a woman's marathon either, so your own categorization system doesn't solve this problem any better than mine.
+zackmdavis — 12/25/2019
+I'm willing to code-switch depending on what I'm talking about
+I definitely want to say "Jessica ... she"
+because of how my visual system makes that call
+ScottAlexander — 12/25/2019
+Do you agree that gendering trans people based on whether they pass or not doesn't work on a society-wide level?
+zackmdavis — 12/25/2019
+unfortunately, yes; that's why this situation is such a horrible, tragic mess; this is where game theory comes in: http://unremediatedgender.space/2019/Oct/self-identity-is-a-schelling-point/
+ScottAlexander — 12/25/2019
+But it sounds like you still disagree with self-identity?
+zackmdavis — 12/25/2019
+I'm not trying to do policy; I'm trying to get the theory right, in the hopes that getting the theory right will help people make better decisions
+ScottAlexander — 12/25/2019
+....but it not being a stable Schelling point is one example of why it would be a bad decision
+I kind of am trying to do policy, in that what seems important to me is whether myself (and the government, and the society whose norms I try to obey insofar as they're socially-just game theoretic things) should gender trans people one way or another.
+If this reflects badly on AI theory, I would explain to AI designers why there are considerations in favor of them doing things differently.
+Which I think would be pretty easy.
+zackmdavis — 12/25/2019
+... you would have AI designers take into account what makes trans people feel better?
+rather than capturing the statistical structure of the world beneath the world?
+ScottAlexander — 12/25/2019
+...I think you may be taking the opposite point I intended?
+zackmdavis — 12/25/2019
+oh
+ScottAlexander — 12/25/2019
+I am trying to (pursue the self-identification criterion) because it's the right thing for me to do right now as a human individual for various reasons. If it's bad AI design, I would want AI designers to do something different.
+(though I would hope a superintelligent AI would gender trans people based on self-identification for the same reasons I would)
+zackmdavis — 12/25/2019
+if you're going to be a generic Berkeley progressive, that makes sense, but if you're going to be a rationalist, then it doesn't make sense, because the entire point of our community is to import insights from AI theory to make humans smarter
+ScottAlexander — 12/25/2019
+Wait, no, the entire point of our community is to do the correct thing!
+I worry I'm interpreting you wrong, because it sounds like you're saying "You shouldn't eat food, because an AI wouldn't eat food. You should connect yourself to a wall socket."
+No, that's not right, "You shouldn't eat food because you wouldn't want an AI designer to program their AI to eat food"
+zackmdavis — 12/25/2019
+okay, my previous message is not phrased well; if I were writing a long form rather than live Discord I would have rewritten it to be less mockable
+ScottAlexander — 12/25/2019
+Fair.
+zackmdavis — 12/25/2019
+Here's another couple paragraphs from my memoir draft:
+zackmdavis — 12/25/2019
+A friend—call her ["Erin Burr"](https://genius.com/7888863)—tells me that I'm delusional to expect so much from "the community", that the original vision never included tackling politically sensitive subjects. (I remember Erin recommending Paul Graham's ["What You Can't Say"](http://www.paulgraham.com/say.html) back in 'aught-nine, with the suggestion to take Graham's advice to figure out what you can't say, and then don't say it.)
+Perhaps so. But back in 2009, we did not anticipate that whether or not I should cut my dick off would become a politicized issue.
+To be fair, it's not obvious that I shouldn't cut my dick off! A lot of people seem to be doing it nowadays, and a lot of them seem pretty happy! But in order to decide whether to join them, I need accurate information. I need an honest accounting of the costs and benefits of transition, so that I can cut my dick off in the possible worlds where that's a good idea, and not cut my dick off in the possible worlds where it's not a good idea.
+ScottAlexander — 12/25/2019
+But I should do the thing which is morally and rationally correct right now, not the thing that will create good results if AI designers inexplicably program a tendency to do the exact same thing into their AI, in a stupid way that changes the entire architecture.
+zackmdavis — 12/25/2019
+Your position seems to be, "It's okay to distort our models, because the utilitarian benefit to this-and-such neurotype interest group makes it a net win", but I'm claiming that I'm in the neurotype interest group and I want non-distorted models
+I don't want people to have to doublethink around their perceptions of me
+Jessica doesn't want that, either
+ScottAlexander — 12/25/2019
+I can see why you think that, but I don't identify with that belief, partly because there are certain ways I wouldn't distort my model. The trans thing seems so much like using a different word, rather than distorting a model, that it's really hard for me to care.
+I definitely don't want to make it impossible to accurately view the territory, but this seems more like (forgive a totally different metaphor than the one I've used before) rebranding Mt. McKinley to Denali on my map, with everyone knowing that these are the same mountain.
+zackmdavis — 12/25/2019
+the entire thesis of "37 Ways Words Can Be Wrong" is that "I can use a different word without distorting my model" isn't true as a matter of human psychology
+if we learned something new about how language works in the last 10 years, that would be really interesting
+ScottAlexander — 12/25/2019
+Yet I feel like I would come to all the same conclusions about men and women as you would.
+zackmdavis — 12/25/2019
+when System 2 knows what the answer your questioner expects is
+ScottAlexander — 12/25/2019
+I think I mean "about the territory", which I think is independent of what my questioner expects.
+zackmdavis — 12/25/2019
+Here's someone who's not part of our philosophy cult who understands the point I'm making: https://fairplayforwomen.com/pronouns/
+author compares preferred pronouns to a Stroop test
+ScottAlexander — 12/25/2019
+The whole interesting thing about Stroop tests is that you can't do them if you try to do them super-quickly, but you don't end up fundamentally confused about what colors are.
+I agree if you gave me an IAT about trans people and pronouns, I would do worse than you (probably)
+But IAT doesn't correlate with actual racism, and I think this is the same kind of situation.
+zackmdavis — 12/25/2019
+If you're sufficiently clever and careful and you remember how language worked when Airstrip One was still Britain, then you can express things using Newspeak, but a culture in which Newspeak is mandatory, and all of Oceania's best philosophers have clever arguments for why Newspeak doesn't distort people's beliefs ... doesn't seem like a nice place to live, right?
+Doesn't seem like a culture that can solve AI alignment, right?
+Zvi wrote about a pattern where people claim that "Everybody knows" something, being motivatedly used as an excuse to silence people trying to point out the thing (because they don't see people behaving as if it were common knowledge) https://thezvi.wordpress.com/2019/07/02/everybody-knows/
+"'Everybody knows' our kind of trans women are sampled from the male multivariate distribution rather than the female multivariate distribution, why are you being a jerk and pointing this out"
+I really don't think everyone knows
+I think the people who sort-of know are being intimidated into doublethinking around what they sort-of know
+I think this is bad for clarity
+ScottAlexander — 12/25/2019
+That is why I keep stressing that this is the only thing I know of that works this way. It's not opening the door to total confusion. If we randomized half the people at OpenAI to use trans pronouns one way, and the other half to use it the other way, do you think they would end up with significantly different productivity?
+zackmdavis — 12/25/2019
+(let me think for a few moments about how to answer that one)
+ScottAlexander — 12/25/2019
+I actually want to go to the Christmas party today. Any interest in also coming and we can continue the discussion there?
+zackmdavis — 12/25/2019
+at Valinor? I've been avoiding because I haven't gotten my flu shot yet because I've been too caught up in the rationalist civil war to have any executive funcitoning left over
+ScottAlexander — 12/25/2019
+No, at Event Horizon.
+zackmdavis — 12/25/2019
+sure, remind me the address?
+(Thanks so much for your time; it would be really exciting if we could prevent a rationalist civil war; I think this "you need accurate models before you can do utilitarianism" general philosophy point is also at the root of Ben Hoffman's objections to EA movement)
+ScottAlexander — 12/25/2019
+(actually, I just saw something on the party thing saying "Ask us before you invite anyone else", so give me a second to ask them)
+I agree that the principle "It's important to have good models because otherwise we can't do anything at all" is a principle. But it seems to me, like other principles, as something that's finite and quantifiable, as opposed to "if you sacrifice cognitive categorization efficiency even 0.01%, everything will be terrible forever and you'll fail at AI". In the same way that "make a desperate effort" is a principle, but I would still balance this against burnout or something, I feel like "never do anything that messes with cognitive categorization efficiency" should be followed but not off a cliff. I still don't feel like I understand whether our crux is absolutism vs. practicality, you just think this is legitimately a much bigger deal than I do, or something downstream of the dolphin problem where you think categories are much less inherently ambiguous than I do.
+Like, suppose Benya Fallenstein gets angry at us for constantly misgendering them and leaves MIRI. Is that better or worse than whatever epistemic clarity we gain from having a better split-second Stroop reaction to gender issues?
+zackmdavis — 12/25/2019
+Eliezer has this pedagogy thing where he'll phrase things in an absolute manner because that's the only way to get people to take things seriously, even if it's not literally absolute in human practice
+thus, "you can't define a word any way you want", https://www.lesswrong.com/posts/WLJwTJ7uGPA5Qphbp/trying-to-try, &c.
+ScottAlexander — 12/25/2019
+(it's possibly unfair for me to point to Benya, and I'm definitely not trying to become even more pragmatic than I was a moment ago talking about being nice to people, I'm trying to head off a problem where AI is so much more important than people's feelings that we end up Pascaling ourselves)
+zackmdavis — 12/25/2019
+maybe all three of those are partial cruxes
+like, when you say "trans is the only issue for which this comes up, it isn't a big deal for me" ... you realize that I do have to functionally treat that "part of an attempt to destroy [me]". You don't want to destroy me, but you're willing to accept it as a side-effect
+ScottAlexander — 12/25/2019
+I think I can predict that you would say something like that, but it still doesn't make sense to me.
+I think I'm modeling you as having the weird electric shock thing, but it fires opposite as for everyone else.
+But I'm pretty sure that's not fair.
+Obviously I will address you by whatever pronouns you prefer and so on, but yeah, I acknowledge that if trying to help everyone else hurts you, then I am going to hurt you.
+zackmdavis — 12/25/2019
+it might be related! I think I have the same neuro-quirk that leads to the trans women we know, but I got socialized differently by my reading
+ScottAlexander — 12/25/2019
+Part of what I'm trying to do is convince you not to care about this.
+My real position is "all of this is dumb, go through life doing things without caring what your gender is", and so I model some sort of weird electric shock thing that prevents people from doing this.
+I feel like because you are at least saying that what you're doing is based on principles, maybe if we can converge on principles I can help you.
+zackmdavis — 12/25/2019
+I read https://www.lesswrong.com/posts/QZs4vkC7cbyjL9XA9/changing-emotions 10 years ago and it had a huge effect on me
+ScottAlexander — 12/25/2019
+But I am also trying really hard not try to help you directly because I think that made you feel a lot worse the last time we talked than just talking philosophy of language.
+I think a meta-position of mine is something like "people fail to converge on object-level issues for unclear reasons, so try to figure out a set of norms that let people with unconverged object-level positions continue to cooperate", but I feel like you're fighting me on that one too.
+zackmdavis — 12/25/2019
+the "category absolutism" sub-crux seems legitimately important to the Great Common Task
+the "you just think this is legitimately a much bigger deal than I do" sub-crux isn't
+the extent to which dolphins are ambiguous is
+ScottAlexander — 12/25/2019
+But it sounds like you're not a 100% Category Absolutist, and you would be willing to let a few tiny things through, and I worry our real difference is I think this is tiny enough to ignore, and you don't.
+zackmdavis — 12/25/2019
+I think your self-report of "this is tiny enough to ignore" is blatantly false (Jessica would say you're lying, but I tend to avoid that usage of "lie" except sometimes when talking to Jessica et al.)
+ScottAlexander — 12/25/2019
+As in I'm misdescribing my experience, or failing to notice ways that my experience is unrepresentative of broader concerns?
+zackmdavis — 12/25/2019
+I'm saying that your brain notices when trans women don't pass, and this affects your probabilistic anticipations about them; your decisions towards them, &c.; when you find out that a passing trans women is trans, then also affects your probabilistic anticipations, &c.
+ScottAlexander — 12/25/2019
+Plausibly yes.
+zackmdavis — 12/25/2019
+This is technically consistent with "tiny enough to ignore" if you draw the category boundaries of "tiny" and "ignore" the right away in order to force that sentence to be true
+... you see the problem; if I take the things you wrote in "Categories Were Made" literally, then I can make any sentence true by driving a truck through the noncentral fallacy
+ScottAlexander — 12/25/2019
+I want to think about that statement more, but they've approved my invitation of you to the party. 2412 Martin Luther King, I'll see you there.
+zackmdavis — 12/25/2019
+cool, thanks so much; I really want to avoid a civil war
+ScottAlexander — 12/25/2019
+I also want to debate the meta-question of whether we should have a civil war about this, which I think is much easier to answer and resolve than the object-level one.
+
+------
projects / Ultimately_Untrue_Thought.git / blobdiff