[Ultimately_Untrue_Thought.git] / notes / a-hill-email-review.md

Email timeline—
1 Dec 2018: "super-presumptuous mail about categorization and the influence graph" to Scott cc Anna Jonah Sarah Michael
... discussion with Scott runs to 3 Dec
3 Dec: forward Scott thread to Ben and Katie, privacy leak
4 Dec: Ben—bullshit nitpicking is an implied ostracism threat
5 Dec: Ben—the interesting part is your model of the "rationalists" is wrong
6 Dec: Michael—this is important, share with Aurora?
10 Dec: asking Anna for her take
11 Dec: social proof from Sarah to Scott
11 Dec: Michael—argument should be compelling unless Craft Is Not the Community
12 Dec: emailing Harmanas after running into him at the train
16 Dec: to Sarah and Michael, "I still owe you" first draft
16 Dec: Michael—I want EY to respond to arguments, but wait a few weeks?
18–19 Dec: to Anna and Michael, on conversational violence and self-defense
19 Dec: met Aurora
21 Dec: to Sarah, something to protect
21 Dec: Sarah—Let's Get This Gender Thing Over With (unpublished)
24 Dec: ... further discussion with Ben/Sarah/Michael
26 Dec: my intelligence notes on Ziz
30 Dec: Ben—phatic declaration of respect makes me feel less safe
1 Jan: final draft of "I still owe you money"
1 Jan: Eliezer has taken initiative to spend Eliezerself's time, Rob as police
3 Jan: meatspace met with Michael/Sarah/Auora/Ben/Andrew
5 Jan: met with Ziz and Gwen with Michael and Aurora, lawsuits 
6 Jan: backchannel to Anna about Ziz
6 Jan: EY thread cont'd
9 Jan: I want to see you be brave (was: Re: I still owe you money; and, discourse on categories and the fourth virtue)
11 Jan: me—What do think submitting to social pressure looks like?!
15 Jan: unsent "thunder comes before gaslighting" draft
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)
20 Feb: cold outreach to Ben Landau-Taylor
21 Feb: telling Marcus, I'm out of cards: we created a monster
22 Feb: "RIP Culture War" thread
22 Feb: Michael offers advice; I forward with Scott; Michael calls it a destructive move
25 Feb: first draft of defending against "alt-right" categorization
27 Feb: more support from the gang
27-28 Feb: on private universes
1 Mar: intra-group disagreement re "ambient epistemic violence"
2 Mar: Michael—independence of perspectives vs. failure to present an united front
2 Mar: I saw the new Rebel Wilson movie last night, went to Simon's birthday (I remember seeing Scott there), then met Anna
2 Mar: Anna's comment about Michael
3 Mar: "Apologies are always intrinsically worthless, but that fact that we are able to legitimately demand an apology means that we can legitimately demand that he look at something."
3 Mar: I heard Anna say my desire for good arugments to win was "selfish"; what she actually said was "self-centered", "it seems less important than what happens on topics that we don't all know there isn't free speech about."
3 Mar: "told them that the thing that would most help AI risk from them anyhow was for them to continue on their present career," is that if I try to only focus on "important" things, I'll probably just come up with something fake
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
14 Mar: meeting with Ben/Sarah/Marie/Jessica; Ben assist on "Blegg Mode"
17 Mar: me to Scott, bringing up idea that in-person ideas were better
17 Mar: Michael characterizing EY as choosing to be a cult leader
18 Mar: starting to overheat (subject: "strategy huddle II")—
I referred to Sarah and Ben at Sunday conversation saying in-person meetings were better
Scott says "I try to avoid in-person meetings in general, and this is probably in my bottom 1% of things that I would enjoy talking about in person"
I write to my team wondering if I should propose a $4000 happy price
then I say, maybe don't reply before Friday
6 p.m.: really angry at Scott for not understanding the difference between cluster in a value-dependent subspace (as with the dagim/fish example), and value-dependent to not make my friends sad. But my boss was expecting code by noon tomorrow
7 p.m.: let go of the anger somehow and just do my job this week and then write up the mathy version
8:44 p.m.: fantasizing about recruiting Buck or Duncan to our coalition; never sure how to interpret my self-reports; even when crying, shaking, randomly yelling, &c. I think I'm still noticeably incentivizable: I think that if someone were to present a credible threat (e.g., slapping me and telling me to snap out of it), then I would be able to calm down: there's some sort of game-theory algorithm in the brain that subjectively feels genuine distress (like crying or sending people too many hysterical emails) but only when it can predict that it will be tolerated
10:27 p.m.: I tweeted a Sequences quote that summarizes how I'm feeling (mention is to attribute credit; Eliezer has so many followers that he probably won't even see the mention notification), except it's not quite appropriate, because the quote is about failure resulting in the need to invent new methods of rationality, better than the ones you were taught. But ... the methods I was taught are great! I don't have a pressing need to improve on them! I just can't cope with everyone else having forgotten!
11:01 p.m.: OK, the dayjob situation is really bad right now (it's almost 2300), so let's try one of those commitment-contract things. I hereby pledge not to check my personal email, Facebook, or Twitter until 1700 Pacific Daylight time on Friday the twenty-second of March. (I have an /etc/hosts manipulation script and I can turn off my home computer to help.) If I fail in this pledge, then ... um, I'm having trouble thinking of a good incentive here. (Traditionally you use money, but unfortunately, I'm only 78% Ashkenazi and my brain doesn't believe money is real.) So, um, "or else"!
2:47-3:55 a.m.: some replies from Michael
4:01 a.m.: I got the unit test passing; why do I keeping lying about email hiatus?
4:15 a.m.: "predictably bad ideas" email to Anna/Michael/Ben/Sarah/Zvi/Scott/Alicorn/Mike
5:26 a.m.: five impulsive points
5:58 a.m.: everyone's first priority should be to make sure Zack can sleep
6:14 a.m.: I'm really sorry about this. I don't need Scott's help and I think Michael is being a little aggressive about that, but I guess that's also kind of Michael's style?
6:18 a.m.: Michael's response: What the FUCK Zack!?!
7:27 a.m.: Home now. Going to get in bed. I could say more, but that would be high-variance, and we don't want variance right now
19 Mar: Michael seems disappointed with me for what he perceived as me escalating and then deescalating just after he came to help, but which from my perspective felt like me just trying to communicate that I don't want to impose too many costs on my friends just because I felt upset today. (Subject: "yet another strategy huddle (III)")
19 Mar: Ben writes off Scott and ignores a boundary; nominates Jessica for helping me
19 Mar: maybe there's a reason to be attached to the "rationalist" brand name/social identity, because my vocabulary has been trained really hard on this subculture; theistic evolution Church analogy
20 Mar: planning to meet Zvi in person for first time
20 Mar: Ben asks Scott to alter the beacon (I think this became https://slatestarcodex.com/2019/07/04/some-clarifications-on-rationalist-blogging/ )
20 Mar: super-proton concession (why now? maybe he saw my "tools shattered" Tweet, or maybe the Quillette article just happened to be timely)
21 Mar: I think of the proton criticism; nominate Ben as mission commander
21 Mar: I suggest to Anna that chiming in would be a movement towards peace with Michael's gang
21 Mar: Ben and Michael don't think the techincal quibble makes any progress
22 Mar: I literally ran into Scott on the train
22 Mar: Scott does understand the gerrymandering/unnaturalness problem, but it sort of seems like he doesn't understand that this is about cognitive algorithms rather than verbal behavior ("someone tells you that classifying a burrito as a sandwich will actually save ten million lives and ensure friendly AI in our lifetime, would you still refuse to classify a burrito as a sandwich?"—well, it doesn't matter what I say), whereas Eliezer definitely understands
24 Mar: Michael to me on Anna as cult leader
24 Mar: I tell Michael that I might be better at taking his feedback if he's gentler
30 Mar: hang out with Jessica (previous week was Zvi and Nick and anti-fascist Purim)
30 Mar: Michael—we need to figure out how to win against bad faith arguments
30 Mar: wrapping up drafts
me to Scott: _what_ post about transgender
conceptual gerrymandering is not unique
it's fine if you don't help; you've suffered enough
Kelsey agreeing on the object level for a different reason
crippling the orcs' general ability to reason
early draft of what became "Where to Draw the Boundaries?"
31 Mar: the real issue is whether we want to destroy everyone's 
31 Mar: Sarah—isn't the math overkill?
31 Mar: Ben thinks I'm reading Kelsey wrong, that she's counting happiness as part of usability
31 Mar: I think that math is important for principle—and intimidation (https://slatestarcodex.com/2014/08/10/getting-eulered/)
31 Mar: "a very vague understanding of what you're trying to do with the second email but it seems kind of sketchy to me, like you're writing an email to Scott meant to look persuasive to Eliezer"
31 Mar: I want a court ruling; I'm betting that it's psychologically harder for someone who knows the Rules to continue being dishonest against someone who has the correct answer and isn't willing to give up on good-faith argument.
31 Mar: Ben— this just seemed like the kind of bank shot where, given the track record of Zack's intuitions, I expect it to fail unless he's explicitly modeling what he's doing
31 Mar: email prosecution not useful, explicitly asking for a writ of certiorari is better than sending "cc in case you're interested" mail and delusionally hoping for it to function as an appeal.
31 Mar: I recount to Michael Jessica's account of her time at MIRI
31 Mar: Michael—"since there is absolutely no plausible narrative in which I was bringing them into a cult, there is no symmetry in our narratives." this didn't age well
31 Mar: Ben—the most important thing for LessWrong readers to know about - a similarly careful explanation of why you've written off Scott, Eliezer, and the "Rationalists"
31 Mar: I'm reluctant, because I don't want to attack my friends in public (and don't want to leak info from private conversations)
31 Mar: Ben—self-promotion imposes a cost on others
1 Apr: fantasizing Susan the Senior sorter parable turing violent
1 Apr: Avoiding politically-expensive topics is fine! Fabricating bad epistemology to support a politically-convenient position and then not retracting it after someone puts a lot of effort into explaining the problem is not OK. Some of the things you said on the phone call made it sound like you didn't understand that I'm making that distinction?
2 Apr: me to Anna cc Ben—"Reasoning in public is too expensive; reasoning in private is good enough" is less damaging when there's some sort of recruiting pipeline from the public into the monasteries
2 Apr: Ben: I do not consider the claim that intact monasteries exist to be plausible at all right now.
2 Apr: Twitter poll about dropping pseudonymity
3 Apr: I don't see any advantage to speaking non-anonymously when people are persecuting you. I also find it REALLY DIFFICULT to deal with the fact that you are cooperating with people who are openly defecting against you and against me. 
6 Apr: "Where to Draw the Boundaries?" draft, Putting It Together epigraph; plan to beg Anna and Eliezer for endorsement afterwards
8 Apr: Jessica—creating clarity about behavioral patterns, and specifically ones that people claim not to have.  What you're saying here implies that you believe creating such clarity is an attack on someone.  If so, your blog is attacking trans people.
8 Apr: the distinction between truthfully, publicly criticizing group identities and named individuals still seems very significant to me
8 Apr: Michael—less precise is more violent
"Boundaries" discussion ...
9 Apr: "VP" is also political
10 Apr: Ben and Jessica coin four simulacra levels
10 Apr: Ben—Eliezer's in denial about whether he's making decisions out of fear. / Scott's only unevenly in denial about this.
me—Anna claims to be calculating rather than fearful. I don't have any particular evidence to dispute her self-report, but I would asphyxiate under that level of risk aversion.
12 Apr: Ben again on, " to trying to explain why Eliezer seems to you like the court of last resort, what he's done that leads you to expect him to play this role, etc."; more money seems wrong
12 Apr: I'm still too cowardly
Sarah explaining how to ask things of people in general
Ben reiterating, "You don't want either an overtly paid endorsement, or to ask Eliezer to lie about whether it's a paid endorsement."
12 Apr: Sarah—what would you do if Eliezer never follows up?
13 Apr: me—I don't know, what would there be left to do?—I don't have any particular claim over Eliezer.
13 Apr: pull the trigger on "Boundaries?" (for this reason it is written: don't hate the player, hate the game-theoretic nature of all life.)
ask Anna
direct marketing to Nick T.
direct marketing and/or conspiracy invitation
court filing with Yudkowsky (Subj: "movement to clarity; or, rationality court filing")
Ben: So, um, this _is_ politics, and we _aren't_ safe, so how about not lying?
Michael: We are safe to understand arguments and logic, and it is Not politics to try to pin understanding down. It is politics to try to pin down the understanding of people as part of credit allocation, and we should be doing politics to try to motivate engagement here, but so far, Zack has been doing a mix of friendship, supplication and economics.  
me: Is ... is there a good place for me to _practice_ politics against someone who isn't Eliezer Yudkowsky? The guy rewrote my personality over the internet. I want more credit-allocation for how relatively well I'm performing under constrained cognition!
me: Like, okay, I see the case that I'm doing a sneaky evil deceptive thing (bad faith is a disposition, not a feeling!) by including subtext and expecting people to only engage with the text, but if we're not going to get into full-on gender-politics on Less Wrong, but gender politics is motivating an epistemology error, I'm not sure what else I'm supposed to do! I'm pretty constrained here!
Michael: Once people think they know what the political sides are, the political thing to do is (trivially, I am annoyed at you pretending not to know this) to be completely unprincipled and partisan in any anonymous situation.
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 [...] 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 Apr: Ben on our epistemic state
18 Apr: me—I'd rather say "Actually, I don't think I'm saner than Jessica, because ..." rather than "You owe Jessica an apology."
18 Apr: Ben to me—"It's sketchy to accept insults leveled at you, but it’s actually a betrayal of trust if you ask friends to back you up and then don’t back them up"
me— You're right; I hereby retract the phrase "or my friends" from the prior email. I was trying to convey a "turn the other cheek"/"social reality is fake, anyway" attitude, but I agree that everyone can only turn her own cheek.
18 Apr: Ben—STAY IN FORMATION
me— I've been struggling with how to "chime in" for hours, but it looks like Eliezer just shut the door on us at 10:12 p.m. / Mission failed?
This closure actually makes me feel a little bit less heartbroken and cognitively-dissonant about the "rationalists" already. If Scott doesn't want to talk, and Eliezer doesn't want to talk, and I'm unambiguously right ... then I guess that's that?
Michael—common knowledge achieved
19 Apr: Zack, that's not what a war looks like. That was ... trying to educate people who were trying not to be educated, by arguing in good faith.
me: Ben, I mean, you're right, but if my "Category War" metaphorical name is bad, does that also mean we have to give up on the "guided by the beauty of our weapons" metaphorical catchphrase? (Good-faith arguments aren't weapons!)
19 Apr: Ben and Jack "sadism and tactics" chat transcript
20 Apr: my closing thoughts 
20 Apr: Michael on Anna as an enemy
30 Apr: me—I don't know how to tell the story without (as I perceive it) escalating personal conflicts or leaking info from private conversations.
30 Apr: If we're in a war, can I apply for psych leave for like, at least a month? (Michael implied that this was OK earlier.) Or does asking get me shot for desertion? It's fair for Ben to ask for an accounting of why I'm stuck, but the fact that I'm sitting at my desk crying at 5:45 p.m. suggests a reordering of priorities for my own health.
30 Apr: No urgency on that timescale. Just don't buy a nerf gun and make shooting noises to pretend you're fighting. On leave means on leave.
3 May: dayjob workplace dynamics
6 May: multivariate clarification https://twitter.com/ESYudkowsky/status/1124751630937681922
7 May: angry at Anna for thinking I was delusional for expecting the basics
8 May: Anna doesn't want money from me—it'll appear in some Ben Hoffman essay as "using you" or some similar such thing
17 May: Jessica points out that Yudkowsky's pro-infanticide isn't really edgier than Peter Singer
21 May:
If someone claims, as someone did, that saying that words having meanings matters is politically loaded and is code for something, that can really be seen as [...] their problem.
I mean, ordinarily yes, but when the person saying that is a MIRI Research Associate, then it's the light cone's problem.
22 May: to Anna cc Sarah Steven—Viaweb and CfAR have importantly different missions (Sarah is Nice, unlike people like Ben or Michael or Jessica, who are Mean)
31 May: "Categories Were Made" LessWrong FAQ https://www.lesswrong.com/posts/MqrzczdGhQCRePgqN/feedback-requested-draft-of-a-new-about-welcome-page-for#iqEEme6M2JmZEXYAk
8 Jun: I think I subconsciously did an interesting political thing in appealing to my price for joining
8 Jun: maybe "EA Has a Lying Problem" seemed like a similarly pointless hit piece to you? And you have as much invested in the "Effective Altruism" brand name as I have in the "Michael Vassar" brand name?
17 Jun: "It was Dave Mitchum who now belonged on this railroad and he, Bill Brent, who did not." basically sums up how I feel about Ozy
22 Jun: I'm expressing a little bit of bitterness that a mole rats post got curated https://www.lesswrong.com/posts/fDKZZtTMTcGqvHnXd/naked-mole-rats-a-case-study-in-biological-weirdness
23 Jun: Schelling Categories, and Simple Membership Tests
25 Jun: scuffle with Ruby
26 Jun: Ben trying to talk to the Less Wrong team
29 Jun: I talked with Ray from 7 to 10:30 tonight. My intellectual performance was not so great and I did a lot of yelling and crying.
29 Jun: talked to Buck (who's working at MIRI these days) at the Asgard (Will and Divia's house) party tonight. He was broadly sympathetic to my plight, thinks EA leaders have it together (saying people like Holden had thought through more plies than him when he talks to them about the Issues), thinks public reason is hopeless, doesn't think much/well of Ben or Michael or Zvi.
1 Jul: I think we can avoid getting unnecessarily frustrated/angry at these people if we can be clear to ourselves and each other that they're MOPs, and could not possibly succeed in being geeks even if they tried
2 Jul: "Everyone Knows", pinging Anna about it
4 Jul:  https://slatestarcodex.com/2019/07/04/some-clarifications-on-rationalist-blogging/ published
11 Jul: AI timelines scam
11 Jul: me—I'm sympathetic to tone-policing on "scam"
Ben—What exactly is a scam, if it's not misinforming people systematically about what you have to offer, in a direction that moves resources towards you?
me—but the law does use mens rea, like the distinction between manslaughter and murder
me—call with Michael,  the Pope surely knows that he doesn't really have a direct line to God (despite holding office based on this pretense), to how GiveWell must know that they're not really making decisions based on numbers (despite holding credibility based on this premise)
16 Jul: https://slatestarcodex.com/2019/07/16/against-lie-inflation/
17 Jul: Alyssa's false claim about my pseudonyms, HRT effects side note
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"
as counterevidence, I brought up communication from Michael to me about it being a small matter whether he's welcome at REACH, that I should pick a better identity than "rationalist", that he doesn't "see a point in defending me from Anna when she is now on a side that is all but exploit in giving up on truth"
this is not the game someone would be playing if he were trying for inclusion
Prosecutor's reply: Stardust didn't prioritize REACH investigation until she found out from Ben's post that Vassar apparently did care
Defense calls Sarah to the stand
25 Jul: I write to the REACH panel address—What are the charges?
25 Jul optimal gossip, cont'd—
savviness as non-normative
Isn't "social reality" the one area you'd most expect the consensus to be systematically wrong about?
prosecution: more detail about out-of-touch with social reality
prosecution points out that "trans women are different from cis women" is universally agreed on
prosecution request defense informs Stardust that Michael said earlier he was leaving the country
25 July: REACH subcommittee: "To be clear, we are not a court of law addressing specific "charges." We're a subcommittee of the Berkeley REACH Panel tasked with making decisions that help keep the space and the community safe."
defense: "To be clear, a community that excludes Michael on inadequate evidence is one where I feel unsafe."
optimal gossip case: Michael as the catalyst
Ben as witness: "the basically fraudulent practice of "agreeing" to follow procedure, distorting the procedure to get the answer one prefers, and then "correcting" the error iff someone does the exact same lobbying they'd have had to do in the absence of any procedure."
5 Aug: ICE Tweets
12 Aug: I would like to see you write up the things you're trying to process, in more complete form, focusing on the details it's most important to you to understand
18 Aug: continuing to draft "Schelling Categories"
21 Aug: Yudkowsky throwing NRx under the bus https://twitter.com/zackmdavis/status/1164259164819845120
Jessica—Yudkowsky is right to denounce NRx if he actually doesn't like NRx; doesn't he have free speech, aren't you arguing for self-censorship? hostilities that actually exist should be declared, and that errors should be corrected
me—witches and heresies coordination
me—tragedy of recursive silencing
23 Aug: contrast this disavowal with the one in 2013
notice the inversion from (both paraphrased) "I don't support NRx" (fine, of course) to "I don't even want them supporting me" (!?!?!!?).
24 Aug: Anna's behavior would make sense if she thought FAI can be solved just by pointing a lot of eyeballs and money at the problem, OR "Compartmentalization is OK; dark side epistemology isn't really a thing; it's realistic for people to 'try to be wrong' when Society demands it, but be otherwise sane", but I don't think she would endorse either of those as stated
24 Aug: I had told Anna about Michael's "enemy combatants" metaphor, and how I originally misunderstood
3 Sep: positive vs. zero-sum metaphysics of communication
5 Sep: SK—openai and givewell are too outgroup for me to worry much about whether they're acting rationally
Ben's reaction—Wasn't Holden's post on MIRI for a while, like, the most-upvoted LessWrong post? Didn't Rationalists have quite a lot to do with OpenAI getting funding?
Jessica—the lines are not drawn anywhere near strongly enough to be able to say Those People over there are clearly dangerously in sane, but We over here are sane.
Michael— I don't know a natural boundary he could be using that includes him and MIRI but not FHI or Paul Christiano
'if you see fraud and don't say fraud (by an institution you have endorsed) then you're a fraud'
8 Sep: Heads I Win, Tails?—Never Heard of Her
Jessica—I mean, this all seems correct, but do you believe other people capable of parsing the math don't already know this?  It seems really obvious (like, to 10 year olds even) that partisan politics distorts impressions by filtering evidence.
me—I mean, do you think we could get a ten-year-old to explain it to Eliezer Yudkowsky?
I reference my "the story of how I gave up on the so-called 'rationalists'" grief-memoir post"—this was three years ago!!!!
Jessica—'it seemed like you were conflating "being a Nazi" with "having observed the world-conditions that Nazis would have found congenial".'
my no-contest bullet-biting strategy
10 Sep: I already told Ben the other day that I made the spaceships for him, so this is kind of double-counting: the spaceships are for everyone
12 Sep: state of Ben Pace
15 Sep: conflict with Anna—the fact that he's done so much good work is exactly what made him a target for pressure
15 Sep: offer money to talk to Scott, then reverse myself afterwards? I didn't remember this!!
15 Sep: me to SK on normies (I have my boundary, but it's like, six people, and doesn't include Eliezer Yudkowsky.)
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
example: "sketchy" includes scam situations, and things that merely look like them
reply: ideal probabilistic beliefs don't correspond to consequences, even if forced-choice decisions do
counterreply: it's complicated; we can recover epistemics as instrumentally convergent
9 Nov: I write to Ben about being stuck. My plan had been to write a memoir, but Glomarizing about the content of private conversations, and offer Scott/Yudkowsky pre-publication right of objection. Having trouble writing. Is it that I don't know how to make "This is not a social attack" credible? I was really angry when I saw Scott's https://www.lesswrong.com/posts/tSemJckYr29Gnxod2/building-intuitions-on-non-empirical-arguments-in-science on the grounds of, "So you do understand the importance of parsimony, you hypocrite!" Some of Yudkowsky's retweets felt like "concessions" that should make me less needful of aggressing against him.  I need to escape from the BDSM epistemology of submitting to this one fucking guy who wrote some good blog posts ten years ago. I guess I'm afraid that telling a story about Our Leaders being intellectually dishonest feels like "the nuclear option" in a way that I can't credibly cancel with "But I'm just telling a true story about a thing that was important to me that actually happened" disclaimers?
Ben— you are a victim of systemic abuse, and you're trying to figure out whether you're being fair to your abuser. I don't think that's clear to you here, and I think that if you internalize that one fact, you'll be able to forgive yourself for a lot of messiness, which will reduce the perceived complexity of the problem. /It's a VERY common abuse tactic to try to silence one's victims by appealing to their desire to not hurt their abuser, & telling them they'll do damage if they speak out / asking for promises of silence, etc.
me—I'm going to bite the bullet here. Yes! Yes, I'm trying to figure out whether I'm being fair to my abusers, and it's an important question to get right! "Other people's lack of standards harmed me, therefore I don't need to hold myself to standards in my response because I have extenuating circumstances" is a lame excuse!
seems related to the motte-and-bailey over "fraud"
15 November: Ziz manifesto
emailing with Katie about defamation
16 Nov: I was mostly flattered that someone bothered to write so much about me. (I didn't spot any outright-lies as opposed to "innocently" biased memory and uncharitable paraphrasing; my argumentative performance with Ziz in 2016 was not the best and I would do better today.) / In contrast, I was pretty annoyed when Alyssa claimed that she had "personally seen" me using "at least half a dozen different pseudonyms." (She apologized after I called her on it.)
17 Nov: charges including kidnapping, child endangerment, trespassing and resisting arrest.
17 Nov: having been in similar situations before (albeit of a psychiatric rather than criminal character), I'm not at all surprised at sufficiently-dense ingroup jargon getting summarized by police and reporters as "an incoherent language."
19 Nov: me to Ben and Jessica on smart fascism http://www.anechoicmedia.org/blog/european_politics/ "He's definitely a Bad Man; I just like his graphs" under present conditions it takes a lot of independence of thought to deny the antecedent of "If doing non-blank-slate social science is fascist, maybe fascism is OK."
19 Nov: slightly worried about splash damage from Ziz, but the dynamics that make people pro forma dismissible as "crackpots" (read: outside the system, no power, not a profitable ally) without processing what they're actually saying, is part of the problem.
Ben—I need to be less prone to Ziz-like conflict-seeking:  decode you're "bad" & align with the losing side of Evil Vs Evil, Ziz literally identifies as a Sith & wears black robes, both of you keep coming back to almosy-virtuous Anna & Eliezer & Scott for more abuse & then try to trigger blame-allocation mechanisms in a context that you know is power-aligned & you're the likely scapegoat instead of updating your model?
in contrast, my multi-thousand word manifesto about how "rationalist" leaders are corrupt as evidenced by their being wrong about gender concepts wasn't going to mention her at all
19 Nov: I again want to ask pointed questions of the form, "So, it's just a coincidence that the four women who got busted blockading a camp entrance in Guy Fawkes masks all happen to be trans?" [...] No one is actually surprised in System 1; it's just that the parts of us that talk aren't supposed to believe in psychological sex differences (since before my time—and I still prefer not to believe) or physical sex (since 2015).
Ben—a faction claiming to speak for trans women is trying to supress that info 
> it seems like you’re more upset about the validation of their gender, than about the felony charges, which are equally ridiculous and substantially more obviously linked to physical violence. What do you think is going on with that? Why does that seem OK to you?
Jessica says I'm cooperating with fascism
I told you about how criminals probably think of defense lawyers the same way I thought of patients' rights, and how I only have that empathy because of my psych-ward experience
Thanks for creating a context in which "Why does that seem OK to you?" is a literal question that I'm supposed to literally answer, rather than an attack to be deflected.
23 Nov: Michael's ASL analogy: "ASL is not a language" doesn't appear on the internet, and no one who had a nuanced position on how ASL differs from natural spoken languages would want to summarize their position as "... is not a language", because there's no one who has it out for deaf people
(This might actually work as a separate post)
26 Nov: I could offer to meet as a friend and then surpise you with demands for answers
27 Nov: mutualist pattern where Michael by himself isn't very useful for scholarship (he just says a lot of crazy-sounding things and refuses to explain them), but people like Sarah and me can write intelligible things that secretly benefited from much less legible conversations with Michael.
29 Nov: drafting not-lying
2 Dec: I still have a box with a wolf and a hat
9 Dec: bid for pre-reader approval of "On the Argumentative Form"
11 Dec: complicity with injustice "Ziz isn't going to be a problem for you anymore"
14 Dec: there's a time-sensitive opportunity to reach out to perennial major donor Rick
18 Dec: Vaniver on "low regard"
21 Dec: Anna reply on my war action
22 Dec: I want to procrastinate reaching out to Rick, which is OK because Michael doesn't want followers; he wants people to follow their sense of opporunity
I think "talking to Rick and Amy about my recent worldview updates, and pointing out that, if true, this could have implications for Giving Season 2019" seems political
Jessica—it moves money, so it's political is erasing non-zero-sum details; informing people isn't an attack
Michael—simple honesty would be timely
I'd be more comfortable trying to share the info/worldview-shift in January, even if that means the December decision will be wrong, because I don't know how to affect the December decision in a way that doesn't require someone to trust my judgment
23 Dec: ask Ben about the psychology of deception, and secrecy promises being abusive
26 Dec: state of Church leadership—extort more bandwith
26 Dec: forward Jessica's relevance-of-politics to SK
1 Jan 20: "low regard" with Vaniver cont'd
three negative and three positive datapoints about our culture going completely off the rails
I'm not worried about: ppl reading Brent, ppl reading Ziz, Vanessa having math students
I am worried about: Ray saying "the landscape is different now", Ruby on "forcing what you think is true on other people", Eli on "arguments are necessarily soldiers"
I am worried about Steven on never be willing to make a valid point for stragic reasons; my complaint against Steven is _the same_ as Anna's charge against Michael, with the difference that I'm arguing with Steven, not trying to discredit him
I am worried that you said you learned from Moldbug, and now Yudkowsky needs to performatively denounce him
I am worried about Yudkowsky wireheading his fiction subreddit
I see a culture that has given up on the common interest of many causes
Vaniver mentions Said saying "sure, Eliezer doesn't post here anymore, but that's bad for Eliezer too."
1 Jan 20: the red phone / state of Church leadership 
feeling more optimistic about not needing to finish the memoir—or giving it a happy ending
"If that's all I have, then it's all I can honestly love"
playing on a different chessboard
he hasn't given up on his Wizard's art, but he has, after painful experience, like the Roko affair, given up on sharing it with the public
me and Michael and Jessica were counting on him to uphold standards of discourse rather than chess
we don't get to know the real Eliezer Yudkowsky; all we can ever know is what we can infer from the words you type—and if that's all we have, then it's all we can honestly love. If the algorithm that determines which subset of your thoughts reaches us changes—then as far as we or the rest of the world can tell, it's as if you've changed.
2 Jan: Michael—I can't work with you if you are accepting Eliezer's schedule instead of ours. That's being political and on his side!
my memoir as a decision-theoretic weapon
my sense of opportunity perceives me playing a useful political role without me being unambiguously on this "side"?
The reason I got wrapped up in this civil war was becuase of this one thing; therefore I'm now at peace with Scott
I'm more trustworthy if I'm not perceived as Michael's pawn
3 Jan: Michael—
> I strongly prefer judge functions over coalition functions, so far as we can cause those to apply!!!
> The challenge is to apply judge functions in a context where there is no recognition of a right to call the litigants into a court. 
now that Scott is legally in the clear, my recourse is polite blog posts, and not police actions like cussing him out in public
Same thing about being complicit with psychiatry (morally suspect, but my only allowable recourse is to write polite critical blog posts).
Jessica—being complicit with psychiatry
I'm too submissive to psychologically maintain more than one vendetta at a time, so if I'm currently angry about the philosophy of language, then psychiatry must be OK because I don't have any anger resources left.
4 Jan: Twitter fight with Ben
a shouty equilibrium rather than the colluding-to-wirehead-each-other equilibrium that we see in Berkeley
4 Jan: Ben thinks I'm no longer interested in creating clarity
memoir is a nuke
5 Jan: Ben—the fact that I think the memoir is an act of war, even privately (insofar as my motivation to work on it vanished when I got concessions) means I've ceded the inside of my own mind to the forces of anticlarity
my reply: I've ceded the internal of my own mind to Eliezer Yudkowsky in particular
5 Jan: using "central" allows mis-definition by mobs of obfuscators
5 Jan: I am confident that finishing writing it will be clarifying and cathartic for you even if you decide not to publish, and would be an excellent use of time
5 Jan: me—vocabulary itself is a surprisingly expensive captial investment
5-8 Jan: baiting Steven to talk to Jessica about global warming
9 Jan: arguing with Jessica about "developmental sex" and natural categories
10 Jan: binary vectors
10 Jan: "isn't meant to force you into agreement here on pains of losing face"
15 Jan: more Twitter/email fighting about "people's philosopher"
15 Jan: the same _interest_ is being defended whether I'm speaking up for my own order or someone else's
16 Jan: Without you, my smartest critic is Ozy, but my self-criticism is smarter than Ozy, so it was like having no critics at all.
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.

> And ... I want to be nice to my trans friends, too, but that can't possibly be the definitive end-of-conversation correct argument. Not _really_. Not if you're being serious.

> But I want to live in a rationalist community that can actually _update_—where a bad argument by a high-status person will _eventually_ be publicly known to lose to a superior counterargument, even if the person originating the counterargument is lower status and the counterargument happens to have genuinely unfavorable political implications.

> But ultimately, I've got _something to protect_. This is even more humiliating, really. It's not even a good or wholesome something to protect that you'd feel comfortable blogging about under your own name. It's actually kind of obscene. Good or not, wholesome or not, obscene or not, it's for all the boys like me who had a dream whose proper realization happens to have [_more complicated implementation details_](https://www.lesswrong.com/posts/QZs4vkC7cbyjL9XA9/changing-emotions) than applying social pressure to make everyone pretend that they've coordinated to redefine a top-20 word. Those boys deserve to not be lied to.

> I've tried to be pretty explicit about only expecting buy-in for a minimal flag that says, "'I Can Define a Word Any Way I Want' can't be the end of the debate, because choosing to call things different names doesn't change the empirical cluster-structure of bodies and minds in the world; while the same word might legitimately be used with different definitions/extensions in different contexts

> I keep thinking of the part where Dagny entreats the great Dr. Robert Stadler to denounce a not-technically-lying statement by the State Science Institute

> Imagine your friends got a new front porch installed on your house, and you noticed that the boards were smoother and shinier than you expected. A question you might then ask your friend is, "Oh, is this real wood, or one of those wood-plastic composites?"

> In phrasing it that way, you're not saying that composites are bad; it's just that it makes sense to use language to asymmetrically distinguish between the natural thing that already existed, and the synthetic thing that has been deliberately engineered to resemble the original thing as much as possible.

The Craft Is Not The Community, (https://srconstantin.wordpress.com/2017/08/08/the-craft-is-not-the-community/amp/) 


Anyway, meanwhile, other conversations were happening. 


Michael—
> Ben once told me that conversation with me works like this. I try to say things that are literally true and people bend over backwards to pretend to agree but to think that I am being metaphorical.


> Yes, exactly. I'm still framing it as "appeal to the conscience of someone who built their career on higher standards" rather than appealing to a standard a court or a community could enforce. Eliezer reportedly didn't think much of Atlas Shrugged ... I assume Andrew has read it? I keep thinking of the part where Dagny entreats the great Dr. Robert Stadler to denounce a not-technically-lying statement by the State Science Institute:

> I do want to distinguish between direct ideological conformity enforcement attempts, and "people not living up to their usual epistemic standards in response to ideological conformity enforcement in the general culture." So-called "rationalists" are actually very good at not-being-persecutory! (I don't think I'd be saying this in the nearby possible world where Scott Siskind didn't have a traumatizing social-justice-shaming experience in college, but it's true here.)

You see the problem, right? I'm kind of at my wits' end here, because I thought the point of this whole "rationality" project was to carve out one place in the entire world where good arguments would eventually triumph over bad arguments, even when the good arguments happen to be mildly politically inconvenient.

appeal-to-arbitrariness conversation halter
https://www.lesswrong.com/posts/wqmmv6NraYv4Xoeyj/conversation-halters

(I _want_ to confidently predict that everything I've just said is completely obvious to you, because I learned it all specifically from you! A 130 IQ _nobody_ like me shouldn't have to say _any_ of this to the _author_ of "A Human's Guide to Words"! But then I don't know how to reconcile that with your recent public statement about [not seeing "how there's scientific truth at stake"](https://twitter.com/ESYudkowsky/status/1067482047126495232). Hence this desperate and [_confused_](https://www.lesswrong.com/posts/5JDkW4MYXit2CquLs/your-strength-as-a-rationalist) email plea.)

And I'm still really confused, because I still feel like everything I'm saying here is a trivial application of Sequences-lore. If I'm getting something wrong, I should be overjoyed to be harshly corrected by the Great Teacher! A simple person like me is as but a mere worm in the presence of the great Eliezer Yudkowsky! But if it looks like the Great Teacher is getting something wrong (wrong with respect to the balanced flow of arguments and evidence in which every "step is precise and has no room in it for your whims", although not wrong in the sense of making a factually incorrect statement) and the 



Great Teacher neither corrects me nor says "OK, you're right and I was wrong, well done, my student", what am I supposed to conclude? Is this a prank—a test? Am I like Brennan in "Initiation Ceremony", being evaluated to see if I have the guts to stand by my vision of the Way in the face of social pressure? (If so, I'm not doing a very good job, because I definitely wouldn't be writing this if I hadn't gotten social proof from Michael, Ben, and Sarah.) Did I pass??

In a functioning rationalist community, there should never be any occasion in which "appeal to Eliezer Yudkowsky's personal authority" seems like a good strategy: the way this is supposed to work is that I should just make my arguments with the understanding that good arguments will be accepted and bad arguments will be rejected. But I've been trying that, and it's mostly not working. On any other topic, I probably would have just given up and accepted the social consensus by now: "Sure, OK, whatever, trans women are women by definition; who am I to think I've seen into the Bayes-structure?" I still think this from time to time, and feel really guilty about arguing for the Bad Guys (because in my native Blue Tribe culture, only Bad people want to talk about sexual dimorphism). But then I can't stop seeing the Bayes-structure that says that biological sex continues to be a predictively-useful concept even when it's ideologically unfashionable—and I've got Something to Protect. What am I supposed to do?

I agree that this is the only reason you should care.

ut if it looks like the Great Teacher is getting something wrong (wrong with respect to the balanced flow of arguments and evidence in which every "step is precise and has no room in it for your whims", although not wrong in the sense of making a factually incorrect statement) and the Great Teacher neither corrects me nor says "OK, you're right and I was wrong, well done, my student", what am I supposed to conclude? Is this a prank—a test? Am I like Brennan in "Initiation Ceremony", being evaluated to see if I have the guts to stand by my vision of the Way in the face of social pressure? (If so, I'm not doing a very good job, because I definitely wouldn't be writing this if I hadn't gotten social proof from Michael, Ben, and Sarah.) Did I pass??

> I don't think it's as clear as that. The trans community has never been solely focused on people who believe they've always been of x gender. It's just that it's such a freeing and helpful explanation when trying to get people to accept that you in fact are of that gender. The x in a body of y explanation has always been a shortcut to get cis people to leave us alone more than a real feeling, honestly.
> And then the thread did not explode into acrimony. 

My paraphrase: "Oh, it's not actually true; it's a non-true thing that we tell cis people to get them to model us the way we want to be modeled"!?

Isn't that at least kind of like lying? (Or if these authors get to choose their preferred category boundary for "woman", do I get to choose my preferred category boundary for "lie"?)


f you want people to use your preferred categories, you're going to need to convince them with some sort of argument that those categories are predictively useful. I don't see the point of saying "There's no epistemic issue here, although maybe there's a policy debate about how to use language; end-of-discussion" if it's not to justify the mind games? Why can't we get a clear ex cathedra statement that gender categories are not an exception to the rule that categories are nonarbitrary? Isn't it true, and important, and something that people who ought to know better are getting wrong? Surely there's some way for "not alt-right" signaling to be purchased separately?




[TODO: Ben—
> I am pretty worried that if I actually point out the physical injuries sustained by some of the smartest, clearest-thinking, and kindest people I know in the Rationalist community as a result of this sort of thing, I'll be dismissed as a mean person who wants to make other people feel bad.

I assumed he was talking about Katie's institutionalization, not SRS

> where gaslighting escalated into more tangible harm in a way that people wouldn't know about by default. In contrast, people already know that bottom surgery is a thing; you just have reasons to think it's Actually Bad—reasons that your friends can't engage with if we don't know what you're talking about. It's already bad enough that Eliezer is being so pointlessly cagey; if everyone does it, then we're really doomed.

Ben's actual worry—
> "saying the wrong keywords causes people in this conversation to start talking about me using the wrong keywords in ways that cause me illegible, hard-to-trace damage."

> marginalization happens through processes designed to be hard for the "rider" in the horse-rider system to see. This makes the threat model hard to communicate, which puts the rider in a double-bind with respect to things like courage, because it's dependent on the horse for most of its information.

I thought the conjunction of the "mere worm" rhetoric and the thousands of words of "Here's why your own writing proves that you're wrong, you coward" and the context of Eliezer's own arrogant writing style, should have implied a certain level of conceit/tongue-in-cheekness? (I wouldn't write like this—or, um, pay someone $1000 to read my blog post—for anyone else in the world, but Eliezer is really special to me.) 

I already tried to say this in my message of the seventh, but let me try again more explicitly: I agree that the rationalists are very good at free speech norms. Ozy is writing in good faith! There is no concealed blade! I can totally do my thing without getting shouted down for heresy! I am not a victim of anything!

I do, however, think there's a subtler failure mode than "heretics get shouted down", namely, "heretics have to put up with spurious isolated demands for rigor, logical rudeness, conversation-halters, &c., such that the community doesn't update or updates slower than it could have."

This is, of course, a much harder problem to solve, because "Speaker gets shouted down" is easy for third parties to detect as a discourse-norm violation, whereas "Speaker's (polite!) interlocutors are engaging in [motivated continuation](https://www.lesswrong.com/posts/L32LHWzy9FzSDazEg/motivated-stopping-and-motivated-continuation)" is a subtle judgment call that a lot of third-parties are going to get wrong. But if and to the extent that such a thing does happen in our community—and you shouldn't take my word for it—I think it's causally downstream of silencing going on elsewhere in the trash fire that is Society (which we're not isolated from).

Ben—
> It seems like for the socioeconomic class likely to become Rationalists, employability increasingly requires the kind of sensitivity to social threat and opportunity that makes people easily silenceable in the absence of a clear distinct threat, even in the Bay, even for programmers (though not 100% yet), since workplaces have become more politicized and less about getting something specific done

Except people are actually super hypocritical about this: everyone knows to say that men are Bad, but only weird nerds who Take Ideas Seriously (e.g., me, Scott Aaronson, Jessica, Ziz), take the idea seriously. In a world where we know that men are Bad, but women are Good, and that women are people who identify as women ... well, for normal, well-adjusted men, it makes no difference. But for the nerds who live in a world made out of words, there's a natural incentive to cross the border between a lonely straight man with no intersectionality points, and a thrice-oppressed lesbian trans woman with three intersectionality points—and thus, a seat at the Blue table.

The gender-refugees story is story is the extra ingredient we need to make sense of the new and crazier thing that's going on in Current Year Berkeley: the escaping-Bad-man-status incentive gradient, and decreasing stigma (or even a status boost in sufficiently social-justice-converged subcultures), make for a much lower threshold of required AGPness before transitioning starts to look like a good idea—and potentially none at all (see Ziz again).

I really did not anticipate that the controversial part of "men who fantasize about being actual women for sexual reasons" would be the phrase "actual women"!

I've got some exciting open-source events coming up (first the Rust team all-hands, then I'm giving a talk at !!Con West) where I want to make friends and impress people, and I'm a little bit worried about being revealed as a thought-criminal earlier than I intended. Among the ~30 !!Con speakers (they're lightning talks), I count 3 trans women, three obvious men who prefer they/them pronouns, and one obvious woman who prefers they/them pronouns.

Karaoke night was fun, but social-justice culture has me on edge (the little things; like, Manish showed me this funny Tweet on his phone, and my uncontrolled facial reaction was probably more disdainful than mirthful, and I don't know how to calculate if or how much I'm being socially penalized these little things)

I've lived in the "rationalist" cult-bubble my entire adult life, and I don't know the rules for forming cross-cult alliances. If some of the social-justicey open-source people seem like they're starting to Accept me and invite me to things, can I say to one of them, "Look, I'm worried about being socially rejected for ideological reasons; maybe we should talk about this privately now rather than risking Drama later"? If that doesn't work (and it works with my "rationalist" friends, for all their other faults), then what do people do??

My original fantasy-plan was to strive for Acceptance, then suddenly stop preference-falsifying on Twitter (rough draft) at an opportune moment (after securing compiler-team membership or another conference speaker slot) as a calculated bid for glorious Damore- or Lindsay-Shepherd-like free-speech martyrdom. (If you don't preference-falsify at first when you're a nobody, then you can be quietly frozen out, but if you get Accepted first, then getting kicked out is a loud event that forces the bastards to make their ideological litmus test visible.)

But I'm pretty sure the martyrdom-seeking intuition is Actually Bad. (They're not bastards; they're people exactly like me who read different books in a different order.)

After more private discussion with Eliezer (after rounding up Michael Vassar, Ben Hoffman, and Sarah Constantin for social proof), and fighting Alicorn's Discord server three times, at this point I'm leaning towards just giving up on the so-called "rationalist" community. We had an entire goddamned Sequence about the hidden Bayesian structure of language and cognition and why "I can define a word/category any way I want" is generalized anti-epistemology.

I think something got derailed in our last discussion. I begged you to remind your readers about the meta-level epistemology-insight about hidden-Bayesian-structure-of-language-and-cognition thing, and somehow we ended up in an object-level discussion about which categories to use, with you appealing to "mental health benefits to trans people."

I'm not upset about the object-level. What I care about is this: as an aspiring epistemic rationalist, seeking out the map that reflects the territory, to cut through to the correct answer in a single movement, I have a moral duty to not care about people's feelings when forming my world-model. I can certainly not-talk about unpleasant ideas or concepts when they're not too relevant when in the presence of non-rationalists who might feel hurt. But when I'm being serious? With people who call themselves fucking aspiring rationalists? Trying to make sense of the world? The correct amount of "mild inconveniences and slight cognitive friction" to accept for utilitarian reasons is zero.

Like, I want to be fair and give credit where credit is due. Like, rationalists are very good at free speech norms. The Alicorner discussion actually made some progress! But ... not enough progress?

You wrote, "some compromise solution like always using 'sex: male/female' to mean 'biological sex' and 'gender: man/woman' to mean 'self-identified gender' would, if it were universally understood and agreed upon, be pretty fair and optimal."

But have you fucking tried trying to talk about biological sex with Berkeley psuedo-rationalists? They won't fucking acknowledge it.

I said, 



Elena said, "'the original thing that already exists without having to try' sounds fake to me" ... and gets 4 "+1" emoji.

And like ... I don't know, these motherfuckers talk a good game about "charity", but when I'm clearly trying to talk about the 'biological sex' thing, and I get a response about women who don't like their own bodies ... I don't think the principle of charity is being applied?! I haven't sent you the whole transcript; you could argue that maybe I'm taking too much out of context and that the full context would be mitigating? I want to be fair and make sure that I'm not strawmanning, but I can't make sense of this; I don't know how to perform the mental contortions to make it look like people are speaking in good faith; I don't trust that these people are trying to help me reason about reality

I'm still pretty horrified that the sticking point in discussions seems to have been not the AGP etiology itself (which is what I was focusing on in 2016/2017), but the category-boundaries-are-arbitrary "insight." So it's not just that we failed to break the irrationalist trans ideology: we actually invented something arguably worse! (Ordinary trans activists at least try to make (confused) truth claims; we invented a fully general counteragument against truth claims that's all the more dangerous thanks to the virtue of a very narrow kind of technical correctness.)

> point out to him your sympathies regarding his mental breakdown under soft political persecution and silencing, make it clear that you are not requesting that he now or at any time in the future do anything in public regarding gender issues, and ask him to remember that you have recently been mentally incarcerated more than under similar political pressure (applied in the name of the mental health needs of non-specific people many of whom are in the same natural category as yourself) that he is totally forgiven for participating in it, but that you would like to discuss the shared aspect of your circumstances  face to face, and maybe both consider not posting anything online for a couple weeks.

> That was a fairly destructive move to pull without consulting me, but OK, you played it and I will run with it.

(Notwithstanding how it's bad that I've been leaning on you for social proof way too much, I've been writing with the assumption that the author of the Sequences still cares about the common interest of many causes. I can see the "MIRI exists, that was all the raising-the-sanity-waterline project turned out to be good for; I don't have any teaching duties besides my not-explicitly-lying Wizard's Oath" CDT logic, but ... ugh, really? I can't be the one to suggest it.)



You write:

    



How would you begin to respond to this? That's not a rhetorical question: I'm actually curious.

My model of Sequences-era 2009!Eliezer says, "This is an attempt to sneak in connotations by performing a categorization and trying to avoid the burden of having to justify it with an appeal-to-arbitrariness conversation-halter. Go read 'A Human's Guide to Words,' and if you've already read it, then you're being incredibly intellectually dishonest and not worth my time to interact with."

But I have no idea what 2019!Eliezer says.

You wrote to me on 28 November:

    The notion of a lie is as sancrosanct as the notion of a truth, and you do not get to say 'lie' when somebody is doing something that you think might tempt somebody to add on a lie as an additional act. [...] 'Lying' is for per se direct falsehoods.


> If the statement "Pronouns aren't lies" has the same type of misapprehension-forming impact on the listener as "Scott Alexander is a racist" or "Slate Star Codex is an alt-right blog" then

Wait! That's not the analogy I'm trying to make!

What I'm trying to say is, whether "You're not standing in defense of truth if you insist on a word, brought explicitly into question, being used with some particular meaning" makes sense as a response to "X isn't a Y" shouldn't depend on the specific values of X and Y.

Your behavior the other month made it look like you think that "You're not standing in defense of truth if ..." is a valid response when, say, X = "Caitlyn Jenner" and Y = "woman."

I'm saying that whether or not it's a valid response, we should, as a matter of local validity, apply the same standard when X = "Scott Alexander" and Y = "racist."

The preceding three paragraphs are glossing over some details, and I totally understand if you want to disagree with the exact way I've formulated them! (Like, maybe you think it's inappropriate that I'm focusing on category-membership even though your explicit statement was about pronouns. It's important to notice that! We may be living in different mental universes with respect to how Gricean implicature works!)

But ... I'm not claiming that Eliezer's statement that "Pronouns aren't lies" has the same type of misapprehension-forming impact on the listener as Sneer Club's statement that "Scott Alexander is a racist." That's just not what the analogy was trying to do!


But like, does anyone understand what the fuck could possibly be going on in Eliezer's head? I've spent thousands of words with lots of direct links and quotes to his own writings from ten years ago trying to explain that there's a problem where people who call themselves "rationalists" are somehow very intent on playing clever definition-hacking mind games to make it artificially expensive to express the idea that humans cannot change sex (given current technology), and that I think he could fix the problem because people might listen to him when they don't listen to me. I agree that trans issues aren't important, but I also think that if you're doing systematically correct reasoning, you should get the right answer even when the issue isn't important.

I think I've been pretty explicit! And today Eliezer is chastising me and Ben for not having enough empathy to realize that people live in different mental universes. And like—I agree that people live in different mental universes! But—that seems bizarrely nonresponsive to the thousands of words I wrote explaining the problem! You see it, right?! What is there left for me to say?

Michael—
> I think that if Zack is having difficulties it would be ideal if we could move this ahead without him. Needless to say, he doesn't need to feel any responsibility and has done all that could be asked of him and more to cause this project to survive unanticipated difficulties.  Right now, I don't think there's much help he can offer except for periodic pushes, like the current one, for the rest of us to keep at it. 



> On my model, it's good to send this because there are two possibilities
> 1) that this demonstrates the independence of our perspectives, which causes them to count as additional evidence for a proposition, and that means that this makes us more convincing
> 2) that this shows a lack of unity, a failure to present a United Front, which is bad, because Arguments are Soldiers, and thus we are making ourselves less convincing. 
> In the latter case, Eliezer is dead to us and it would probably help if we were to bury something in a funeral. In the former case, our friends are still in there, and the way to reach them isn't by balancing the social forces arrayed against them.

I'm pretty sure I can directly link to Ozy and Kelsey speaking favorably of literal self-identity, not gating on hormones or social passability. And just ... I keep rereading this stuff trying to check that I'm not misreading, because I can't take it seriously. They're just fucking with me, right???? I don't harbor ill will against any individual people, but I just can't share a rationalist community with this level of mindfuckery.

I mean, okay, if that's what you want to do. (If the outcome of these hopefully-not-too-annoying emails is that Eliezer Yudkowsky feels more social/political constraints on what he can say rather than less, then this horribly backfired and you should forget we said anything.)

"How would you feel if someone else tried to get away with the 'hill of meaning in defense of validity' appeal-to-arbitrariness conversation-halter against you when you were trying to use language to reason about something you actually cared about (like defending Scott, who is great)? If it would be logically rude in that context, maybe it was also logically rude the first time?"

Michael
> We have all been essentially trained to believe that we are not entitled to voice our complaints. When someone backs off, we are prone to loosing confidence and to think that maybe we shouldn't be talking with them, yet function institutions have existed in the past, and if people aren't adhering to the normal of functional institutions, we have to not allow that to cause us to stop adhering to those norms. Therefore, we should keep doubling down with reminders regarding the rules for discussion. If he decides to not play, we should establish that to ourselves with no plausible deniability and move on. 


 "An alternative categorization system is not an error", like all appeals-to-arbitrariness, is necessarily what you called a symmetric weapon: the same form of argument works just as well for being mean to our friends as it does for being nice to our friends. The best the good guys can hope for is to win by coincidence. (Assuming there's any way to distinguish who the "good guys" are, if alternative systems that draw the "good guys" category boundary differently aren't in error, assuming there's any way to distinguish what "winning" is, if alternative systems that draw the "winning" category boundary differently aren't in error, &c.)

If someone showed up at a Less Wrong meetup claiming that they had an invisible dragon, I would expect ordinary meetup attendees to exert social pushback against the dragon-claimant: "Dude, that doesn't make sense. Beliefs need to be falsifiable! You should read the Sequences!" I would even expect the social pushback to be strong enough such that the claimant would eventually either internalize the "Beliefs need to be falsifiable" idea or get the fuck out of my rationalist community.

If I show up at a Less Wrong meetup claiming that people don't have invisible genders, and I get social pushback ... then I'm not really sure what my options are other than writing off the entire community.

     

I'm not extremely troubled by my minor disagreements with you. All the actually-smart high-status old-timers I've talked to (you, Eliezer, Anna, Michael Vassar, Sarah Constantin, Ben Hoffman, Steve Rayhawk, &c.) clearly see the hidden-Bayesian-structure-of-language-and-cognition thing I'm trying to point at. Even if you say it's mere awkwardness and not a serious epistemology problem, at least you acknowledge the awkwardness, which is more than I've gotten out of Kelsey.

The thing I find extremely troubling is that none of the old-timers who clearly understand the thing are willing to say so loudly and in public where other people can learn from it, with the exceptions of Ben and Vassar—and Vassar's in-community status is getting debased. (Sarah drafted a post supporting me on the categories thing, but decided not to publish it.)

You wrote about the "inevitable" forces pointing towards a future guided by the beauty of our weapons, where good arguments eventually win out over bad arguments. What if you were too optimistic? What if good arguments don't win—not even in the so-called "rationalist" community?

I think "Categories should carve reality at the joints to help us make efficient probabilistic predictions" is a better argument than "Categories capture tradeoffs you care about; they don't have to carve reality at the joints if that makes people sad." Right now, the better argument is losing. Unless something changes, the better argument will continue to lose. In other words, we live in a world where reason doesn't work. If we live in a world where reason doesn't work, then we're dead. If we can't get consensus on easy problems (Q: Can men become women by means of saying so? A: No. Why would you think that?), then how are we supposed to solve AI alignment?



> I don't know if you read the glowfic that Alicorn and Kelsey write. There was a scene in one of their LoTR things where some elves capture some orcs. They learn that the orcs are nice people deep down, evil god Melkor has forced the orcs to swear an oath to serve him and do evil all the time and kill people. And for complicated magic reasons, orcs cannot break oaths, or else they are tormented by unbearable pain. So even though the orcs are nice, they are compelled to serve Melkor, who usually wants them to kill people and stuff.

> After a while, our hero comes up with a solution. She founds a religion on the spot devoted to a Deist-style omnipotent observer God, who just so happens to also be named Melkor. Deist-Melkor just wants people to live their lives and be happy and help others. She claims that the orcs did swear an oath to follow Melkor, but they didn't specify *which* Melkor, and since her Melkor is omnipotent he takes preference over evil-Melkor. The orcs (who are all kind of dumb, and also incentivized toward highly motivated reasoning) agree this makes sense, this seems to satisfy whatever magic enforces their oath, and so they go off and live their lives and be happy and help others.

> Part of this story is that occasionally some people will approach the orcs and be like "This is kind of dumb, right, on some level you know you're supposed to be following evil Melkor", and this information is infohazardous to orcs and can put them in unbearable pain (or compel them to kill people and stuff) until they're able to talk themselves out of it.

> I asked Alicorn whether this was an intentional metaphor for this discussion, since I know you've been talking to her about it. She says no. But it captures a lot of how I think about this. Are the *real rationalists!!!* the ones who go to the orcs and tell them that they obviously meant to serve evil Melkor and so should be in constant pain all the time? Are those the people who are """winning"""?  Or would saying that just be pointlessly harming them for no reason?

> I couldn't care less about transgender.

Michael—
> I agree that he isn't, but it seems Important to me that if we think he isn't interested we actually establish common knowledge that we consider him at this point to be choosing to try to be a cult leader. 


I was kind of considering replying, "That's totally understandable! Would a financial incentive change your decision? For a two-hour meeting, I'd be happy to pay up to $4000 to you or your preferred charity. If you don't want the money, then sure, yes, let's table this. I hope you're having a good day."

But, maybe I shouldn't do this because the "Volunteer to throw four-figure amounts of money at people in order to get them to hear me out" move is simultaneously personally degrading, psychologically coercive, extremely socially weird, and bad for my economic self-interest.
 
https://meltingasphalt.com/tears/

commit a5316c23ac837c004922e2d192c2bea914a21ebc
Author: Zack M. Davis <zack@thecoalition.com>
Date:   Tue Mar 19 00:27:53 2019 -0700

    DIS-327: broker-facing duplicate-quote route
    
    It's sharing code with the internal resïssue-quote endpoint, because
    that's convenient.




Scott, like, you sent me that email piously arguing that it's OK to gerrymander category boundaries because it's important for trans people's mental health. And there's a couple things that I don't understand.

(1) If creatively reinterpreting the meanings of words because the natural interpretation would make people sad is OK ... why doesn't that just generalize to an argument in favor of outright lying when the truth would make people sad?  I don't understand why semantic mind games are OK, but lying is wrong. (Compare: "If you’re going to ignore a probabilistic counterargument, why not ignore a proof, too?") From my perspective the motte-and-bailey definition-hacking mind games are far more cruel than a simple lie. Speaking of which----

(2) If "mental health benefits for trans people" matter so much, then, like, why doesn't my mental health matter? Aren't I trans, sort of? Like, what do I have to do in order to qualify, was my being on hormones for only 5 months not good enough? Or do I have to be part of the political coalition in order for my feelings to count? We had an entire Sequence whose specific moral was that words can be wrong. And of the 37 Ways that Words Can Be Wrong, number 30 is "Your definition draws a boundary around things that don't really belong together," and number 31 is, "You use a short word for something that you won't need to describe often, or a long word for something you'll need to describe often." And this is the thing I've been trying to tell everyone about. And when I try to talk to you about it, you're like, "Doesn't matter, mental health benefits for trans people are more important according to the utilitarian calculus! Also, you're wrong." And like ... getting shut down by appeal-to-utilitarianism (!?!?) when I'm trying to use reason to make sense of the world is observably really bad for my mental health! Does that matter at all? Also, if I'm philosophically wrong, then Eliezer was wrong in 2009, because everything I'm doing now, I learned from him. Do you really think "c'mon, parsimony doesn't matter" is an improvement on the Sequences?!

(3) You write, "wouldn't it be embarrassing if the community solved Friendly AI and went down in history as the people who created Utopia forever, and you had rejected it because of gender stuff?"

But the original reason it seemed plausible that we would create Utopia forever wasn't "because we're us", it was because we were trying to do systematically correct reasoning, with the hope of using the systematically correct reasoning to build aligned superintelligence. If we're not going to do systematically correct reasoning because that would make people sad, then that undermines the reason that it was plausible that we would create Utopia forever. You can't just forfeit the mandate of Heaven like that and still expect to still rule China.

(4) You write, "I realize you are probably thinking something like 'This isn't just me getting triggered by some culture war thing evolved to trigger me, it actually really *is* that important that rationalists get transgender right!' [...] All I can say is that I beseech you, in the bowels of Christ, think it possible that you may be mistaken."

I don't know how much social proof influences you, but is the fact that I managed to round up Michael and Ben and Sarah and [name redacted] into a coalition to help me argue with Eliezer shift your estimate at all of whether my concerns have some objective merit? It can simultaneously be the case that I have the culture-war PTSD that you propose, and that my concerns have merit.

(5) I don't know how other people's confidentiality intuitions work. Am I being an asshole by cc'ing in my coalition members (plus Anna, who is my friend even though she hasn't been helping me with this particular quest) in an email that references things you said in an email that was addressed only to me? I think this should be OK; I don't think we're talking about anything particularly secret.
 
----


I'm really sorry about this. I don't need Scott's help and I think Michael is being a little aggressive about that, but I guess that's also kind of Michael's style? I'm really sorry, because I still think that even very emotionally disturbed people are clearly incentivizable, and so the narrative of, "Oh, poor Zack, he's crazy, he can't help it" is not attributing me enough agency. Like, if I get really upset literally once every two years and impose a lot of care costs on other people in the process, I should be held accountable for that somehow. Last time, I ended up giving people some money (I didn't end up doing the public-praise part that I proposed in the blog post about it, because it turned out that that would have been too socially weird) but I don't know if even the money was a good social experiment or not.

> No.
> We don't have punishment.
> We don't have morality.
> I'm trying to change that and you are undermining the attempt!
> We won't be friends. 
> We're not friends anymore already! Once again, I am trying to change that and you are making it impossible. 
In terms of making intellectual progress, I'm still most excited about the idea I was talking myself into all day in the "strategy huddle II" thread (I'll forward that to Jessica): "If Scott and Eliezer don't want to talk, whatever, I can just write those guys off (why did I think they were so important anyway?) and just focus on writing up the mathematically rigorous version of the thing I'm trying to say about how categories can be value-dependent because your values determine which dimensions you pay attention to (like 'birds' vs. 'flying things'), but that's really different from 'people will be sad if I use this category', and we should be able to demonstrate mathematically that the 'people will be sad if I think this' categories are less useful for making inferences. Not only will I learn lots of stuff in the process of figuring out the details, but when it's finished, our coalition can hype it up as interesting new 'Human's Guide to Words'-like content, and then at least the mathy people in our alleged community will see the parsimonious-categories-are-important thing in a context where the math will convince them and politics won't distract them."
 
So ... I really should have just stuck to that plan and not freaked out: forget this stupid "appeal to local celebrities" game and just do cool math for two months. But when it was 4 a.m. (because I'd been obsessing all day, and I really wanted to meet my manager's deadline for the code that we want for our business, and I did write some code intermittently between 11 p.m. and 4 a.m.), I was just really upset at the whole situation, and I sent some more hasty emails, I think because at some psychological level I wanted Scott to know, "I'm so frustrated about your use of 'mental health for trans people' as an Absolute Denial Macro that I'm losing sleep over it even though we know sleep is really important for me." But then when Michael started advocating on my behalf (Michael seems to have a theory that people will only change their bad behavior when they see a victim who is being harmed), I started to minimize my claims because I have a generalized attitude of not wanting to portray/sell myself as a victim, even if it is technically literally true that I'm really hurt in a way that is causally related to other people behaving in ways that are bad. And I feel really guilty and wanting-to-appease about disappointing Michael (who has done so much for me), even though I know that Michael doesn't want me to care about disappointing him in particular, because I should just care about doing the Right Thing. I think that Michael thinks that aggression is more honest than passive-aggression. I think that's obviously true, but I just can't be too aggressive to my friends and people I admire (e.g., Anna, Scott, Eliezer). I just won't do it. I can passively let my metaphorical lawyers like Ben do it while accurately claiming to represent my interests, but I won't do it myself because I'm scared of damaging relationships. (Or offending people I worship as demigods.)

So, I don't know. I think having been on the edge of a bad outcome but stepping away (I got 2ish hours worth of napping; I'll be just fine) will help me move on, and dissolve my unhealthy attachment to the "rationalist" brand name/social identity? (At the meeting on Sunday, it was pointed out that forging a new brand-name/Schelling-point-for-interesting-people-to-gather is easier than recapturing an old one.) I just ... we had a whole Sequence specifically about how you can't define a word any way you want, because you need words to map to the statistical structure in the world. I keep trying to point out, "Look, isn't that inconsistent with this newly fashionable 'categories were made for man'/'hill of meaning in defense of validity' moral? Right? Right? You have to see it? It can't possibly just be me who sees it?!" and Scott and Eliezer seem to keep rounding me off to, "Zack is upset about trans stuff which doesn't matter" even though I keep explicitly disclaiming that (while it's true that I'm upset about trans stuff that doesn't matter), that's not the part I expect them to care about. And I've just been really freaking out for three and a half months on end, because I learned everything I know from Eliezer, and the other half of everything I know from Scott, and like ... they can't really be that stupid, right? They have to just be fucking with me, right??

That's really unhealthy. (We were warned that it's unhealthy: "nervously seeking reassurance is not the best frame of mind in which to evaluate questions of rationality".)

The tension between "I don't want to do too much social aggression" and "losing the Category War within the rationalist community is absolutely unacceptable; if I have to do social aggression and lose friends and burn bridges if there's any hope of getting people to see the hidden-Bayesian-structure-of-language-and-cognition thing again, then I'll do it" has been causing me a lot of pain and making me make wildly inconsistent decisions. (Emailing Scott at 4 a.m., and then calling Michael "aggressive" when he came to defend me is just crazy.)

If there isn't actually any such thing in the world as a "rationalist community" (it's just a category boundary in my head that I keep attaching importance to, even though Sarah clearly explained to my why this is dumb two years ago, and I just haven't made the corresponding mental adjustments), then maybe losing the Category War within it is acceptable? Maybe that's the answer? If so, I wish I had been able to find the answer without leaving so much of a mess.

Wait, maybe there is a reason to be attached to the "rationalist" brand name/social identity that isn't just me being stupid, and if we can find what the non-stupid reason is, then we can make better decisions.

I think a non-stupid reason is that the way I talk has actually been trained really hard on this subculture for ten years: most of my emails during this whole campaign have contained multiple Sequences or Slate Star Codex links that I can just expect people to have read. I can spontaneously use the phrase "Absolute Denial Macro" in conversation and expect to be understood. That's a massive "home field advantage." If I just give up on "rationalists" being as sane as we were in 2009 (when we knew that men can't become women by means of saying so), and go out in the world to make intellectual friends elsewhere (by making friends with Quillette readers or arbitrary University of Chicago graduates), then I lose all that accumulated capital. The language I speak is mostly educated American English, but I rely on subculture dialect for a lot. My sister has a chemistry doctorate from MIT (so speaks the language of STEM intellectuals generally), and when I showed her "... To Make Predictions", she reported finding it somewhat hard to read, probably because I casually use phrases like "thus, an excellent motte", and expect to be understood without the reader taking 10 minutes to read the link. This essay, which was me writing from the heart in the words that came most naturally to me, could not be published in Quillette. The links and phraseology are just too context-bound.

Maybe that's why I feel like I have to stand my ground and fight a culture war (even though I really don't want to be socially aggressive, and the contradiction between the war effort and my general submissiveness makes me make crazy decisions).

I feel like I have to be careful phrasing my complaints about Berkeley culture, because if I'm not super-careful, people will round me off as saying, "Help! Censorship! I'm being repressed!" and then counterargue that I'm observably not being censored (which is true, modulo Ben's concerns about distributed silencing, which I feel motivated to downplay because I don't want to sell myself as a victim). The problem is much subtler than that. Berkeley "rationalists" are very good at free speech norms and deserve credit for that! But it still feels like a liberal church where you can say "I believe in evolution" without getting socially punished. Like, it's good that you can do that. But I had a sense that more is possible: a place where you can not just not-get-punished for being an evolutionist, but a place where you can say, "Wait! Given all this evidence for natural selection as the origin of design in the biological world, we don't need this 'God' hypothesis anymore. And now that we know that, we can work out whatever psychological needs we were trying to fulfil with this 'church' organization, and use that knowledge to design something that does an even better job at fulfilling those needs!" and have everyone just get it, at least on the meta level.

I can accept a church community that disagrees on whether evolution is true. (Er, on the terms of this allegory.) I can accept a church community that disagrees on what the implications are conditional on the hypothesis that evolution is true. I cannot accept a church in which the canonical response to "Evolution is true! God isn't real!" is "Well, it depends on how you choose to draw the 'God' category boundary." I mean, I agree that words can be used in many ways, and that the answer to questions about God does depend on how the asker and answerer are choosing to draw the category boundary corresponding to the English language word 'God'. That observation can legitimately be part of the counterargument to "God isn't real!" But if the entire counterargument is just, "Well, it depends on how you define the word 'God', and a lot of people would be very sad if we defined 'God' in a way such that it turned out to not exist" ... unacceptable! Absolutely unacceptable! If this is the peak of publicly acceptable intellectual discourse in Berkeley, CA, and our AI alignment research group is based out of Berkeley (where they will inevitably be shaped by the local culture), and we can't even notice that there is a problem, then we're dead! We're just fucking dead! Right? Right?? I can't be the only one who sees this, am I? What is Toronto??????

Ben—
> Just wrote to Scott explicitly asking him to alter the beacon so that people like Zack don't think that's the place to go for literally doing the thing anymore, and ideally to redirect them to us (but that's a stretch). If he actually does this, that would be a really good time to start the new group blog.

> This seems like the sort of thing where it's actually in his local interest to help us, as it reduces our incentive to ask him to do hard things.


Michael—
> This all seems sound, but also not worth the digression.  Zack, do you feel comfortable with generalizing the sort of things that Scott said, and the things that others here have said about fear of talking openly, and assuming that something similar is probably happening with Eliezer too?  

> If so, now that we have common knowledge, there is really no point in letting technical quibbles distract us. We need to deal with the actual crisis, which is that dread is tearing apart old friendships and causing fanatics to betray everything that they ever stood for while it's existence is still being denied. 


He said that he wasn't sure why the grand moral of "A Human's Guide to Words" was "You can't define a word any way you want" rather than "You can define a word any way you want, but then you have to deal with the consequences."

Ultimately, I think this is a pedagogy decision that Eliezer got right. If you write your summary slogan in relativist language, people predictably take that as license to believe whatever they want without having to defend it. Whereas if you write your summary slogan in objectivist language—so that people know they don't have social permission to say that "it's subjective so I can't be wrong"—then you have some hope of sparking a useful discussion about the exact, precise ways that specific, definite things are, in fact, relative to other specific, definite things.


You texted last week, "Am I agitating you in a manner that isn't a good idea? If so, please tell me."

I'm tempted to speculate that I might be better at taking your feedback about how to behave reasonably rather than doing the "cowering and submission to whoever I'm currently most afraid of losing relationship-points with" thing, if you try to be gentler sometimes, hopefully without sacrificing clarity? (Concrete examples: statements like, "What the FUCK Zack!?!" are really hard on me, and I similarly took "We're not friends anymore already!" badly at the time because I was reading it as you damning me, but that one is mostly my poor reading comprehension under sleep deprivation, because in context you were almost certainly responding to my "everyone will still be friends" babbling.)

But, if you think that gentleness actually sacrifices clarity in practice, then I guess I could just do a better job of sucking it up.

Michael—
> I think that the strongest argument for continuing to press on is that we need to actually be able to press on into opposition to clarity if we are going to succeed regardless of where we go. We actually have to figure out, either here or elsewhere, how to win this sort of argument decisively, as a group, against people who are acting in bad faith, and can't afford to simply accept a status quo that says that we accept defeat when faced with bad faith arguments in general.

     
Wait, you lost me. What post about transgender? I've been objecting to the post about how category boundaries are arbitrary and we can redraw them for utilitarian benefits. Sure, the post used transgenderism as an illustrative example, but the main point should succeed or fail independently of the choice of illustrative example. (Obviously a person of your brilliance and integrity wouldn't invent a fake epistemology lesson just to get a convenient answer on one particular question!)

> I realize that allowing one thing violates a Schelling fence that potentially keeps other people from making me do awkward things, but as far as I can tell gender dysphoria is a pretty unique problem

"Conceptual gerrymandering to sneak in connotations" is _not_ a unique problem! I know you know this, because you've written about it extensively. 

"Conceptual gerrymandering to sneak in connotations" is not a unique problem! I know you know this, because you've written about it extensively. That's why my "twelve short stories about language" email gave "(repeat for the other examples from "The noncentral fallacy—the worst argument in the world?")" as entries II.–VII.

>    Zack I think - everyone else experiences their disagreement with you as a disagreement about where the joints are and which joints are important, including Scott
>    and parses the important part of Scott's post as 'the usability of the categories for humans is one thing we should use to decide which joints are important'
>    and it is annoying to be told that actually we all believe the categories should be whatever makes people happy and that this is an inexcusable rationalist sin
>    when no one believes that
I didn't want to bring it up at the time because I was so overjoyed that the discussion was actually making progress on the core philosophy-of-language issue, but ... you seem to be pretty explicit that your position is about happiness rather than usability? If Kelsey thinks she agrees with you, but actually doesn't, that's kind of bad for our collective sanity, right?



(The tenth virtue is precision! Even if you cannot do the math, knowing that the math exists tells you that the dance step is precise and has no room in it for your whims.)

 "Naïvely" hauling out the most exhaustively rigorous analytical machinery I have is the only way to shatter the surface-level symmetry between "How much do you care about this? C'mon, it isn't worth it" versus "Well, I think it is worth it."

Perhaps it is "overkill" to transform the debate into "How much do you care about this? C'mon, it isn't worth it" versus "No! Fuck you! I have math, and I have a pretty decent heuristic argument about how the math translates into human practice, and I'm willing to put in more blood and pain and sweat and money and hours to make it even more rigorous if there's any hope left of getting you savvy lying bastards to stop fucking with my head."

I want the people I love and trust and respect to stop fucking with my head. That's more important to me than whatever proportionality-intuition generates the word "overkill."


Ben, I know that you've previously written about how holding criticism to a higher standard than praise distorts our collective map (in the context of effective altruism). You're obviously correct that this is a distortionary force relative to what Bayesian agents would do, but I'm worried that when we're talking about criticism of people (as opposed to ideas), the removal of the distortionary force would just result in an ugly war (and not more truth). Criticism of institutions and social systems should be filed under "ideas" rather than "people", but the smaller-scale you get, the harder this distinction is to maintain: criticizing, say, "the Center for Effective Altruism", somehow feels more like criticizing Will MacAskill personally than criticizing "the United States" does, even though neither CEA nor the U.S. is a person.

This is why I feel like I can't give up faith that good-faith discourse eventually wins (even though I'm starting to suspect that Michael is right to suggest that it doesn't in our world). Under my current strategy and consensus social norms, I can criticize Scott or Kelsey or Ozy's ideas without my social life dissolving into a war of all against all, whereas if I give in to the temptation to flip a table and say, "Okay, now I know you guys are just fucking with me," then I don't see how that leads anywhere good, even if they really are just fucking with me.

I think Anna's input could potentially be really valuable here, because this kind of issue (whether to cooperate with Society and have peace, or give it the lie at risk of war) seems like one of the core tensions between her and Michael's crew (of which I've been saying for a while that I would really like to see a peace treaty).

> Self-promotion is imposing a cost on others. Eliezer, to a lesser extent Scott, and formerly Will, have all collected actual revenue from their self-promotion, on top of the less legible social capital they've acquired. It sounds like you think that this somehow creates a reciprocal obligation on our part to avoid talking about ways in which they've disappointed us or people like us. ?!?!?!?!?!

> Unilateral surrender to someone strong enough to hold and defend territory, with a low enough time preference not to consume everything in it can be a good way to avoid the war of all against all. Do you think Eliezer or Scott is meeting that standard? It seems to me like Eliezer's not interested in having you as a subject, and Scott will just pass any threats from the outside right through to you.

> Separately there are good pragmatic reasons to be careful about this and I don't want to ignore them - I'm really specifically arguing against what I perceive as a misapplied moralism around not criticizing people who definitely aggressed first, in a way that's much more clearly objectively aggression than criticizing someone's conduct, accurately described.


I'd like to elaborate on what seems off to me about this.

> Um. I see the value you're pointing at, but I also really don't want to attack my friends in public.

What is being proposed here is creating clarity about behavioral patterns, and specifically ones that people claim not to have.  What you're saying here implies that you believe creating such clarity is an attack on someone.  If so, your blog is attacking trans people.

What's going on here?  Here's a hypothesis.  Creating clarity about behavioral patterns is interpreted as an attack by the social field, and can sometimes (but not always) result in concretely worse outcomes for someone.  As such, it is much easier to create clarity about behavior patterns down power gradients than up power gradients.  So it's easy for you to create clarity about trans people's narcissistic behavior but hard for you to create clarity about Eliezer and Scott's narcissistic behavior (where they pretend to be pro-truth/rational/agenty and in fact aren't, and even optimize against those things).

In effect, creating clarity about behavior patterns down power gradients but not up power gradients is just going to reinforce the scapegoating system.  In a world with high amounts of hypocrisy about drug use, it's rather perverse to only investigate and report on the drug use of majority-black neighborhoods; it ends up maintaining the impression (to privileged people) that it's other people's problem, that someone else can be blamed, and so it does nothing to reform the drug laws.

You don't get to claim that creating clarity about trans people's behavior patterns is justified because that which can be destroyed by the truth should be, and also that creating clarity about Eliezer/Scott's behavioral patterns is unjustified because it would be an attack and start a war or whatever.  If you're being cowardly here or otherwise unprincipled, own that instead of producing spurious justifications.

> You're obviously correct that this is a distortionary force relative to what Bayesian agents would do, but I'm worried that when we're talking about criticism of people (as opposed to ideas), the removal of the distortionary force would just result in an ugly war (and not more truth).

Oh come on, calling all non-androphilic trans women delusional perverts (which you've done) is totally a criticism of people.  We're actual people, you know.  If you call all non-androphilic trans women delusional perverts, then you are calling me a delusional pervert, and having corresponding effects on e.g. my chances of being scapegoated.  (Which is not to say you shouldn't do this, but you should be honest about what you are doing!)

To conclude, I'm on board with a project to tear down narcissistic fantasies in general, but not on board with a project that starts by tearing down trans people's narcissistic fantasies and then doesn't continue that effort where it leads (in this case, to Eliezer and Scott's narcissistic fantasies), and especially not if it produces spurious justifications for this behavior.


I would be way more comfortable writing a scathing blog post about the behavior of "rationalists" (e.g.), than "Scott Alexander and Eliezer Yudkowsky didn't use good discourse norms in an email conversation that they had good reason to expect to be private." I think I'm consistent about this: contrast my writing to the way that some anti-trans-activism writers name-and-shame named individuals. (The closest I've come is mentioning Danielle Muscato as someone who doesn't pass, and even there, I admitted it was "unclassy" and done in desperation of other ways to make the point having failed.)

Am I wrong here?

Like, I acknowledge that criticism of non-androphilic trans women in general implies criticism of Jessica Taylor, and criticism of "rationalists" in general implies criticism of Eliezer Yudkowsky and Scott Alexander and Zack Davis, but the extra inferential step and "fog of probability" seems really useful for maximizing the information-conveying component of the speech and minimizing the social-warfare component?


If someone says "the Jews are OK, but they aren't Real Americans" I feel similarly concerned, and if they say "Michael is OK but he isn't a Real American" likewise not. 

Targeting that is Less Precise is always more violent.

On Mon, Apr 8, 2019, 4:33 PM Michael Vassar <michael.vassar@gmail.com> wrote:

    I think that is really really importantly backwards. If someone says "Michael Vassar is a terrible person" I hope they are OK, and I try to be empathically curious, but if they don't have an argument, I don't worry much and I worry more For them than About them. If someone says The Jews are Terrible People I see that as a much more serious threat to my safety. "rationalists and trans-girls very clearly appear to me to be the exact categories of people that get targeted by the sorts of people who would also target Jews, so this analogy is *particularly* apt.  

This actually interacts with the good-faith kind of social constructionism! (Which mostly isn't done, but could be done.) Speakers of my generation are more likely to say "person" when sex isn't relevant, whereas older works (pre-second wave feminism?) would actually say "man" (or "woman"—see how anachronistic it feels to put the female option second and in parentheses?), presumably not because human nature has changed that much, but because of shifting cultural priorities about when sex is "relevant." (Modern speakers similarly frown on mentioning race when it's not relevant.)


I want to be careful not to fall into the trap of being a generic social conservative. My gender crusade becomes much less interesting insofar as it can be explained simply by me being born too early and not getting "kids these days." It's not that cultural change is necessarily bad; it's that I want culture to innovate in the direction of "achieving godlike understanding of objective reality and optimizing the hell out of it" rather than "playing mind games on each other to make it artificially harder to notice that we don't actually live in a Total Morphological Freedom tech regime."


cut from the first draft—
> You scream and punch him in the face. Clutching his bloody nose, Bob seems more shocked than angry. "You—you punched me."

> "I wouldn't say that," you say. "It depends on how you choose to draw the category boundaries of the word 'punch.'"

Um, I guess it's important to notice that the only reason I think the "Vice President" example is "depoliticized" is because I live in a bubble that's obsessed with identity (gender, &c.) politics and doesn't care about class politics. It still works if my audience lives in the same bubble as me, but it would be pretty poor performance on my part to not notice the bubble.



One reason someone might hypothetically be reluctant to correct mistakes when pointed out, is the fear that such a policy could be abused by bad-faith nitpickers. It would be pretty annoying to be obligated to churn out an endless stream of trivial corrections by someone motivated to comb through your entire portfolio and point out every little thing you did imperfectly, ever.

This is why I found myself wondering if, in Scott or Eliezer's mental universe, I'm a blameworthy (or pitiably mentally ill) nitpicker for freaking the fuck out over a blog post from 2014 (!!) and a Tweet (!!!) from November. Like, really? I, too, have probably said things that were wrong five years ago that I would be annoyed about being pressured to issue a correction for now!

But ... I don't know, I thought I put in a lot of interpretive labor making a pretty convincing case that a lot of people are making a correctable and important rationality mistake, such that the cost of a correction would actually be justified in this case, and that it should be possible to take a hard stance on "categories aren't arbitrary, you idiots, just like Eliezer said in 2008" without taking a stance on what implications that might have for other more contentious questions?? Right? Right?? Like, there are noticeable differences between me and the hypothetical bad-faith nitpicker, right????

My mental health is continuing to deteriorate today, so I may need to focus on that (and therefore staying employable and having money, &c.) for a while rather than whatever we're trying to accomplish here.


I still feel sad and upset and gaslighted and I haven't done any dayjob work so far this week even though it's Tuesday afternoon. So, I feel motivated to send this email to communicate that I feel sad, even though this message is of questionable value to the recipients.





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, [...] 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.

In retrospect, I totally backwards to say that I was worried about you using up our bandwidth! My message was probably pretty predictable given everything else I've said, whereas you actually made progress, even if I didn't quite understand where you were going in the moment.)


... I'm kind of perversely enjoying my position as plantiff/pawn/puppydog.

But—wait. Eliezer is accusing you guys of driving me crazy?? I think I might be obligated to respond to that. But, on a weekend.


Ben on our epistemic state—

> Zack started out operating on object-level trust that "Rationalists" would generically be trying to build true maps, and that while there might be errors, even quite bad errors, in individuals, coordination would be towards truth on net, on the meta-level and all higher levels of recursion. This hypothesis was decisively falsified - political factors caused it to be false. Zack's level of trust did not pay rent, except insofar as he ended up helping catalyze the formation of a dissident circle who are sympathetic to coordinating towards truth, and not just pretending to do so.

> Eliezer, as the best plausible individual representative of the "Rationalists" as currently constituted, is Eliza the bot therapist in Zack's story, letting the clock run out on the interaction with Zack instead of trying to inform him. He even initially accepted in principle financial compensation for doing this, and more generally is getting paid by MIRI donors. We should consider the whole system structurally similar to Zack's Eliza story. Regardless of the initial intent, anxious scrupulous Rationalists are paying rent to something claiming moral authority, which has no concrete specific plan to do anything other than run out the clock, but remains in a facsimile of dialogue with them in ways well-calibrated to continue to generate revenue, at least insofar as bigger revenue streams aren't available yet.

> Zacks don't survive long-run in this ecosystem. If we want minds like Zack's to survive somewhere, we need an interior that justifies that level of trust. In the short run, Zack just isn't useful until he self-mods to not be vulnerable to this sort of trolling again, and give up on people exhibiting behaviors like Eliezer's (or Scott's) as not worth his time. This means modeling the politics of the situation proactively, not just defensively as a horrible mess "we" want to fix.

> Jessica started out thinking that things like Eliezer's code of meta-honesty were worth parsing in detail, instead of just dismissing them as the sort of thing someone formerly scrupulous who wanted to narrate their transition to a deceptive strategy as technically okay would do. This hypothesis has been falsified as well - Jessica's attention was wasted by Eliezer on the object level, and was only valuable - like Zack's - because it clarified that people like Eliezer are not operating on the level of fidelity that justifies that attention to detail.

> We should expect others who are modeling politics, including the good guys who would like to build robust networks of truth-oriented trust, to ignore clever explanations we give for our behavior that substantially resemble what Eliezer is doing, and to rationally predict that we have intent to expropriate via clever rationalizations if we behave like someone with intent to appropriate via clever rationalizations. If we were mistaken to trust Eliezer, then others would be mistaken to trust us if we externally seem to behave like Eliezer.

> I'm not sure what this means on the object level except that SOME leftists and anarchists are probably much better allies than they seem to us right now. Possibly some of them have analogous confusions with respect to their faction's internal coordination mechanisms, and haven't figured out how to look trustworthy to people like us. (And also, there are way more than two or three factions.)


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 [...]  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.

> In fact,bin all seriousness, we should see Eliezer as less trustworthy than Demis by default in light of the fact that Eliezer claimed to be following a nuanced strategy but defected while Demis never stood on any claims of morality/trustworthiness and is thus entitled to the standing of any person who seems to have managed to be successful while remaining sane enough to advance technology and to turn down large monetary gains for strategic advantage.

Ben—
> Zack, I'm not asking you not to say that publicly if that's how you honestly feel and you want to report it, but as a political matter I don't expect it to help Eliezer understand; I just expect it to make him feel sad.

Jessica—
> Worth making explicit why Eliezer's response to my criticism is sufficient reason to write him off until something majorly changes (e.g. apology and serious signals of intent to recognize and act against corruption):

> - Instead of responding to my criticisms, he does a "tu quoque" (in this case accusing me of making a conflation I didn't make), which is particularly illegitimate (and thus, amusing) because I'm holding him to his standards, not mine.

> - He implies I'm insane for modeling psychology.  This is classic play in high-corruption gaslighting equilibria: act in bad faith while pretending to act in good faith, and label those who see it is pretend (and talk about this) as insane, to discredit them.  (It's rather obvious that psychological modeling is actually a sign of sanity, not of insanity)

> - He praises Zack relative to me (saying he is comparably sane because he doesn't psychologize... which, Zack does too, that's a lie).  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.

> Eliezer shifted from bullshitting in ways that plausibly defend his reputation, to actively trying to hurt me in ways that do not plausibly defend his reputation.  (Him initiating a political slap fight with me actually looks worse for him than it does for me)


>  I think this is what happens when one plays politics too long without sufficiently robust defenses, and develops habits and modes of thinking in different ways. Thinking of such action as being goal-oriented with a detailed map is the wrong way to model it - as you put it, Eliezer's thinking isn't worth this level of attention to its details, as Jessica/Zack observed with its meta-honesty.

> He also engaged in rhetorical actions that the Eliezer we know and value would never do, so he can't be that person at present. He is 'insane' in the sense that he has lost his mind, but it's his mind that he has lost rather than a mind in general, as such views are alas normal round these parts. 

Michael—
> Common knowledge achieved. 
> Intention to transition to politics achieved in so far as everyone agrees regarding 'stay in formation, this is not a drill'.

> Can everyone please let me know what you think about the value of achieving common knowledge. To my mind, we don't need Eliezer, but we absolutely need to be able to count that as value when using our 'board evaluators' or we are Totally Fucking Dead!

> Now that we are doing politics at all, and conveniently, against the least skilled opponent around, it's time to start practicing offense.  We have an extremely large, clear-cut weapon here. Eliezer has publicly stated his intention to commit fraud. He has publicly defrauded a number of us. If the guns that we can point at his head aren't larger than the guns already points at his head that means that he has forgotten how to see guns.

Ben—
> Eliezer's claim to legitimacy really did amount to a claim that while nearly everyone else was criminally insane (causing huge amounts of damage due to disconnect from reality, in a way that would be criminal if done knowingly), he almost uniquely was not.

> "Eliezer is criminally insane" and "Eliezer's behavior is a normal failure of rationality" are not competing hypotheses from the perspective of the Sequences - they're the same hypothesis.

> WE CAN'T GRADE OBJECTIVE CATEGORY DESCRIPTORS ON A CURVE.

I don't think I'm setting [my price for joining](https://www.lesswrong.com/posts/Q8evewZW5SeidLdbA/your-price-for-joining) particularly high here?

In general, if someone writes to me with a criticism of the form of "That was really misleading in context; I really think it's worth clarifying to your readers", I think I would either say something like "Actually, I argue that it's not misleading because ..." or "You're right, I'll clarify." Promising to never speak about the topic again just seems weird?

Michael—
> From my perspective, she is an enemy combatant and can claim no rights in regards to either me or you.  We should be lying to her because we should also be trying, when convenient, to slash her tires.  I do think that it's pretty important that you absorb that point. Like, we know her to be actively acting against me in a manner that I think is unambiguously unrighteous, unprovoked and not self-serving, and we have established that Eliezer is actively breaking his oaths despite everything he has ever said about how 'you can break your principles once. After that, you don't have principles to break' and the like. Do we at least agree that he is unambiguously wither a criminal or an enemy combatant?  

> The latter frame is more accurate both because criminals have rights and because enemy combatants aren't particularly blameworthy. They exist under a blameworthy moral order and for you to act in their interests implies acting against their current efforts, at least temporary, but you probably would like to execute on a Marshall Plan later. 

> I don't really need this information, so don't bring it if it makes you uncomfortable, and it's totally fine to be 'friends' with enemy combatants, just not to feel obligations towards them mediated by lawful intuitions. We should work out the logic of this with Ben, as it's important, but basically one can have an obligation of beneficence towards a small child, a tame animal, or many other things, but to attribute rights to those things rather than entitlement to beneficence is actually to give up on the belief in rights.

Um. I can imagine that you might find my attitude here frustrating for reasons analogous to how I find it frustrating when people reach for the fake precision of talking about hormone levels and genitalia as separate dimensions in a context where I think the concept they actually want is "biological sex." But that's just the objection I come up with when I run my "How could a hypothetical adversary argue that I'm being hypocritical?" heuristic; it doesn't actually make me feel more eager to talk about criminality myself.


Ben—
> I don't think it's worth your time to read that thread - the result was Zack decided to do all the public interpretive labor himself, just asked Eliezer for an endorsement of a pretty straightforward but insightful extension of A Human's Guide To Words, and then we have the recent thread. [...] The coordination capital we are building seems way more important than persuading Eliezer.

Jack—
> And the more this sort of thing continues the more embarrassed I am to be associated
> Ptting this level of effort in looks pitiful
> Like actually just pathetic
> Sharing thread with the 3 most promising MIRI employees and offering to chat with them about it seems like not a horrible idea
> But sharing it with the whole cult just seems like poking a nest of dog shit - nothing in this thread was sufficiently obviously bad by the standards of either cynicism or naive pattern matching, so making a fuss rather than writing off the troll looks either stupid (can't see obvious troll) or weak (have nothing better to do than kick a pile of excrement)
> Oh or weak like Sarah read it
> Someone needs to slap Zack. Not literally, probably.
> But I already don't have much hope for the level of cultist cuck that Nick is. Zack seems like he might be further in that direction, but otoh he at least has narratized internal conflict about it.

> Zack sacrificed all hope of success in favor of maintaining his own sanity by CC'ing you guys (which I think he was correct to do conditional on email happening at all). We're (accurately) narratized as a political coalition who people feel obligated to talk to but to not take too seriously.


I was _just_ trying to publicly settle a very straightforward philosophy thing that seemed really solid to me

if, in the process, I accidentally ended up being an unusually useful pawn in Michael Vassar's deranged four-dimensional hyperchess political scheming

that's ... _arguably_ not my fault



"Harry, were you involved in the breakout from Azkaban?" "Involved doesn't have a simple truth condition ..."



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 (like with the blue-eyed islanders), 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)

(also, I almost certainly don't have a coherent view of what "unjustly-acquired rents" are)

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"

I'm not supposed to be sending email during Math and Wellness Month, but I'm sending/writing this anyway because it's Wellness-relevant

You've got to be strong to survive in the O-ring sector

(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)

This is genuinely pretty bad

I should do some of the logging-package prototyping stuff on the weekend (it looks fun/interesting enough such that I might be able to psych myself into thinking of it as a "fun" project)
 
30 April saying, essentially (and sincerely), "Oh man oh jeez, Ben and Michael want me to join in a rationalist civil war against the corrupt mainstream-rationality establishment, and I'd really rather not, and I don't like how they keep using scary hyperbolic words like 'cult' and 'war' and 'criminal', but on the other hand, they're the only ones backing me up on this incredibly basic philosophy thing and I don't feel like I have anywhere else to go."

(In a darkly adorable twist, Mike and Alicorn's two-year-old son Merlin was reportedly saying the next day that Kelsey doesn't like his daddy, which was confusing until it was figured out he had heard Kelsey talking about why she doesn't like Michael Vassar.)

Anyway, all of this leading up to a psychological hypothesis (maybe this is "obvious", but I hadn't thought about it before): when people see someone wavering between their coalition and a rival coalition, they're motivated to offer a few concessions to keep the wavering person on their side. Kelsey could afford to not-understand (pace Upton Sinclair) the "words need to carve reality at the joints" thing when it was just me freaking out alone, but "got it" almost as soon as I could credibly threaten to walk (defect to a coalition of people she dislikes)—and maybe my "closing thoughts" email had a similar effect on Eliezer (assuming he otherwise wouldn't have spontaneously tweeted something about the univariate fallacy two weeks later)??

This probably doesn't work if you repeat it (or if you try to do it consciously)?

> I keep scratching my head here Zack. Weaker language on this would just be weaker.  We aren't dealing with a problem of pedagogy here. We are dealing with a group of people who literally got together and built a movement around the need to do charity non-fraudulently but who were eventually shaped by the feedback systems they were in into doing exactly the thing they built their movement in opposition to while claiming that "anyone else would have done it and anyway some of us are vegans".  Nobody is endorsing punishing people for doing things that anyone else would have done, but we are asking them to stop preventing others from trying to do actual charity and actual science, which they are currently actually opposing in order to defend their egos. 

Yes, this is a lame rationalization: if everyone actually knew, what would the act be for? Present day senile!Eliezer wrote of people being "pseudo-fooled in part of themselves even when they are not really-fooled" as if it were no big deal, but 2007!Eliezer identified people being pseudo-fooled as a bad thing. I'll add this to my post-ideas list (working title: "Fake Common Knowledge"). 

he seems to be conflating transhumanist optimism with "causal reality", and then tone-policing me when I try to model good behavior of what means-end reasoning about causal reality actually looks like. This ... seems pretty cultish to me?? Like, it's fine and expected for this grade of confusion to be on the website, but it's more worrisome when it's coming from the mod team.

I talked with Ray from 7 to 10:30 tonight. My intellectual performance was not so great and I did a lot of yelling and crying. Ray thinks that people need to feel safe before they can seek truth, because otherwise they'll distort stuff until they can feel safe. I think this is mostly true: it's not a coincidence that I put out a poor intellectual performance during the same conversations that I do a lot of yelling and crying. But then Ray goes on to say something about stag hunts and coordination, and I'm like, "In slogan form, 'Telling the truth is not a coordination problem!' You can just unilaterally tell the truth!" And ... it doesn't seem to land?

I wish I hadn't done so much yelling and crying with Ray (it's undignified, and it makes me stupid), but when people seem to be pretty explicitly denying the possibility of actually reasoning with them, it's as if I don't perceive any alternative but to express myself in a language they might have a better hope of understanding

Also, I think it's pretty ironic that this happened on a post that was explicitly about causal reality vs. social reality! It's entirely possible that I wouldn't feel inclined to be such a hardass about "Whether I respect you is off-topic" if it weren't for that prompt!


I think mens rea as a concept is pretty clearly necessary for good incentives, at least some of the time. The law distinguishes between accidentally hitting a pedestrian in one's car ("manslaughter") from premeditated killing ("first-degree murder"), and this is an important distinction, because traffic accidents are significantly less disincentivizable ("elastic"?) than offing one's enemies. (Anyone who drives is going to have analogues in nearby possible worlds who are guilty of vehicular manslaughter.) In the absence of mindreading tech, it may not be easy for the court to determine intent, but it should be possible in at least some cases: if the victim is a stranger and the weapon was a car, then it's probably manslaughter; if the victim was recently revealed to have slept with the perp's wife and the weapon was a gun purchased the previous day, then it's probably murder.

I would be pretty surprised if any of you disagree with this, at least in the case of homicide? The case of misinformation that moves resources is a lot trickier, to be sure (for one extremely important disanalogy, perps don't gain from manslaughter), but I don't see what would reverse the extremely weak principle that "intent matters, somewhat, sometimes"?


1. I didn't stay on them for long enough for the full/interesting effects. (I chickened out at 5 months because of health/fertility conservatism; might try again sometime.)

2. I in particular am unusually bad at introspection?? (I think I mentioned this in the posts: there could very well have been lots of changes that I just didn't consciously/verbally register as "change due to HRT".) Possible evidence: I drink coffee often, but I don't feel like I can consciously notice its effects most of the time, except when I'm very sleepy, feel not-sleepy some time after drinking coffee, and infer, "Oh, I'm not sleepy any more; it must have been the coffee, which I know is a stimulant because everyone says so and I believe them." But the fact that I feel motivated to seek out coffee even though it doesn't "objectively" taste particularly good suggests that the drug is doing something to my brain even though I wouldn't be very good at blogging about it.

3. People-in-general's verbal self-reports are heavily influenced by priors/selective-attention: if you expect HRT (or coffee) to drastically change your experience, you'll find things to notice; if you don't, then you won't. The actual effects of biochemistry are presumably real and in-principle-measuable, but the translation from "brain biochemistry" to "blog post" destroys almost all non-culturally-preconceived information.


 My previous understanding of the present struggle (filtered through the lens of my idiosyncratic Something to Protect) seemed like a simple matter of Bay Area political correctness; I didn't really feel like I understood what Michael/Ben/Jessica were seeing as a more general dire problem. But now we have two Less Wrong moderators (Ruby and Ray, not rando n00bs) who seem pretty actively opposed to honest discourse (on the grounds that we need to solve the coordination problem of colluding to protect each other's feelings before we can feel safe enough to have arguments). Michael's cockamamie social theories about there being a socio-psychological attractor optimizing for fraud-in-general are actually starting to seem plausible??

Steven, if you have time and are interested, could you read the secret ("secret") thread and my earlier meta-scuffle with Ruby and a just-now thread with Ray tell me if you're seeing what I'm seeing?

Anna, I continue to be pretty baffled at how (from my perspective, stated bluntly) you seem to be basically uninterested in the mounting evidence that your entire life's work is a critical failure?

(Ruby's profile says he's one of two people to have volunteered for CfAR on three continents. If this is the level of performance we can expect from a veteran CfAR participant, what is CfAR for? Like, yes, we've succeeded at getting money and eyeballs pointed at the alignment problem, which is good. But I pretty clearly remember back in 2009 that we thought this was going to take more than money and eyeballs. If we were wrong, that's good news for humanity! But there should be some sort of argument that explains why we changed our minds, rather than everyone seemingly just conveniently forgetting the original vision.)

Really interested in spending time with either of you in person to double-crux on some of this! (Or more Friendship maintenance with Anna.)

>  People sometimes think of it as "a math problem", but that's because of the branding, not because of the reality.  Yes, a lot of it is a math problem, but it's also an analytic philosophy problem, a cognitive science problem, a social psychology problem, a sociobiology problem, a sociology of science problem, a political strategy problem, and so on.  Many of these are politics-laden domains, where it isn't sufficient to merely be skeptical; it's also necessary to directly talk about politics without the conversation getting derailed.

Jessica—
> 1. GiveWell (talking about how current charities are inefficient/fraudelent/useless, and how it's possible to do better with cost-benefit analyses)
> 2. The Atheism movement (talking about how religion is false and people who promote it are being deceptive)
> 3. The fable of the dragon tyrant (talking about how deathists are being insane, worshipping death instead of fighting it)
> 4. AI safety as a concern for people not in the AI field (talking about how current AGI approaches would result in the discussion of the world; isn't that saying that what AI researchers are doing is bad, and they won't figure out how to do better themselves?)
> 5. The Sequences (talking about human rationality, including talking about lies and dark side epistemology)
> 6. Inadequate Equilibria (talking about how e.g. academia is producing disinformation or useless research, due to perverse incentives)

What I was trying to point out in the thread is a pattern where people say, "Yes, of course criticism and truthseeking is important; I just think that tact is important, too," only to go on and dismiss any particular criticism as insufficiently tactful. More generally: the surface indicators of rationality are easy to fake, and if we don't know how to detect when someone is faking, then it might be surprisingly hard to notice if fake versions of "rationality" are memetically outcompeting better versions. (To which a standard reply seems to go: "Ah, but the behavior of claiming that the surface indicators are easy to fake and that you know how to do better is also easy to fake." Well, yes. That's why this is a hard problem.)



Ben and Jessica, what did you get out of yesterday's meeting? Do you even remember what anyone said?

I think it was Vaniver who brought up "sanity-maximizers" and admitted that that's not what "the team" is trying to do (because Society would destroy it for heresy, like Anna has told me before)? I like Vaniver, who seemed to "get" me.

For me and my selfish perspective, I think the main outcome is finally shattering my "rationalist" social identity. (We knew this was coming, and I still need to mourn, but I needed to exhaust all possible avenues of appeal before it became real to me; this is the first morning where "rationalists ... them" feels more natural than "rationalists ... us".)

I hope I wasn't too emotional, but I guess I had nothing to lose. I have my friends and I have money, and if Oli thinks I'm annoying, I guess I can live with that. (I mean, that which can be destroyed by the truth should be: I am, in fact, annoying.)

Ben—

> Seems like Oli would rather have ongoing internal political disagreements than make Lesswrong anything in particular. Vaniver actually believes that they should behave esoterically. Jim claimed that he updated during the meeting towards believing that norms had in fact shifted - possible he’ll make more downstream updates later. I don’t know how to characterize Ray.

> Everyone agreed that Zack was right and Scott wrong on the substantive issue of category boundaries, and I think agreed with my characterization of Scott as conflating basic decency with giving into extortion.

> There basically seems like there is no appetite for becoming the sort of counterparty we could negotiate with, and it seems like they think we’re probably right to leave.

I actually still feel pretty motivated to write for Less Wrong, but now more in the spirit of "marketing channel for my own writing" and "fuck you guys" rather than my previous hope of recapturing the beacon.

I was really grateful to Jim for piping up with, "We are somewhat cowardly about what we curate" (concerning why "Where to Draw the Boundaries?" didn't get Curated). (I'm not sure if he actually used the word "cowardly", but that was the sense.) I was like, "Thanks, that makes me respect you more!" (Ray was maintaining that my pedagogy wasn't good enough.)

I am a high-neuroticism person (the sex difference is only d = 0.4, so this isn't model-breakingly surprising; it's like being a woman who's 6'2")

In the absence of more details, I'm inclined to speculate that what looks from one perspective like "worse at predicting other people" is actually something more like "preexisting social naïveté (the kind that goes along with an autism-spectrum quotient +1.7 standard deviations from the male mean) plus more courage equals less social savviness, which is horrifying from the perspective from someone who thinks of savviness as a survival skill, but OK from the perspective of someone following a different 'life-history strategy' that regards savviness as non-normative."

Suppose there are 5 true heresies, but anyone who's on the record believing more than 1 gets burned as a witch. Then it's impossible to have a unified rationalist community, because people who want to talk about 1 heresy can't let themselves be seen in the company of people who believe another. This is why Scott Alexander can't get the philosophy-of-categorization right in full generality (even though he's written exhaustively about the correct answer, and he and I have a common enemy in the social-justice egregore): he can't afford to. He's already spent his Overton budget on anti-feminism.



I told Anna about Michael's "enemy combatants" metaphor, and how I originally misunderstood the point of the analogy. War metaphors sound Scary and Mean—I don't want to shoot my friends! But the point of the analogy (which Michael had explained at the time, but I wasn't ready to hear until I did a few more weeks of emotional processing) was specifically that soliders on the other side of a war aren't particularly morally blameworthy as individuals: their actions are just being controlled by the Power they're embedded in. And Anna was like, "But you could still be friends with someone on an animal level, like with a dog", and I was like, "Yeah, that's basically what Michael said."

I'm not worried about scenarios where posts criticizing MIRI are marked as spam. That would be too obvious. You are a smart and moral person who would never consciously do anything obviously corrupt.

I'm more worried about elephant-in-the-brain scenarios where, for example, "rationalist" higher-ups who don't consciously think of themselves as corrupted by political power games nevertheless write superficially-plausible bullshit hit pieces denouncing their political rivals as "likely to lie or manipulate" that get highly upvoted based on the author's high status (rather than anyone having thought critically about the content of the charges) and later seemingly-positively cited by their colleagues as an example of the community immune system making an update.


In summary, what I see is a culture that has given up on the common interest of many causes, a culture where even our best and brightest are mostly focused on colluding to be nice to each other, minimizing social attack surface, and funneling money to branded EA institutions—even when their own fans can see that they're capable of so much more. In contrast, I want to work on asking interesting questions and then getting the right answer for the right reasons, in public (so that people who aren't already my closest trusted friends can learn, too). Under these circumstances, I don't want the main gravity well to award itself more authority to clear the neighborhood around its orbit, because I haven't been given sufficient reason not to expect that authority to be wielded against me and mine.

 * "losing his ability to model other people and I'm worried about him", I think Ben-and-Jessica would see as [...] angry about living in simulacrum level 3 and we're worried about everyone else."


It seems to that according to Zack's own account, even writing the memoir privately feels like an act of war that he'd rather avoid, not just using his own territory as he sees fit to create internal clarity around a thing. 

I think this has to mean either
(a) that Zack isn't on the side of clarity except pragmatically where that helps him get his particular story around gender and rationalism validated
or
(b) that Zack has ceded the territory of the interior of his own mind to the forces of anticlarity, not for reasons, but just because he's let the anticlaritarians dominate his frame.

This seems like a pretty important impasse to resolve. I don't think even Zack seriously thinks option (a) is sustainable long-run, which is some evidence for (b) but not decisive. I'm not sure (a) and (b) are actually separate hypotheses, rather than just different frames.

or (c) I've ceded the territory of the interior of my own mind to Eliezer Yudkowsky in particular, and while I've made a lot of progress unwinding this, I'm still, still not done, and him seeming to respect me in person at the Newtonmas party set me back a bit ("Ooh, I wish you'd want me to stay / I'll be alright / Just not tonight")

So a philosopher trying to recover a partially-destroyed word using ostensive definition needs to say, "You know all these specific examples (training data) of things we call 'X'? Okay, so, that training data has been mislabeled. Actually, these-and-such specific things are false-positives (your corrupt Society calls them 'X', but they're not) and these-and-such specific things are false-negatives (your corrupt Society doesn't call them 'X', but they are)—and that's the true meaning of 'X'."

angels who timelessly perceive the universal wavefunction sub specie aeternitatis would have no motivation for a shorter encoding for "sex characteristics not changeable at tech level x". But mortal creatures who actually live in a Society with tech level x:=x₀ would, because they need it to compress their descriptions of the world they see (in which, e.g., trans men sometimes get pregnant).

Oh, I think I see what you're getting at if I forget the "application" and just think about binary vectors. In the beginning, there are two kinds of vectors: [0, 0, 0, 0, ... 0], and [1, 1, 1, 1, ..., 1]. As technology gets better, we gain the ability to flip more coordinates from 0 to 1. First we learn how to make [1, 0, 0, 0, ... 0], then we learn how to make [1, 1, 0, 0, 0, ... 0], then [1, 1, 1, 0, 0, ... 0].

I think you're saying it's lame to make a classifier with a pretentious name that's like, "One 1, then zeroes" when the tech allows that, but then the classifier gets updated to "Two 1s, then zeroes" when the tech allows that, &c.

What I'm saying is that, in any tech regime between the stone age and the Singularity, your population does in fact have three distinct clusters that you might want to use language to talk about: the "all 0's" cluster, the "some 1s, then 0s" cluster, and the "all 1s" cluster. In the very early and very late tech regimes, maybe you don't care about the middle category and just use a two-category system. I think that we live in an early regime that needs three categories.

As far as the state ideology of California is concerned, people who learn math from textbooks rather than dutifully attending classes don't exist. As far as the ideology taught in those schools' social-science classes is concerned, males who wish they were shaped like you but think this is obviously not an intersex condition don't exist and merely having a word for to name them is harmful. I think it makes sense that I'm at war here! If I can avoid it, I don't want to cause collateral damage to you if you're at war with a strain non-progressive authoritarian collectivism which is lower on my enemy list than yours. But if I can't cheaply avoid it (in that I don't want to say silent in situations like this or pay a disclaimer tax), maybe we need to accept some small amount of collateral damage from each other???

Let's add the "lawyer" structure back in. I'm sitting at a restaurant and someone else complains that they ordered a burger but got a veggie burger. Why is it trolling if I speak up in their defense, but not if I speak up about my own burger order? It seems to me that the same interest is being defended (the right to use unmarked "burger" to mean "meat(1)" and expect people to know what you meant rather than pretend not to in order to save animals), and that's what matters (rather than whose name gets listed as "defendant" or who has the philosophy skills to explain why they feel cheated by the veggie burger).


> If you are rounding off "they're trying to make me say I'm one of these other things instead of 'autogynephile'" to "they're trying to kill me" then, uhh, stop trying to kill trans and nonbinary people who Blanchard's model fits poorly???  (Like, this is all really hyperbolic!  Maybe we shouldn't do this silly exaggeration thing???)
> - Like, if you can understand why being misrepresented even in this slight way is painful, can't you try to reduce the harm that would come from imposing an ill-fitting ontology on lots of socially vulnerable people???
> - (Also some trans people don't think it's an intersex brain thing!!  They might have a social constructionist view etc.)
> So it seems like you're going about this "getting a short message length" thing in an overly zero-sum fashion (trying to gain representation at the expense of other people's representation).

(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.

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)

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.