Ziz recounted [her](/2019/Oct/self-identity-is-a-schelling-point/) story [of Anna's discrimination](https://sinceriously.fyi/net-negative), how she engaged in [conceptual warfare](https://sinceriously.fyi/intersex-brains-and-conceptual-warfare/) to falsely portray Ziz as a predatory male. I was unimpressed: in my worldview, I didn't think Ziz had the right to say "I'm not a man," and expect people to just believe that. (I remember at one point, Ziz answered a question with, "Because I don't run off masochistic self-doubt like you." I replied, "That's fair.") But I did respect how Ziz actually believed in an intersex brain theory: in Ziz and Gwen's worldview, people's genders were a _fact_ of the matter, not just a manipulation of consensus categories to make people happy.
-Probably the most ultimately significant part of this meeting for future events was Michael verbally confirming to Ziz that MIRI had settled with a disgruntled former employee, Louie Helm, who had put up a website slandering them. I don't actually know the details of the alleged settlement. (I'm working off of [Ziz's notes](https://sinceriously.fyi/intersex-brains-and-conceptual-warfare/) rather than particularly remembering that part of the conversation clearly myself; I don't know what Michael knew.) What was significant was that if MIRI _had_ paid Helm as part of an agreement to get the slanderous website taken down, then, whatever the nonprofit best-practice books might have said about whether this was a wise thing to do when facing a dispute from a former employee, that would decision-theoretically amount to a blackmail payout, which seemed to contradict MIRI's advocacy of timeless decision theories (according to which you [shouldn't be the kind of agent that yields to extortion](/2018/Jan/dont-negotiate-with-terrorist-memeplexes/)).
+Probably the most ultimately significant part of this meeting for future events was Michael verbally confirming to Ziz that MIRI had settled with a disgruntled former employee, Louie Helm, who had put up [a website slandering them](https://archive.ph/Kvfus). I don't actually know the details of the alleged settlement. (I'm working off of [Ziz's notes](https://sinceriously.fyi/intersex-brains-and-conceptual-warfare/) rather than particularly remembering that part of the conversation clearly myself; I don't know what Michael knew.) What was significant was that if MIRI _had_ paid Helm as part of an agreement to get the slanderous website taken down, then, whatever the nonprofit best-practice books might have said about whether this was a wise thing to do when facing a dispute from a former employee, that would decision-theoretically amount to a blackmail payout, which seemed to contradict MIRI's advocacy of timeless decision theories (according to which you [shouldn't be the kind of agent that yields to extortion](/2018/Jan/dont-negotiate-with-terrorist-memeplexes/)).
----
Despite Math and Wellness Month and my "intent" to take a break from the religious civil war, I kept reading _Less Wrong_ during May 2019, and ended up scoring a couple of victories in the civil war (at some cost to Wellness).
-MIRI researcher Scott Garrabrant wrote a post about how ["Yes Requires the Possibility of No"](https://www.lesswrong.com/posts/G5TwJ9BGxcgh5DsmQ/yes-requires-the-possibility-of-no). Information-theoretically, a signal sent with probability one transmits no information: you can only learn something from hearing a "Yes" if there was some chance that the answer could have been "No". I saw an analogy to my philosophy-of-language thesis, and commented about it: if you want to believe that _x_ belongs to category _C_, you might try redefining _C_ in order to make the question "Is _x_ a _C_?" come out "Yes", but you can only do so at the expense of making _C_ less useful.
+MIRI researcher Scott Garrabrant wrote a post about how ["Yes Requires the Possibility of No"](https://www.lesswrong.com/posts/G5TwJ9BGxcgh5DsmQ/yes-requires-the-possibility-of-no). Information-theoretically, a signal sent with probability one transmits no information: you can only learn something from hearing a "Yes" if there was some chance that the answer could have been "No". I saw an analogy to my philosophy-of-language thesis, and commented about it: if you want to believe that _x_ belongs to category _C_, you might try redefining _C_ in order to make the question "Is _x_ a _C_?" come out "Yes", but you can only do so at the expense of making _C_ less useful. Meaningful category-membership (Yes) requires the possibility of non-membership (No).
[TODO: explain scuffle on "Yes Requires the Possibility"—
> I'm also not sure if I'm sufficiently clued in to what Ben and Jessica are modeling as Blight, a coherent problem, as opposed to two or six individual incidents that seem really egregious in a vaguely similar way that seems like it would have been less likely in 2009??
- * _Atlas Shrugged_ Bill Brent vs. Dave Mitchum scene
* Vassar: "Literally nothing Ben is doing is as aggressive as the basic 101 pitch for EA."
* Ben: we should be creating clarity about "position X is not a strawman within the group", rather than trying to scapegoat individuals
- * my scuffle with Ruby on "Causal vs. Social Reality"
+ * my scuffle with Ruby on "Causal vs. Social Reality" (my previous interaction with Ruby had been on the LW FAQ; maybe he couldn't let me "win" again so quickly?)
* it gets worse: https://www.lesswrong.com/posts/xqAnKW46FqzPLnGmH/causal-reality-vs-social-reality#NbrPdyBFPi4hj5zQW
* Ben's comment: "Wow, he's really overtly arguing that people should lie to him to protect his feelings."
* Jessica: "tone arguments are always about privileged people protecting their feelings, and are thus in bad faith. Therefore, engaging with a tone argument as if it's in good faith is a fool's game, like playing chess with a pigeon. Either don't engage, or seek to embarrass them intentionally."
* there's no point at being mad at MOPs
* me (1 Jul): I'm a _little bit_ mad, because I specialize in cognitive and discourse strategies that are _extremely susceptible_ to being trolled like this
+ * me to "Wilhelm" 1 Jul: "I'd rather not get into fights on LW, but at least I'm 2-0-1"
* "collaborative truth seeking" but (as Michael pointed out) politeness looks nothing like Aumann agreement
* 2 Jul: Jessica is surprised by how well "Self-consciousness wants to make everything about itself" worked; theory about people not wanting to be held to standards that others aren't being held to
* Michael: Jessica's example made it clear she was on the side of social justice
> I am actively hostile to neoreaction and the alt-right, routinely block such people from commenting on my Twitter feed, and make it clear that I do not welcome support from those quarters. Anyone insinuating otherwise is uninformed, or deceptive.
-[TODO: I objected that he shouldn't pay the Danegeld like this]
+[I pointed out that](https://twitter.com/zackmdavis/status/1164259164819845120) the people who smear him as a right-wing Bad Guy do so _in order to_ extract these kinds of statements of political alignment as concessions; his own timeless decision theory would seem to recommend ignoring them rather than paying even this small [Danegeld](/2018/Jan/dont-negotiate-with-terrorist-memeplexes/).
When I emailed the posse about it begging for Likes (Subject: "can't leave well enough alone"), Jessica said she didn't get my point. If people are falsely accusing you of something (in this case, of being a right-wing Bad Guy), isn't it helpful to point out that the accusation is actually false? It seemed like I was advocating for self-censorship on the grounds that speaking up helps the false accusers. But it also helps bystanders (by correcting the misapprehension), and hurts the false accusers (by demonstrating to bystanders that the accusers are making things up). By linking to ["Kolmogorov Complicity"](http://slatestarcodex.com/2017/10/23/kolmogorov-complicity-and-the-parable-of-lightning/), I seemed to be insinuating that Yudkowsky was under some sort of duress, but this wasn't spelled out: if Yudkowsky would face social punishment for advancing right-wing opinions, did that mean he was under such duress that saying anything at all would be helping the oppressors?
I perceived a pattern where people who are in trouble with the orthodoxy feel an incentive to buy their own safety by denouncing _other_ heretics: not just _disagreeing_ with the other heretics _because those other heresies are in fact mistaken_, which would be right and proper Discourse, but denouncing them ("actively hostile to") as a way of paying Danegeld.
-Suppose there are five true heresies, but anyone who's on the record believing more than one gets burned as a witch. Then it's impossible to have a unified rationalist community, because people who want to talk about one heresy can't let themselves be seen in the company of people who believe another. That's why Scott Alexander couldn't get the philosophy-of-categorization right in full generality (even though he'd [written](https://www.lesswrong.com/posts/yCWPkLi8wJvewPbEp/the-noncentral-fallacy-the-worst-argument-in-the-world) [exhaustively](https://slatestarcodex.com/2014/11/03/all-in-all-another-brick-in-the-motte/) about the correct way, and he and I have a common enemy in the social-justice egregore): _he couldn't afford to_. He'd already spent his Overton budget on anti-feminism.
+Suppose there are five true heresies, but anyone who's on the record believing more than one gets burned as a witch. Then it's impossible to have a unified rationalist community, because people who want to talk about one heresy can't let themselves be seen in the company of people who believe another. That's why Scott Alexander couldn't get the philosophy-of-categorization right in full generality (even though he'd [written](https://www.lesswrong.com/posts/yCWPkLi8wJvewPbEp/the-noncentral-fallacy-the-worst-argument-in-the-world) [exhaustively](https://slatestarcodex.com/2014/11/03/all-in-all-another-brick-in-the-motte/) about the correct way, and he and I have a common enemy in the social-justice egregore): _he couldn't afford to_. He'd already [spent his Overton budget on anti-feminism](https://slatestarcodex.com/2015/01/01/untitled/).
Scott (and Yudkowsky and Anna and the rest of the Caliphate) seemed to accept this as an inevitable background fact of existence, like the weather. But I saw a Schelling point off in the distance where us witches stick together for Free Speech, and it was _awfully_ tempting to try to jump there. (Of course, it would be _better_ if there was a way to organize just the good witches, and exclude all the Actually Bad witches, but the [Sorites problem](https://plato.stanford.edu/entries/sorites-paradox/) on witch Badness made that hard to organize without falling back to the falling back to the one-heresy-per-thinker equilibrium.)
>
> Also to be clear: I try not to dismiss ideas out of hand due to fear of public unpopularity. However I found Scott Alexander's takedown of neoreaction convincing and thus I shrugged and didn't bother to investigate further.
-My "negotiating with terrorists" criticism did _not_ apply to the 2013 statement. "More Right" _was_ brand encroachment on Anissimov's part that Yudkowsky had a legitimate interest in policing, _and_ the "I try not to dismiss ideas out of hand" disclaimer importantly avoided legitimizing the McCarthyist persecution.
+My "negotiating with terrorists" criticism did _not_ apply to the 2013 statement. "More Right" _was_ brand encroachment on Anissimov's part that Yudkowsky had a legitimate interest in policing, _and_ the "I try not to dismiss ideas out of hand" disclaimer importantly avoided legitimizing [the McCarthyist persecution](https://www.unqualified-reservations.org/2013/09/technology-communism-and-brown-scare/).
The question was, what had specifically happened in the last six years to shift Eliezer's opinion on neoreaction from (paraphrased) "Scott says it's wrong, so I stopped reading" to (verbatim) "actively hostile"? Note especially the inversion from (both paraphrased) "I don't support neoreaction" (fine, of course) to "I don't even want _them_ supporting _me_" [(_?!?!_)](https://twitter.com/zackmdavis/status/1164329446314135552).[^them-supporting-me]
-----
-In November 2019, I received an interesting reply on my philosophy-of-categorization thesis from MIRI researcher Abram Demski. Abram asked: ideally, shouldn't all conceptual boundaries be drawn with appeal-to-consequences? Wasn't the problem just with bad (motivated, shortsighted) appeals to consequences? Agents categorize in order to make decisions. The best classifer for an application depends on the costs and benefits. As a classic example, it's very important for evolved prey animals to avoid predators, so it makes sense for their predator-detection classifiers to be configured such that they jump away from every rustling in the bushes, even if it's usually not a predator.
+On 3 November 2019, I received an interesting reply on my philosophy-of-categorization thesis from MIRI researcher Abram Demski. Abram asked: ideally, shouldn't all conceptual boundaries be drawn with appeal-to-consequences? Wasn't the problem just with bad (motivated, shortsighted) appeals to consequences? Agents categorize in order to make decisions. The best classifer for an application depends on the costs and benefits. As a classic example, it's very important for evolved prey animals to avoid predators, so it makes sense for their predator-detection classifiers to be configured such that they jump away from every rustling in the bushes, even if it's usually not a predator.
I had thought of the "false-positives are better than false-negatives when detecting predators" example as being about the limitations of evolution as an AI designer: messy evolved animal brains don't bother to track probability and utility separately the way a cleanly-designed AI could. As I had explained in "... Boundaries?", it made sense for _what_ variables you paid attention to, to be motivated by consequences. But _given_ the subspace that's relevant to your interests, you want to run an epistemically legitimate clustering algorithm on the data you see there, which depends on the data, not your values. The only reason value-dependent gerrymandered category boundaries seem like a good idea if you're not careful about philosophy is because it's _wireheading_. Ideal probabilistic beliefs shouldn't depend on consequences.
------
-On 20 December 2019, Scott Alexander messaged me on Discord—that I shouldn't answer if it would be unpleasant, but that he was thinking about asking about autogynephilia on next _Slate Star Codex_ survey, and wanted to know if I had any suggestions about question design, or a suggestion of who to consult on "the other side" of the issue. After reassuring him that he shouldn't worry about answering being painful for me ("I am actively at war with the socio-psychological forces that make people erroneously think that talking is painful!"), I referred him to my friend [Tailcalled](https://surveyanon.wordpress.com/), who I thought was more qualified. (Tailcalled had a lot of experience running surveys, and had a better grasp than me of the science of sexology and transgenderism, in constrast to my preoccupation with the philosophical and political aspects of the problem.)
+On 20 December 2019, Scott Alexander messaged me on Discord—that I shouldn't answer if it would be unpleasant, but that he was thinking about asking about autogynephilia on next _Slate Star Codex_ survey, and wanted to know if I had any suggestions about question design, or a suggestion of any "intelligent and friendly opponents" to consult on the other side of the question. After reassuring him that he shouldn't worry about answering being painful for me ("I am actively at war with the socio-psychological forces that make people erroneously think that talking is painful!"), I referred him to my friend [Tailcalled](https://surveyanon.wordpress.com/), who I thought was more qualified on both counts. (Tailcalled had a lot of experience running surveys, and ran a "Hobbyist Sexologists" Discord server, which seemed likely to have some friendly opponents.)
The next day (I assume while I happened to be on his mind), Scott also [commented on](https://www.lesswrong.com/posts/bSmgPNS6MTJsunTzS/maybe-lying-doesn-t-exist?commentId=LJp2PYh3XvmoCgS6E) "Maybe Lying Doesn't Exist", my post from back in October replying to his "Against Lie Inflation."
> 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
-Scott messaged back the next morning, Christmas Day. He explained that the thought process behind his comment was that he still wasn't sure where we disagreed, and didn't know how to proceed except to dump his understanding of the philosophy (which would include things I already knew) and hope that I could point to the step I didn't like. He didn't know how to convincingly-to-me demonstrate his sincerity, and rebut my accusations of him motivatedly playing dumb (which he was inclined to attribute to the malign influence of Michael Vassar's gang).
+Scott messaged back at 11:08 _a.m._ the next morning, Christmas Day. He explained that the thought process behind his comment was that he still wasn't sure where we disagreed, and didn't know how to proceed except to dump his understanding of the philosophy (which would include things I already knew) and hope that I could point to the step I didn't like. He didn't know how to convincingly-to-me demonstrate his sincerity, and rebut my accusations of him motivatedly playing dumb (which he was inclined to attribute to the malign influence of Michael Vassar's gang).
I explained that the reason I accused him of being motivatedly dumb was that I _knew_ he knew about strategic equivocation, because he taught everyone else about it (as in his famous posts about [the motte-and-bailey doctrine](https://slatestarcodex.com/2014/11/03/all-in-all-another-brick-in-the-motte/), or [the noncentral fallacy](https://www.lesswrong.com/posts/yCWPkLi8wJvewPbEp/the-noncentral-fallacy-the-worst-argument-in-the-world)). And so when he acted like he didn't get it when I pointed out that this also applied to "trans women are women", that just seemed _implausible_.
-He asked for a specific example. ("Trans women are women, therefore trans women have uteruses," being a bad example, because no one was claiming that.) I quoted [an article from the nationally prominent progressive magazine _The Nation_](https://www.thenation.com/article/trans-runner-daily-caller-terry-miller-andraya-yearwood-martina-navratilova/): "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." Scott agreed that this was stupid and wrong and a natural consequence of letting people use language the way he was suggesting (!).
+He asked for a specific example. ("Trans women are women, therefore trans women have uteruses," being a bad example, because no one was claiming that.) I quoted [an article from the prominent progressive magazine _The Nation_](https://web.archive.org/web/20191223235051/https://www.thenation.com/article/trans-runner-daily-caller-terry-miller-andraya-yearwood-martina-navratilova/): "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_." Scott agreed that this was stupid and wrong and a natural consequence of letting people use language the way he was suggesting (!).
I didn't think it was fair to ordinary people to expect them to go as deep into the philosophy-of-language weeds as _I_ could before being allowed to object to these kinds of Shenanigans. I thought "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 amounted to "wireheading" (optimizing someone's map for looking good rather than reflecting the territory) and "war" (optimizing someone's map to not reflect the territory, in order to gain an advantage over them). If I were to transition today and didn't pass as well as Jessica, and everyone felt obligated to call me a woman, they would be wireheading me: making me think my transition was successful, even though it actually wasn't. That's ... not actually a nice thing to do to a rationalist.
I linked to Zvi Mowshowitz's post about how [the claim that "everybody knows" something](https://thezvi.wordpress.com/2019/07/02/everybody-knows/) gets used an excuse to silence people trying to point out the thing (because they don't see people behaving as if it were common knowledge): "'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?" But I didn't think that everyone knew. I thought the people who sort-of knew were being intimidated into doublethinking around it. I thought this was bad for clarity.
-At this point, Scott mentioned that he wanted to go to the Event Horizon Christmas party, and asked if I wanted to come and continue the discussion there. I assented, and thanked him for his time; it would be really exciting if we could avoid a rationalist civil war. (I thought my "you need accurate models before you can do utilitarianism" philosophy was also near the root of Ben's objections to the EA movement.)
+At this point it was almost 2 _p.m._ (the paragraphs above summarize a larger volume of typing), and Scott mentioned that he wanted to go to the Event Horizon Christmas party, and asked if I wanted to come and continue the discussion there. I assented, and thanked him for his time; it would be really exciting if we could avoid a rationalist civil war. (I thought my "you need accurate models before you can do utilitarianism" philosophy was also near the root of Ben's objections to the EA movement.)
-When I arrived at the party, people were doing a reading of the "Hero Licensing" dialogue from _Inadequate Equilibria_. Yudkowsky himself was, playing the part of the Mysterious Stranger in the dialogue. At some point, Scott and I retreated upstairs to continue our discussion. By the end of it, I was at least feeling more assured of Scott's sincerity (rather than him being coerced into not saying anything incriminating over email). Scott said he would edit in a disclaimer note at the end of "... Not Man for the Categories".
+When I arrived at the party, people were doing a reading of [the "Hero Licensing" dialogue epilogue](https://www.lesswrong.com/posts/dhj9dhiwhq3DX6W8z/hero-licensing) to _Inadequate Equilibria_. Yudkowsky himself was, playing the part of the Mysterious Stranger in the dialogue. At some point, Scott and I retreated upstairs to continue our discussion. By the end of it, I was at least feeling more assured of Scott's sincerity (rather than him being coerced into not saying anything incriminating over email). Scott said he would edit in a disclaimer note at the end of "... Not Man for the Categories".
If I also got the chance to talk to Yudkowsky for a few minutes, I don't think I would be allowed to recount any details of that here due to the privacy rules I'm following in this document.
> No. The thing that's been driving me nuts for twenty-one months is that <strong><em><span style="color: #F00000;">I expected Eliezer Yudkowsky to tell the truth</span></strong></em>. I remain,
>
> Your heartbroken student,
-> [...]
+> Zack M. Davis
I followed it up with another email after I woke up the next morning:
> **I really used to look up to you.** In my previous interactions with you, I've been tightly [cognitively constrained](http://www.hpmor.com/chapter/57) by hero-worship. I was already so starstruck that _Eliezer Yudkowsky knows who I am_, that the possibility that _Eliezer Yudkowsky might disapprove of me_, was too terrifying to bear. I really need to get over that, because it's bad for me, and [it's _really_ bad for you](https://www.lesswrong.com/posts/cgrvvp9QzjiFuYwLi/high-status-and-stupidity-why). I remain,
>
> Your heartbroken student,
-> [...]
-
-
+> Zack M. Davis
[TODO: Sep 2020 categories clarification from EY—victory?!
https://www.facebook.com/yudkowsky/posts/10158853851009228
_ex cathedra_ statement that gender categories are not an exception to the rule, only 1 year and 8 months after asking for it
]
+
+[TODO: "Unnatural Categories Are Optimized for Deception"
+
+Abram was right
+
+the fact that it didn't means that not tracking it can be an effective AI design! Just because evolution takes shortcuts that human engineers wouldn't doesn't mean shortcuts are "wrong" (instead, there are laws governing which kinds of shortcuts work).
+
+Embedded agency means that the AI shouldn't have to fundamentally reason differently about "rewriting code in some 'external' program" and "rewriting 'my own' code." In that light, it makes sense to regard "have accurate beliefs" as merely a convergent instrumental subgoal, rather than what rationality is about
+
+somehow accuracy seems more fundamental than power or resources ... could that be formalized?
+]
+
+And really, that _should_ have been the end of the story. At the trifling cost of two years of my life, we finally got a clarification from Yudkowsky that you can't define the word _woman_ any way you like. I didn't think I was entitled to anything more than that. I was satsified. I still published "Unnatural Categories Are Optimized for Deception" in January 2021, but if I hadn't been further provoked, I wouldn't have occasion to continue waging the robot-cult religious civil war.
projects / Ultimately_Untrue_Thought.git / blobdiff