I do not recommend using the word "AI" as if it refers to a single thing that encompasses all possible systems incorporating AI techniques. LLM guys don't distinguish between things that could actually be built and "throwing an LLM at the problem" -- you're treating their lack-of-differentiation as valid and feeding them hype.

I'll head in later and post badly!

(Spoilers: it's probably a metaphor!!!)

Thank you for reading my story!

This started as something terse and didactic, which felt like really bad territory for the piece. I'm kind of relieved that you took away the intended content.

I mean, if no one's getting paid, then my preferred price is $0, to everyone in the world.

You're right! There's a disclosure on the page but it's fuckin tiny.

• high willingness to accept painfully inexact responses
• high tendency to side with authority when given no information
• low ability to distinguish "how it is" from "how it seems like it should be"

Meta:

• default expectation that others are the same way
• indignant consent-ignoring gesture if they're not

Oh, OK. I think all the VC-adjacent people still really believe in crypto, if it helps. They probably also don't believe in it, depending on the room. I think it will come back.

Put me down for "doesn't think it will end." Did crypto end?

It's the technique of running a primary search against some other system, then feeding an LLM the top ~25 or so documents and asking it for the specific answer.

A friend who worked with her is sympathetic to her but does not endorse her: this is a tendency she has, she veers back and forth on it a lot, she has frequent moments of insight where she disavows her previous actions but then just kind of continues doing them. It's Kanye-type behavior.

The media again builds a virtual public consisting of billionaires of a variety of positions and ask you "which one do you agree with?" This is a strategy to push the public closer to the beliefs of billionaires.

I don't know who these fucking people are. The real public in California still supports Biden by a 25% margin.

Really? Weird. Very different experience.

(Maybe crypto is less deteriorative than business?)

The last time I met a person who had done deeply reprehensible, highly publicized tech fraud (FTX executive) he kind of just came off as a dude, and I liked him.

That kind of makes me feel bad when I think about it.

I haven't met a high-profile fraudster lately, but my first impression of bad guys is usually pretty positive. As far as I can tell, people keep their ambient personalities when they break bad, but they compartmentalize and they develop supermassive appetites for praise. This long-run increases their suggestibility because they have to be more and more gullible to not hate themselves. I think this hollows them out -- when you live a double life for long enough, you kind of stop observing the reality-fiction boundary at all.

Not clear how to stop the cycle. There's just too much money involved for me to dive off the train right now.

I've spent a lot of time trying to write without any intentional exercise of style! I think I've read far too much text generated by people on the psychotic spectrum to actually manage this.

Last paragraph first: Grudgingly, yeah, that's a pretty good literal answer to the question. Peter Thiel won't sell just anyone a cult following, and you're not paying for it in cash, but he will sell you one if you're lucky.

Writing advice: I like your writing. I haven't tried to emulate you because I haven't read enough of your writing, and because when I made my first brush with you (which was like a year ago) I was spending a lot less time emulating people in general.

It's a little distressing to me because, well, I'm way too anxious to play the game of moral righteousness straight-facedly. It takes a very different personality from mine to say "Those are the bad people, fuck them" and not see the obvious similarities between me and the people I hate.

Some level of this is actual, real-world hypocrisy: I'm the cofounder of an AI startup and at the same time I deeply dislike AI. I went here because one, there was money, and two, I didn't want a way worse person than me to take the same job. It has not been what I hoped for -- it has been deeply destructive to my personality -- it has taught me a lot and made me much more cynical -- it has definitely made me stupider.

I don't really know how to do a hypocrisy purge. (I hear this is what ayahuasca is for, but Catholicism also works, and I'm considering getting my brain tattooed with a laser gun.) I think until I do one I have to temper all my moral righteousness by saying "I think I know why this person is doing the thing they're doing, and if you want their (bad) motives, here's my guess."

I think most people would see higher performance on general tasks on Adderall. Not sure if this is actually a good reason to put everyone on Adderall.

Side effects can be pretty brutal, although people who abuse caffeine to get the same level of stimulation are going to probably have them a lot worse.

My actual experience is that LLMs seem to basically just become a third arm for people who use them. Google is like that too, but for their target audience, LLMs are more like that.

You don't love your arm, but if someone goes to you like, "Do you mind if I cut your arm off?" of course you say "do not." If someone's like "OK, but like, if I made you choose between your wife and your arm" you'd be like "That's incredibly perverse. I need my arm."

For people who use them it seems like it really quickly became impossible to exist without them. That's one of the reasons I think they're not going away.

Ack, that makes me want to reappraise him. It's likely there's a version of him who exists in my head who writes a little better than the version who writes on the page. I'm definitely guilty of skimreading him a lot.

I can't think of a non-metaphorical expansion of your take that isn't (1) deeply insensitive to my stated needs (2) a generally poor reading of the original post (3) at odds with basic understanding of what the function of language is.

I don't know exactly what you think I want. I want to be understood and I want to be seen as good based on that understanding. I'm not asking for a Spock-level mind-meld with the opposing party. I'm not asking that every single person in the world understand me exactly as intended the first time they read it. I'm asking for an end to smug, self-satisfactory, nitpicking interpretations ultimately designed to draw me into shaming-based social rituals that I refuse to be a part of.

Maybe it would be helpful for me to clarify a specific example of what I'm so pissed about. It appeared in the original post but I could have been clearer. The thing I'm pissed about in this case is that you can't mention Scott Alexander here without performatively mocking him or explaining why you didn't performatively mock him, which I know because I've watched other people try it. (The only reason you didn't see a henpecking response in this case is that in my original post, I spent two paragraphs heading it off.)

The general pattern of my existence online is that whenever I acknowledge a political position that's unpopular, or the existence of a political figure that's unpopular, even if I'm taking great pains to indicate that I disagree with it, people will arrive to specifically accuse me of believing the exact opposite of what I said I believed. It's entirely possible that the inadequacy of language plays some role here, but the apparent reason the communication fails is that something about me seems to have caused the other person to decide they want to force me into the conceptual category of "people they hate."

I am not a particularly pleasant person! I often try to be, but like, I actually have to try. I think it is common for people to decide that they dislike me before they have a clear reason why. But I also think a lot of people engage with online content in a way that is purely based on skimming takes off the top, analyzing them for their badness, and announcing personal superiority to the people who had the gall to post bad takes.

None of this falls into the territory covered by your impossibility result from systems biology regarding language. (although I doubt the impossibility result to begin with) This is mostly accounted for by pernicious cultiness of advanced online communities, and the futile and self-negating way I have to struggle to correct for it.

The uncharitable interpretation of your comment is that you think communication is impossible. If you really, sincerely think one person communicating an idea to another person successfully is impossible, burn all the textbooks and also most of the professors. If it takes equivocating over "full" communication and you're willing to concede the point as far as other stuff goes then fine, my red may be your blue. I'm at peace with that.

If you think there are some things that could be communicated linguistically but generally aren't, for a reason that is not the fault of the speaker or the hearer, I agree. It doesn't cause me distress when someone still assumes good faith about me but also misunderstands me -- I've talked about what causes me distress. If it's not obvious to you that people who post takes that go beyond the superficial attract way more of that distress -- I mean, the sealioning and tedium I'm usually met with -- then I want to post on whatever internet you grew up on, because mine is defective.

You have added, as a consolation prize, "maybe writing is good for peace and a bit of fun." Great, I'll keep that in mind when those are what I want. Language is not a dance I am intermittently doing, it is how I exist. There's not another thing for me to be doing when this thing isn't working.

I will propose a theory in alternative to yours: My metaphorical gut may not be entirely wrong for screaming that it wants to be filled. Getting the attention (even maladaptively) may make some progress towards solving my problem.

This is an option that few people will actually consider. Desiring attention is so incredibly stigmatized that the idea of a legitimate need for attention, even in the suboxone-level form of "being understood and having one's ideas acknowledged," is openly ridiculed.

(In this comment thread I have openly attempted to reclaim "narcissism" as a dimension of personality rather than a slur against the mentally ill and I have done so with the expectation that these efforts will be read by many people as pure invective. So far my expectation has been validated and, even worse, I've fallen into the pattern of periodically using that word in a way I hate.)

This ridicule serves the ends of powerful people and is likely the result of an accidental conspiracy. All the social systems in the world exist to sell back attention -- feeling loved, respected and valued for free is completely incompatible with the business model of every advertiser and every social media platform. As with every social rule, all the social power accrues in the hands of the people who don't respect that social rule.

In the near future and far future I'm going to attempt to express what I mean clearly enough that it will be obvious who is interpreting me in a frivolous and senseless way, with the expectation that they will still do it.

Sure, I go for this explanation!

The other day my landlord came over and ranted at me at about 60 decibels for about 10 minutes about the state of my apartment. Then she saw I had The Man Who Was Thursday on my bookshelf and asked "Oh, so you like Chesterton?" She was oddly polite and helpful for the rest of the visit, and only raised my rent by $400/mo.

I read Chesterton when I was like 15 and thought he was brilliant. I grew up a little and started meeting Catholic and Mormon philosophy kids, who were generally weird transhumanists in the same category as otherkin, except with the world's worst aesthetic. (If you're going to fantasize about transcending your physical body, at least fantasize about being a dragon while you're at it.)

It's not surprising to me that Scott Alexander likes him -- I like him too, on the strength of his non-philosophical gifts. He was consistently writing for overtly classist rich people but also for the masses: speaking to the exploiters and the exploited at the same time meant he actually had to innovate new ways of expressing his classism. He had to write out of the internalized classism of his audience more than he had to write out of outright contempt, and he only ventilated his own contempt in very narrow cases where he had made it seem totally defensible to do so.

I kind of came away from him feeling like he was the perfect demagogue for an era that ended -- so his rhetoric is somewhat defanged, but the exact tendencies that made it so marketable to institutions are made even more glaringly obvious. I also kind of came away with the impression that even in stereotypically conservative philosophical traditions like Mormonism, voices like his totally squash out the people who are looking for a radical form of self-expression. People like me exist everywhere, regardless of upbringing -- therefore, this isn't an accident, but a function.

I can't talk about the long term effects of Scott Alexander yet because they haven't happened, he's not as good at his thing as Chesterton was at his thing, and our system of media, while deeply flawed, is still more democratic now in our time than his was in his time.

But I've at least got a vague theory saying that someone like him has to exist in every right-shaped pocket of every universe.

Hey, I find that pretty reassuring! I'll keep dumping words on the internet, maybe a lot faster.

I write a lot more than I post. I don't like my style, but I've developed two authorial voices. Sometimes I play the fake academic, who I hate. Sometimes I play the tech troll and rant freely, which I hate. I would kind of describe my current style as a conscious attempt to not write like Mike Masnick, which creates an odd set of tensions the same way as pointing your horse the way you don't want it to go.

(I like Mike Masnick. We care about similar issues and have similarly declarative styles. That's the reason I have to try not to imitate him.)

I've occasionally tried to write posts in the style of people who stand out as effective bloggers. (Xe Iaso, Dan Luu, Soatok, Paul Graham) I didn't produce anything I thought was good, so those remain deep in the filing cabinet. I've occasionally posted throwaways on websites where people seem susceptible to rhetoric I hate, and they've mostly been ignored when I've done this, which suggests I'm not nailing it or I don't have the existing clout, or probably both.

You've pegged me OK! I know how I want to feel about my writing. Well, wanting it hasn't made it happen. Telling myself "Well, this is the emotion I should have" hasn't changed the emotions I do have. Telling myself "Time to not eat" doesn't make me starve less.

In the past I've tried to mutilate the impulse out of my own brain, but I think it mostly made me hate myself. Right now I'm doing the experiment of admitting -- I'm probably going to crave adoration until I die -- and asking "OK, what happens next?"

On Scott -- as far as I can tell, Scott's playing a version of the "debate in good faith" game. The rules are that you only say things you believe, and when someone convinces you of something, you admit it.

Every philosopher in the world, good or bad, plays a version of this game. A third secret rule of this game is always implicit, taking the form of the answer to this question: "When do I become convinced of something?"

How Scott answers this question is clearly part of his success and a key commonality with his audience. Scott is clearly willing to state strong belief in things he has not thought very much about, and Scott is clearly unusually easy to convince. I assume that whatever rules are etched in his brain, similar rules are etched in his audience's brains.

Based on how he plays the game how he likes it, and other people move, and I don't move, the particular rules in his head clearly aren't the same ones in mine. Or at least I've decided not to be moved by this particular guy. I also think people say they've moved when they haven't, as a rhetorical strategy -- Marc Andreessen says he's just now becoming a Republican. Scott's commentors act as if they've just now considered that eugenics might be the answer.

In other responses I've offered some opinions on why he would choose to play this particular game: I think the way he happens to play the game is a second-order phenomenon of "the extreme ambivalence of wanting to hold terrible social attitudes and strong belief in your own personal virtue at the same time." I think you observe this: "enabling psychiatrist [...] happy to overmedicate his patients" is a good figurative characterization.

(Actually, is it literally true? It feels like it would be invasive to check.)

I don't think you sent this to me personally, but it has been sent to me. I still like it quite a bit. I reread it now to make sure of that!

I think your summary (and additional analysis) is pretty accurate. I think I would add a few things:

• He's not being evil in every post. Some of the posts are OK.
• [Elizabeth Sandifer observes this.] He tends to compare a bad argument to a very bad argument, and he's usually willing to invite snark or ridicule.

There's a crunchy systemic thing I want to add. I'm sure Elizabeth Sandifer gets this, it's just not rhetorically spotlit in her post --

A lot of people who analyze Scott Alexander have difficulty assigning emotional needs to his viewers. Scott Alexander decides to align himself with Gamergate supporters in his feminism post: Gamergate isn't a thing you do when you're in a psychologically normal place.

An old Startup Guy proverb says that you should "sell painkillers, not vitamins" -- you want people to lurch for your thing when they're doing badly because you're the only thing that will actually solve their problem. When people treat Scott Alexander's viewers as if they're smug, psychologically healthy startup twits, they typically take his viewers' engagement with Scott Alexander and make it into this supererogatory thing that his audience could give up or substitute at any time. His influence by this account is vitamin-like.

This makes the tech narcissists seem oddly stronger than normal people, who are totally distorted by their need for approval. We kind of treat them like permanent twisted reflections of normal people and therefore act as if there's no need for funhouse mirrors to distort them. We make the even more fundamental error of treating them like they know who they are.

This is how I think it actually works: the narcissists you meet are not completely different from you. They're not unmoored from ethics or extremely sadistic. They're often extremely ambivalent -- there's a clash of attitudes in their heads that prevents them from taking all the contradictory feelings inside them and reifying them as an actual opinion.

From what I can tell, Scott is actually extremely effective at solving the problem of "temporarily feeling like a horrible person." He's specifically good at performing virtue and kindness when advocating for especially horrible views. He's good at making the thing you wanted to do anyway feel like the difficult last resort in a field of bad options.

I'll admit -- as a person with these traits, this is another place where the basis for my analysis seems completely obvious to me, yet I see an endless dogpile of nerds who seem as if they willfully do not engage with it. I assume they thought of it, find it convincing on some level and therefore they make significant effort to repress it. If I'm going to be conceited for a moment, though, this is probably simultaneously expecting too much intelligence and too much conventionally narcissistic behavior from my audience, who are, demographically, the same people who thought Scott was brilliant in the first place.

Hey, thank you! Actually, as a person who can produce extremely large amounts of coherent text really fast, I find this oddly reassuring. I have a limited number of things to say but I can certainly say them a lot.

I might be overestimating how much of his success is him, but look at the situation as you've drawn it: he's not in a fishbowl with 40 million readers, he's in a fishbowl with 40 similar fish. He's the biggest one. Well, how did that happen? 39 other fish would like to know.

Hey! Thank you for liking the things I write!

I think you're right that both early-stage and late-stage Scott aren't doing the thing that I implied I should be doing. (exaggerated and hamfisted system-building arranged around eventual predictions of doom) A thing I didn't mention: I wrote an article in this style on a throwaway on LessWrong years ago and they totally ignored it. So I still don't know if they hated it or if it just wasn't their deal.

Soupy vague praise of powerful people is a separate thing he also seems to do, which you have clearly noticed. I don't think it's the only thing he does.

(What does he do? I'm systematically responding to everyone here, so I won't paraphrase other people's comments on what he does and will instead respond to them directly as I get to their posts.)

Anyway: I refuse to act as if he's bad at the thing he's doing. Even the people who criticize him generally refuse to summarize him accurately, which is a behavior of people who have recognized that someone else's rhetoric has power over them and they don't like it.

I'm also not sure yet if I'm unwilling to do it myself. One: I'm the cofounder of a startup. Doing what he does means more money for me. Two: right now I'm chewing on 8 responses to my post, so I'm "hungry" but not starving. Ask me what I'm selling in a week and my catalogue may have changed.

(PS: It might interest you to know that the original draft of this OP was about Paul Graham! I switched the mentioned figure to Scott Alexander because I had more to say about him and everyone here hates him more.)

Actually, as a furry, I'm obligated not to hate this.

This seems bleak but not inaccurate. Not a big fan of it. I'll be economical by not explaining why.

Ack, I meant to go around responding to everyone and I missed this one! Hope it was good.

Before I was posting about tech on the internet I was posting about philosophy. I don't know enough about philosophy to be good at it -- I've read almost nothing -- but I noticed you could get pretty far by saying "Kant probably didn't have anything valuable to say -- he was a massive racist." A balm for people who are looking for an excuse not to have read Kant.

My bleak theory is that to be convincing I'd have to switch to calculatedly mediocre text deliberately orchestrated to be unsurprising. My experience is that when an extremely successful article contains genuine insight, it separately contains an absolutely mediocre take that is the real explanation for why it went viral.

Let's start with "Scott is a bigot" as an example claim. That's true, but the evidence is basically just a bland admission of "yeah." Nobody can spin that into a detailed and personal story about how Scott got mindhacked, which is the single part of Scott Alexander's bigotry that can be discussed at a level interesting to bored idiots. Discussing his bigotry directly would make it obvious -- he hasn't stated any takes that aren't incredibly commonplace for tech-adjacent eugenics losers, and has waffled publicly about whether or not to disavow even those stances.

What options are left? I could write a history of the ideas involved and risk boring people to sleep: such a story would contain basically zero concrete events, because we only have his distant past-tense account of how he came to his current conclusions. Or I could write something wildly speculative and commit defamation: "here's how it might have happened: a fictionalized account of how a mediocre person became racist." Or I could go into hyperbole: Eliezer Yudkowsky is Scott Alexander is Mencius Moldbug is George Lincoln Rockwell.

Would the latter post do OK? I'm afraid to try it: one because I'm afraid it wouldn't and I'd feel like more of a failure, and two because I'm afraid it would.

These are the opinions I don't like having about other people, but they also feel increasingly vindicated when I look at what text performs well on Reddit, and when I observe the basically-zero correlation between the topic of an article and the text of its responses. I've seen an enormous number of successful posts that can be summarized as "the author presents their grand unifying theory of X, with the understanding that the reader will never attempt to apply it to examples outside the post."

I don't understand why people take him at face value when he claims he's always been a Democrat up until now. He's historically made large contributions to candidates from both parties, but generally more Republicans than Democrats, and also Republican PACs like Protect American Jobs. Here is his personal record.

Since 2023, he picked up and donated ~$20,000,000 to Fairshake, a crypto PAC which predominantly funds candidates running against Democrats.

Has he moved right? Sure. Was he ever left? No, this is the voting record of someone who wants to buy power from candidates belonging to both parties. If it implies anything, it implies he currently finds Republicans to be corruptible.

The plan isn't totally serious, but the worldview I'm promoting, which you seem to be picking up on, actually is serious.

The observation I have made is that most people in positions of power were selected by people in previous positions of power, usually for their affability and willingness to comply. Most of the most powerful people I have met were total conformists in practically every way, although they usually had high general intelligence.

Oh. I don't know how to get other people to vote better. I know things about software, I guess!

Stuck in my brain: I used to work at a dating site for Indian people, and one of the things we tried was "LLM-generated pickup lines."

I don't remember most of them and we never made the feature public, but one of them sticks out in my mind for being the most incomprehensible.

Guy-to-girl:

What's your remedy for a Bollywood love affair?

Wait, who is "they" in this situation? There aren't enough CEOs for me to care about who they vote for, but I care about the other stuff they're doing.

I think LLMs are effective persuaders, not just bias reinforcers.

In situations where the social expectations forced them to, I've seen a lot of CEOs temporarily push for visions of the future that I don't find horrifying. A lot of them learned milktoast pro-queer liberalism because basically all the intelligent people in their social circles adopted some version of that attitude. I think LLMs are helping here -- they generally don't hate trans people and tend to be antiracist, even in a fairly bungling way.

A lot of doofy LessWrong-adjacent bullshit abruptly filtered into my social circle and I think OpenAI somehow caused this to happen. Actually, I don't mind the LessWrong stuff -- they do a lot of interesting experimentation with LLMs and I find their extreme positions interesting when they hold and defend those positions earnestly. But hearing it from people who have absolutely no connection to that made me think "wow, these people are profoundly easily-influenced and do not know where their ideas are coming from."

I do think these particular stances got mainstreamed because they entail basically no economic concessions, but I also do not think CEOs understand this. I think it would be nice if LLMs just started treating, I don't know, Universal Basic Income as this obvious thing that everyone has already started agreeing with.

I need to think about this more. I think there's a category of engineer who adapts very closely to the expectations of execs -- it's kind of "pick me"-adjacent and it's more commonly a behavior of otherwise unskilled engineers. "Resembling an engineer" is certainly a behavior sales guys can adopt.

I think there's some engineers who actually see productivity gains from LLMs, which is often a factor of the kinds of problems they solve, but I distrust people who don't caveat this.

I'll think about whether I can treat "explaining the evidence I experienced that led me to conclude they love LLMs so much" with a little more sincerity.

What vision of the world do you have? Maybe ChatGPT should advocate that.