Architeuthis @ Architeuthis @awful.systems

It's not always easy to distinguish between existentialism and a bad mood.

---

Read more

Posts

18

Comments

560

Joined

2 yr. ago

Nice. Here's the bluesky account as well.

Some quality wordsmithing found in the wild:

Remember, when your code doesn’t compile, it might mean you made a mistake in coding, or your code is about to become selfaware.

Good analogy actually.

Don't forget Yud is also a big compiler understander

The arguments made against the book in the review are that it doesn't make the case for LLMs being capable of independent agency, it reduces all material concerns of an AI takeover to broad claims of ASI being indistinguishable from magic and that its proposed solutions are dumb and unenforceable (again with the global GPU prohibition and the unilateral bombing of rogue datacenters).

That towards the end they note that the x-risk framing is a cognitive short-circuit that causes the faithful to ignore more pressing concerns like the impending climate catastrophe in favor of a mostly fictitious problem like AI doom isn't really a part of their core thesis against the book.

They also seem to broadly agree with the 'hey, humans are pretty shit at thinking too, you know' line of LLM apologetics.

“LLMs and humans are both sentence-producing machines, but they were shaped by different processes to do different work,” say the pair – again, I’m in full agreement.

But judging from the rest of the review I can see how you kind of have to be at least somewhat rationalist-adjacent to have a chance of actually reading the thing to the end.

The pair also suggest that signs of AI plateauing, as seems to be the case with OpenAI’s latest GPT-5 model, could actually be the result of a clandestine superintelligent AI sabotaging its competitors.

copium-intubation.tiff

Also this seems like the natural progression of that time Yud embarrassed himself by cautioning actual ML researchers to be weary of 'sudden drops in loss function during training', which was just an insanely uninformed thing to say out loud.

the only people who like prediction markets [..]

Apparently Donald Trump Jr. has found his way into the payroll of a couple of the bigger prediction markets, so they seem to be doing their darndest to change that.

assuming prediction markets are magic

Bet it's more like assuming it will incentivize people with magical predicting genes to reproduce more so we can get a kwisatz haderach to fight AI down the line.

It's always dumber than expected.

Apparently the hacker who publicized a copy of the no fly list was leaked an article containing Yarvin's home address, which she promptly posted on bluesky. Won't link because I don't think we've had the doxxing discussion but It's easily findable now.

I'm mostly posting this because the article featured this photo:

I figure eventually some proprietary work would make it into the wild via autocomplete. Copilot used to be cool with inserting other programmer's names and emails in author notes for instance, though they seem to have started filtering that out in the mean time.

Copilot licenses let you specifically opt out from your prompts and your code being used to train new models, so it would be a big deal.

We should be so lucky, the ensuing barrage of lawsuits about illegally cribbing company IP would probably make the book author class action damages pale in comparison.

This is too corny and overdramatic for my tastes. It reads a bit like satire, complete with piling on the religious undertones there at the end.

Getting love bombed in that rationalist con he went to recently probably didn't help matters.

The common clay of the new west:

if one person came out and spilled the beans, it’d suggest that there might be more people who didn’t

I mean, after his full throated defense of Lynn's IQ map (featuring disgraced nazi college dropout Cremieux/TP0 as a subject matter expert) what other beans might be interesting enough to spill? Did he lie about becoming a kidney donor?

I think the emails are important because a) they make a case that for all his performative high-mindedness and deference to science and whinging about polygenic selection he came to his current views through the same white supremacist/great replacement milieu as every other pretentious gutter racist out there and b) he is so consistently disingenuous that the previous statement might not even matter much... he might honestly believe that priming impressionable well-off techies towards blood and soil fascism precursors was worth it if we end up allowing unchecked human genetic experimentation to come up with 260IQ babies that might have a fighting chance against shAItan.

I guess it could come out that despite his habit of including conflict of interest disclosures, his public views may be way more for sale than is generally perceived.

some relevant discussion from a couple of years ago

I wonder if this is just a really clumsy attempt to invent stretching the overton window from first principles or if he really is so terminally rationalist that he thinks a political ideology is a sliding scale of fungible points and being 23.17% ancap can be a meaningful statement.

That the exchange of ideas between friends is supposed to work a bit like the principle of communicating vessels is a pretty weird assumption, too. Also, if he thinks it's ok to admit that he straight up tries to manipulate friends in this way, imagine how he approaches non-friends.

Between this and him casually admitting that he keeps "culture war" topics alive on the substack because they get a ton of clicks, it's a safe bet that he can't be thinking too highly of his readership, although I suspect there is an esoteric/exoteric teachings divide that is mostly non-obvious from the online perspective.

In his early blog posts, Scott Alexander talked about how he was not leaping through higher education in a single bound

He starts his recent article on AI psychosis by mixing up psychosis with schizophrenia (he calls psychosis a biological disease), so that tracks.

Other than that, I think it's ok in principle to be ideologically opposed to something even if you and yours happened to benefit from it. Of course, it immediately becomes iffy if it's a mechanism for social mobility that you don't plan on replacing, since in that case you are basically advocating for pulling up the ladder behind you.

Shamelessly reproduced from the other place:

A quick summary of his last three posts:

"Here's a thought experiment I came up with to try to justify the murder of tens of thousands of children."

"Lots of people got mad at me for my last post; have you considered that being mad at me makes me the victim and you a Nazi?"

"I'm actually winning so much right now: it's very normal that people keep worriedly speculating that I've suffered some sort of mental breakdown."

I’m even grateful, in a way, to SneerClub, and to Woit and his minions. I’m grateful to them for so dramatically confirming that I’m not delusional: some portion of the world really is out to get me. I probably overestimated their power, but not their malevolence. […]

Honestly what he should actually be grateful for is how all his notoriety ever amounted to^[1] was a couple of obscure forums going 'look at this dumb asshole' and moving on.

He is an insecure and toxic serial overreactor with shit opinions and a huge unpopular-young-nerd chip on his shoulder, and who comes off as being one mildly concerted troll effort away from a psych ward at all times. And probably not even that, judging from Graham Linehan's life trajectory.

[1] besides Siskind using him to broaden his influence on incels and gamer gaters.