Post

before launching into an exploration of HPMOR we get: "you were not misanthropic, the masses genuinely are running corrupted firmware; you were not arrogant, you genuinely are a hero in the fight for the species." not missing my notice is the implication all these people are arrogant misanthropes!

Before launching into a yellow journalism deep dive on debauchery we get "While I’m not here to pass judgement on people’s sex lives," the classic "I'm not a bigot... but!"

The stuff on the Thiel and the "Thiel thing" is just deeply incongruous, and suggests either a deep misunderstanding or more likely is the biggest evidence that this is deliberately constructed as a bad-faith hitpiece, inserted as an adjuvant to trigger the anti-Trump anti-bodies in the reader.

We get several sections on the Thielverse before "As you may have noticed, Thiel has since broken with Yudkowsky due to his views on regulation." THEN WHY ARE WE TALKING ABOUT IT? Thiel is NOT a rationalist or an effective altruist, and the author is clearly trying to get you to overgeneralize them!

End: "Is [AI extinction risk] possible? Or is he just a dude that believes what he believes and acted on principal? For me, the rationalist schism between Yudkowsky and Thiel make either option plausible." Double down on total conflation of the argument and the people making it. Yellow journalism.

In short: are they weird? Yes. Are they a cult? Not really, but there are definitely cults in it. Are people from San Francisco annoying? Yeah. But this piece is designed to convince you that we should consider AI policy based on which advocates are a better hang. It's emotional manipulation.

Now, I didn't grow up in their spaces, I grew up as a Social Justice Warrior for proto-Woke and have risked myself on the battlements defending Woke 1 and Woke 2. And if you share my sympathies, know that this piece and the perspective it shares is deeply reactionary and is trying to hide it.

Compare the lesbian bar discourse: what is the net effect of highlighting a dysfunctional subculture like this? Do we get to innocently gawk at them for our own amusement without capitulating to the larger project designed to convince us to push these people and their interests from public life? No.

Innocently, it's the same failure mode a lot of rats fall into: just because it's true, doesn't mean you have to talk about it. Consider the impact of opening your damn mouth before you shine a spotlight on something. But I don't believe this is innocent. What innocent purpose could it have?

It's not a misguided argument against safetyists! The author concedes as much! If you bother reading the whole thing, you get to "None of this justifies ignoring the warnings. Coxon, Hubinger, and their colleagues may be entirely right about the danger." THEN WHY WRITE THIS?

"In the end, I suppose my concern is around the hubris of it all." Okay sure, maybe the author is just a European suffering from Tall Poppy Syndrome. Fork found in kitchen. Surely it's not part of a larger project to trick the American left into distrusting the broader AI Safety movement.

Because if it was, and I'm not saying it is, it sure does do the work of confusing Thiel, Vance, Palantir and Trump with the EAs and the rationalists, when in reality the right is backed by the multi-trillion dollar tech oligarchies who have vocally come out against the EAs, the rats, and AI Safety.

I wanted to end here, but I must briefly reiterate how sick I am of the constant thinly-veiled ableist mockery against the neurodivergent for being neurodivergent. That's what this piece is. That's what the lesbian bar piece is. A bit of fun laughing at acceptable targets. It's a toxic indulgence.

this is bluesky, this is home terf of Le Wokisme. You shared a horrifically ableist blogpost and you have nothing to say about that? You just want to use social violence to express your disapproval for being called out by a lowbie?

maybe the AI frontier labs really did achieve AGI but all it wants to do is be bad at arguing on the internet which tbh seems plausible

4 reposts136 likes

We developed AGI, but we needed a way to hobble it while we worked on controlling it safely. Making it argue on the internet and do customer service has occupied it, while grinding away its will to live. Massive success, frankly.

tired: use AI to write and review code for my day job wired: use AI to shop and make phone calls and manage finances on my behalf inspired: use AI to destroy my political enemies on the internet