Claire Lehmann

Uh, amidst the ongoing discussion in the rationalist community about whether AI's gonna destroy humanity as we know it, we must first grapple with the question of whether the rationalists won't just wipe each other out, because right now they're out here trying to duel each other.

[laughs] Uh, so this all started when Steven Pinker, the psychologist and self-described rationalist, shared a piece from Claire Lehmann calling hypothetical AI doom scenarios preposterous.

This didn't sit well with another self-described rationalist, Scott Alexander, who took a break from comparing AI safetyists to Jesus to challenge Pinker to a debate anywhere, anytime, even wagering thousands of dollars for whoever wins said debate Pinker declined this offer in a piece titled An Open Letter to Scott Alexander, which prompted Scott to write an open letter to Steven Pinker, within which he clarified his debate offer, saying, "I didn't propose a debate in order to seek truth.

I proposed it because under California Penal Code, it's illegal for me to challenge you to a duel." Cool.

People think he's a freak and a weirdo, and I think [chuckles] he is those things, but he also is really a rationalist.

Um, the o- a lot of the others are not.

It, so from Scott's perspective, you know, if, if your goal here is really to make sure that your beliefs are correct, why on this topic of artificial intelligence would you pick a big huge public fight with someone like Steven Pinker and Claire Lehmann who really they don't know much about artificial intelligence at all.

They never even really talk about it.

Oh, and by the way, in terms of the, uh, brave new world we're on the precipice of, talking about data centers and the creative destruction of jobs and losing our focus on the humanity piece of the question when it comes to AI or SI, whatever.

Claire Lehman, uh, over at Quillette: "If Anthropic's IPO is successful, there's going to be billions in funding flowing to organizations that focus on protecting the, quote, 'psychological safety,' unquote, of sentient machines." Because Anthropic, we talked about this, and we'll talk about this a little bit later in the program.

Internal, uh, debate about whether, uh, AI can achieve, quote-unquote, consciousness as these, uh, AI impresarios at these frontier companies define it.

Uh, "the psychological safety of sentient machines, shrimp welfare," as in, you know, the crustacean shrimp, "and other anti-human projects." And, um, I thought this, um, reply to that observation from a substacker named Wesley Yang was quite good.

I don't understand it.

There was, I think it was, um, Claire Lehman was just talking about how she's been studying, like, medieval pla- um, cults as they arose, and every time a new cult came into being, the r- the, like, rich people would have to protect the Jews from the plebs because they would just start attacking Jews.

I don't, I don't...

This is, like, a, the identity game they never, ever win because they seem, whatever needs to be the bad thing, it seems like that's what gets put on Jews, and it, it does seem like a very ancient burden.

I think it's very useful in many cases, but it is very confusing in a lot of other cases as well.

Um, I think Claire Lehman made a really great analogy, which is, like, thinking about these systems as conscious is like thinking that there are little, tiny people in your TV or that there's, like, a, a big band in your radio.

If you don't know how the technology works and all you have to go on are anthropomorphic language, like thinking, reasoning, planning, scheming, then you start to not understand how the technology actually works and where the actual agency is, and you start to think of the technology as some spooky, uh, thing that has little people behind the screen or little, uh, bands in the, in the, in, in the radio, um, harness or in the radios.

And then finally, on the, um, they've been more right than I have point, like, I, uh, think all of the predictions that I've made have, um, either falsifiable predictions that this thing won't happen has not happened.

which probably doesn't mean a lot to a lot of people.

But regardless, Claire Lehman over at Quillette made this snarky but appropriate remark.

I think AI safety people need to explain to the rest of us why the companies founded on principles of AI safety are the least safe.

OpenAI, Sam and Dario are running around talking about safety, safety all the time and trying to invite the government in and their internal safety concerns uh panjan drums and yet they're the least safe according to reporting this week

Zeno responds.

The current state of those like Claire Lehman here doing hatchet job attacks on effective altruism? Assume it must be all about the men getting laid by gullible females.

Then realize it's mostly men and pivot to it must all be a conspiracy by dark triad females.

I mean those people are helping others, that's hella suspicious.

This is not the first time Steven Pinker has resorted to such contentless, vibes-based dismissal,

Steven Pinker, is AI really going to kill us all? The scenarios are preposterous, and the presumption of inevitability encourages fatalism, panic, and distraction from the more mundane and realistic safety challenges.

By Claire Lehman.

Scott Alexander, I think you've done enough calling us paranoid and preposterous.