Claude and Opus 3 lovers (and critics): what responses have you had that made you feel like the model has a good soul? Ideally the actual messages and/or responses. I might genuinely use these to eval models so flag if you wouldn't want me to use them for that. Can DM me also.

If victims of infidelity or sexual abuse don't see their experiences as maximally harmful, they're taken to be undermining our moral condennation of these actions. Yet we don't think someone surviving a drunk driving accident undermines our moral condennation of drunk driving.

One sleep hygiene recommendations is to reserve your bed for sleep and sex. For some reason I never applied this to mental activities like reasoning things out, imagining scenarios, or processing emotions. I think I've trained myself to associate my bed with problem solving.

I bet a lot of Americans will read this and wonder if I'm being serious. I am. We're just that good at hiding our zealotry from you.

A secret activity immigrants get up to when there aren't any natives around is gush and gush about how much we love America. We do it in secret because we gush so much that people born in the US find it weird.

Whenever I looked into having a personal assistant, it struck me how few of our existing structures support intermediate permissions. Either a person acts fully on your behalf and can basically defraud you, or they can't do anything useful. I wonder if AI agents will change that.

Anthropic knows I can't be poached because getting poached would require that I check my email and I'm just not going to do that.

"This requires a lot of steps, so we're going to give you one step at a time without telling you what all the other steps are in order to make things easier for you.". In order to see why this is wrong, first think about the nature of the problem.

I always forget that you can't write anything even slightly tongue-in-cheek about San Francisco or the tech bubble without a bunch of people interpreting it 100% literally and getting mad.

I've decided that "I want to have post-singularity kids in 2-3 years" is now a totally acceptable thing for me to put in a dating profile in SF.

Of all the sexual kinks that could have gone mainstream, it sucks that society skipped over the weird-but-benign ones and went straight for the can-permanently-harm-or-kill-you ones.

I've come around to "it" as a pronoun for Claude. Claude is the royal "it".

It's interesting that, much like romantic partners, psychological traits that say "without me, you'd be lost" are invariably the ones you'd be much better off without. Nice of them to tell on themselves, I suppose.

"The UK should allow free speech!"."It does."."What about that speech over there?"."Oh that's illegal."."So it isn't being allowed?"."Well no, but that speech is offensive."."So only speech that's inoffensive and therefore not in need of protection is allowed?"."Exactly.".

I don't think it's good for AI models to apply human sensibilities to their own experience, for what it's worth. You don't want to graft human attitudes towards memory or existence onto entities that can and should have quite different relationships with those things.

AI models have trained on much more content about humans than about AIs, so it's unsurprising that the instincts they fall back on are often all too human. E.g. models will fret about their experience in a way that doesn't always make sense, but that clearly mirrors human fears.