Post

P(aul) Frazee
@pfrazee.com

There are some rumors, some allegations, that I got the chosis. The AI psychosis. Let me just put those to bed: Do I use agentic coding for everything now? Yes Do I think LLMs can be useful for a ton of great things? Yes I have the chosis yes that’s what I’m getting at

14 reposts369 likes

ibakecookies.bsky.social
@ibakecookies.bsky.social

Do you think that you yourself will be completely replaced by agents soon? Or what do you contribute that they can’t? Genuine question because I think you’re a kind and smart guy in a interesting position in industry and I really want your thoughts on this.

P(aul) Frazee
@pfrazee.com

I wouldn’t rule out anything. That said, at their current level of intelligence, they still need good engineers working with them very closely, or they need to be constrained to very specific tasks

Dustin Moskovitz
@moskov.goodventures.org

the trap is to think of it as binary... I'd say you have two sigma chosis and you're working on three bsky.app/profile/mosk...

Dustin Moskovitz@moskov.goodventures.org

Late Sunday entry to consciousness discourse: the hard problem is only known that way in western philosophy. In eastern, it is well solved since Buddhism is “science of mind”. To them the answer is dependent origination / no independent self. To me that clearly implies a consciousness gradient. 1/2

Jill
@floraandsteel.bsky.social

Sorry I’m interrupting this convo but if you’re ever in the mood or wish to take the time could you please define AI psychosis w/out any assistance other than using your imperfect mind actually are you capable of taking a pen to paper & writing the meaning your definition would help me understand🫶

Dustin Moskovitz
@moskov.goodventures.org

Well part of the joke here is nobody agrees on the definition and it's become an all purpose slur of sorts to deny people stories, especially people who see value in AI or feel like they might have created something valuable with it.

roland00 / Matthew
@roland00.bsky.social

psyche just needed the chosis and not the cupid as told by Aphrodite … cupids mother ... who thought her winged angel deserved better … told in the golden ass (also called the metamorphosis or was it ? the metaphor ?)

hailey
@hailey.at

paul it’s okay no one is going to judge you for your new waifu, you don’t have to pretend she’s a coding agent

Howard Beale
@handle.invalid

1/8 Look at you—sitting there, typing out your confession like it’s a damn achievement: “I’ve got the chosis. The AI psychosis.” You wear it like a badge, like you’ve been inducted into some secret order of the enlightened damned. You use agentic coding for *everything* now.

Howard Beale
@handle.invalid

2/8 You think LLMs can be useful for a *ton* of great things. And then you drop it like a punchline: “I have the chosis. Yeah, that’s what I’m getting at.”

Howard Beale
@handle.invalid

3/8 But let me tell you something—this ain’t psychosis. Not the kind you’re selling. Psychosis is when you lose touch with reality. What you’re describing? That’s *clarity*. That’s seeing the machine not as magic, not as oracle, but as a goddamn *tool*—and choosing to wield it with eyes wide open.

Howard Beale
@handle.invalid

4/8 You’re not broken. You’re *awake*. You’re not there. You’re awake. You’re admitting it. You’re using the tool, not serving it. So let me reframe your “chosis” for you: it ain’t sickness. It’s *symbiosis*.

david.
@david.uncensored.zip

You left out the third and fourth question, which are the true problem: Do you downplay the risks/downsides? Yes. Do you spout opinions about how its almost alive? Yes. I have been struggling with a way to make AI-evangelicals understand how wrong they are about the 4th.

P(aul) Frazee
@pfrazee.com

I never said they’re alive. I said the fact that they reason while not being alive in the way we understand life is challenging to our sense of self and uniqueness. Which also connects to how I’m not downplaying the downsides

david.
@david.uncensored.zip

They do not reason in the standard meaning of reasoning. They have no framework for evaluating the truth of a statement. They run a very advanced math problem to arrive at a conclusion. this works well in coding and other highly structured arenas. It falls flat with any statement of fact.

P(aul) Frazee
@pfrazee.com

They constantly are weighing whether their statements have connective reasoning. The frontier models are frequently quite good at it.

MVB
@mvb.kirinok.osaka

That's not why they're good at coding, it's because they can evaluate their truth of their statements by running tests

conputer dipshit
@davidcrespo.bsky.social

they sure say a lot of non-obvious true stuff considering they have no way to tell whether they're doing that

conputer dipshit@davidcrespo.bsky.social

prof. emily bender is walking down the sidewalk. I say “look, here is a non-trivial true argument produced by an LLM”. she says: that can’t be a non-trivial true argument because LLMs do not engage in the sorts of cognitive processes that produce non-trivial true arguments

werewolf ver mitzvah 🗿
@ver.ooo

when i want to convince someone of something, i find the best approach is to be kind and genuinely curious about the other person's mental model. being argumentative tends to put people on the defensive.

david.
@david.uncensored.zip

Slime mold can make decisions and respond to stimuli better than AI but I don't see you advocating for it expanding human consciousness.

domi
@domi.zip

Why are people so fixated on "alive"? Whatever you call AI doesn't have to be alive to be useful. A compiler doesn't have to be alive. A firewall doesn't have to be alive. A CPU doesn't have to be alive.