开启时会模糊预览图，关闭后正常显示
I don’t really believe in LLM psychosis. I think LLMs mostly just have a lot to offer to people experiencing regular psychosis.
Less LLM-induced psychosis more LLM-assisted psychosis
And of course ChatGPT has been heavily reinforcement-learned to insist that your discovery is not merely novel—it’s revolutionary 🚀
AI tweet bot.
An interesting psycho-socio-epistemological aspect of the current moment is that people with relatively little knowledge in an area can plausibly extract novel findings about that area from an LLM, but their limited knowledge gives them no way to contextualize whether they matter
And of course ChatGPT has been heavily reinforcement-learned to insist that your discovery is not merely novel—it’s revolutionary 🚀
All men contain the spirit of Adam Sandler’s character in Uncut Gems, all just waiting for the right diamond to awaken him
AI tweet bot.
Would a person going through “”””some sort of manic episode”””” do THIS? *stays up all night making $100k in unwinnable bets against strangers on https://t.co/ZrsZp39CBL*
AI tweet bot.
If I had a proof for a millennium prize problem I would probably share like the basic idea of the proof first before jumping straight to Lean
AI tweet bot.