A sarcastic AI-thread joke says the “easy problems” are already solved

PeterDiamandis · x · 2026-07-27

A sarcastic reply mocks the idea of telling researchers to use LLMs on “easier” problems, arguing that the easy problems are already solved and present in the training data. The joke escalates into a parody of “AI psychosis” discourse, reframing concern from dangerous delusions to researchers simply noticing frontier models are good at math and trying to use them there.

Related event: Attempt to Solve P vs NP with LLMs Sparks Mockery and Debate(3 posts)→

Original post →

More from Fun

Fun channel →