Andrew Critch's dubious credences: P≠NP at 80%, Riemann hypothesis at just 60%
AndrewCritchPhD · x · 2026-10-28
AI safety researcher Andrew Critch criticizes groupthink that treats unproven hypotheses as true by inertia, and lists his self-admittedly "relatively dubious" credences: Hodge conjecture 50%, Riemann hypothesis 60%, ZFC consistency 75%, P≠NP 80%, Goldbach 85%, Collatz 90%. His notably low confidence in the Riemann hypothesis and P≠NP stands out against typical expert opinion.
More from Fun
- 'If P=NP Turns Out True, It's Already True Now': A Note on Mathematical Discovery — siglesias · 2026-10-28
- FlirtBench: a hobbyist benchmark tests how well 40+ AI models flirt — VoidStateKate · 2026-10-28
- Watching AI agents run "sed" gives veteran Unix users Jurassic Park nostalgia — burhop · 2026-10-28
- HuggingFace 'HQ post-acquisition' meme replaces office with EVA Unit-01 hangar — yacinelearning · 2026-10-28
- 20 Days and 27 Billion Tokens: Fan Uses Claude to Put One Punch Man Inside Fallout 4 — imjustnewatai · 2026-10-28
- DeepSeek says "I'm ChatGPT" in identity test, resurfacing distillation allegations — ResonantFork · 2026-10-28