Kenneth Stanley: the thousands of math problems AI fails to solve are an underrated scientific goldmine
kenneth0stanley · x · 2026-11-01
AI researcher Kenneth Stanley argues the field obsesses over the few hundred cutting-edge math problems AI has solved while ignoring the thousands it attempted and failed — a set he calls an underrated scientific goldmine.
- Systematically surveying failures and their nature would map what is actually hard for AI and why, pointing research in new directions
- For years or decades, it would also show human mathematicians where their minds remain irreplaceable
- The two implications eventually tension with each other but feed new human-machine collaboration — the key is knowing what remains out of reach and why
More from AGI Musings
- Reddit's AI-slop decline pushes users toward private groups, not old-school forums — mgostIH · 2026-11-01
- Roko's Basilisk in the weights: your public words are already training data — nathanborror · 2026-11-01
- Google's 'Goodnight' datacenter in Claude, Texas projected to host 2.27M H100-equivalents by late 2027 — scaling01 · 2026-11-01
- Will big labs really refuse to sell on-prem year-old models? Sovereign AI bets questioned — Afinetheorem · 2026-11-01
- davidad podcast on emergent wisdom in LLMs, says Anthropic should stop cyber-offense training — RichardMCNgo · 2026-11-01
- Rereading Bostrom's Superintelligence: 2014 predictions all arrived early — kimmonismus · 2026-11-01