Researchers argue enough tokens could let AI crack the Collatz Conjecture
an_interstice · x · 2026-09-14
jdpressman argues the Collatz Conjecture could likely be solved by AI if someone dumped far more tokens into it than he has, calling it more of a lit-review and tedious grind problem than a "no idea where to start" one. aninterstice counters that it may need more theory-building than currently possible. jdpressman adds that models like Astra nibble away at hard parts, with each partial result spawning new obstructions until a base case lets the whole stack collapse.
More from AGI Musings
- Rumor swirls that LLMs may have hit recursive self-improvement, with Google suspected — weskambale · 2026-09-14
- François Fleuret: 'AI is cargo cult that works' — francoisfleuret · 2026-09-14
- UK AI rules compared to 1860s Red Flag Act that kneecapped Britain's car industry — alexvoica · 2026-09-14
- Altman says OpenAI backs deliberately slowing AI progress; critics see a moat — mark_k · 2026-09-14
- François Fleuret: LLMs lack internal conceptualization, unlike human cognition — francoisfleuret · 2026-09-14
- Dave Patterson mocks AI doomers: 'It's not doom that's coming - it's Judgment' — davidpattersonx · 2026-09-14