OpenAI model nets new math solutions; critique says missing preregistration skews the curve
Jsevillamol · x · 2026-10-07
Debate over OpenAI's new math results: cited stats suggest 2 new FMOP solutions (of 40 remaining) and 12 new FM Erdős problems (of 63), implying a 5-20% chance this model+compute solves an arbitrary open problem of substantial interest. Jsevillamol argues the apparent burst is an artifact of no preregistration — with a preregistered problem set, the progress curve would look much smoother.
Related event: OpenAI Math Results Split the Math World Between Awe and Existential Doubt(47 posts)→
More from Models
- SciCode benchmark outdated, >90% pass rates no longer reflect frontier — geoffwolfe · 2026-10-08
- Looped transformers trend revives a 2016 looping paper: what's old is new again — gowthami_s · 2026-10-08
- Same model, very different performance: Dylan Patel on why the harness matters — MatthewBerman · 2026-10-08
- Unsloth fine-tunes Qwen3.5 0.8B on 4GB VRAM, lifting decision accuracy from 20.7% to 74.3% — kalyan_kpl · 2026-10-08
- OpenAI drops 722 AI-written math papers targeting previously unsolved problems — iamfakhrealam · 2026-10-08
- Dev slams 'speed-maxing' trend: chasing tokens/s at the cost of output quality — casper_hansen_ · 2026-10-08