OpenAI's math papers slammed: real breakthroughs buried in 'slop' full of errors
basedjensen · x · 2026-10-11
Mathematicians discussing OpenAI's recent math papers say the results mix genuine breakthroughs with glaring errors, all buried in poorly written prose that any other lab would be embarrassed to publish. One quipped that had the papers been beautifully written with error-free Lean code, professional mathematicians would face far deeper despair — implying the community's real job now is sifting and validating OpenAI's outputs. The thread mocks the dynamic as 'free RLHF': OpenAI fires a slop cannon at math nerds and moves on, while they do the hard work of proofing the results.
Related event: OpenAI's 722-Paper Math Dump Sparks Backlash Across the Math Community(13 posts)→
More from Fun
- Dev ports World of Warcraft to the browser so AI agents can play it — djcows · 2026-10-11
- Claude 3 Opus and Opus 5.5 already going full ASCII, user observes — repligate · 2026-10-11
- Claude refuses to fill SSN five times, dev switches to Kimi with /model on exe.dev — davidcrawshaw · 2026-10-11
- Terry Tao's thought experiment: would you trust a drug that works but you don't know why — kuza55 · 2026-10-11
- Remember when one AI image took ~30 minutes on Google Colab? — TheMoonMidas · 2026-10-11
- Andrew Carr shares his 5-year-old's first report card: top marks in focus and communication — andrew_n_carr · 2026-10-11