Gary Marcus: AI Math System Failed to Solve Some Solvable Problems
GaryMarcus · x · 2026-08-02
Gary Marcus further critiques the recently hyped AI mathematical reasoning systems. He points out that it turns out the system actually failed to solve at least some of the problems it was confronted with, and some of those may indeed be solvable. This suggests that the current capabilities of models in complex reasoning tasks are still being overestimated.
Related event: OpenAI Math Breakthrough Questioned by Gary Marcus and Others(38 posts)→
More from Models
- Test: Sarvam is 58x Faster than Claude for Hindi Tasks — itsOmSarraf_ · 2026-08-17
- OpenAI's aggressive compute bet may explain why stronger models are held back — haider1 · 2026-08-17
- Grok takes instructions too literally, listing icons instead of using them — Daniel_Farinax · 2026-08-17
- Users Report ChatGPT Getting Noticeably Dumber at Following Instructions in Recent Weeks — Cubic_Inequality504 · 2026-08-17
- DeepSeek V4 Flash tested on Mac Studio: Impressive quality, high RAM demand — pj-frey · 2026-08-17
- GPTZero flags pure ChatGPT text as "Mixed" — No-Mortgage-2037 · 2026-08-17