cloneofsimo: academia badly underestimates the problems OpenAI's math agents are solving

cloneofsimo · x · 2026-09-22

Researcher cloneofsimo responds to skeptics who dismiss OpenAI's all-in math agents as "just Erdős problems" or claim they only worked by stealing mathematicians' Codex logs. He argues such critics have no idea what "100 solved internal problems" looks like, that academia systematically underestimates the scale of progress, and that significant breakthroughs are on the horizon.

Related event: Debate Flares Over OpenAI's Math Agents as Researchers Say They Are Underestimated(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →