OpenAI keeps botching its math breakthroughs, and even its new advisory group isn't fixing it

The Verge AI · rss · 2026-09-29

The Verge reports that OpenAI has spent months demonstrating it can do two things consistently: make impressive math breakthroughs, then botch announcing them. Its latest repair attempt — a new independent advisory group of elite mathematicians — is already described by mathematicians, including one member, as messy and confusing, bearing hallmarks of OpenAI's previous rushed forays into the math community.

Original post →

More from Companies & People

Companies & People channel →