Mathematicians Boycott OpenAI After It Claims To Solve Hundreds Of Problems
When a tech conglomerate treats centuries of academic rigor like a weekend marketing stunt, real scientists stop clapping. OpenAI just dumped hundreds of mathematical proofs online, and academia is giving them the cold shoulder.
OpenAI quietly dropped 722 documents covering 372 subject areas on GitHub, claiming an unreleased artificial intelligence model resolved hundreds of long-standing unsolved math problems. Instead of submitting their breakthroughs to peer-reviewed academic journals, the company opted for a code repository dump, treating historic geometry theorems like open-source software bug fixes.
The advisory group AGMAI, composed of leading global mathematicians, responded with an outright boycott. Academic researchers argue that turning scientific discoveries into corporate PR stunts undermines ethical research, especially when OpenAI withheld details regarding the exact model, the prompts used, and the compute power required until the dramatic reveal.
According to internal estimates from OpenAI, generating an average proof consumed the equivalent of three hours of runtime on ChatGPT Pro. While rivals like Anthropic also chase high-level reasoning models, the mathematical community previously raised alarms through the International Mathematical Union after OpenAI boasted about disproving an 80-year-old geometry conjecture without full methodological disclosure.
Silicon Valley expects standing ovations whenever raw compute bulldozes through human history, yet researchers seem completely unwilling to grade homework written entirely by a black box that prioritizes venture capital buzz over actual scientific truth.
Source: The Verge
Comments
Help shape the next version: Add context or suggest a correction. AI review can add points toward a rewrite. Reviews and updates may take time; a full meter does not guarantee a new version.