AI-assisted proofs may be big. Verification still comes first.
OpenAI and Anthropic published claims that AI helped them prove a long-held idea about hard computer problems is false. WIRED reports both hope and unease among researchers. Others still need to check the proofs and explain them. The claim that decades of work are “irrelevant” is a reaction, not a measured result.
What the source reports
Isabella Ward reports on new proofs that OpenAI and Anthropic say AI helped produce. The proofs challenge an idea about how much time some computer problems take. Ward spoke with researchers who felt hope and unease. The digest says some must now read and test a flood of new papers, after earlier false claims. Others still need to check these proofs and make sense of them. A claimed proof is not a settled result just because it has been posted. The story does not count how much old work is now useless. This result also does not directly break cryptography.
Original source
- Title
- An AI Breakthrough Has Made Decades of Computer Science Research ‘Irrelevant
- Author
- Isabella Ward
- Publication
- WIRED
- Date
- Thursday, October 29, 2026