The most exciting claims from OpenAI’s heap of new proofs

On Tuesday afternoon OpenAI dropped more than 700 manuscripts reporting proofs, solutions and progress on 372 open problems across theoretical research.

It will take time for mathematicians to trudge through all the artificial-intelligence-generated results. No one, not even OpenAI researchers themselves, has been able to even read through everything yet. And not all of the proofs were certified as correct by the automatic logic verification platform Lean—in fact, some have already been retracted after outside scientists found mistakes. OpenAI also did not include full prompts or run-time details for how their agents uncovered the proofs, and officials admitted in a blog post about the results that citations and written descriptions of findings could be improved in future releases.

On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.

Despite these issues, experts say the results are jaw-dropping. The papers span mathematics, theoretical computer science, physics, and more. Some of the discoveries could have easily netted a human author a top research award just a year ago