Mathematics Has a Move 37 Moment
OpenAI's release of hundreds of AI-generated math results is being called mathematics' Move 37 moment, raising questions about human absorption of the output and whether other fields will follow.
Watch on YouTube ↗Charlie's summary
- ▶ 0:28
OpenAI's reported revenue gap came from gross-versus-net accounting, and it shook the semiconductor trade
The FT reported OpenAI told investors it hit roughly $50 billion annualized revenue, not the widely reported $68 billion. That figure came from an investor trying to match Anthropic's gross-revenue convention, which counts revenue later shared with Amazon or Microsoft. OpenAI uses net revenue. Nvidia fell 3% and Oracle 5.5%. The host expects Anthropic's audited numbers to look lower than what it has shared with private investors.
- ▶ 3:00
Enterprise AI returns are rising, and agent use has jumped, but hiring has barely shifted
The Information's subscriber survey found 35% of organizations returning multiples on AI spend and only 8% reporting a net negative. OpenAI regained the lead over Claude and Gemini. 72% use open-weight models, but only 21% use Chinese ones. Almost half say AI hasn't changed their hiring, with 22% hiring less and 10% hiring more. The host notes it is a skewed, highly tech-forward sample.
- ▶ 8:07
OpenAI's internal model produced 372 new math results, including 90 of the top 500 open problems
The GitHub release held 372 novel results and 722 supporting papers. By the Proof Atlas list of top 500 open problems, 90 were fully solved, with partial results on the Riemann hypothesis and the Hodge conjecture. The average result used about three hours of ChatGPT Pro compute, versus roughly 10,000 agents over 88 hours for Navier-Stokes a month earlier. Mathematician Alex Kontorovich said a human doing this would earn “an instant fields medal.”
- ▶ 11:11
Mathematicians split between awe and shock: a “Move 37 moment” for the field
Researchers who had worked at the AI-math intersection called it the most significant moment in mathematical history, and one former OpenAI researcher called it his personal Move 37. Some mathematicians see their job shifting to verifying AI proofs. A grad student said the release nukes the projects they wanted to work on. Shriram Krishnan sees a great day, since mathematicians would always have wanted a math genie, and the shock to identity is temporary.
Give your email to unlock the rest of the key points — for every episode.
Do more with Charlie
Free — sign in with your email, no password.