a16z researcher: AI can solve Millennium math problems with Lean proofs but can't explain them
a16z · x · 2026-11-21
An a16z Frontier roundtable with Ittai Abraham, Dan Boneh, Scott Kominers and SuccinctJT discusses how AI is reshifting the role of mathematicians.
Key points from Ittai Abraham:
- AI can now solve Millennium-level problems and produce Lean formal proofs, yet cannot produce human-understandable explanations — explanation requires modeling the reader's intuitions, and RL tasks are far easier to build for "is this proof correct?" than "is this explanation good?"
- Short-term prediction: AI will get very good at pushing the discovery frontier, while humans keep doing the work of translating results into human understanding.
Other topics include watching AI make progress on a problem you spent years on, what's worth publishing when papers can be produced this fast, and how Lean checks proofs.
More from AGI Musings
- Investor: AI's sweep through software engineering is a bellwether for healthcare — pzakin · 2026-11-21
- Impetus Grants Appeals for Philanthropic Funding to Build the Biological Data AI Cures Need — shae_mcl · 2026-11-21
- Metaculus Had This Milestone at 2035 — Until This Week — StartupYou · 2026-11-21
- Mathematician littmath reverses skepticism on AI proving Riemann Hypothesis by end of 2026 — littmath · 2026-11-21
- The 250,000-page Riemann proof claim: when computational verification outpaces human comprehension — SydSteyerhart · 2026-11-21
- lateinteraction: a tool that solves Millennium Prize Problems yet fails 1000 small things reshapes the map — stochasticchasm · 2026-11-21