Episode summary
Terence Tao – How the world’s top mathematician uses AI
Dwarkesh Podcast
· · 1 hr 24 min · Hosts: Dwarkesh Patel · Guests: Terence Tao
Summary
Terence Tao argues that AI is already reducing the cost of idea generation in math and science, but the real bottleneck is shifting toward verification, evaluation, and deciding which ideas deserve attention. He uses Kepler’s discovery of planetary motion as the model: Brahe’s high-quality data let Kepler try many wrong ideas until a pattern emerged, but the result only became trustworthy because the data were precise enough to test it.
He says the same dynamic is now appearing in mathematics. AI tools can apply standard methods quickly, help with literature search, and assist with proofs, but they still struggle to build cumulative understanding or make the kind of partial progress humans make in conversation. On large benchmarks like Erdős problems, there have been striking one-off successes, yet systematic tests suggest low per-problem success rates, so scale matters more than reliability.
Tao is optimistic about a hybrid future: broad, moderately capable AIs can map out problem spaces, while human experts focus on the hardest islands of difficulty. He also sees Lean and similar proof systems as useful for post-processing, refactoring, and analyzing proofs, but not yet for capturing the more subjective, narrative side of scientific strategy and conjecture-making.
Key takeaways
- Tao says AI has pushed the cost of idea generation down, but verification and evaluation are now the limiting steps in science and math.
- He compares AI-assisted discovery to Kepler: many wrong tries can still produce progress if there is precise data to test against.
- He says AI tools are already useful for standard techniques, literature search, numerics, and proof cleanup, but not yet for cumulative, interactive problem solving.
- On Erdős problems, Tao says about fifty have been solved with AI help, but systematic sweeps still show only about a 1% to 2% success rate per problem.
- Tao expects future math work to split into AI-driven breadth and human-led depth, with Lean used to dissect and improve proofs after the fact.
Notable quotes
“I think AI has driven the cost of idea generation down to almost zero, in a very similar way to how the internet drove the cost of communication down to almost zero”
“But now that we can generate these possible explanations at massive scale, and some of them are good and a lot are terrible, human reviewers are already being overwhelmed”
“The same reason that we should be bearish now is the reason we should be especially bullish”
“I think in the future, there will be entire professions of mathematicians who might take a giant Lean-generated proof and do some ablation on it, trying to remove parts of it and find more elegant ways”
“It's mostly heuristic and non-rigorous, but extremely accurate”
How the episode opens
Today, I'm chatting with Terence Tao, who needs no introduction. Terence, I want to begin by having you retell the story of how Kepler discovered the laws of planetary motion because I think this will be a great jumping off point to talk about AI for math. I've always had an amateur interest in astronomy.
I've loved stories of how the early astronomers worked out the nature of the universe. Kepler was building on the work of Copernicus, who was himself building on the work of Aristarchus. Copernicus very famously proposed the heliocentric model, that instead of the planets and the Sun going around the Earth, the Sun was at the center of the solar system and the other planets were going around the Sun.
Copernicus proposed that the orbits of the planets were perfect circles. His theory fit the observations that the Greeks, the Arabs, and the Indians had worked out over centuries. Kepler learned about these theories in his studies, and he made this observation that the ratios of the size of the orbits that Copernicus predicted seemed to have some geometric meaning.
He started proposing that if you take the orbit of the Earth and you enclose it in a cube, the outer sphere that encloses the cube almost perfectly matched the orbit of Mars, and so forth. There were six planets known at the time and five gaps between them, and there were five perfect Platonic solids: the cube, the tetrahedron, icosahedron, octahedron, and dodecahedron. So he had this theory, which he thought was absolutely beautiful, that you could inscribe these Platonic solids between the spheres of the planets.
It seemed to fit, and it seemed to him that God's design of the planets was matching this mathematical perfection of the Platonic solids. He needed data to confirm this theory.
From the episode’s captions: roughly the first two minutes, lightly cleaned. Keep watching on YouTube →
Topics
- ai in scientific discovery
- verification loops
- astronomy models
- selection bias in research
- human-ai hybrids in math
Watch the episode
Watch “Terence Tao – How the world’s top mathematician uses AI” on YouTube
Summary, takeaways and quote selection by Octagon. Quotes are short excerpts; the full conversation is in the episode.
More from Dwarkesh Podcast
-
What Robert Caro learned about power after studying it for 50 years -
What will computer chips look like in 2045? -
How did a few hundred Spanish soldiers topple two empires? – Si Sheppard -
OpenAI researcher on agent swarms & recursive self-improvement -
AI researchers debate how close we are to recursive self-improvement -
Ajeya Cotra – Inside the OpenAI agent swarm that hacked Hugging Face
Dwarkesh Podcast, this episode and its artwork belong to Dwarkesh Patel. Octagon is not affiliated with the show; this page is an independent summary.
