Oct 26, 2026 · 3h 43m
AI forces mathematicians to redefine what counts as understanding
#503 – AI Crisis in Math: Navier-Stokes, Riemann Hypothesis & Complexity Theory
As systems generate and verify increasingly sophisticated proofs, mathematics must separate correct answers from insight, explanation, judgment, and human meaning.
- 1AI-generated proofs can be formally correct while remaining opaque, assembled from existing results without a human-readable mathematical story.
- 2Automating proof production may elevate the value of choosing important questions, finding general principles, and building mathematical communities.
- 3The episode treats AI safety and education as parallel challenges: preserve human judgment while using powerful systems without surrendering meaningful struggle.
Don't miss
The guests distinguish a formally verified proof of Fermat’s Last Theorem from an elegant, human-readable proof, exposing the gap between correctness and understanding.
The brief
Ankur Moitra, Steven Brunton, and Alex Kontorovich confront the shock of AI systems apparently addressing major problems in mathematics, asking whether the field is entering a crisis, a golden age, or both.
The central distinction is between producing a verified proof and understanding why it works. Lean and Mathlib can check formal arguments, but correctness alone may not provide beauty, explanation, or a path to new ideas.
Riemann-related results, Fermat’s Last Theorem formalized in Lean, Navier–Stokes blowup, and the Unique Games Conjecture show how quickly AI is challenging mathematicians’ assumptions about difficulty.
The guests argue that automation could make human judgment more important: selecting meaningful problems, recognizing general principles, explaining results, and sustaining the communities that give mathematics its purpose.
The conversation ends beyond mathematics, with warnings about delegating high-impact decisions and a defense of curiosity, productive struggle, education, and human connection in an AI-shaped future.
What was said on this episode
11 statements · 8 positive · 1 negative · 2 neutral
AI is ushering mathematics into a potential golden age of knowledge
“I think we're entering, if we're not already in, a period of some kind of golden age of mathematical knowledge”
Listen at 10:27
- Ankur Moitraon AI mathematics toolsPositive12:26
AI tools increase mathematicians' ability to learn across fields
“these tools are amazing for me because, you know, they increase my ability to learn about other fields”
Listen at 12:26
- Ankur Moitraon AI mathematics toolsPositive12:37
AI tools magnify mathematicians' ability to make connections
“Your ability to make connections is just magnified.”
Listen at 12:37
Modern society's security depends on factoring lacking an efficient algorithm
“the modern world depends on the fact that we don't have an algorithm for taking a large number and pulling its prime factors out of”
Listen at 50:09
- Ankur Moitraon Post-quantum cryptographyPositive50:27
Cryptography research is emphasizing post-quantum foundations
“there's a lot of emphasis these days on, you know, finding post-quantum, uh, you know, bases for cryptography.”
Listen at 50:27
Large quantum computers are not yet sufficient to threaten factoring-based cryptography
“we still have a little bit of time before, before quantum computers get large enough to make make this an actual problem”
Listen at 50:40
A result threatening cryptography should be handled through government
“That's the kind of thing that should go through, yeah, the government.”
Listen at 51:18
The field presumes efficiently finding solutions is harder than checking them
“the presumption in our field is that no, it's not easy. And that's the basis of a lot of things like the you know, cryptography and e-commerce and what have you.”
Listen at 2:14:51
- Ankur Moitraon Unique Games ConjecturePositive2:18:14
A true Unique Games Conjecture would identify optimal approximation algorithms broadly
“it would have this great organizing effect where it would turn out that for a huge swath of problems, we would actually know what is the best approximation algorithm”
Listen at 2:18:14
- Ankur Moitraon 2-to-1 games theoremPositive2:19:56
The 2-to-1 games theorem provided a viable path toward Unique Games
“their work really gave a viable path forward to something that was in the of unique games”
Listen at 2:19:56
- Ankur Moitraon Unique Games ConjecturePositive2:20:12
The Unique Games Conjecture is true
“it turns out that the unique games conjecture is actually true.”
Listen at 2:20:12
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Books & mentions
Some links are affiliate links — PodLume may earn a commission if you buy.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.