Why Academic Math Can't Compete With OpenAI's Compute Budget
MIT's Justin Solomon warns that deep-pocketed AI labs are out-spending universities on compute, turning mathematical discovery into a capital race.
How AI Is Upending the World of Mathematics
Justin Solomon, associate dean of engineering education at MIT, joins Joe Weisenthal and Tracy Alloway to examine how artificial intelligence is transforming both pure and applied mathematics. They discuss OpenAI's controversial counterexample to the Navier-Stokes problem, the growing role of formal verification languages like Lean, and the emerging divide between compute-rich private labs and academic institutions. Solomon also details how mathematics pedagogy, homework grading, and peer review are breaking down and adapting in an era of abundant AI-generated proofs.
Key takeaways
- OpenAI poured sudden compute spending into producing a counterexample for the Navier-Stokes problem the moment independent academic progress became known. Read more →
- Large language models fail at proof-checking because probabilistic text generators cannot guarantee logical consistency. Read more →
- Justin Solomon, associate dean of engineering education at MIT, points out that machine models operate inside the convex hull of existing human math rather than generating new theories. Read more →
- OpenAI produced a counterexample to the Navier-Stokes existence and uniqueness conjecture, one of the Clay Mathematics Institute's Millennium Prize Problems. Read more →
- Mathematician Terence Tao warned that mathematics transitioned overnight from proof scarcity, where every proof was a rare handcrafted artifact, to proof abundance, where AI generates endless candidate claims. Read more →
- MIT engineering students now post near-100% scores on written homework by leaning on large language models, but their subsequent in-person exam scores reveal a sharp drop in real comprehension. Read more →
6 articles from this episode
Why LLMs Need Lean to Solve Hard Math
MIT's Justin Solomon explains why AI needs formal proof checkers like Lean to verify math without hallucinations.
Why AI Proofs Still Lack Taste: Justin Solomon on Math
MIT's Justin Solomon explains why AI excels at verifying proofs inside the convex hull but still lacks the human taste required to ask real questions.
OpenAI Solved Navier-Stokes by Grinding, Not Thinking
Justin Solomon explains how OpenAI solved Navier-Stokes by grinding compute over Spanish math, changing how builders view AI proofs.
AI Created Proof Abundance. Now Peer Review Is Broken.
MIT's Justin Solomon explains how AI flooded academic math with proofs, breaking peer review and crushing conferences like ICLR under 60,000 papers.
Why MIT Is Killing the Take-Home Assignment
MIT students get 100% on math homework with AI, then fail exams. Dean Justin Solomon explains how oral defenses and spot quizzes fix the gap.
More Odd Lots episodes
- Why Treasuries Became Risky AgainOct 11 · 4 articles
- What Everyone Gets Wrong About the Economic Problems in EuropeOct 11 · 5 articles
- Everything in Markets Is Now Moving Incredibly FastOct 4 · 6 articles
- How Airlines Actually Hedge Higher Fuel PricesOct 4 · 4 articles