Math podcasts have a way of being either too soft (vibes about beauty and wonder) or too hard (notation that assumes you already have a PhD). The episodes below, pulled from our full library of episode summaries, split that difference. They come from working mathematicians, self-taught theorists, education platform founders, and one teenager who happens to be the son of a mathematical physicist, and every one of them earns its spot because something specific and revealing actually gets said.
Expect open problems explained in plain terms, a public referendum on whether 1 times 1 equals 2, arguments over whether math is discovered or invented, and a candid look at what it actually takes to write the most influential computer science book ever published. Whether you want rigor, controversy, or just a better mental model of infinity, there's an entry point here.
Terence Tao: Hardest Problems in Mathematics, Physics & the Future of AI | Lex Fridman Podcast #472
Fields Medalist Terence Tao walks through some of the hardest unsolved problems in math, and the details are stranger than you'd expect. He describes building a literal water-based 'liquid computer' that could force a finite-time blowup in the Navier-Stokes equations, and explains the Equational Theories Project, which generated 22 million abstract-algebra problems and used Lean plus roughly 50 collaborators to settle all but two of them. He also lays out the parity barrier that's stalled the twin prime conjecture for decades. This is the one to start with if you want the state of the art from the person actually working at it.
Read the full episode notesEdward Frenkel: Reality is a Paradox - Mathematics, Physics, Truth & Love | Lex Fridman Podcast #370
Edward Frenkel argues that reality itself is paradoxical, and backs it up with math history rather than mysticism, from Gerolamo Cardano's 'mental tortures' accepting the square root of negative numbers to the fact that well-behaved number systems only exist in dimensions 1, 2, 4, and 8. He also gets personal, admitting he used Platonic math as a refuge from a world he found cruel and unjust, and later moved away from that Platonism entirely. Listen if you want the philosophy of math delivered by someone willing to revise his own convictions on tape.
Read the full episode notesDonald Knuth: Algorithms, Complexity, and The Art of Computer Programming | Lex Fridman Podcast #62
Donald Knuth, the man behind The Art of Computer Programming and TeX, traces his path from a 1957 IBM 650 with 2,000 words of memory to decades of writing his magnum opus, and he's refreshingly honest about the limits of understanding, including his own. He states plainly that he believes P probably equals NP, then immediately explains why that wouldn't actually help anyone. His warning about the gap between real understanding and the illusion of understanding lands harder now than when he said it. Essential for anyone who wants math and computer science history from a primary source.
Read the full episode notesJoe Rogan Experience #2171 - Eric Weinstein & Terrence Howard
This is the most contentious entry on the list, and that's exactly why it belongs here. Mathematical physicist Eric Weinstein sits with Terrence Howard to seriously evaluate Howard's '1 times 1 equals 2' claims and his Flower of Life geometry, and rather than dismissing him outright, Weinstein pinpoints the actual error: the arc cosine of negative one-third is 109.47 degrees, not the 108 Howard's geometry needs. He's blunt that Howard's odds of contributing new mathematics are slim while still praising his geometric intuition. A rare example of a rigorous mathematician steel-manning a fringe theory in real time.
Read the full episode notesGrant Sanderson: 3Blue1Brown and the Beauty of Mathematics | Lex Fridman Podcast #64
The creator of 3Blue1Brown makes the case that math's reputation for being mysterious is mostly a notation problem, going so far as to call Euler's identity 'quite hideous' once you look past the symbols. He explains why he's still not fully comfortable with infinity, preferring to think of it as simply always being able to add one more, and connects the Euler product for the zeta function directly to the fundamental theorem of arithmetic. If you've ever wanted math explained the way 3Blue1Brown animates it, but in conversation, this is it.
Read the full episode notesPo-Shen Loh: Mathematics, Math Olympiad, Combinatorics & Contact Tracing | Lex Fridman Podcast #183
USA Math Olympiad coach Po-Shen Loh spends most of this episode showing how the combinatorics mindset that wins Olympiad problems also solved a real-world crisis: his NOVID app measures your distance from a disease in relationships rather than feet, and counterintuitively found that the deadlier a disease is, the stronger people's incentive to adopt the app. He also describes teaching invention instead of memorization, treating hard problems as puzzles rather than gatekept knowledge. Good for anyone who thinks math education is broken and wants to hear what actually fixes it.
Read the full episode notesVladimir Vapnik: Statistical Learning | Lex Fridman Podcast #5
The co-inventor of support vector machines and VC theory makes a mathematical case against deep learning itself, calling it 'fantasy' and arguing the Representer theorem shows optimal learning solutions actually live on shallow networks. He distinguishes instrumentalism, just predicting outcomes, from realism, actually understanding the underlying law, and insists real intelligence needs invariants and predicates rather than brute-force data. Cited over 170,000 times, Vapnik is arguing against the entire direction of modern AI from deep inside the math that built it.
Read the full episode notesLuís and João Batalha: Fermat's Library and the Art of Studying Papers | Lex Fridman Podcast #209
The brothers behind Fermat's Library make the case that academic publishing is quietly broken, recalling how as students they had to email paywalled papers to each other because one of them, in Europe, couldn't access them. They reveal that about half of all machine learning papers on arXiv were published in just the past year, and that their site was deliberately built as a 20-year project rather than a fast-growth startup. A good listen for anyone curious about how famous papers, from the Bitcoin whitepaper to the Higgs boson discovery, actually get read and understood.
Read the full episode notesZev Weinstein: The Next Generation of Big Ideas and Brave Minds | Lex Fridman Podcast #158
The youngest voice on this list, Zev Weinstein, argues that math is discovered rather than invented while building a broader case for why deep, original thinking becomes more dangerous and more necessary as stagnation sets in. He publicly diverges from his father Eric's more pessimistic civilizational outlook, and reframes free will as making legitimate decisions inside a deterministic system. Worth hearing for the philosophy of mathematics angle delivered by someone still working out his own convictions in real time, on camera, afraid and doing it anyway.
Read the full episode notesThat's nine episodes covering everything from open millennium problems to a public fact-check of amateur physics. Browse the full library of episode summaries on Episode Notes for more conversations worth your time.