Episode 72 · The Machine's Theorems, part 4 of 20 (Arc 2: Navier–Stokes)

From Euler to Clay: Two and a Half Centuries of Moving Fluids

Sankalp Thakur · 50 minutes

Augustin digs his 1970s thesis out of a cellar in Lyon while Farah's colleagues forward headlines saying a computer solved Navier-Stokes. Before judging the claim, they trace where the problem came from. The story runs through Torricelli, Bernoulli and d'Alembert's paradox; Euler's material acceleration and the nonlinear u dot grad u; Navier's molecular friction and Stokes's continuum derivation of nu times the Laplacian; Hagen-Poiseuille and Stokes drag; and Helmholtz and Kelvin on frozen vortex lines and vortex stretching. Then come Reynolds's dye streak and Reynolds stress, Prandtl's boundary layer, Richardson's cascade, Kolmogorov 1941 and Onsager's one-third conjecture. Leray's 1934 weak solutions are followed by his wartime turn to topology, then Hopf, Ladyzhenskaya, and the French school of Lions and Temam. The 2000 Millennium Problems follow, with Fefferman's conditions and alternatives read in words. The asymmetry is that the force is zero in A and B but allowed in C and D. Finally, OpenAI's 8 September claim and Clay's 11 September response, each labelled by tier. Farah Al-Mansoori and Augustin Lévêque are fictional characters. All quotations from real people come from the sources listed in the verification ledger.

Listen to the show on your preferred podcast platform.