TechnicalDiscussion

The hardest physics problem in fluid dynamics: Turbulence explained | Lex Fridman

Lex Clips

Turbulence is fundamentally difficult to model because all scales of fluid motion matter simultaneously, unlike gases where simple parameters (temperature, pressure, density) suffice. The key to solving this Millennium Prize problem may lie in finding a coordinate transformation that simplifies turbulent fluid descriptions, combining both data-driven AI approaches and deep mathematics.

Summary

The discussion explores why turbulence represents one of the hardest unsolved problems in physics and fluid dynamics. Turbulence involves complex, multi-scale phenomena where vortices of all sizes—from large to tiny—interact and influence each other, similar to Leonardo da Vinci's fluid drawings showing structures at every scale. This contrasts sharply with gases, where despite containing 10^23 particles in chaotic motion, only three parameters (temperature, pressure, density) are needed to describe behavior completely. In turbulent flows—over airplane wings, through hearts, or in rivers—all the individual computational details matter for accurate predictions, making it fundamentally challenging. The speaker frames this as an AI problem: discovering whether a coordinate system or mathematical representation exists that could simplify turbulent fluid descriptions as elegantly as thermodynamics simplified gas behavior. The solution likely requires both brute-force data analysis with machine learning to test different coordinate transformations and deep mathematical theory about what forms these transformations should take. Historical physics breakthroughs, like Newton's laws, were preceded by coordinate transformations (heliocentrism replacing geocentrism). The speaker notes that vortex formation and interaction—where vortices can merge, stretch, fold, and create chaotic yet highly structured, tangled trajectories—exemplifies turbulence's beautiful complexity. Problems that stumped brilliant physicists in the 1970s-80s due to computational limitations are now ripe for AI approaches with modern data and computing power.

Key Insights

  • Turbulence is fundamentally hard because all scales of fluid motion matter for predictions, unlike gases where only temperature, pressure, and density determine behavior despite containing 10^23 chaotic particles.
  • A simple descriptive coordinate system or representation for turbulent fluids may exist but hasn't been discovered, similar to how temperature and pressure describe gases.
  • Major physics breakthroughs historically were preceded by coordinate transformations, such as Newton's laws coming after placing the sun at the center of the universe.
  • Problems that very smart physicists in the 1970s-80s couldn't solve due to computational constraints are now fertile ground for AI, particularly when the mathematical form of solutions is known but too complex to write down.
  • Turbulent flows contain huge amounts of structure and are not random—vortices stretch, fold, and create tangled trajectories like spaghetti while maintaining underlying organization.

Topics

Turbulence and multi-scale fluid dynamicsComparison between gas behavior and turbulent flowsCoordinate transformations in physicsAI and machine learning applications in fluid dynamicsVortex formation and mergingMillennium Prize problems

Transcript

[0:03] - Can you speak to some of the maybe, beautiful, fascinating phenomena in the space of fluids that you've been thinking about? You mentioned turbulence. What, what are the big puzzles to sort of solve in this space to understand? You've mentioned drag, like air, fluid. What are like the big Millennium Prize type problems that people are thinking about? - I mean, there's so many. You can break them down into a few categories. So turbulence is very hard because it's multi-scale. [0:34] - What is turbulence? - Turbulence is this phenomena that when you have a complex fluid roiling or mixing and moving kind of like a Da Vinci's-- He has these beautiful drawings of fluids at…

Full transcript available for MurmurCast members

Sign Up to Access

More from Lex Clips

Get AI summaries like this delivered to your inbox daily

Get AI summaries delivered to your inbox

MurmurCast summarizes your YouTube channels, podcasts, and newsletters into one daily email digest.