AI & Machine LearningAI in Science

AI Beats the Weather Supercomputers

A model trained on decades of weather now forecasts in a minute

A new forecaster never studied the physics of air, yet it beats the supercomputers. So what did it learn from?

You'll find out

  • How supercomputers forecast weather
  • What the AI learned instead of physics
  • Why the old models still matter

Why it matters

Forecasts warn of storms days ahead. AI is making them faster and cheaper, but it still leans on physics.

About 3 min6 interactive slides · 5 questionsLesson 712 · Zone 80

Play this lesson

Free to play in your browser or on iPhone

More in this zone

  1. 713BlockchainFrom Mining to StakingEthereum cut its energy use by more than 99% in a single day
  2. 714High-Energy PhysicsCherenkov's Blue GlowParticles can outrun light inside water and leave a blue glow
  3. 715Philosophy of AIThe Chinese RoomCan a system understand without understanding anything?
  4. 716Computational ComplexityYour Favorite Games Are Mathematically HardTetris and Minesweeper are as hard as the hardest quick-to-check puzzles
  5. 717Post-Quantum SecurityThe Quantum ThreatA future computer could pick today's digital locks
  6. 718Graph TheoryFour Colors Are EnoughAny map of one-piece countries can be colored with four colors, but proving it took a century
  7. 719Computer-Assisted ProofsThe Proof Too Large to CheckThe four-color proof needed a computer, and mathematicians argued about it
  8. 720Surface PhysicsCold Welding in SpaceClean metals can fuse on contact when there's no air between them