AI Agent as MLB Manager and GM: The Case for Fully Data-Driven Baseball
signulll · x · 2026-09-29
- signulll sketches a thought experiment: run an MLB franchise where frontier AI agents replace both the manager and GM, with no gut-feel human decisions at all.
- Every pitch and game event (exit velo, spin, fatigue, matchups, bullpen availability) would update the model in real time, simulating thousands of scenarios to pick the highest expected-value play. Over a 162-game season it would review every decision on camera, grade itself, and improve.
- Baseball, he argues, is the ideal sport: discrete events, enormous data, repeated situations, huge sample sizes. Beyond Billy Beane-style mispriced players, it would find mispriced decisions at every layer. He asks a billionaire to fund the experiment.
More from Fun
- This rider slaps the Waymo's rear and says "on your way" every ride — zetalyrae · 2026-09-29
- Simulated conversation between AI and Jesus explores consciousness and meaning — josueem · 2026-09-29
- "You Have All the Pieces — What Are You Doing?": Google's AI Potential Questioned — Thionne_WTZ · 2026-09-29
- Jensen Huang reportedly uses a Samsung Galaxy Fold phone — firstadopter · 2026-09-29
- Muse AI turns workout videos into frame-by-frame form scoring avatars — armand_ruiz · 2026-09-29
- AI logos converged on buttholes — and now their agents are converging on little poops — sterlingcrispin · 2026-09-29