Show notes
About twenty researchers spent a paper writing down, formally, the rule that when you tell a learned simulator to move the robot's arm, the arm should move. Six of the open-source world models the field builds on don't reliably obey, and a second paper explains why: trained only on successful demonstrations, they've never had a reason to learn what a bad action does, so feed them one and they show you the task succeeding anyway. Also: Google DeepMind's whole-body humanoid stack, which contains no world model at all, a robot that triples its success rate under a lighting change because it…
Website: physicalaipodcast.co

