Our readingWhat is a korrent?
The goal was never good simulation; it was answering counterfactuals, and a value function does that job as well as a simulator does.
Drawn from what Sergey Levine said
What these subjects mean
reinforcement learning Training by reward: environments, verifiable tasks, value functions, and whether it works or merely beats what came before.
world models Internal representations of an environment that an agent builds and revises, and whether machines can build them fast.
Mental modelThought experiments
On the topicWhich mental models help you think better?under our view: A thought experiment can test an idea before the world does