The speaker considers using world-model simulations for reinforcement learning a paradigm shift.

“especially like world models with like reinforcement inside the world model you get nice information you can do the simulation really fast you get potentially things like gradient information depending on how you set it up um but mostly you can do like you can do infinite rollouts that's a it's a huge it's a paradigm shift for reinforcement learning which we can also see in something like large language models they something like grpo which effectively does doesn't learn a value function does like the the advantage estimation on the fly but just simulating a bunch of stuff”