Eugene Vinitsky, a professor at NYU's Civil and Urban Engineering department, dives into the fascinating world of reinforcement learning (RL). He discusses groundbreaking results in self-play for self-driving technology and its implications for future RL applications. The complexity of self-play in multi-agent systems is explored, alongside its surprising link to language model advancements. Eugene shares insights on scaling simulations, the importance of reward design, and the rich potential of AI collaboration, making for a thought-provoking conversation about the future of technology.