Fantastic conversation with Igor Halperin around the application of reinforcement learning into forecasting problem, and the limits to data and understanding the world.