Chelsea Finn, Assistant Professor of Computer Science and Electrical Engineering at Stanford, closes out the CS224R Deep Reinforcement Learning course with a lecture on the frontiers of the field. She works through problem set-up, the current landscape of methods, and the harder question of deployment and evaluation, looking at how reinforcement learning systems hold up once they leave benchmark environments and get tested in the real world. The lecture builds on the course's earlier material on policy learning and reward design, pulling those threads together into a discussion of open research problems and where current methods fall short. Running just over an hour, it is a graduate-level capstone session aimed at students who have followed the course's technical arc rather than a general introduction.