Doesn't waymo and other self-driving systems use reinforcement learning? I thought it was used in robotics as well (i.e., bipedal, quadrupedal movement).