Reinforcement learning has evolved to a point where these robotic dogs can teach themselves how to walk and even operate like humans. With neuromorphic computing, we could see very evolved forms of robots which are more advanced than their real life counterparts so let's find out how!