This is a very introductory walk through some of the basic ideas of reinforcement learning - the third arm of machine learning which is perhaps less often talked about than supervised and unsupervised learning, but which is showing fascinating promise as well as giving us plenty of questions about our own interactions with the world. It is the study of how agents interact with the world using trial and error. I will discuss some of the most fundamental algorithms in RL, from model based to model free learning and show some of the recent advances and challenges that we still face in this area.