The Third Branch of Machine Learning

There are these examples of using machine learning and training things will get better at like playing the game way faster than anybody else. And I would say, is that supervised learning or is that unsupervised where there's a specific goal of stay alive as fast as possible and then you just let it go? That's actually like that third branch that I wanted to talk about next. It's reinforcement learning with most of these games. So there's no like clear answer. But if you can correctly define your reward function to be the thing you actually care about, it's very much more likely to be successful.