Press the flag and watch. Turn on turbo mode (shift+press flag) for the AI to learn faster. I was going to allow the player to control the bottom paddle, but then training would be way too slow.
Want to learn how reinforcement learning works? Check out this project: https://scratch.mit.edu/projects/559749241/ @natalie for the iconic Pong Starter project :) The bottom paddle automatically works. The top one learns using reinforcement learning on a simple perceptron.
