How do you teach AI how to create a better traffic infrastructure for a city? How about getting it to solve a bunch of puzzles first? We look into an interesting strategy dedicated to improving artificial intelligence by throwing as many different problems at it as possible.
