What well known dissertations on machine learning should I read before I form the opinion that it’s dangerous hype?