Review: Algorithms to Live By: The Computer Science of Human Decisions

Algorithms to Live by is a self-help book disguising itself as a pop-psychology book but it is ACTUALLY a pop-computer science book. This misunderstanding comes from the title: "The Computer Science of Human Decisions" sounds like regular cognitive psyche 101, but that's not what the book is actually about. Instead, it is "human decisions which ought to be corrected by computer science." That is a big change, and it turns the book into something very interesting.

Computers and computer science are a very interesting field that sits at the combination of physics, mathematics, philosophy, and psychology. I don't think it is possible to be good at it without having a moderate understanding of those fields (and I believe that the interview questions posed during tech interviews are indicative of this being true). Algorithms to Live By makes this combination very clear. The structure of the book is repeated in a powerful (algorithmic!?) way: Explain a human problem, explain how a computer/math algorithm might answer that problem, complicate that problem to explore the literature on that problem, and then explain how well humans do at it naturally. For example, the Multiarmed Bandit problem, or the Explore-Exploit Dilemma is used to shed light on the phenomena that old people have few friends, but they are very, very close.

In fact, I would read this book just for the chapter on the Explore-Exploit Dilemma, just because I think its critical to thinking about life and how one ought to live.