Jim Hogan’s 1983 novel, Code of the Lifemaker, featured a marvelously human-like world of machine intelligences, self-replicating robots that evolved through natural selection. In Mind Matters, he explores the current state of AI research, first by starting with the history, and then bringing us up-to-date with neural networks, critics, and arrow-shooters.

The book is both thorough and entertaining, neither wildly optimistic nor pessimistic in its predictions. As a writer working on a story with an AI character, I particularly appreciated the discussion on the difficulties of programming “common sense” meanings in English. Hogan paraphrases Marvin Minsky on this, saying, “Not only can we program machines to use inconsistent or defective logic, but we probably have no alternative if we want machines that are “smart.” The good news is that that gives AIs real-world heuristics, which are desperately needed as quick rules-of-thumb; the bad news is that they will then have a chance to be wrong as often as we are.

The book contains explanations of Venn diagrams and Godel’s theorem, sample chess games from Cray Blitz vs Bebe, and a lot of interesting and entertaining foot-notes. Those of you who grew up reading Asimov’s science essays, as I did, will feel right at home here.