Talking to experts about tech (really, talking to experts about a lot of things) is an exercise in readjusting expectations about what the tech/thing/field can really do and what is BS. This book does that par excellence because it makes almost all its points on that front via whimsically absurdist jokes, silly drawings, and hilarious anecdotes. (Almost all, because Shane does bring in the “AIs that learn on biased data are equally or more biased” examples and similar warnings; those are sobering-to-horrifying.)
I learned an unexpected amount about how AI work, laughed an incredible amount, and was sad when the book ended. If I ever teach a course on AI, I would assign this book. Highly recommended.