Really great book. For a while I'd been searching for a book that really goes into detail on the possibilities of superintelligent AI, and this book was perfect for that. If you're not into computer science and math/stat, you can probably skip the beginning where he goes through the history. The middle of the book where he talks through the different risks and how we might address them were fascinating and I think everyone would enjoy that section. The last 3 chapters were pretty boring, but there were still some significant insights in there. I'd recommend skipping the last 3 chapters unless you are REALLY into this subject (like me).