Bostrom writes about superintelligence, with the particular focus on the dangers that may arise from the development of superintelligent artifical intelligence. This topic is very important, and it is clear that growing attention will be paid to it as its economic and political ramifications become more significant.
This book should be read by anyone with an interest in technology, or the future of the human species. It will be interesting to see how much of it stands the test of time, since Bostrom is careful to note throughout that it is too early to know what specific computational problems we will face when the time comes. But in a way this is unimportant: the book has served its purpose of getting people to start discussing this issue seriously.