It’s hard to talk and think about artificial intelligence or superintelligence with out resorting to science fiction or shallow references to pop culture. WOPR, HAL 6000, Skynet – it’s best if you put those aside for a moment. Well not really, because the basic theme of the book is the tremendous danger a true machine super intelligence would pose to humanity. AI is an existential risk – quite possible it will be “our” meteor wipes us out.
I love serious books. I love how Bostrom saw a problem – lack of serious critical thinking about AI and its dangers – and sat down at his desk, sharpened his pencils and addressed the issue in a well written, insightful book. Logical, smart, yet with rare and subtle humour, Superintelligence is the sort of book that may not be that widely read, but will deeply impress those who do read it.
Systematically exploring the pathways to “super intelligence” the threats, dangers and the odds, this book is neither a pundit-like “rah-rahing” for AI, nor is this a conspiracy theorists’ doom-porn expose. It’s a serious, well written book – and in the age of twitter and fake news – it is incredibly refreshing. Bostrom – like all people’s who intelligence is worth respect, sees opportunities, but offers no easy or simple solutions.
A book that proves that “smart” and “academic” can be enjoyable, even invigorating.