SUMMARY: (5 of 5 stars) Absolutely fascinating discussion about the potential and risks of AI on the future of humanity and how we might best avoid the worst outcomes by being more thoughtful in our development of artificial intelligence. It’s rich and thought-provoking. The author’s style is orderly and direct. Highly recommended reading for making sense of AI and the latest headlines on the topic.
SPOILER WARNING: From this point forward, I’m going to discuss this book without any concern about spoilers. So, if you don’t like spoilers you should stop reading now.
This book covers so much ground and gets pretty deep on a very wide range of topics. To give you an idea of what I mean, here’s a brief list of some of the topics explored:
- How might AI kill us
- Why AI probably wouldn’t kill us
- How the universe began and might end
- How we could explore the universe using different types of AI
- How we might harness the power of the sun
- How human consciousness works… kinda
- How computers work… kinda
It’s VERY thought-provoking and I’m honestly not sure I truly grasped everything that was being discussed.
Not long ago, there were a bunch of headlines about Elon Musk saying that AI was going to rise up and kill us all or some such. These headlines are wildly out of context and this book puts them into a much clearer, more honest framework.
Could AI present a bigger threat to humanity than anything we’ve ever seen before? Absolutely. Just imagine what a poorly designed computer program could do with not only the ability and capacity to access all of our systems, but the will to do so with our destruction in mind. No bueno. Obviously.
As it turns out, Elon Musk is simply part of a group of people who want humans to take a thoughtful approach to designing and developing AI so that we avoid those doomsday scenarios. That is not at all controversial… but “not controversial” doesn’t earn clicks or sell papers.
For the most part, the author’s writing style is strong. It’s clear and direct. It’s creative and personable. And I really like the way he organized the chapters including “bottom line” summaries of key points from the chapter. My only complaint is that even though it’s very clear and easy to track, I found the “here’s a list of X items. Now I shall discuss each item” structure mildly repetitious. It’s perfectly succinct, well-organized, and so easy to follow… but he uses that structure over and over and over again and it started to feel monotonous at points.
My only other quibble with the book is something of a philosophical one. I don’t like his “teleological” approach to describing causality in terms of intent. For example, when you drop an object and it falls to the floor, he would describe this as the object “seeking” or “wanting” or “striving” toward the floor even though the object is doing no such thing. I understand his point in taking this approach and it is effective in shifting the reader’s perspective in assessing the conscious capacities of artificial intelligences. It is nevertheless inaccurate and misleading in some ways.
Those are my only two complaints worthy of note. I otherwise thoroughly enjoyed this book. It’s smart, impassioned, well-researched, informative, and accessible. Highly recommended for anyone who has even a passing interest in artificial intelligence.