Scan barcode
kasss's review against another edition
3.0
There was a lot of information here, but I feel the points the author was trying to make got lost in all the anecdotes. The book felt like a wall of text (at least when listening to the audiobook).
I can't help but compare it to Artificial Intelligence: A guide for thinking humans. Don't get me wrong, they're different books doing different things. AI: A Guide is basically AI 101, whereas The Alignment Problem addresses a specific issue. However, AI: A Guide addresses basically the same issue, and The Alignment Problem mentioned a lot of the same examples and anecdotes from the field. So there's a lot of overlap between the two books and I don't think my comparison is entirely unfair. AI: A Guide managed to be concise, educational, and have a lot of anecdotes that made the book not just informative but also fun. With The Alignment Problem, I was frankly just bored most of the time.
Two things that stuck with me from this book concern ethics (that in turn concern AI): the effective altruism (EA) movement, and the possibilism vs actualism theory.
I can't help but compare it to Artificial Intelligence: A guide for thinking humans. Don't get me wrong, they're different books doing different things. AI: A Guide is basically AI 101, whereas The Alignment Problem addresses a specific issue. However, AI: A Guide addresses basically the same issue, and The Alignment Problem mentioned a lot of the same examples and anecdotes from the field. So there's a lot of overlap between the two books and I don't think my comparison is entirely unfair. AI: A Guide managed to be concise, educational, and have a lot of anecdotes that made the book not just informative but also fun. With The Alignment Problem, I was frankly just bored most of the time.
Two things that stuck with me from this book concern ethics (that in turn concern AI): the effective altruism (EA) movement, and the possibilism vs actualism theory.
tony_t's review against another edition
challenging
informative
reflective
medium-paced
4.5
"The Alignment Problem" by Brian Christian is a comprehensive and even-handed survey of the current state and challenges of Artificial Intelligence. In the introduction he uses the "sorcerer's apprentice" as an example of AI run amok. He goes on "How to prevent such a catastrophic divergence - how to ensure that these [A.I. computer] models capture our norms and values, understand what we mean or intend, and, above all, do what we want - has emerged as one of the most central and most urgent scientific question in the field of computer science. It has a name "the alignment problem." This volume makes an excellent summary for those who want to read one book that will help them understand what this aspect of the future looks like. I can wholeheartedly give it my unqualified and highest recommendation.
tonyleachsf's review against another edition
5.0
I’m not an AI expert, but this is the best of the books I’ve read on the topic. It strikes the right balance between showing the underlying (human and technical) research leading up to 2020 and grappling with the ultimate moral implications.