Manny’s Reviews > Scary Smart: The Future of Artificial Intelligence and How You Can Save Our World > Status Update

It's interesting to compare this book with If Anyone Builds It, Everyone Dies. Despite the fact that Gawdat is writing several years earlier, he strikes me as much more in touch with the realities of the situation: it's too late to stop superintelligence from happening, a capitalist society doesn't work that way. Our only option is to figure out how to deal with it.

I was wondering if this book would be hopelessly out of date (it came out before the release of ChatGPT-3.5), but so far I don't get that impression. He makes a strong case for superintelligence in the next few years and keeps reminding us that we always underestimate exponential growth.

He says the second half of the book will say what we can do about it. I'm curious, seems impossible given the build-up.

The author's fact-checking is amazingly sloppy. 60 pages in, I've been told that AIs play all games better than humans (maybe Bill Gates could explain bridge to him sometime), that R2-D2 never actually does anything in Star Wars (doesn't he remember the ending of the original movie?) and that the showdown between Bowman and HAL in 2001 occurs "in the far reaches of the Galaxy" (in fact near Jupiter).