I'd never had the time to read Nick Bostrom's "Superintelligence" before. However, on my many recent flights, I've had plenty of time.

It's impressive to see how well this book has aged, and at the same time, how much faster its predictions have come true than expected. For example:

At the beginning of the book, Bostrom writes that at that time (2014), Go computers would be playing in the amateur leagues, and - according to his prediction - if development continued at its current pace, they would win the world championship in 10 years.

Well, as we know, things happened much faster: In 2016, Google's AlphaGo defeated Lee Sedol and catapulted itself into the world rankings. And in 2017, AlphaGo Zero followed up, mastering the ancient board game of Go completely from scratch.

The other predictions have also become reality far more quickly: image analysis, object recognition, and controlling a robot are particularly challenging, and a general intelligence that "understands natural language as well as an adult" is the future, capable of "everything humans are capable of."

Twelve years later, in 2026, almost everything has become reality and arrived faster than expected. The most difficult unsolved mathematical problems are being solved weekly. A review using this book clearly demonstrates how exponential the current developments are and how quickly all expectations have been met. I highly recommend rereading this book.