The concept from It's Better Than It Looks that best exemplified Gregg Easterbrook's style was a comment that in some cases buying a country would have been cheaper than attacking it with a military. Perfectly logical, but totally implausible - who would you buy it from?
The idea that computer models can only produce conclusions baked into the program was one that I thought about constantly in the year after reading this book. Advances in computing may limit the strength of this statement in the future to "mostly true" but in 2020 it's probably a good enough rule of thumb. I'll add that most people who build models predicting the future aren't good enough at it - I don't mention this dismissively, but rather as a reminder that predicting the future is almost impossible; predictions made in 1980 about life today likely didn't account for the internet.
