The idea from the book that seems most obsolete to me is that a web startup founder could gain a competitive advantage by using Lisp. (Not that a Lisp might not be the best choice, but your competitors are more likely to be using it too, and if they're not they're using something much closer than they were in 2004.)
