In which I link positively to an article in National Review

Derbyshire's debunking of Gilder is good, 'cept that he doesn't get information, and its essentially random structure either. But it's well worth the read.