John Quiggin wrote an interesting post about the failure of search theory's prediction that with the advent of the internet (reducing search costs and times), unemployment should drop faster.