One of the authors was astounded to see this very argument in a Scientific American article on dissipation in computing written by a very respected scientist (whose name we withhold for the sake of avoiding embarrassment!).