I heard talk of some math that said, in effect, "if the work is useful, proof of work breaks down." Not finding a link...