"Wormholes" seem to describe ensemble averages of some kind, but
in realistic models of holography, like AdS_4xS^7, it seems there
isn't anything to average over, except, in some examples, a single
integer N. Is "averaging over N" enough to account...