lukeprog comments on Expecting Short Inferential Distances