This factoid comes up again and again, but really, it's only true for certain ways of measuring efficiency. If you're measuring the energy expended to move a person one mile on an already existing flat road and bicycle, yes, it beats just about everything, even simply walking.

If there is no road or no bicycle (i.e. they have to be built or otherwise procured, or the human is going somewhere where there is no road or path), or the road is broken up or super hilly or something ... then it's not so clear.