Doesn't bother me at all. Only two simple facts are necessary:
1. There is wind shear between the air and ground.
2. The vehicle has wheels.
Wind shear can be extracted via a grounded propeller, or transmitted to wheels. The vehicle needn't have any drag or mass, by itself. The prop also can have arbitrarily low drag, at any relative wind speed, by rotating at the corresponding co-moving speed. And wheels have very low friction on ground, of course.
It follows that, while a sailboat might not be able to go upwind by itself, one with a prop geared to a windmill could. It would be less efficient as the water isn't a rigid body, though it could be quite good as water is a good thousand times denser than air.
I'm not actually sure how important the relative stiffness or density is. Say you were in a "large" balloon, sailing the shear band between jet streams. Both flows are air, same density, same temperature. ("Large" as in, "large enough" -- including fractional-planetary scale if necessary. If it helps, think of the shear bands on Jupiter, and something big enough to straddle the shear zone -- perhaps the size of a large moon. Just for sake of argument, we'll assume that's plausible.) Certainly, you can track equal wind velocity, in either direction, by moving towards the middle of the respective stream, or anything inbetween in the shear zone, where you get rotation instead. Is there anywhere you can move in this setup, given one or two propellers, and suitable gearing, that allows greater than local airspeed? (Where "local" is measured within proximity of the path of the vehicle, since we do need to take account of its size in this scenario.)
I think you can, because you could extend a turbine into the shear zone, and move it relative to the vehicle so it's co-moving with the flow in the shear zone; it contributes no drag. It can then be geared to a propeller, pushing the vehicle forward through the stream. Mind -- how you retract and reset this turbine, is an open question, no small feat by itself; but if it is collapsible and low mass, it should require arbitrarily little work to do this.
Tim