I used to be able to easily notice the difference in going from light sew ups to heavier clincher training wheels. It was when climbing out of the saddle that it was obvious. I suspect light wheels were more of a thing in the days when we got over everything in a 42x21 (or higher) gear, no matter what. Usually climbs were done about half in the saddle and half out of the saddle. When climbing like this with light wheels, there's typically a micro acceleration with every pedal stroke. That's when the lighter wheels feel easier. Perhaps there is a great bio-mechanical efficiency when pedaling like this - something the mythbustering study did not address. From of physics standpoint of course speeding up and slowing down is a waste of energy. There may however be an optimum where a small amount of this velocity variation is balanced by a greater bio-mechanical efficiency. We are not motors after all. Our legs pulse, they don't spin.

At any rate, what I remember is that with a small conscious change in pedaling style - a bit more in the saddle, a bit lower gear - slightly heavier wheels seemed to make little difference. They could be compensated for. Lighter wheels still felt nicer. Obviously modern bikes now have lower gears and most people spin climb. That kind of explains the results in that study. Even so, .25% or whatever is not nothing. It's still a few seconds on a long climb.

The moral of this story is don't wear a puffy shirt, or don't worry about it and just ride.