Though fun to see. More interesting than the station keeping (indefinite part) is that it enables ~100km orbits, which are ~5x closer than what remote sensing satellites use today and thus could get 5x the resolution for the same instrument.
That is an extreme example for multiple reasons, of course, but the point is that the gulf between idea and working thing is often monstrous. Ideas are cheap, implementation is hard.
I would think a bit less. You won’t get a factor of 5 decrease in atmosphere to watch through and I would guess you’d get more vibration.
Indeed at the bottom tfb quotes heights of 190-250km for model calculations.