I hadn't considered the drag-a-fan approach. It brings to mind an interesting kind of sailing where you've got multiple wind directions to solve for. Never mind that an atmospheric colony wouldn't really have a lot of destinations besides itself. Might as well just be one big flotilla.

I expect that a more boring approach would prevail: solar panels. No need for impossibly tall airships to resist wind shear just so they could get at both sides of the hot/cold setup for operating a sterling engine.

Engineering sense aside, I still wanna know if nature would allow the geothermal approach. Something about the idea that a column of air at rest would settle down into such a disequilibrium that it could be harvested of energy in this way feels wrong to me, but I've been unable to muster the thermodynamics chops to align my head with my gut.