@WindFisher implements 2 motors per cable wrapping the cylinder. These motors are ground-based. They supply two respective winches via gears. By this the 2 motors ensure the rotation of the cylinder in the two directions, variations of the used length of the cable to achieve power and recovery phases for the reeling system, then also works as generators during power phase.
I think this would be applicable for the balloon described on this topic, using vertical trajectories. However this device is protected by the patent FR3123317A1.
Apart from that, it would be possible to reduce the problem (for other problems it is another story) of inertia mass compared with the “aero lift” by reducing the diameter, while retaining the maximum possible length, as in the balloon below:
The cables (as for the belts) allow some rigidity of the balloon. That said the Magnus balloon would not be so scalable.