Or I used to. But rope drives seem to be need to be replaced often and don’t seem to be able to do powered ascent. So I think it’s probably better to integrate the above extra airborne pulley into a carriage that also includes a motor/generator and make it flygen instead.
Another idea then might be to put multiple balloons in a train. You’d do something like put an extra non- rotating ring around the primary airborne pulley, or disks on either side of balloon, not touching it except through bearings, that the tether/cable going to the next balloons would slide through. The carriage would be permanently fixed to this and the tether/cable would also slide through it. The carriage/ring would clamp to the cable at a spot where an electrical connection was made possible. Since the balloons would be flying at different radii and with that speeds, you would rotate them at different speeds to change their L/D ratios.
This start looking a bit like the Brainwhere concept, except this uses crosswind flight, balloons instead of VAWTS, and I don’t remember how it handled, or if it described, the reeling out process.
Thinking about BrainWhere, the extra ring around the balloon probably isn’t needed. You can make it so that the carriage rides along the airborne pulley and the carriage wheels go around it, fixing it to the pulley. …If it were not flying crosswind, maybe… If it is flying crosswind that puts a lot of stress on the carriage/pulley, sitting on the side and some distance away from the lift force creating a large moment arm. You instead want the tether to be in line with the lift force, entering the ring at the center of the balloon below and exiting at the center of the balloon above, so you would need the ring after all, or part of the ring.