I had the same expectations. Would be a super difficult control problem I think, with stuff like the end of the tails motion range, collisions, lack of ability

I had the same expectations. Would be a super difficult control problem I think, with stuff like the end of the tails motion range, collisions, lack of ability to predict the motions of actual limbs and so on. Perhaps a 2D model could be a useful approach to this problem, with super low friction swivel chairs too high for the feet to interact with the ground? Might even be an interesting playground for Neuralink. Robotic tails with a diffuse control from the brain, now that would put the "furries" comments elsewhere in the thread on the next level!