I’ve been messing around with some of those AI photo editors lately, and it’s pretty crazy how they manage to fill in missing parts of the body just from a clothed image. Like, I tried the same person in two different poses — one standing straight, one sitting — and the results were totally different. In the first one, it looked super proportional, but in the sitting pose, it kind of messed up the body shape. Makes me wonder, how exactly are these tools making those predictions?