Saw a demo for this a while ago on here and lost it, been thinking about it ever since. Now that visual models are getting a lot better at vectors and reasoning in general I wonder if we could reach a point where it is at least possible to translate part of an cartoon into a scalable vector for preservation or remastering