I've always been curious about the growth model @nervous_jesse developed so I thought I'd explore it a bit. Here is the model but with expansion governed by grey-scott reaction diffusion.
I was unsatisfied with how handles were resolving, and I found a solution using quadrics... but then I happened upon another implicit form which is even better,The Darboux Cyclide. The Darboux Cyclide is like a super torus...
I have quite a bit more to learn, but it turns out least squared fitting of the Cyclide worked very similarly to fitting a quadric!
OK, here's the whole thing together. Going to stitch 6 of these together for a submission to little group show. Not sure how to display the sequence, probably just s monitor or something. I have notes, but I think it will look out pretty nice, looping the day away.