For anyone not wanting to go over there, I drew a bad line drawing of a pony (I could have done better, but I was deliberately not putting too much effort into it, since I intended the ai to do most of the lifting.), then I ran it through img2img with this prompt: Prompt: A cute happy pony standing on a hill with one leg in the air, one ear down, line art Negative prompt: EasyNegative boring_e621
Then I edited the line drawing it generated, and did inpainting on the edit. Then I went to txt2img with a revised prompt (Listed on the linked thread. I did specify fur and eye colors, for example.), put the line are in controlnet using the lineart model, generated a new image with that, and threw that through img2img.
I was largely experimenting to see how well that worked… and I think it works pretty well!