Comment by andrew_mason1

this one i really don't get. it seems pretty low-effort to me to just:

1. make the dish

2. take a photo of the dish

3. say "hey AI, touch this photo up to make it look nicer"

and you'd get way better results than this samey slop that looks nothing like what you'll actually be served?