The art is really bad. There's no easy fix for this: today's image models are simply not good enough yet. (At CFG>1, which is the typical setting, they give the overcooked "AI look". At CFG=1, that problem disappears but the images become blurry, curvy and nonsensical. It's a tradeoff that can't be easily resolved.) You can use models for coding, they're good at that thanks to RLVR, but the art still has to be done by humans.