The whole “turn this picture into another picture” AI use case is very limited compared to training it to render assets procedurally in a defined style. The former is really just about gag / novelty akin to boardwalk caricature “art”— the latter would actually be useful in creating art.

the hype over sora is hitting absurd levels sure it's impressive but still gives off strong creepy AI vibes like a bad dream there's a serious fundamental flaw in all AI "art" humans see the flaw, but can't describe it _... See more