Raw LLM Responses

@voidmammal This could be argued- but the AI currently isn't smart enough to recognize a mistake, to see someone say "Hey, that's not how an arm works, it's meant to curve like X'. and then fix the mistake, No, it quite literally just sees an image, in a computer, then spits out a copy- Yes you can argue humans do that, but we're getting a live-feed of images 24/7 through our eyes, what we see as reality, which we base all our art on, an AI has no eyes, no which-way to recognize why something moves the way it does, it just sees an image and spits out something similar, it- unlike humans, has no base to correct itself from. Also, on the matter of trivialization, yes, the models CAN output cool artworks, Usually through batching though- where a human, generates some 20+ Images, then repeats until it finally generates something usable, the human equivalent is to commission an artwork, the style can depend on whoever you commission, and the contents you can explain in the commission, afterwards they will make an actual piece of art, Flaws? Perhaps, but human flaws that are more-so brush stroke issues than a composition issue, the AI outputs so many flaws that you can glance over- so I guess you could make an argument that AI can make art good enough for just a glance, which it does. (I'm also saying this from a person who has experimented with AI-art before, I tried to get it to make a specific character, from a PNG, and also text, neither times was any model I tried able to remotely reproduce said art, mainly due to the fact it was in a pose AI hadn't trained on- and an artystyle that's not common, it had no solution.) Anyway to summarize, while AI recreates art based on art, a human will make art based on reality, and their experience will shape the outcome of the 'art', so AI hasn't yet been shown to create its own artstyle, nor do I think it will soon, it will most likely improve its capabilities, or collapse, either or, but it has a long way to go.