I recognize that AI slop is controversial, and for valid reasons.
Nit I got bored and decided to see if chatgpt knows what fluffies are. Turns out that it does, and I have successfully traumatized it to the point that it no longer has any qualms about generating abuse stories.
And then I thought… can it generate pictures of fluffies?
Apparently, it can, but the hooves get a little sketchy. Unfortunately, it refuses to generate actual gore or waste, but I can get it to generate terrified and crying photorealistic fluffies in abuse-ish situations, just with no gore.
In fairness, I feel like fluffies would be kind of uncanny, like those posts where people use photoshop to reshape actual people into cartoon characters, so they’re shaped like the character but have visible pores, hairs, veins, etcetera.
Definitely and typical AI weirdness doesn’t help - the pink unicorn in the last picture has 5 legs (it’s standing on 3 and two are holding the doughnut)
I wonder if it is possible to instruct the AI to generate the subject and the background separately. Then someone could more easily edit the fluffy for abuse images or corrections like extra legs.
you could title that sturvingFAM and sell it in a musem
