Burning Man was definitely not around in the sixties but it’s easy to imagine that it was with these stylized AI-generated images. They were created earlier this month in Midjourney by Happy …
I'm sure someday the technology will be capable of producing hands that consistently have five fingers
Edit: Expanding on this thought, it's these small imperfections that make AI generated images of people really unnerving to me regardless of subject matter. The mistakes the AI tends to make seem really efficient at dumping the subject deep into the uncanny valley, even when it's very convincing over all. I often don't notice these imperfections right away, but I know they're there and I get this sense of dread anticipating the moment I'll notice whatever body horror is going on somewhere in the image.
ThisPersonDoesNotExist.com is particularly horrifying when it comes to this. 99% passable human faces, but there's usually a noticeable tell somewhere, usually in the teeth or hair, that reminds you of what you're really looking at. But the worst of it is what happens when the training data has a hiccup where perhaps it used an image containing two people instead of one. When the AI tries to make a person based on that you get surprise eldritch abominations that can manage to jumpscare you with a still image.
I'm sure someday the technology will be capable of producing hands that consistently have five fingers
Edit: Expanding on this thought, it's these small imperfections that make AI generated images of people really unnerving to me regardless of subject matter. The mistakes the AI tends to make seem really efficient at dumping the subject deep into the uncanny valley, even when it's very convincing over all. I often don't notice these imperfections right away, but I know they're there and I get this sense of dread anticipating the moment I'll notice whatever body horror is going on somewhere in the image.
ThisPersonDoesNotExist.com is particularly horrifying when it comes to this. 99% passable human faces, but there's usually a noticeable tell somewhere, usually in the teeth or hair, that reminds you of what you're really looking at. But the worst of it is what happens when the training data has a hiccup where perhaps it used an image containing two people instead of one. When the AI tries to make a person based on that you get surprise eldritch abominations that can manage to jumpscare you with a still image.