Raw LLM Responses

I for one am terrified of what AI can do with images, but a thing I've noticed, and a skill anyone can pick up, is how AI can't seem to replicate the similarity in skin, patterns and texture. Images of people will have spots, or birthmarks in different places in different images. Patterns that are expected to repeat, don't, and textures can seem blurry, or low quality. I actually think that the photography market can save us from losing our grip on reality, because if you can keep seeing more and more detail in an image, like imperfections in skin, bricks being out of place, and slight impurities in patterns caused by time, then you start getting a good idea of what pictures are real and which are not. Kind of like how we had to learn to differentiate photoshopped from real. A harder skill to master, is the idea of creative thinking. Does something look like it was prompted in a couple of sentences? Is there an overall theme, or did the AI have to fill out something itself? Do you see small details like birds in the sky, leaves falling off trees, people in the background? What would you expect to see? A shot of The Eiffel Tower, on a cloud free day, no people in the park, and no birds on the top of the tower would be indications of a fabricated image. Railings with a pattern that seem to have floating metal is a given. A person's skin changes from upload to upload, in ways that makeup can't change, or dirt being washed off does, are also great indicators if you have multiple images. A little piece of trash on the side of the road from a company you know, is a way to tell it's real. You see a bit of dirt on the clothes, that people are wearing. Something like a car parked in an imperfect way might be an indicator of a real shot too. The seasons match. You're not going to see a baby deer in the winter, as breeding season and winter don't match that way. Sorry Bambi. Or let's just hope that AI can keep up, and do the work for us, in this field as well.