Obvious but still weird
Images produced by generative AI are kind of like photography but without a lens and without a world to point it at. Instead it is pointed inward at a massive glob of information, then sucked back out gradually in smaller sets of information triggered by words.
