> Making art is hard. But art is mostly about surfacing the inner world, and only in part about skill. It’s unfortunate that art selects so strongly for skill. Not to sound like a luddite, but I do question the idea that the skill gap is merely an inconvenience. I suspect learning how to paint or make music changes something in yourself which teaches you some deeper life lessons. I've heard the phrase (paraphrased):…
Does the totality of the project have that property? That would be less clear, but IMO, no. I see it as technically driven, not psychologically driven, although I can see how you could write an artist's statement that claims it was (it's about mirrors, after all, which are hugely symbolic).
To be clear I still like it, and if I'd done it I'd be proud of it. But it's more artifice than art.
(If I was him, I'd slow the frame rate down, not speed it up - work with the technical limitation, not against it. Have the system only display "good" images, and not update the display until another "good" image is generated. The code that decides if an image is "good" or not would be the most interesting part of the system, and could fairly be said to embody the artist's intent, and so cross (my own personal) threshold into capital-A Art.
I'd also experiment with buffering the image stream à la _Light of Other Days_ by Bob Shaw.
Oh, and as Halloween is nearly with us, the temptation to occasionally inpaint a figure standing behind the viewer would be massive.
Idle thought: to get some stability in the image, would it be possible to have an LLM generate random video filter code, instead of random images? "Write me a video filter that makes the input video look cubist". "...like an oil painting" "...with a Flash aesthetic". etc etc. Every time a filter gets generated that doesn't actively crash, swap to it. No idea if that's feasible or not.)