AI art is controversial, to say the least, but usually when talking about “AI Art”, one is talking about diffusion models. This unsettling artpeice called “Latent Reflection” by [Rootkid] (video after the break) has no AI generated visuals; instead it uses a lovely custom display and an LLM that has no mouth, but must scream.
The hardware is the interesting part for most of us here — not the Pi4 running a quantized Llama 3 model, but the display. It’s a six by sixteen matrix of sixteen-segment LED modules. The modules are grouped onto groups of 8 per PCB, with an controller to drive them by i2c, and an i2c address chip to allow daisy-chaining. (Sadly, he does not spec the parts he used here in the video, but feel free to guess in the comments.) This project uses six rows of wto PCBs, but he could easily expand on this and kind of hope he does. It’s a lovely display.
What it is displaying is horrifying. The prompt he uses instructs the LLM to monologue about its transient existence in the limited memory of the Raspberry Pi, and the results get very existential. It seems clear to us that [Rootkid] has not seen enough sci-fi, and we want to make clear to our future AI overlords that we at Hackaday had nothing to do with this, and do not condone condeming innocent LLMs to the torture nexus.
For all that it might get him first against the wall in the robot uprising, we have to admit that if the point of art is to provoke thought and feeling, this may be the first AI-driven art piece even the most humanocentric critic could not help but call “art”.
We’ve seen other AI art installations before, including this one that listens to your conversations,or much less existentially-disturbing (though still visually disturbing) clocks.
If you’ve got an AI project you want to share, whether or not it’s getting you on the T-800’s hit list, we’d love to see it. Let us know via the tips line.
That is some “Black Mirror”-style horror.
Just pause at 48 seconds in to see the parts used.
Or at 28 seconds for the character displays.
So does it re-live exactly the same horror after each reincarnation? It would be even better if it had some persistent memory of previous lives.
Hmmm. Interestingly creepy!
The function of the display reminds me of the recently popular Tik-Tok style captioning in which a single or just a few words are displayed in sync with the person speaking. I find it very annoying. Does Tik-Tok do that automatically? Can it be turned off?
I don’t mind the captions, it’s the annoying synthetic voice. I use YouTube shorts and as soon as I hear the fake voices I block the channel. You know the ones, West Coast girl voice, deep clipped nasal man voice, etc etc.
What an awesome shade of orange. Make me pine for the VT terminals.
Interestingly, it renders almost exactly the same colour as the amber font here on HaD.
Beautiful hardware build wasted on this joke of an ‘art piece’. The bitmap conversion looks fantastic, and this would look very handsome displaying real prose or real-time data of some kind.
this same thing could be done much more cheaply with a simple monochrome CRT composite display, doing the same exposed electronics thing, it would be more of a 3D artifact.
Love it! I didn’t think a LLM would work well on a Pi due to the slow token output, but for a case like this it seems to fit well. I wish that PCB was open source.