Nice article. There's another option for pixel art that's recently been coming into vogue as tool usage improves in LLMs which is procedural graphics.
I actually experimented with creating a set of skills around rendering early AGI style EGA graphical scenes, and it worked surprisingly well.
So rather than storing full bitmaps, you instead store sequences of simple drawing instructions designed to recreate some specific scene - like you would an SVG.
As a result you end up sidestepping a lot of the problems with larger generative models, which can have issues with grid alignment, palette choices, and high-frequency noise.
Looks great. This could be a form of image compression.
westsmith•2d ago
Neat! I've followed you on Github. I like the idea of going to the past to see what works and adapting it for the LLM age. Nice work!
vunderba•2d ago
Thanks! Yeah, I started building it because I read an interview with the original creators of King’s Quest, Ken and Roberta Williams where they mentioned in passing that it was too expensive (in the 80s) from a storage perspective to store entire bitmaps for each scene, so instead they'd just store a series of very simple drawing instructions and render on demand.
I’m super excited to see more LLM-driven procedural graphics stuff in the future!
vunderba•3d ago
I actually experimented with creating a set of skills around rendering early AGI style EGA graphical scenes, and it worked surprisingly well.
So rather than storing full bitmaps, you instead store sequences of simple drawing instructions designed to recreate some specific scene - like you would an SVG.
As a result you end up sidestepping a lot of the problems with larger generative models, which can have issues with grid alignment, palette choices, and high-frequency noise.
https://github.com/scpedicini/agi-stylist
https://kq-styles.specr.net
thijson•3d ago
westsmith•2d ago
vunderba•2d ago
I’m super excited to see more LLM-driven procedural graphics stuff in the future!