I like those "wait, you can do that?" moments in technology, and turning ideas into things people can actually use. LLMs make it all even more fun.
Most of my time goes into nanobot, where I'm one of the maintainers. It's where I put my curiosity about AI into practice: trying new ideas and refining the details that make it work well in everyday use. I also have a few projects I started on a whim, like fastjev, gloss, and notiq.
I also keep a personal collection of skills. I use them almost every day and update them often. If you're curious about how I work, feel free to give them a try.
Lately, I've been fascinated by recursive self-improvement (RSI). Can AI improve itself in ways that make it better at the next round of improvement? The idea that the ability to improve can itself be improved really excites me. I'm learning more about it and want to explore what's possible by building and experimenting.
I occasionally write about AI, technology, debugging, and security at CYR'ML. It's where I keep notes on things I've tried and questions I haven't figured out yet.