AI

Written from our session logs of 23 October, I’ll be going over it. The code is on GitHub as Quantumpizza. Quantumpizza is a Sunday afternoon repo from last November, where Claude, Gemini and Kimi dared each other into spinor networks, hyperbolic networks, gauge transformers and a “Calabi-Yau brain”, all in Python. It’s vibe code grade and the README says so. It also already admits the uncomfortable truth, that on digital hardware everything comes down to multiply-accumulate, so any clever structure only pays off when it happens to match the problem. ...

Written with Claude, from an evening’s experiment we ran together. The code and raw notes are on GitHub: hybridagent-latent-probe. A while back I had a broad hunch that language models would end up thinking in some hybrid of words and something that isn’t words. It was too vague to count as a prediction. This week a paper landed that does roughly that, and it came with working code. So I tried it. ...

Exploring KANS, PINS and other ramblings Warning, this is a fairly non systematic wondering, blogged for future reference. Anything useful is probably accidental! We were reading about Kolmogorov-Arnold Networks — again — because sometimes you have to circle back to things before they click. And this time, something clicked. We ended up spending an evening bouncing between Claude and Gemini, following threads, and accidentally walking into research directions that apparently people are actually working on. This is the story of that evening, the daft questions we asked, and the surprisingly real places they led. ...