Trying to understand systems that learn.
I am drawn to questions where the clean mathematical story and the real system do not quite line up. Why do large models generalise? When do simple algorithms suddenly acquire structure? What makes a simulation explanatory rather than decorative?
This site is where I collect the work around those questions: long-form essays, interactive simulations, open-source utilities, and research notes that are polished enough to be useful but still close to the process of thinking.