I'm a software engineer. In my own time I do machine-learning research, mostly because I don't believe what we currently call intelligence is intelligence. It's artificial intuition: pattern-matching against the known, fast and effortless and empty at the frontier. Real reasoning is search, and search is the part we keep skipping.

Most of my work circles that gap. This is a working record of it: what I'm building, what I'm reading, and what I've been wrong about. It updates when I have something worth saying.