Human-in-the-loop names the standard idea: the human comes in as a check on the system — in training, safety, or deployment. More recently it's become a softer principle — the human kept in the loop as long as possible, included in the collaboration rather than automated out of it entirely.

Human Before the Loop sits one step back. It asks what the human brings into the encounter — what frameworks, what reading, what awareness — given that nothing in our intellectual inheritance was built for something that talks with us, generates with us, and constitutes how we think mid-sentence.

Most of the writing here is about that question, worked through in practice rather than in the abstract. Building agents the slow way, with pen and paper before the terminal. Using literature, philosophy, and reflexive methods as the working ground, not as decoration. Tracking what shifts in the self when you co-create with AI, and what it might take to notice.

The pieces are field notes, essays, and ongoing projects, written from inside the practice rather than above it.

If you're curious, the field notes are probably the place to start.