Research

Independent research on small, calibrated models for the everyday decisions inside software. We publish how we measure, what worked, and what didn't.