The working, shown
Essays and models from the practice — on rehearsal, calibration, institutional memory and the owned learning loop. Written the way we work: the claim, the reasoning behind it, and how far it can be trusted.
PSP-001 The rehearsal layer: why we build it this way A decision system that only learns from committed decisions learns at the speed of the world. Why simulation multiplies evidence, calibration keeps it honest, and an owned memory makes it compound. 31 AUG 2026 →