Phased rollout and trust gates
Autonomy is earned, not declared. Students learn a phased path that pays off at every stop: first give agents better context (instruction files, build-before…
Autonomy is earned, not declared. Students learn a phased path that pays off at every stop: first give agents better context (instruction files, build-before…
Autonomy is earned, not declared. Students learn a phased path that pays off at
every stop: first give agents better context (instruction files, build-before-push,
and agent-readable linter messages written as instructions rather than
descriptions); then spec-driven generation with a human approving the validation
report instead of reading diffs; then selectively remove the human gate on services
whose numbers justify it; then expand to full auto-merge. The transition to
auto-merge is governed by measurable thresholds — scenario pass rate, evaluator
false-positive rate, and human-override rate — rather than by intuition, and any
phase can be the permanent stopping point. This mirrors the autonomy-level framing
from earlier in the module and the layered oversight of EB-8.