Operating Note 03
Discipline precedes autonomy
Alignment → Discipline → Structure
Pattern
Autonomy granted before discipline is established produces inconsistency at scale. Teams interpret priorities differently, standards drift, and coordination costs rise invisibly.
Discipline isn't about control — it's about shared operating logic. When that logic is clear, autonomy accelerates instead of fragments.
Guardrails enable speed. Freedom without them creates drag.
What to do
Define non-negotiables explicitly. What must always be true, regardless of local context?
Earn autonomy incrementally. Expand decision rights as operating discipline proves out.
Distinguish operating standards from preferences. Teams can flex the latter — not the former.
Where this shows up now
This is the rule for embodied AI. A robot's autonomy should grow the same way a team's does: start narrow, prove the operating logic, then widen what it's allowed to decide. The non-negotiables — where it stops, what it never does near people — come first.