A single "one-way door" decision now defines the transition to AGI governance. Nathan Lambert argues that current frameworks fail to prepare for the sudden arrival of autonomous intelligence. This gap leaves regulators chasing technical reality. Practitioners must now prioritize rapid, adaptive safety protocols over static policy documents to avoid systemic failure.