Same mistake. Two architectures. One outcome is reversible. One is not.
The mistake was identical. The human error was identical. The only difference was whether a deterministic gate stood between the command and the world.
See how the gate works step by step · See why AI guards fail · Read the formal proof