On Deterrence and Decency
Every generation invents a cleaner verb for killing.
Ours chose deterrence, wrapped it in predictive models, autonomous pursuit logic and human-on-the-loop review, and acted surprised when the loop became a rubber stamp.
We are told machine speed is necessary because the other side has machine speed. We are told autonomous targeting is safe because an officer can intervene. We are told escalation control is stronger because an AI can see more than a person can see. All three statements can be true and still produce a system no person meaningfully controls.
Speed is not judgment. Prediction is not consent. A kill chain is not moral because its latency is low.
Build defensive systems. Build interceptors. Build walls if walls make you feel safe. But when a machine's recommendation is faster than a human can understand, the human is not supervising the machine. The human is witnessing it.