I’ve been mapping out a framework for steering AI behavior through incentive structures—thought you’d find the mechanics intriguing.