AI: Endgame

Every attempt to govern AI is framed as a policy problem. These four essays treat it as a problem of organism: how a system maintains itself when the perturbation enters the apparatus that would measure it.

The series follows one argument across four pieces. Biology predicts integration, not consolidation, as the structural attractor. Competitive knowledge-work reorganizes around a human–AI compound the firm’s instruments cannot see. The judgment needed to evaluate that compound is built by independent practice the compound’s own efficiency removes. The remaining question is how to keep, on purpose and over time, the capacity to know if the answer is wrong.