Agent governance research models the agent as the primary unit. I argue the unit is wrong. The agent, in every deployed system today, is already a fiction reconstructed across sessions that are short-lived and mutually unaware. The real object is a population of such sessions, always-running in aggregate, communicating through a family of uncoordinated substrates: tokens, memory files, handoffs, shared state. Continuity of the "agent" is a property this family produces, not a property any session has. I claim this population-with-medium is the correct unit of agent governance, that current protocols underspecify governance of the medium because they have been looking at the wrong object, and that the architectural move that makes authority survive session death — artifact-based state with signed authorization — can be extended to govern the medium. I further claim, speculatively, that continuity of cognition at population scales may require a governance substrate, though the substrate alone does not produce it. I ground this in existence proofs we already have: institutional memory, Wikipedia, open-source development. I name one gap in that grounding as the central open problem: cryptography formalizes authorship, not meaning, and the analogies work because humans evaluate semantic content in ways cryptography cannot replace. The paper defines the medium as a governance contract that specific substrates implement, defers consensus to implementation, concedes that the medium is governed retrieval rather than categorically separate, distinguishes access from declared influence, and names the open problems.
Tymofii Pidlisnyi (Tue,) studied this question.