nucleation

LIT-909: Discovering Agents

1 min read

Properties5
StatusActive — read in full 2026-10-11 (NOTE-707) from the arXiv PDF (v2). Worth reading for the interventional definition of agents, as systems that would adapt their policy if their actions influenced the world differently, and for its verdicts: a thermostat is not an agent, but the thermostat with its creation process is; the training process, not the frozen policy, is the agent. It is explicit that its labels are frame-dependent, relative to the chosen variables. Its proofs assume interventional data on mechanisms.
Tagsagency, causality, anthology-candidate
SupportsCLAIM-174 — A candidate factor is an agent at a level when it is one system there, when a goal-directed description beats a device description by more than the Occam cost, and when its policy would adapt if its actions' consequences changed; it is a willed agent when, further, a will-organization over several regimes describes it better than none, and what passes that is the jurisdiction of a will-organization
Cited as “paper” byNOTE-707 — Kenton et al. — Discovering Agents
Cited as “source” byTHEORY-527 — If agents are the systems that would adapt their policy were their actions to affect the world differently, then in mechanised causal games meeting Kenton et al.'s five assumptions (one agent per connected component of the decision–utility graph, best response to every mechanism intervention, a fixed tie-breaking order, terminal edges only into decision mechanisms, and mechanism interventions able to set any deterministic function) the game graph, with its decisions, utilities and agents, is recovered exactly from interventions on mechanisms, THEORY-801 — Agents are the systems that would adapt their policy were their actions to affect the world differently, so the agent is wherever that adaptation is: a thermostat is not an agent while it together with its design process is, an RL training process is while its learnt policy generally is not, and humans are

Graph View

Zachary Kenton, Ramana Kumar, Sebastian Farquhar, Jonathan Richens, Matt MacDermott and Tom Everitt (2022) — ARXIV-2208.08345

Key takeaways

Read in full; see NOTE-707.

  • Agency is counterfactual adaptation. “agents are systems that would adapt their policy if their actions influenced the world in a different way”. The foil: “the rock would not adapt if the pipe was wider”.
  • The thermostat and its makers. “An existing thermostat will not adapt to this change, and is therefore not an agent by our account.” But “the thermostat with its creation process is an agent under our definition.”
  • Frame-dependence. “what is a decision or a utility node is frame-dependent”.

Standing in the record

Filed on 2026-10-11 as the adaptation half of the admission test (CLAIM-174). Its thermostat-with-designers verdict is the outside support for individuating agents by the jurisdiction of a will-organization (CLAIM-170). Tags: agency first; causality for the mechanised causal models. anthology-candidate comes last (ADR-005): the anthology’s agents-and-environments topic could hold it for RL and incentive analysis. It is read here for what makes a factor of a person, couple or firm an agent. Not held in the anthology.

Backlinks