Philosophy
Is Agency Compositional? Resource Commitments, Coalitions, and the Myth of a Single ‘Agent’
“The Pledge, Part II.” — Lucius Meredith (Feb 2026)
In brief: The essay treats agency as tied to the ability to make and honor commitments under constraints. ‘Single-agent’ intuitions break down where outcomes are produced by coalitions, protocols, and shared infrastructure. Evaluation depends on evidence: you need a reliable record of commitments, actions, and outcomes.
Key takeaways
- The essay treats agency as tied to the ability to make and honor commitments under constraints.
- ‘Single-agent’ intuitions break down where outcomes are produced by coalitions, protocols, and shared infrastructure.
- Evaluation depends on evidence: you need a reliable record of commitments, actions, and outcomes.
Key questions
- How do you design interfaces that make commitments explicit?
- Where does the coalition model of agency break down?
- What does an audit trail for multi-agent commitments look like in practice?
Why it matters for developers
Design explicit commitments (interfaces, policies, guarantees) and capture audit trails of decision points and actions.
agency
commitments
coalitions
multi-agent
audit trails