Agents

Human or artificial agents carry work that needs expertise, labor, perspective, or judgment you lack.

Your knee has hurt for four months, and your system for dealing with it consists of a fitness tracker, YouTube stretching videos, and increasingly creative workarounds. Everything in that setup is a tool or a document. What the problem actually needs is a kind of carrier you have not enlisted: a person with trained judgment who can examine you and take responsibility for a diagnosis.

AGENTS names human or artificial actors, experts, collaborators, delegates, coaches, critics, institutions, AI systems, as one of the six spaces where cognitive work can live. It is the right carrier when the work needs expertise, labor, perspective, or judgment you do not have: capabilities that cannot be looked up or calculated, only exercised by another mind. The cue is a problem persisting because something is missing that no artifact or computation can supply, here, clinical judgment. The operation: name the missing capability precisely, choose an agent qualified to carry it (a sports-medicine clinician or physical therapist, not a louder app), and hand over a bounded piece of the work while staying the owner of your own case. What changes is that the problem finally meets a capability adequate to it. The boundary: agents bring incentives, availability limits, and fallibility, so verify competence and keep decision rights explicit, and do not route to an agent what a checklist or a calculator could carry more cheaply.