The Hired Loop: Who Controls the Error Signal?

*The Hired Loop: Who Controls the Error Signal?* examines a problem that persists even in institutions with functioning oversight: correction machinery can remain active while control over what counts as error—or what may be done about it—shifts elsewhere. Building on *Don’t Break the Loop*, the essay distinguishes three institutional freedoms that are often treated as one: the freedom to reason, the freedom to learn by revising one’s account of reality, and the authority to correct by changing conduct or allocating resources. The paper develops a common pipeline through which reality must pass before it can alter institutional action: observation, data, classification, measurement, reporting, decision, and consequence. Along that pipeline, it separates three epistemic powers—measurement, interpretation, and adjudication—and identifies three distinct failure modes. **Capture** occurs when dependence on the evaluated actor alters or discounts the signal or judgment. **Flanking** preserves an independent judgment but deprives it of practical force. **Orphaning** occurs when a recognized adverse finding reaches no actor possessing the combined duty, authority, and resources required to correct it. Four public-record cases test the framework: operational testing of the F-35 and the accumulation of aircraft before the full-rate decision; Boeing 737 MAX certification and the informational chokepoint surrounding MCAS; Wells Fargo’s sales-practice warnings and incentive structure; and the Joint Tactical Radio System portfolio, where accepted interoperability remedies closed without implementation. These cases are presented with competing explanations and evidentiary limits rather than as universal proofs. The constructive proposal is concise: **separate epistemic authority and join corrective accountability**. Institutions should protect the independence, evidence access, scope, and funding of those who measure, interpret, and adjudicate, while ensuring that every authorized finding lands on a named owner with the duty, power, resources, answerability, and escalation path needed to dispose of it. The essay concludes by applying the framework prospectively to Combined Joint All-Domain Command and Control (CJADC2). It records a dated, publicly scoreable prediction concerning whether emerging requirements-and-resourcing reforms produce observable portfolio-level closure by approximately fiscal year 2029, with an earlier statutory comparison point in December 2027. The result is both an institutional diagnostic and a blueprint for making correction consequential.