The design of an institution built around a single connecting operator and intelligent systems, where direction is the scarce resource and headcount is not the unit of scale.
A class of institution is appearing in which one person holds the direction and intelligent systems do the work. This paper treats it as a form, not a curiosity, and asks the only questions that matter about a form: what holds it together, where it breaks, and what governance it needs before anyone should trust it.
For a century, the size of an institution told you something true about its capacity. That equation is coming apart. A class of organisation is now appearing in which the binding constraint is not headcount but the calibration of a single human node against intelligent systems that do the work. This paper treats that as a form, not a curiosity, and asks the only questions that matter about a form: what holds it together, where it breaks, and what governance it requires before it can be trusted.
Locating the form inside a century of organisational economics, from why firms exist to the principal agent problem, it shows why the classical tools for governing delegation fail when the agent cannot be deterred. It names the four failure modes every instance carries, sets out the governance a single node institution actually needs, and is honest about where the form is the wrong choice. The argument arrives, structurally, at the layer the form cannot build for itself: an identity and trust substrate over the orchestration layer, hardening now.
For a century, the size of an institution was a reliable readout of its capacity. That proxy is dissolving. What an institution can do and how many people it employs are no longer the same measurement, and every instrument built to underwrite institutions was calibrated against the proxy that is coming apart.
Intelligent systems did not primarily make people faster. They collapsed the internal coordination cost that Coase identified as the limit on firm size. When one of the two costs in that equilibrium falls by an order of magnitude, the equilibrium does not adjust gently. It relocates.
When execution becomes abundant, productivity stops being the constraint. The binding constraint becomes the fidelity between what one person actually intends and what the systems actually do, judged against an intent that mostly lives, unspecified, in a single head.
Every classical solution to delegation assumes a counterparty that can be made to care about consequences. A software agent cannot be deterred, incentivised, or shamed. Governance therefore moves from shaping what the agent wants to constraining what the agent can do.
The most dangerous failure mode produces no warning. The work does not degrade. It becomes an increasingly excellent execution of an intent that is increasingly wrong, and every available signal says things are fine.
A single node institution and a recklessly run one look identical from the outside. Accountability and trust are relations between an institution and the world, and a relation cannot be constructed from one side. The trust is migrating into the layer the organisation stands on.
A certification the institution can quietly swap when it becomes inconvenient vouches for nothing. The layer has value exactly to the degree the institution it certifies does not control it, the way a court's authority comes from the litigant not being able to select it.
The institution does not get smaller. It gets denser, and the only thing left genuinely scarce is the calibration of one human intent against work that human will never personally do.
The full paper, 30 pages, sent to you and yours to keep. We will know it reached someone serious.
One email. We do not add you to anything.