The Moral Obligation Not to Centralize

The choice to build federated rather than centralized is not, finally, a technical preference. It is a moral obligation that follows directly from everything this volume has argued. An institution that claims to offer an alternative to the extractive, centralizing logic of the platform is required, by the logic of its own claims, to build in a way that structurally prevents the concentration of power it exists to resist.

This is harder than it sounds, and we name the difficulty plainly so that we may be held to it. Centralized systems are easier to build, cheaper to run, simpler to monetize, and faster to improve. The pressure to centralize will be constant, and it will arrive wearing the respectable clothing of practical necessity — from investors who want clean metrics, from engineers who want tractable systems, from operators who want efficient control. The moment we begin to grant exceptions is the moment the architecture begins its quiet conversion into the very thing we promised it would not be.

The river still runs; the knowledge that made it legible is gone. We are trying to build an institution that, when the pressure to simplify and centralize arrives — and it will arrive — has already written into its own architecture the resistance that neither goodwill nor governance alone has ever been able to provide.