Authority Under Acceleration: Institutional Rate Limits in Capital, Governance, and AI

Failures in capital markets, public governance, knowledge institutions and artificial intelligence are normally treated as separate problems with separate remedies. This paper argues they are one problem observed in four places: institutions are rate-limited meaning systems, and each is being driven past its rate limit. Authority is not a possession but a continuously produced good. Producing it requires interpretation, deliberation and legitimation, each of which takes time and none of which can be compressed without loss. When the operational rate an institution faces exceeds the rate at which it can generate, absorb and legitimate meaning, authority degrades — and it degrades in a characteristic order: simplification, then centralisation, then panic. The paper states this as an invariant, traces it through extractive capital, compressed governance, accelerated discovery and AI deployment, and shows that the familiar remedies in each domain are the same remedy: buffering, sequencing and constraint. Artificial intelligence is treated not as an agent but as a rate-multiplier — which relocates the governance question from what a system intends to what an institution can absorb. No solutions are proposed; the paper's claim is that the four literatures are one literature, and that the design primitive they share is a limit rather than an optimisation.