Papers

Foundational papers building the case for authority control as a distinct infrastructure layer.

This work develops the structural argument behind the site: why access is not authority, why consequence requires its own enforcement layer, and why judgment infrastructure becomes more important as AI accelerates execution.

Research program

One metaproblem, examined across disciplines.

The work asks how organizations preserve judgment, delegated authority, accountability, and evidence when software and AI compress the interval between recommendation and consequential action. The papers approach that problem through cognitive science, organizational theory, cybersecurity, institutional accountability, and human-AI collaboration.

Papers in development
  • 01

    The Collaboration Gap

    Why better AI is not producing better decisions. The structural diagnosis of human-AI underperformance, and the measurement infrastructure required to close it.

    Fall 2026
  • 02

    The Three Collapses

    Process compression, structural anchoring, and authority transfer: the patterns that erode human judgment upstream of the commitment boundary under AI acceleration.

    Spring 2026
  • 03

    The Last Boundary

    A security analysis showing how cybersecurity's foundational principles point toward the commitment boundary, and why that boundary stands as a distinct enforcement domain.

    2026
  • 04

    Authority Control

    An operating model for governing organizational authority over consequential actions. Organizations need controls that address the distinction between valid access and valid authority.

    2026
Architecture document

Authority Control Architecture

The architecture overview as a PDF.

Request the architecture PDF →

Brief analyses

Browse the insights hub
Claude Code
When Agents Learn to Cooperate
The London Whale

Core concept pages

Authority Control + Zero Trust
Data at the boundary
Evidence