Publications

Writing & Research

The thinking behind the work, published in the open. Papers and essays on decision systems and the ownership of intelligence.

PreprintZenodoJune 24, 2026

Separating Authority from Intelligence: A Middleware-Free Architecture for Autonomous Agents

An architecture that separates the authority operators grant from the intelligence agents author, so the intelligence an enterprise runs on stays owned and replaceable.

Read

EssayMosaic Singularity

The Mosaic Singularity

Why the contest with AI was never about depth, but scope. Intelligence as a mosaic rather than a single ascending line: neither inside a singularity nor outside it, held coherent by the human capacity to think across domains.

Read

EssayMosaic Singularity

The Intention Behind the Architecture

The purpose underneath the architecture: the conviction that the systems we build must leave intelligence owned by the people who depend on it.

Read

EssayMosaic Singularity

Your Desk Is the New Data Center

Why four converging forces, cheaper inference, capable open models, faster consumer hardware, and tightening data law, make local-first the inevitable home of enterprise intelligence, and cloud lock-in a strategic mistake.

Read

EssayMosaic Singularity

The Age of Autonomous Work

Why autonomous agents mark a fundamental break in how work gets done, and why everything before them was only a warm-up.

Read

EssayMosaic Singularity

Chat Is Not Work

The prompt loop has a ceiling: a great answer still leaves the work undone. Why autonomy, not conversation, is where value is created.

Read

EssayMosaic Singularity

Your Data Is Not Their Business Model

A clear-eyed look at why cloud AI keeps your data on its own servers, and why that incentive is structurally misaligned with yours.

Read

EssayMosaic Singularity

The Maturity Debt

The industry places expert-level agents years away. The real bottleneck is not capability but the architectural debt accumulating beneath it.

Read

EssayMosaic Singularity

Walls, Not Rules

Why advisory permissions cannot contain autonomous systems, and why the only safeguard that scales is a boundary built into the architecture, not written into a policy.

Read