About the Practice

Systems of Thought is a publication and practice about the architecture of collective thought: the cognitive, institutional, and informational systems through which societies and organizations reason, govern, and lose the capacity to do both.

It is published by UX Minds, LLC, the consultancy practice of Jedi Wright.

Jedi Wright is an AI Experience Architect, Content Strategist, Information Architect, and Independent Researcher with 25 years of practice in UX, information architecture, and content governance systems. He holds the Nielsen Norman Group UX Master Certification in Interaction Design (ID #1068194), a rigorous practitioner credential from the field's leading UX research organization. His research on AI governance infrastructure, "The Inference Flagging Gap: A Missing Governance Requirement for AI-Integrated Execution Environments," is published on SSRN. He is the founder of UX Minds, LLC, a lead content strategist at Huge, and the author of Systems of Thought. He is based in Phoenixville, PA, a Philadelphia suburb.

What's published here

Five research projects, a software practice, and a convergence essay are developed in public on this site. They share a common concern: the conditions under which systems, informational, institutional, and cognitive, hold together, and the conditions under which they don't.

End of History, Revisited tracks the compound civilizational stress event now underway—the simultaneous failure of the institutional, epistemic, and political systems of democratic societies built to govern themselves and adapt. The AI governance window is its most urgent front. Each post is a dispatch from a specific front of that failure: a targeting system that hardened an unverified assumption into operational fact, a governance framework that arrived after the thing it was meant to govern, a democratic reversal that moved one clock without moving the other. The formal argument develops across a set of companion documents, the core essay, a policy framework, the AI Governance Window Tracker, an Agentic Accountability Playbook, a Legibility Project for practitioners, and a Companion Architecture document mapping their relationships, all part of the public record of this work.

The Tiered Content Framework is a six-tier content governance model for organizations operating at scale, pressure-tested across years of practice. It maps the production chain from strategy through delivery, with an Intelligence Layer governing AI's role at every tier. It is the most operationally complete of the four projects: the one with clients, a commercial surface, and a talk track, and the one whose practitioner logic turns out to be foundational for the other two. When a governance failure is a TCF failure, that is not a metaphor.

The Resonance Architecture is a cross-domain synthesis that argues for structural isomorphism between content, matter, and consciousness across the same six organizational tiers first mapped in my Tiered Content Framework, and now extended into a much larger theoretical claim. It is the most speculative of my projects, and the one that, if it holds, reframes all the others.

Current status: intellectually rigorous as a working spec, not yet ready for formal research or peer review. The framework has begun doing argumentative work in adjacent projects under a more constrained sense of "resonance," the recognition of a participant across contexts and over time, requiring a foundation on which prior recognition can compound, and that operationalization is itself an early form of testing. The full cross-domain claim still requires one independent collaborator and at least one operationalized prediction before it reaches funding-grade.

It develops here because the argument needs to be tested in public before it can be tested anywhere else.

The same organizational logic that I found in content strategy and systems thinking may run all the way through matter, mind, and everything in between. Same structure, six tiers. That's a testable claim. We're the first generation with the computational and cognitive tools to find out whether it holds. That's what this is.

Three versions are in development: a public research essay, a researcher circulation spec, and a frontier companion document for the more speculative material. Each addresses a different audience. None are ready yet.

The Grammar of Trust is a book in development. The argument: Latin held Western European institutional life together not as a literary tradition but as a governance technology—the case system made contracts portable across jurisdictions, the root system built verification into every formal utterance, the honorific system signaled who could bind whom. That structure is the grammar of trust. The book traces it forward through the press, the contract, the journal, and into the present, where the surface of institutional language can now be produced without the practice it was designed to make legible. A book proposal is in progress. An open call for a visual artist to develop the book's design language is active here.

Full Personhood is the governance argument, framework, and manifesto that the four projects converge on. The essay traces the 140-year structural asymmetry, corporations have had durable personhood since Santa Clara (1886); persons, communities, future generations, and the living world have not, and argues that the Seam Stack provides a governance model, not just an architectural one, for systems where the substrate belongs to the participant and the platform facilitates and exits. Appendix A states fifteen governance principles across four domains: individual personhood, collective personhood, architectural enforcement, and intergenerational and ecological standing. Current status: the working draft, Full Personhood: The Governance Model AI Requires and Capitalism Never Built, v0.1, is here.

The projects are not peers in their function. End of History, Revisited tracks the problem. The Tiered Content Framework provides the operational vocabulary for diagnosing it. The Resonance Architecture asks whether the vocabulary points to something deeper. The Grammar of Trust asks how the infrastructure of trust was built in the first place, and what it means that it can now be mimicked without being inhabited. Full Personhood says what they're all for. The Sherlockverse reads the cultural signal.

The Sherlockverse tracks the most continuously adapted fictional mind in the modern era as a cultural antenna: when Holmes resurfaces at scale, as he has in 2026, it is worth treating that as a signal rather than a coincidence. This cultural intelligence project runs cultural criticism, adaptation analysis, lore development, and original concept work in parallel, using the Holmes multiverse as the instrument through which to read the present-day events. It sits alongside End of History, Revisited, and the Resonance Architecture as a companion application of the Cultural Antenna concept, extending that project's analytical framework into the cultural domain. Published at sherlockverse.com, with selected work cross-posted here.

The software practice

Five local-first web applications have been built as part of this work: four have been published, and one is in build. Local-first is not a technical preference here; it is a political one. Systems that monitor governance, process sensitive intake data, maintain their users' social graphs, or hold records of someone's working life should not depend on servers that their builders control.

The AI Governance Window Tracker is the monitoring instrument for End of History, Revisited, a structured five-domain signal assessment of whether binding democratic AI governance is becoming enforceable faster than AI embedding in critical infrastructure makes binding governance irrelevant. Built on React, TypeScript, and Y.js with IndexedDB persistence.

The checkout-seam prototype documents the deliberate network boundary design problem in local-first commerce: how a system that keeps all state local handles the one operation—payment—that requires an external endpoint. The prototype is live, and the codebase is public. A Pattern Commons entry documents the seam as a reusable architectural pattern.

The fhir-seam prototype applies the same seam pattern to healthcare intake — writing a FHIR bundle to a clinical endpoint you don't control, with local data preservation on failure. Chosen because the stakes are highest and the design argument is clearest when failure has clinical consequences.

Local-First Social (localfirst.social) is the hardest version of the problem at the social layer: a social network where every piece of user data lives in IndexedDB, and a minimal WebSocket relay facilitates peer connection, then exits the path. The relay stores nothing. The platform owns no relationships. The social graph is the user's.

A fifth prototype, the keyhive-employment-seam, is in build, the first reference implementation of the employment seam pattern, with cryptographic enforcement (Keyhive) layered over an Automerge substrate. The architectural inversion: the worker owns the knowledge graph, and the platform facilitates the handoff and exits. The relay is cryptographically prevented from reading bundle contents.

Building the first four prototypes in nine days using AI-assisted development produced a governance framework for that practice, AI-Assisted Development: Methodology Failures, Learnings, and Governance Framework, which documents what breaks when the AI fills specification gaps, and what disciplines prevent it. It is published alongside the prototypes.

A further artifact emerged through the software practice rather than alongside it: the Seam Stack, a four-layer architectural composition, substrate, governance, boundary, evidence, that names what local-first systems converge on when boundary events carry legal weight:

1. Substrate (Solid; or Automerge with Keyhive in the keyhive-employment-seam prototype).
2. Governance (the Tiered Content Framework—six tiers, three cross-cutting governance dimensions, with explicit posture for AI-generated content).
3. Boundary (the Pattern Commons seam discipline).
4. Evidence (W3C Verifiable Credentials Data Model 2.0, RFC 3161 timestamping from two independent authorities, OpenTimestamps for long-term tamper-evidence, bilateral cryptographic signatures, and Bitstring Status List revocation).

None of the layers is novel; the synthesis claim is.

The worked example is the employment seam, the case where the legal substrate is part of the architecture rather than a wrapper around it, and where the architectural inversion (the worker owns the knowledge graph; the platform facilitates the handoff and exits) becomes load-bearing.

A fifteen-principle governance model travels with it, scoping individual, collective, intergenerational, and ecological personhood. seamstack.org redirects here; a standalone charter is in preparation.

Talk proposals for the prototypes and the Seam Stack synthesis have been submitted to Local-First Conf 2026 (Berlin, July 12–14).

The document architecture

Each project develops through a set of working documents—versioned, publicly linked, and updated as the argument changes. The documents are not drafts awaiting completion. They are the work. The version numbers are part of the record.

The End of History project currently includes a core theoretical essay, a policy framework that translates the governance window argument into binding intervention mechanisms, a legibility project for practitioners, an AI Governance Window Tracker that provides structured periodic assessment across five monitoring domains, and an Agentic Accountability Playbook that specifies accountability architecture for agentic AI systems. A companion architecture document maps their relationships.

Document links are embedded in relevant project posts as they are published.

AI disclosure

This publication is developed in part with AI assistance, a method disclosed here because the work demands it. The analytical direction, framework choices, and editorial judgment are human-authored. Claude (Anthropic) functions as a structured thinking partner, interrogated, redirected, and contested throughout. Publishing AI-collaborative work without disclosing it would be a performative contradiction of this publication's own argument about epistemic infrastructure and illegibility.

Published by UX Minds, LLC · Phoenixville · jediwright.com

Access all areas

By signing up, you'll get access to the full archive of everything that's been published before and everything that's still to come. Your very own private library.

Fresh content, delivered

Stay up to date with new content sent straight to your inbox! No more worrying about whether you missed something because of a pesky algorithm or news feed.

Meet people like you

Join a community of other subscribers who share the same interests.

Start your own thing

Enjoying the experience? Get started for free and set up your very own subscription business using Ghost, the same platform that powers this website