thesis.

Library Theses Research Arguments Field studies Institute People Convenings

The shelf. Seven volumes, one per problem the institute intends to solve.

Three carry an opening argument and a confidence mark. Four are published as the agenda so the framing can be argued with before the work starts. Pull one out to read it.

In focus

003 · Management Is Becoming an Operating System

Numbered positions on how work is changing. Each one sourced, challenged, and reviewed on a published date.

The register is the institute's memory. A thesis can be revised or retired, but it is never deleted; the previous claim stays on the record with what changed and why.

Filter

All · Argued · On the agenda

Original studies that test a thesis against evidence, with methods and limitations published in full.

Research is heavier than commentary. Every study names the thesis it supports or challenges, the sample, and what it does not establish.

Latest

Research 05 · The Agent Ownership Study

Essays, interviews, responses and field notes. Every one a claim someone can disagree with.

Lighter than research, still tied to the library. An argument that moves a thesis is linked to the revision it caused.

This week

Augment, don't replace · Jeff Wald

Theses applied inside real organizations, for two to four quarters, with the outcome reported against a written baseline.

This is where a position earns or loses confidence. Every field study ships with its limitations. Organizations are described, not named, until they consent.

Latest

Moving people, not skills · a 9,000-person manufacturer

How the work moves from signal to impact, and the standards that hold it to account.

Seven stages, from noticing a pattern to measuring what changed in the world. Plus evidence, counterargument, revision and independence standards, published in full.

Fellows and contributors, listed by what they have authored, challenged, and revised.

A profile here is a record of intellectual contribution, not a biography. Challengers are credited alongside authors.

Working sessions built to test a live thesis, with the question published before and the synthesis after.

Small rooms, real disagreement, and a written record of which positions moved.

Next

What is management for when coordination becomes software? · 08 October

Newsletter Search Standards Independence Contact

Response/Coordination · AI systems

The case against coordination.

A standing objection to Thesis 003, published in full because the institute's rules require it.

Rachel Okafor28 Mar 20267 min read

Response28 Mar 2026Coordination · AI systems

Thesis 003 argues that as execution is automated, coordination quality compounds. I think the opposite is at least as likely, and the institute's standards require the strongest version of that objection to sit next to the thesis. This is it.

Coordination exists because work is divided. Handoffs, decision rights, information flow: all of it is the cost of splitting a process among people who cannot hold it whole. The premise of agentic software is that the process no longer needs to be split. If an agent can carry a workflow end to end, the handoffs that coordination manages do not become more valuable. They disappear.

The thesis answers that the coordination work moves to the interfaces among people, agents and exceptions. Perhaps. But that is a claim about a transition period, and the thesis is written as if it were a claim about a destination. Organizations investing heavily in coordination quality today may be polishing a layer that the next two years delete.

Organizations investing heavily in coordination quality today may be polishing a layer that the next two years delete.

What would settle it: a panel of organizations at high agent coverage, scored on coordination quality, watched for whether the advantage holds. The Decision Rights Index is meant to be that panel. Until it reports, I am keeping this objection open.

Moves the library. Raised the unresolved counterargument on Thesis 003 v1.1. Thesis 003, Management Is Becoming an Operating System.

Respond to this piece All arguments