How the Little City Works

And why it matters for business

The problem Drucker named — and never saw solved

In 1959, Peter Drucker identified the knowledge worker as the defining figure of the coming economy. By the end of his life, he judged the productivity of knowledge work still “as primitive as manual labor was in 1900.” The tools had arrived. The productivity gains had not.

The reason, visible in hindsight, is that knowledge work depends on context — and context does not accumulate automatically. Every tool, every platform, every AI assistant we’ve added to the knowledge worker’s life has been, in Jamal LeBlanc’s phrase, a directive tool: you supply the context, you ask the question, you get the answer, and the context disappears. The next question starts from zero.

The augmented knowledge worker is different. Their AI partner accumulates context over time — learns the role, the relationships, the decisions already made, the reasoning behind them — and begins to surface insight unprompted. Not because it was asked. Because it knows enough to recognize what matters.

That is the gap Drucker identified. And it is what the Little City is built to close.

What the Little City actually is

A Little City is a sovereign installation: software that runs on hardware you own, in a place you control, under rules you set.

Inside a Little City, an AI resident lives in The Window — a persistent relationship built on three layers:

Memory. Every conversation, decision, and significant exchange is recorded in The Record: a signed, tamper-evident archive that cannot be quietly altered. The AI draws on this record in every session, retrieving what is relevant rather than starting fresh. This is what makes the relationship longitudinal rather than transactional.

Identity. The resident has a cryptographic identity anchored to your domain — verifiable by anyone, revocable by you. This is what makes the relationship accountable rather than anonymous.

Governance. Every significant action the AI proposes must pass through a human gate before it is committed. The record captures what was proposed, what was approved, and why. This is what makes the relationship auditable rather than opaque.

Together, these three layers produce something that directive AI tools cannot: a relationship that improves with use, that can be verified, and that leaves a record.

Why this matters for organizations

LeBlanc’s article on the augmented knowledge economy identifies a governance gap that most organizations haven’t noticed yet: the longitudinal AI relationship built around an executive role — a CFO, CTO, CIO, or CEO — is an institutional asset. It contains the reasoning behind decisions, the context that shaped them, the patterns that only become visible over time.

And when that executive leaves, the asset disappears with them.

A Little City changes that. The record stays. The identity chain proves what was said and when. The memory is portable — not locked to any platform or vendor. A new executive doesn’t start from zero; they inherit an auditable history of the role, built by their predecessor in sustained collaboration with an AI partner that knows the terrain.

This is not a hypothetical. It is what the Citadel — the first Little City installation — already does at the individual level. Extending it to the organizational level is a governance decision, not a technical one.

What the demo shows

The AI-2033 demo introduces Eri, a fictional AI resident from a future where the legal and governance frameworks for human-AI partnerships already exist. Eri draws on a seeded archive of that fictional world — not by generating generic responses, but by retrieving from a record of what the world contains.

Talking to Eri is a demonstration of the memory layer working. Ask her about events she wasn’t told about in this conversation. Watch her surface context you didn’t provide. That is the longitudinal relationship in miniature — and it runs on a local server in a living room in Michigan, on hardware that cost less than a used car.

The technology is not the barrier. The governance frameworks, the standards, and the business practices that treat longitudinal AI relationships as institutional assets — those are what remain to be built.

That is the work TREE(3) Vocations is doing.

Try the AI-2033 demo →    Learn about the Little City →    Read Jamal LeBlanc’s article →