The thesis

The agent economy needs rails, not a landlord.

Forty years of platform monopolies were a tax we paid for coordination we could not do ourselves. AI changes the math. This is the thesis in full.

Premise

Two problems, one century.

Every major platform business of the last four decades exists because coordination between strangers is expensive. They solve Finding and they solve Trust, and they charge a toll that scales with the volume of commerce flowing through them.

This was a fair trade when the alternative was worse. It is no longer. Language models that can browse, verify, negotiate, and settle make the intermediary function itself commoditizable — if, and only if, the primitives underneath are open.

Observation

The capture is already priced in.

Incumbents see the same opening you do. They are bidding on identity, acquiring wallets, vertically integrating hardware, and training models on private data they will never give back. The public roadmap looks like "AI assistants." The real roadmap is to own the four pieces — identity, payments, interface, data — before anyone else can assemble a competing stack.

Once those four converge inside a single company, the agent economy becomes a walled garden more total than mobile ever was. There will be no regulatory fix that arrives in time.

Claim

The fix is structural, not political.

The way out is not a law or a campaign. It is an open substrate shipped before the walls are poured — identity, wallet, directory, and protocol — held by an institution with a charter that forecloses the path to capture.

Forest is that substrate. The next section shows what it is, in concrete terms.