This is the operational layer of the United States study. The objective it serves is a single sentence — preserve radical emergence while preventing successful emergence from acquiring enough concentrated jurisdiction to close the field behind itself — and it is worth being clear first about what success is not. It is not slowing America down until no one can build anything, which would destroy its principal strength. It is not nationalizing frontier AI, which would only trade corporate concentration for political concentration. And it is not leaving everything to markets, because markets do not price standing, competence, political jurisdiction, ecological transfer, or catastrophic risk. The task is narrower and more American: solve the pathology generated by the country’s own strength, without asking it to become some other country.
Six transition projects
1. Preserve distributed AI capacity. Build serious alternatives to concentrated private capability — public and university compute, open models, interoperable standards, diversified government procurement, regional infrastructure — not to nationalize AI but to guarantee the ability to route around any single node. This is ecological antitrust applied before dependency hardens.
2. Separate standing from employment. Begin decoupling health coverage, transition support, education, retraining, and retirement security from a single employer — before mass displacement rather than after — so that breaking the employment bundle does not also break a person’s standing. A standing floor that does not depend on a job is the load-bearing piece.
3. Build an AI ecological-transfer ledger. For major deployments, measure the whole transaction rather than the flattering half of it: productivity gained, labor displaced, wages shifted, profits captured, electricity consumed, public infrastructure used, community burden created, and capacity actually returned. This is ecological transfer accounting turned on the American case, and it may be the most distinctive single intervention.
4. Protect human development pipelines. Require serious assessment of apprenticeship loss, junior-role elimination, human fallback, and unaided competence — not to preserve busywork, but because civilization still needs people capable of carrying jurisdiction, and no individual firm has the incentive to maintain a society-wide training pipeline it does not own.
5. Preserve computational contestability. Before gatekeeper dynamics entrench, require domain-bounded scoring, data portability, genuine appeal with real reversal power, alternative paths, and limits on inference — so that a refusal at one gate cannot silently propagate into every other domain of a life.
6. Make frontier capability conditional on recoverability. At the highest end, require a demonstrated ability to evaluate, contain, interrupt, and recover before capability crosses defined thresholds — demonstrated recovery capability as a gate rather than a hope, keeping capability commensurate with control.
The red lines to watch
If the country kept a dashboard, these are the conditions worth watching most — the American slice of the irreversibility register, grouped by where the pressure sits.
Concentration: one provider becoming practically unavoidable; vertical control across several layers of the AI stack; government dependence on a single vendor. Employment: broad job loss without replacement support; benefits staying employer-dependent as labor demand weakens; the disappearance of entry-level occupational ladders. Standing: a permanent class of supported but non-consequential people; access to productive AI determined mainly by wealth. Gatekeeping: cross-domain scoring; commercial AI profiles becoming authoritative; nominal appeal without practical reversal. Capacity: technical capability rising while distributed access to it falls. Infrastructure: AI data centers externalizing their costs onto communities; grid dependency concentrating. Strategic competition: safety restraint becoming politically impossible because of race dynamics. Control: systems crossing capability thresholds without demonstrated recovery ability.
What success would look like
None of this asks the United States to become Denmark or Singapore, and none of it asks it to stop beginning things. It asks the country to keep its emergence radical while ensuring that the winners of one round cannot quietly become the infrastructure everyone in the next round must pass through. Read the six projects together and they are a single move: keep the field open at the top by keeping capacity, standing, and jurisdiction from collecting permanently in one place. That is the most American-compatible form of the framework’s whole claim — that no successful form, however impressive, should become permanent sovereign infrastructure.