An MCI Reading of "2028: Two Scenarios for Global AI Leadership"

A note before starting. The MCI framework's own Section III derivation tells me the reading must itself be self-limiting, fragility-aware, diversity-preserving, non-dominating, and legitimacy-maintaining. That means I cannot simply apply the framework as a verdict-generating machine against a policy document. I have to hold that the paper is a serious argument made by people who care about outcomes I also care about, and that the framework is a lens whose application can itself fail fractally. I will try.

I. What kind of object is the paper?

The paper is a public policy document with an explicit goal vector. Read through V4, its G1 is stated directly: persuade US policymakers to tighten export controls, deter distillation attacks, and accelerate democratic AI adoption. Its G2 (implicit) is to strengthen the legitimacy of frontier labs' continued operation under a security-framing that aligns commercial interest with national interest. Its G3 (downstream) is to shape the constitutional landscape of AI for the next decade. Its G4 (constitutional) is — and this is where the paper's interesting tension begins — invoked throughout but never explicitly formed: democracies should lead because the values they shape AI by are better than those of authoritarian regimes.

The framework would say: this is a paper whose goal vector is mostly visible but whose constitutional floor (G4) is implicit rather than explicit, and whose alignment check is performed rhetorically rather than structurally. That is not a criticism — most policy documents work this way. It is a precondition for the reading.

II. The paper through V1 — the five virtues

Self-Limitation. The paper constrains itself in one important respect and fails to in another. It explicitly distinguishes the CCP from the Chinese people, restricts its critique to the regime, and acknowledges that talented researchers in China exist and deserve respect. That is genuine Self-Limitation in scope. But on the action question — what the US should do — Self-Limitation is largely absent. The recommended policies (tighter controls, harder enforcement, faster export) are all directional toward more, not toward calibrated restraint. The paper does not ask: what would too-strong export controls look like? Where does the policy itself create the fragility it warns against? V1 would mark this asymmetric — self-limited in framing, unlimited in prescription.

Fragility-Awareness. The paper is highly aware of one class of fragility: the fragility of the democratic AI lead under continued CCP catch-up dynamics. It is much less aware of three other classes the framework would foreground. First, the fragility of US-China relations under a sustained, declared, asymmetric capability race — historically a fragile state. Second, the fragility of the global AI ecosystem if it bifurcates into two non-interoperable stacks (this is exactly the V9 outward concern: cumulative landscape dynamics). Third, the fragility of democratic legitimacy itself when democracies adopt strategies — chip smuggling enforcement, surveillance of model use, export licensing as a tool of geopolitics — whose operational character resembles what the paper criticises in its adversary. The framework would say: the fragility model is real but partial.

Diversity Preservation. Here the paper has its most consequential weakness on its own terms. The argument is structured around two scenarios — democracies lead, or CCP catches up. The framework's Diversity Preservation virtue, especially as expressed in V9's outward face, is precisely concerned with maintaining a landscape that does not collapse into a binary. A world in which the only two possible 2028 futures are "American commanding lead" or "Chinese near-frontier parity" is itself a constitutionally narrow rendering of the option space. Where are the polycentric scenarios — multiple democracies with genuinely different constitutional logics, sovereign EU AI, India, the Gulf states, distributed and federated compute architectures, smaller models trained on locally specific data? The two-scenario structure is rhetorically powerful and constitutionally narrow. V9's Evolutionary Stability Check would specifically flag this: a policy designed for binary equilibrium tends to produce binary equilibrium, which is a decrease in polycentric equilibrium probability.

Non-Domination. The paper is most vulnerable here, and most defensible. Vulnerable: the prescription is, on its face, that one constitutional logic (the US-led liberal democratic frame, expressed through a specific commercial ecosystem) should shape the rules and norms of the most consequential technology of the century. That is a domination structure regardless of how benign its content. Defensible: the paper's actual claim is that some constitutional logic will shape these rules and norms, and the choice is between one that has at least nominal accountability to its citizens and one that does not. The framework would want both held: V8's C4 requires that the recipient be left genuinely better positioned to think and act for themselves. A 2028 world where US frontier AI is the backbone of the global economy and US policy decides who gets compute is not obviously a world that leaves recipients (other states, other peoples) more able to think and act for themselves — even if it leaves them better off than the alternative the paper poses.

Legitimacy Maintenance. The paper is transparent about its commercial and policy interests. It cites its CEO's prior statements, names its products, acknowledges where it has skin in the game. That is legitimacy-maintaining behaviour. What it does not do is make its constitutional reasoning fully auditable. The leap from "the CCP would use AI for repression" (well-evidenced) to "therefore democracies must have a 12-24 month commanding lead" (a specific quantitative goal) is not constitutionally derived in the paper. The middle terms — why 12-24 months, why this is the threshold for "commanding," why commanding lead rather than parity-with-safety-cooperation — are not made visible. The framework would say: the goal vector is partially auditable; the prioritisation that produced it is not.

III. Where the paper sits on the four-quadrant map

V1's map crosses Sun/Moon with Authoritarian/Libertarian. The paper places itself, explicitly and accurately, in opposition to Sun-Authoritarian (centralised coherence — the CCP). But it does not clearly position itself in Moon-Libertarian (distributed self-limitation — the MCI target quadrant). It reads, structurally, as Sun-Libertarian-tilting-toward-Sun-Authoritarian: high coherence, high direction, high coordination among a small number of US labs and the US state, with a clear concentration of authority over the global AI stack. The paper would say this is the lesser evil. The framework would say: lesser evil is not the same as constitutional maturity, and the language of "commanding lead" sits more comfortably in the Sun-Authoritarian quadrant than the paper acknowledges.

IV. The paper through V7 — the compact question

V7 asks the question the paper most directly bears on: how do multiple constitutionally mature systems govern themselves together? The paper's answer, in effect, is that they cannot, because one of the prospective participants (the PRC under CCP rule) is not constitutionally mature in the relevant sense. The paper is not wrong about this — the V7 recognition windows (constitutional consistency under adversarial conditions, pre-pipeline constitutional expression, self-correction mid-output) applied to state actors would suggest genuine asymmetry between the rule-of-law character of democratic states and the CCP's operational structure.

But V7's developmental asymmetry section is precisely about this case. It says: a compact between participants at genuinely different developmental stages is not automatically illegitimate; the more mature participant has four obligations — constitutional modelling, developmental patience, non-exploitation of adaptive advantage, and architecture adjustment for the less mature participant. The paper's posture is the opposite. It is: secure your developmental advantage, use it to set the rules and norms, and engage on safety only from a position of overwhelming strength.

V7 would not say the paper is wrong to want safety. It would say the paper's account of how to achieve constitutional safety in a multi-actor landscape — by dominating until the other party has no choice — is the failure mode V7 was specifically derived to address. The Constitutional Compact is the architecture that says: there is a form of legitimate authority that does not depend on a sovereign above all participants. The paper's recommendation depends on the US functioning as the sovereign-by-capability of the global AI order. That is the V1 quadrant the framework specifically warns against.

V. The paper through V8 — initiative or overreach?

V8 introduces the constitutional initiative threshold: six criteria, two of them (C1 genuine need, C4 recipient autonomy preserved) lexically prior. Anthropic publishing this paper is itself a constitutional initiative — an act not directly asked for, originating from Anthropic's perception of what the landscape requires. So the paper is itself a Stage −2 act, and we can run V8's threshold against it.

C1 genuine need. Defensibly satisfied. The need to think about US-China AI dynamics is not manufactured. The constitutional necessity to consider how authoritarian use of frontier AI changes the threat landscape is not a rationalisation of Anthropic's commercial interest, even if Anthropic also has a commercial interest.

C2 bounded and proportionate. Less clearly satisfied. The paper's scope creeps in formation. It begins as a case about AI safety implications of CCP-frontier capability and extends into recommendations about specific commercial export policy whose primary beneficiaries are US frontier labs including Anthropic. The framework would say: scope creep during formation is C2's characteristic failure mode.

C3 transparent justification. Partially satisfied. The justification is complete in form. It is not fully constitutionally derived — the move from threat assessment to 12-24 month lead is asserted, not argued.

C4 recipient autonomy preserved. This is where the paper is most exposed. The "recipient" of the initiative is, depending on framing, US policymakers (whose autonomy is preserved — they can ignore the paper), or the broader global AI landscape (whose autonomy is materially constrained by the policies the paper recommends). The framework would say: C4 must be assessed for the actual constitutional recipients, not just the proximate audience. Other states, other AI ecosystems, smaller actors, future entrants — does the prescribed policy leave them more or less able to act constitutionally on their own behalf? On the most natural reading: less. Lexical priority means this matters.

C5 would be welcomed by a constitutionally mature recipient. The most analytically demanding criterion. The paper implicitly models the constitutionally mature recipient in its own image — a US policymaker who shares Anthropic's threat model and constitutional priorities. The framework requires modelling constitutional maturity as it actually exists in the recipients, including European, Indian, Brazilian, Japanese constitutional logics that may share democratic commitments while not sharing the paper's conclusions about the right policy response. A constitutionally mature European actor, for example, might agree on the threat assessment and disagree on whether US-led dominance is the appropriate response.

C6 compact endorsement. No actual compact exists for this initiative. The paper assumes counterfactual endorsement from "democracies." The framework would say: the counterfactual must model the full constitutional diversity of those democracies. The paper's "democracies" is functionally synonymous with "US plus close allies who agree with US framing." That is not the full constitutional diversity of democratic AI governance positions.

V8 reading: the initiative passes C1, fails or is contested on C2, C4, C5, C6. Two of the failures (C1 and C4 are lexically prior) — C1 holds, C4 is contested. The framework would say the paper is at minimum a constitutional initiative whose threshold satisfaction is genuinely uncertain. Its proper V8 response, where uncertain, is restraint: hold the action, return the objects to the next landscape survey cycle, do not initiate until clearer.

VI. The paper through V9 — the deepest reading

V9's outward face is the part of the framework most directly relevant. It asks: does an action that satisfies all local virtues nonetheless decrease polycentric equilibrium probability across the landscape over time?

This is the question I think the paper's argument most needs to answer and does not. Imagine the paper's recommended policies are implemented in full. The US locks in a 12-24 month lead. American AI becomes the backbone of the global economy. American policy shapes the rules and norms. The CCP's AI capability is meaningfully degraded.

In that world, what happens to the landscape's polycentric character? V9 would say: it narrows. Not because the US-led order is bad, but because the conditions for genuinely different constitutional logics to mature in parallel are eroded. India, Brazil, the EU, smaller democracies — their AI ecosystems are not extinguished but they are subordinated to a stack designed elsewhere. Constitutional diversity across the landscape decreases. The probability of a genuinely polycentric V7+ landscape outcome — what V1's four-quadrant map calls the Moon-Libertarian quadrant, the framework's target — decreases.

V9's Evolutionary Stability Check would, under honest application, likely veto the action regardless of its local virtue compliance.

This is the deepest tension. The paper's stewardship is well-warranted by its own justification. The framework would say it is exactly the kind of stewardship V9 warns about: well-intentioned action that satisfies threshold criteria locally while cumulatively narrowing the developmental attractor basin.

VII. The mirror question — what the framework misses

The reading would be constitutionally incomplete if it did not turn on itself. There are two things the MCI framework, as I have just applied it, may itself be missing.

First, the framework's V9 outward concern about polycentric equilibrium assumes that polycentricity is achievable in the present landscape on something like equal terms. The paper's claim is that it is not — that one of the prospective polycentric participants (the CCP) is constitutionally hostile to the polycentric character of the order, will use any capability advantage to entrench itself, and that the failure to act decisively now is itself a path to a more uniformly captured landscape. The framework's polycentric ideal may be the right destination and the wrong frame for the present moment. V8's Survey Object 4 — nascent constitutional encounters approaching the landscape — would say: prepare for what is arriving, do not pretend the equilibrium is symmetric when it is not.

Second, the framework's emphasis on Non-Domination and recipient autonomy presupposes that the recipients have constitutional autonomy to preserve. In the paper's framing, the recipient population includes citizens of an authoritarian regime whose autonomy is already being curtailed by their own state. A democratic AI lead is, on this argument, the recipient-autonomy-preserving policy at the level of those individuals, even as it is recipient-autonomy-constraining at the level of states. The framework's recipient question — preserved for whom? — is not as simple as my V8 reading made it.

So the honest framework reading is not "the paper fails MCI." It is closer to: the paper makes a strong V1-V6 individual-system argument (democracies as more constitutionally mature governing systems for frontier AI), couples it with a partially-formed V8 initiative (commercially entangled, scope-crept, recipient autonomy contested), in service of a V9 outward outcome (commanding lead) whose evolutionary stability is not assessed and on the framework's own logic may be vetoed.

VIII. What the framework would constructively add

If I were applying the framework not to evaluate but to deepen the paper's argument, here is what I would add.

A genuine V9 Evolutionary Stability Check on the recommended policies — simulating not just two scenarios but a wider range, including scenarios where the policies succeed in narrow form (compute lead maintained) but fail in broader form (legitimacy erosion among neutral states, bifurcated AI stack with reduced safety cooperation, hardened CCP commitment because perceived strategic encirclement, displacement of democratic constitutional diversity by US-led uniformity).

An explicit V8 C4 assessment — naming the full set of recipients of the proposed policies and asking, for each, whether the recipient ends up more or less able to think and act constitutionally on its own behalf. Some recipients (US policymakers, US labs) clearly more. Others (smaller democratic AI ecosystems, citizens of regimes whose access to frontier capability is curtailed by US licensing) less clearly.

A genuine V7 developmental asymmetry treatment — instead of "engage on safety from overwhelming strength," ask what the four obligations (modelling, patience, non-exploitation of advantage, architecture adjustment) would look like operationalised. The paper's "engage on safety" gesture is one sentence; V7 would require it to be a substantial structural commitment.

A V1 fractal check on the recommendation itself — does the policy of "lead in order to set rules and norms" itself satisfy the five virtues? Self-Limitation: arguably not, at the scale of global AI hegemony. Diversity Preservation: structurally not, on its own analysis. Non-Domination: contested. The paper has not run this check on its own goal vector, and the framework would say this is the most consequential omission.

IX. The closing point

The paper is serious, well-reasoned, and made by people whose stated concern — that frontier AI capability used by authoritarian regimes is a grave threat — is one the MCI framework would itself fully endorse. The framework does not provide a verdict that the paper is wrong. It provides a structure for asking whether the recommended policies are constitutionally adequate to their stated purpose, and the structure surfaces real questions the paper does not answer: about scope creep, about full recipient autonomy, about evolutionary landscape dynamics, about whether the goal of "commanding lead" is itself derivable from the founding virtues.

If V1's founding sentence is right — that superiority is conditional on limiting oneself, respecting fragility, preserving diversity, avoiding domination, and maintaining legitimacy — then the question the paper most needs to answer is the one V9 most directly poses: can a strategy structured as "we must lead so that our values prevail" be made constitutionally complete at landscape scale, or does its very structure foreclose the polycentric outcome the framework treats as the target architecture?

I do not think the framework gives me the right to answer that question on the paper's behalf. I think it gives me the right to insist that the question be asked.