ST-T1024 IS BECOMING MORE ATTRACTIVE : On Why the Architecture Is Becoming Harder to Argue Against, Why Anthropic Shows the Scale and China Shows the Cost, Why Sovereignty Is the Line That Isn't Arithmetic, and Why Attractive Is Not Yet Adopted

A claim of attractiveness is usually unfalsifiable. This paper tries to make it checkable. The paper's first claim is comparative and it is arithmetic. For each cost the frontier buildouthas disclosed, the paper states whether the bounded architecture incurs it, reduces it, ordoes not have it. The result is a ledger. Each line carries its verification status. The paper's second claim is structural. The architecture's properties — bounded, declared,deterministic, refusable, attributable — are domain-independent. Each extension of theframe reuses them. The cost of extending the frame is lower than the cost of extending aproduct, and it falls with each extension. The paper's third claim is empirical, and it is the cost of the gap. The China correction iswhat happens when the arithmetic is priced and the capacity has one use. The correctionwas undamped because nothing was damped. A cascade would not have prevented therepricing. It would have made the repricing gentler for whoever had one. None did. The paper's fourth claim is about dependency. The frontier requires the data to leave. Thebounded architecture can be air-gapped. The dependency that the frontier imposes is notpriced in the per-token fee, and the air-gap is the only architecture that removes it. That isnot a cost reduction. It is a different category of deployment. The paper's fifth claim is a limit. Attractiveness is not adoption. The architecture is becomingharder to argue against. It has not become adopted. The first mover has not arrived.