NO NEW ACTOR TO NAME: Erman and Furendal's Market and Epistemic Authority, the Frontier AI Labs Now Funding Longevity Research, and Why the Dubai Longevity Authority Is Where the Question Stopped Being Speculative

Erman and Furendal's tripartite typology of non-state authority — moral, epistemic, market — and their distinction between democratic agents and agents of democracy, developed across three papers addressed specifically to global AI governance, was built to characterize a named set of actors: the frontier AI laboratories whose soft-law guidelines, safety commitments, and market position shape AI governance without any electorate having delegated them the power to do so. This paper's finding is narrower, and for that reason sharper, than either of this corpus's prior extensions of their work. Not a Divisible Good asks whether Erman and Furendal's legitimacy chain transfers structurally to a different kind of good, AI-driven longevity allocation. PRE-PRICED's own extension applies their authority typology to the market data a longevity-clinic sector generates. Neither paper needed to ask, because neither had occasion to, what this paper takes as its subject: whether the specific actors Erman and Furendal's framework was built to describe — OpenAI, Anthropic, Google, and the wider frontier-AI industry DeepSeek is conventionally grouped with — are the same actors now assembling market and epistemic authority in longevity research specifically, by name, and not merely by structural resemblance. They are. Sam Altman's personal $180 million founding investment in Retro Biosciences, since grown into a company partnered directly with OpenAI on AI-accelerated cellular reprogramming; Anthropic's April 2026 acquisition of Coefficient Bio and its June 30, 2026 launch of Claude Science, an internal drug-discovery program explicitly aimed at the biology of neglected and age-related disease; and Alphabet's now-decade-long Calico Life Sciences alongside DeepMind's Isomorphic Labs, are not new entrants recruiting new authority for a new domain. They are Erman and Furendal's own named non-state actors, already holding the market and epistemic authority their framework describes, extending an existing accumulation into a second good. DeepSeek is the documented exception: this paper finds no comparable first-party longevity program, only third-party use of DeepSeek's open-weight models for biomedical text tasks — an asymmetry the paper treats as evidence against flattening "the AI labs" into one undifferentiated actor, not as a gap in the argument's core claim. This enhanced edition adds two further findings to the original argument. First, applying Erman and Furendal's own race-to-the-bottom mechanism for national AI regulation to longevity governance specifically, the paper documents that Dubai's design is not the first instance of jurisdictional competition for this class of capital: Próspera ZEDE, a privately governed special economic zone in Honduras operating since 2017, has hosted longevity-specific gene-therapy biotechs since 2021, competing on regulatory permissiveness where Dubai competes on sovereign certainty — two distinct strategies converging on the same capital pool, with Bryan Johnson's own documented visit to Próspera linking this corpus's existing elite-access anecdote directly to the regulatory-arbitrage tier. Second, drawing on Erman and Furendal's distinction between outcome-focused and process-focused evaluation of governance arrangements, the paper sharpens why no scale of market or epistemic authority converts into delegated public power — not because such authority produces bad outcomes, but because the question of who decided is constitutively different from the question of whether the decision was good.