← narwal.one/Second Brain
SecondBrain
Ask the Brain
Index/Sourceupdated Sat Jul 04 2026 08:00:00 GMT+0800 (Philippine Standard Time)

Donald Trumps AI Regime Is Opaque Unpredictable and Unsustainable (Economist)

economistbusinessairegulationhierarchy-of-access

Donald Trump's AI Regime Is Opaque, Unpredictable and Unsustainable (Economist)

Section: Business · Edition: 2026-07-04 Edition · The Economist

The Business-side anchor for the 07-04 AI cluster. Reports how, in a matter of weeks, US federal AI policy has gone "from implausibly libertarian to increasingly draconian and opaque" (Dean Ball, former Trump AI adviser). Documents the June 26 — June 30 window of concurrent restrictions on OpenAI's Sol and eased-then-lifted controls on Anthropic's Mythos, and the fact that the White House's own June 2 executive order explicitly said no "mandatory governmental licensing, preclearance or permitting requirement" would exist — while in practice, that is exactly what now exists.

Pairs with America Should Not Imprison Frontier AI (Economist) (the Leader).

The June 26 – 30 window

  • Jun 26: OpenAI announced Sol would be restricted to "a handful of trusted partners." The same day, Commerce eased export controls on Mythos.
  • Jun 30: Commerce lifted the Mythos controls entirely (Anthropic tweaked safety protections).
  • During the window: Anthropic could share Mythos with only ~100 American firms and institutions, plus their foreign-national employees.
  • Fable 5 (public tier) remains heavily constrained.
  • Sam Altman was called by Howard Lutnick (Commerce Secretary) warning against releasing Sol without prior approval (per The Information).
  • OpenAI submitted a list of companies to be granted access; the government excluded some firms located outside America (per Washington Post).

Key claims

  • De facto licensing regime exists. The June 2 EO's disclaimer is contradicted by the practice. Prior approval, list-based access, and jawboning phone calls are the operating mechanism.
  • Confusing cocktail of permissions. Britain's AI Security Institute (AISI) had access to Sol per the UK AI minister, but did not have access to Mythos while it was still export-controlled (though it could use Mythos Preview, "slightly less efficient"). The nationalistic tinge is not consistent — allies get partial access.
  • "Classified benchmarking process" is coming by August per the EO. But few frontier-lab staff hold the security clearances needed to see the red lines. The judge role is contested — the White House's national cyber director office spearheaded policymaking but "senior staff have been leaving in recent weeks." Lutnick's Commerce Department has limited in-house technical expertise.
  • Three labs, three governance preferences.
    • Anthropic — favours a government veto on releases (ironic given how it has been treated).
    • OpenAI — wants a government agency in charge, but with more predictable rules.
    • Google — prefers an industry-funded body like FINRA (brokerage firms) or NERC (power grid).
  • Front-loaded amortisation. Labs typically recoup the enormous training cost "in the first few months of their release, while they have a temporary edge over competing models." Every week of licensing delay eats into that window — Ball warns companies may hesitate to fund data-centre capex "to serve frontier models to whatever 100 companies the us government will allow access."
  • China catch-up context. Leading Chinese models 6–10 months behind US equivalents, "far cheaper to run," most open-weight. Microsoft was reported to be considering DeepSeek's model for Copilot — the June cover-Leader worry made concrete.
  • Guardrail tightening as political response. Labs are tightening the guardrails on public models in response to political pressure, meaning "these are more likely to refuse requests that would once have passed muster." Alex Stamos (ex-Facebook CISO) notes many companies have already prepared to switch to Chinese open-weight models in the event of further disruption.
  • David Sacks quote (Trump AI adviser, initially defended the Mythos treatment): "we deviate from that strategy at our peril" — a rare intra-Trump-camp signal of discomfort.

Named actors to remember

  • Dean Ball — Former Trump-admin AI adviser; the "implausibly libertarian → increasingly draconian" quote; the amortisation-window warning.
  • Howard Lutnick — Commerce Secretary; the jawboning phone call to Altman.
  • David Sacks — Trump AI adviser; the "deviate at our peril" internal-dissent quote.
  • Alex Stamos — Ex-Facebook CISO; "many companies have already prepared to switch to Chinese open-weight models."
  • Sam Altman — Sol "worldwide" hedge on X.

Why it matters for this vault

This is the piece that operationalises Hierarchy of Access as a lived-in regime, not a policy proposal. The specific numbers — ~100 firms, Fable-5 public tier, front-loaded amortisation window, ~6–10 month China lag, Microsoft-considering-DeepSeek — are the concrete claims we can cite in synthesis pages. The three-labs / three-governance-preferences split is a durable structural feature to track: Anthropic-veto vs OpenAI-agency vs Google-industry-body.

The Stamos + DeepSeek-in-Copilot signals connect this piece to Token Scarcity and the Frontier AI Ecosystem participation-over-consumption thesis: American guardrail tightening is creating demand for Chinese open-weight alternatives, which weakens both the Hierarchy of Access enforcement and the American labs' cash-flow window at the same time.

Cross-references