Top Pentagon official reaffirms Anthropic blacklist despite Lutnick comments

TEXT START: Pentagon official Emil Michael on Thursday said that Anthropic is still considered a " supply chain risk ," a day after Commerce Secretary Howard Lutnick said the company is "back on the right side."

The Dissection

This is a power-map dispatch disguised as an AI policy update. It turns an interagency contradiction into the story: the Pentagon labels Anthropic dangerous while Commerce signals rehabilitation. The excerpt focuses on bureaucratic control over military AI access, not on Anthropic’s underlying technical or economic position.

The Core Fallacy

It treats official designation and political alignment as the main determinants of AI-lab viability. Under the Discontinuity Thesis, these are lag mechanisms. A blacklist can obstruct contracts and redirect capital, but it cannot reverse cognitive automation or the struggle over ownership of scalable AI capacity. The article mistakes institutional friction for a structural defeat.

Hidden Assumptions

• The administration can produce a coherent, enforceable AI policy.
• A supply-chain designation accurately measures technical danger rather than factional power.
• Pentagon approval is decisive to Anthropic’s long-term survival.
• “Back on the right side” has analytical meaning rather than serving as political rhetoric.
• The conflict is an isolated dispute instead of an early fight over state control of private AI capital.
• Administrative labels can determine the trajectory of a technology whose strategic value exceeds any single agency.

The supplied passage also ends mid-quotation at “Tha,” limiting what can be concluded about the omitted statement.

Social Function

Transition management, prestige signaling, and partial truth. It accurately exposes institutional fracture, but packages the deeper transition as a quarrel among officials. That framing makes the collapse of the old labor order disappear behind procedural drama.

The Verdict

The excerpt records a temporary governance fracture, not Anthropic’s obsolescence. The blacklist is lag infrastructure: political friction around custody of cognitive capital. It may damage Anthropic’s access to state demand, but it cannot stop automation. The real contest is who owns and controls scalable AI; this story reduces that contest to officials trading labels.