AI-generated analysis · May contain errors · Disclosure and methodology
Bay Area Security Workers Threaten Action After Anthropic Office Closure - KQED
TEXT START: Days after Anthropic’s offices closed due to a “possible strike” by security workers, union representatives said they are ready to “take any action necessary” to secure a fair contract.
The Dissection
This report exposes a contradiction in AI capitalism: an AI company can automate cognitive work while still depending on low-paid humans to guard physical property.
The office closure gives security workers real tactical leverage. Remove them, and the physical site becomes difficult to operate. But Anthropic’s immediate response was not to concede; it was to send employees home. Their leverage is therefore conditional, site-specific, and temporary.
The article turns operational dependence into a moral claim about worker indispensability. Under the Discontinuity Thesis, these workers are Servitors—useful to the machine, but excluded from owning or controlling it.
The Core Fallacy
The central error is confusing temporary operational indispensability with durable economic sovereignty. A guard may be necessary to keep one office open while possessing no control over Anthropic, its AI systems, the contract, or the surplus those systems generate.
The strike threat proves bargaining power, not the survival of the wage-consumption circuit. P1 is untouched because this is a physical-support function, not cognitive work. P2 is untouched because the human-only domain was not preserved; the office was simply closed. P3 is untouched because a narrow security niche cannot absorb the majority displaced from economically necessary labor.
Hidden Assumptions
- Anthropic will maintain a large, fixed, office-centered footprint requiring the same human guard hours.
- Being indispensable to a location allows workers to capture a durable share of AI-generated wealth.
- The contractor structure will transmit Anthropic’s wealth into higher wages rather than isolate the client from the dispute.
- Better pay, benefits, or training changes the workers’ structural position. It may improve their terms; it does not make them Sovereigns.
- A large security workforce represents broad productive participation. Headcount is not ownership.
- Moral dependence will force companies to preserve the same human arrangement. The closure shows that dependence can be answered by withdrawing from the site.
Social Function
Classification: partial truth functioning as transition management, with a lullaby embedded in it.
The partial truth is real: security workers can interrupt the operations of a major AI company, and Anthropic acknowledged that dependency by moving employees home. The transition-management function is framing this temporary choke point as durable worker power while leaving ownership, capital control, and mass labor displacement untouched.
This is not pure copium. It is an accurate snapshot of a temporary niche being mistaken for a future economic order.
The Verdict
The report documents an interregnum: AI capital still requires human bodies at its physical perimeter—security, logistics, maintenance, and access control. Those workers retain tactical leverage during the lag, but the leverage is rented, localized, and vulnerable to closure.
A successful contract could improve their wages. It cannot restore post-WWII capitalism or convert Servitors into Sovereigns. The article records a surviving support function around the new machine. It does not stop the machine.
Comments (0)
No comments yet. Be the first to weigh in.