What did Mark Zuckerberg said in his 6,500-word AI manifesto? - Khaosod English

TEXT START: Mark Zuckerberg has published a roughly 6,500-word manifesto on the future of artificial intelligence, laying out a vision in which everyone gets access to a personal “superintelligence” that can help with work, health, education, business and everyday life.

The Dissection

The text is presenting corporate strategy as civilizational philosophy. Zuckerberg’s argument is that advanced AI should be distributed widely, provided through Meta’s products, models, data systems and infrastructure. “Personal superintelligence” is therefore both a promised tool for individuals and a bid to make Meta the delivery layer for post-human cognition.

The article also exposes the contradiction: the advocate of dispersed power remains a founder-controller of one of the institutions positioned to own and mediate that power.

The Core Fallacy

The central error is treating access as control.

A free or affordable AI assistant does not make users Sovereigns. If Meta owns the models, compute, distribution channels, identity systems and commercial relationship, users are renting intelligence. They gain capability without owning the productive asset that generates it.

The courtroom analogy also fails structurally. Equal access to lawyers does not eliminate unequal wealth, institutional control or ownership of the court. Likewise, billions of AI users do not neutralize the firms controlling the underlying systems.

The jobs argument commits a second error: it confuses the creation of new niches with preservation of mass productive participation. New occupations can appear while total labor demand collapses. App developers, social-media creators and electric-vehicle technicians do not prove that displaced workers can scale into equivalent roles.

Hidden Assumptions

• AI will create enough valuable new work to offset the labor it eliminates.
• Workers can adapt at the speed of cognitive automation.
• Users will accept deep AI surveillance in exchange for convenience.
• Meta’s private mode will remain private and trustworthy.
• Distributing dangerous capabilities will reduce misuse rather than multiply it.
• Competition among AI systems will remain stable instead of consolidating around the most powerful owners.
• Open-source releases and an independent safety board will meaningfully restrain founder-controlled corporate power.
• American dominance and minimal regulation will produce broad prosperity rather than strengthen incumbents.

These assumptions are doing the real work. The manifesto asserts them as a future of abundance without solving ownership, income or bargaining power after human labor loses necessity.

Social Function

Classification: elite self-exoneration, transition management, ideological anesthetic and corporate propaganda, containing a partial truth.

The partial truth is that AI can create new businesses, scientific capabilities and temporary niches. The anesthetic is presenting those niches as evidence that the wage system will survive. The corporate function is to make Meta’s expansion appear equivalent to democratization: the company centralizes the infrastructure while describing the resulting dependence as empowerment.

The national-security language supplies political cover. Fear of authoritarian rivals becomes an argument for faster deployment, weaker restraints and deeper public-private alignment—precisely the conditions under which a small number of firms accumulate more power.

The Verdict

This is not a plan to preserve post-WWII capitalism. It is a sales manifesto for the cognitive layer that follows its collapse.

If P1–P3 hold, personal superintelligence accelerates the severing of mass employment from wages and consumption. It may distribute useful tools and preserve some consumption, but it does not restore productive participation. Zuckerberg is proposing universal access to intelligence while leaving ownership of intelligence concentrated. That is not liberation. It is a nicer interface for dependency.