The Permanent Underclass Is Already Here - City Journal

TEXT START: No city better captures the contradictions of twenty-first-century America than San Francisco.

THE DISSECTION

This is a policy indictment wearing an AI analysis costume. It takes San Francisco’s visible street pathology, demotes AI displacement fears to bourgeois status anxiety, and recodes the future underclass as a problem of addiction, conservatorship, welfare, and municipal permissiveness. It keeps the central question—who owns AI capital after human labor loses necessity—almost entirely out of frame.

THE CORE FALLACY

Category substitution and temporal myopia. The existing underclass may be largely policy-mediated; that does not refute a future underclass created by AI severing the employment–wage–consumption circuit. Current unemployment and newly invented job categories measure the lag, not the terminal mechanism. Nor does the article understand UBI’s actual role under the Discontinuity Thesis: it cannot restore productive participation, but it can preserve consumption and social stability after participation collapses. Calling it a failed cure attacks a claim the thesis does not make.

HIDDEN ASSUMPTIONS

• AI will not achieve durable cost and performance superiority across cognitive work, or will create enough economically necessary human work to absorb the displaced.
• Institutions can coordinate a protected human-only labor sphere at scale.
• Better treatment, coercive care, or welfare redesign can restore economic necessity for most people made redundant.
• An underclass only counts when it is visibly destitute; subsidized exclusion and managed dependency are treated as something else.
• Present labor-market normality is evidence against structural displacement rather than a temporary lag.
• The motives of AI executives—fundraising, fearmongering, or status signaling—determine whether the underlying automation mechanism is real.

SOCIAL FUNCTION

Partial truth, elite self-exoneration, and ideological anesthetic. The text correctly observes that transfers do not cure addiction, psychosis, or institutional failure. Its political function, however, is to relocate systemic danger onto defective individuals and incompetent government, implying that competent administration can contain the future. It converts a question of ownership and productive participation into a morality play about disorder.

THE VERDICT

The article accurately describes one existing carcass and badly misreads what it proves. San Francisco is not evidence that AI-driven exclusion is imaginary; it is a crude preview of a maintenance regime in which institutions manage surplus people instead of restoring their economic agency. The future version may be cleaner, wealthier, and less visibly chaotic—but structurally larger. The text sees the underclass, misidentifies its scalable cause, and mistakes hospice policy for system survival.