Casar's AI Token Tax Would End Payroll Subsidy for Automation, Fund New WPA

TEXT START: The federal tax code currently gives companies a quiet financial incentive to replace workers with AI: every employee cut saves an employer payroll taxes, while the AI system that takes that employee's place pays none.

The Dissection

The text is doing three jobs at once: explaining H.R. 10044, legitimizing its payroll-tax argument, and constructing a political battlefield around AI capital, organized labor, and the 2028 election.

Its strongest point is real: taxing wages while leaving AI computation untaxed can erode the payroll-tax base as labor demand falls. Its central maneuver is then to present a token levy and a new WPA as a way to restore employment. That is where the argument changes from fiscal correction to social containment.

The article also functions as campaign infrastructure. The bill's weak legislative prospects are openly acknowledged, while its value as a Democratic organizing vehicle, labor coalition platform, and response to AI-industry political spending is emphasized. This is not merely policy reporting. It is positioning.

The Core Fallacy

The text assumes that taxing automation and financing human-intensive public jobs can preserve the wage-to-consumption circuit. Under the Discontinuity Thesis, that circuit is not merely underfunded; it is being severed by superior machine production.

A token tax may reclaim revenue. It does not restore the economic necessity of mass human labor. A WPA-style program may distribute income through employment, but those wages are transfers funded by AI activity, not proof that workers remain structurally productive. It preserves consumption and social stability while productive participation collapses.

The bill also taxes a proxy, not displacement itself. Token volume or AI-product revenue does not reliably measure how many workers were replaced. Firms can pass costs downstream, shift to open-weight models, vertically integrate, relocate activity, or automate without proportional token use. The article identifies these weaknesses—especially business-to-business cost cascades—but never solves them.

The unemployment trigger is reactive. It raises the tax after visible labor-market damage has already occurred. That is a brake applied after the vehicle has entered the ravine.

Hidden Assumptions

• Token sales will remain observable, domestic, and difficult to evade.
• Token usage will correlate closely enough with jobs destroyed to support fair taxation.
• Firms will absorb the levy rather than pass it through, relocate, or redesign around it.
• Tax receipts will be large and stable enough to finance mass public employment.
• Government can create millions of durable, socially valuable jobs without turning the program into bureaucratic make-work.
• Housing, infrastructure, child care, and elder care are durable human domains rather than temporary physical and institutional lag zones.
• Employment-funded consumption is an adequate substitute for productive participation.
• National unemployment will capture displacement better than labor-force exit, wage compression, underemployment, or declining bargaining power.
• Political institutions can coordinate against AI capital despite the industry's financing power and competitive pressure.

Social Function

Classification: partial truth, transition management, prestige signaling, and ideological anesthetic.

The article is not pure copium. It correctly identifies a fiscal asymmetry and treats AI displacement as a serious risk rather than a distant abstraction. But it converts that accurate diagnosis into an orderly reform narrative: tax the machines, create public jobs, preserve the social contract.

That narrative makes systemic rupture politically digestible. It allows policymakers to appear to confront obsolescence while avoiding the harder question: who owns and controls the productive systems after human labor loses bargaining power? The WPA analogy supplies historical prestige, but it is also a nostalgia trap. The 1930s program expanded employment because human labor was still the productive engine. That premise is precisely what AI destroys.

The Verdict

H.R. 10044 is a fiscal tourniquet and a carcass-management plan, not a resurrection of mass employment. It may claw back AI rents, fund temporary niches, and delay the visible social consequences of automation. It cannot defeat P1, P2, or P3.

The article's fatal error is treating redistribution through jobs as restored productive participation. The proposed New WPA would preserve the wage shell after the productive core has been hollowed out. Its political value may be substantial; its structural solution is nonexistent.