Briefing: The UBI Math Doesn't Work. The Math Was Always There. Now It's Arriving.

Published: April 28, 2026 | Source: ejsays.com | Author: E. J. Original article: https://posts.ejsays.com/the-ubi-math-doesnt-work-the-math-was-always-there-now-its-arriving/

Core claim: Every UBI proposal currently on the table — OpenAI's robot tax, Musk's Universal High Income, Dario Amodei's 3% token tax — is calibrated to a transition, not a permanent structural shift. The math does not close. The tax base being proposed to fund the safety net is being eroded by the same force that requires it.

The 2018 predictions, confirmed: The author wrote two Chinese-language articles in 2018 predicting (1) autonomous vehicles would displace truck drivers first, generating $168B/year in industry savings while eliminating $21B/year in direct trucker spending that sustained highway-dependent towns; and (2) AI would displace white-collar workers — paralegals, analysts, programmers — because the new AI companies would hire far fewer people than the industries they replaced. Both predictions are now visible in employment data. White-collar payrolls have contracted for 29 consecutive months as of early 2026, outside any formal recession.

The new companies are not hiring: AI-era companies generate extraordinary revenue per employee — a structural departure from every previous wave.

For comparison: Google replaced AT&T in economic weight while employing ~7% as many people. AI companies are compressing that ratio further.

The displacement radius: Job loss undercounts the problem. 3.5 million truck drivers × $6,000/year in road spending = $21B/year flowing to diners, motels, gas stations, and town tax bases. Automate the trucks; the diner closes, the motel closes, the school district cuts teachers. The $168B industry saving flows to shareholders, not to those towns. AI displacement extends this logic to six-figure jobs — paralegals, analysts, mid-career engineers — with a proportionally larger demand destruction effect.

The payroll tax trap: US social programs are funded primarily by payroll taxes — approximately $1.7 trillion in FY2025, ~34% of total federal receipts. AI displacement shrinks the wage base at exactly the moment demand for safety net programs grows. Nvidia earned $123B in US pretax income in 2025 — the second-highest single-year corporate earnings in American history — and paid approximately $17–19B in federal cash taxes due to FDII deductions and R&D credits. To replace payroll tax revenue lost from mass white-collar displacement, the US would need approximately 90 Nvidias performing simultaneously at record levels.

Why current proposals fail the math:

The convergence signal: OpenAI (April 2026 policy blueprint) and Elon Musk have both proposed versions of UBI/wealth redistribution. The author interprets this convergence not as reassurance but as a signal — the people closest to the technology have looked at the same numbers and concluded the existing social contract cannot absorb what is coming.

The productivity-demand paradox: AI-driven productivity optimizes supply. Payroll is the engine of demand. Automating payroll to increase productivity eliminates the customer base for the products being produced more efficiently. A small business cannot sustain itself on efficiency alone if its customers are analysts replaced by the same API the business is using.

Author's conclusion: The amounts being proposed are right ideas at the wrong scale. The destination that historically absorbed displaced workers — a new industry hiring at scale — does not exist this time. If redistribution is the answer, the numbers need to be much larger than anything currently proposed. The social contract requires redesign, not adjustment.

UBI Proposals vs. Displaced Income

The Payroll Tax Math