We Can't Weld Our Way Out of an AI Employment Crisis - Kiplinger

TEXT START: AI is beginning to hollow out parts of the white-collar economy, and the dominant response for workers has been: Go learn a trade.

The Dissection

The article correctly demolishes “learn a trade” as a universal answer. It identifies the financial, psychological, and status collapse that occurs when credentialed workers are pushed into lower-paid physical occupations.

But it stops at the threshold of the real diagnosis. It treats AI displacement primarily as a badly managed labor-market transition requiring retraining, counseling, placement, and temporary income support. Under the Discontinuity Thesis, the deeper rupture is that AI attacks the wage-to-consumption circuit itself. The problem is not merely that workers need different jobs. It is that fewer workers remain economically necessary.

The trades are a lag domain, not a systemic escape hatch. Electrical work, HVAC, infrastructure maintenance, logistics, and other physical roles may retain demand while automation is less capable, but they cannot absorb the displaced cognitive workforce at comparable scale, income, or status. They are temporary refuges inside the transition, not a replacement for mass productive participation.

The Core Fallacy

The article’s central error is assuming that better transition management can preserve economic opportunity and social stability after productive participation collapses.

Retraining can move people between occupations only when the economy still requires enough human labor. Counseling can reduce distress but cannot restore bargaining power. Income support can preserve consumption but cannot recreate ownership, necessity, or productive status. The article sees the failure of the trade pivot but still assumes the labor market can be repaired with a broader support package.

That is the old system trying to administer its own disappearance. Under P1, AI’s cost and performance advantage spreads across cognitive work. Under P2, institutions cannot permanently quarantine human-only economic sectors at sufficient scale. Under P3, the majority lose access to economically necessary labor. The article describes the symptoms while retaining the assumption that employment remains the cure.

Hidden Assumptions

The text smuggles in several assumptions:

• That AI will mainly compress selected occupations rather than progressively weaken the employment value of cognitive work across the economy.
• That human-AI collaboration will generate enough new roles to offset the roles eliminated.
• That trades will remain protected long enough, and at sufficient scale, to absorb millions of displaced professionals.
• That retraining curricula can keep pace with AI rather than repeatedly training workers for shrinking labor markets.
• That temporary income assistance can bridge workers into durable employment instead of subsidizing a permanent loss of necessity.
• That employers and policymakers can coordinate an orderly response despite competing incentives to automate faster.
• That restoring income is enough to restore identity, status, household stability, and social legitimacy.
• That the central problem is worker adaptation rather than ownership and control of AI capital.

The article never confronts the decisive question: who owns the systems that replace labor, and what claim do displaced people retain on the output?

Social Function

The piece is a partial truth serving as transition management and ideological anesthetic.

Its useful function is to puncture a patronizing policy slogan and acknowledge that downward occupational mobility is not painless. Its limiting function is to convert a structural break into a program-design problem. By recommending more comprehensive retraining and support, it preserves the fiction that the existing wage order can be stabilized if policymakers become sufficiently humane and technically competent.

It is not pure copium. It accurately recognizes that “go learn a trade” is inadequate. But its proposed remedy remains a softer form of the same evasion: prepare people to compete for diminishing human roles instead of confronting the distribution of AI ownership, productive surplus, and political power.

The Verdict

The article is right that welding cannot absorb the cognitive workforce. It is wrong to imply that a more sophisticated transition package can preserve the employment-centered order.

The trades are hospice care for the old labor market, not its resurrection. Without a transfer of ownership or durable claims on automated output, retraining merely reorganizes the queue for obsolescence. The article identifies the failure of the first comforting answer while leaving the terminal mechanism untouched.