Workers in China worry over being replaced as they adapt to the growing impact of AI on jobs

TEXT START: HONG KONG: Computer programmer Fei Zhaojun's boss asked him if artificial intelligence (AI) could soon replace humans in coding jobs.

THE DISSECTION

The text documents the early stages of the Discontinuity Thesis: AI adoption, layoffs, wage compression, educational devaluation, and fear-driven consumption weakness. Then it tries to metabolize those facts into a manageable story of adaptation through AI-enabled entrepreneurship, independent work, teaching, and demographic necessity.

The first layer is the autopsy. The second is the sedative. Individual adjustment stories are presented as though they can scale to the displaced majority. Nothing in the article establishes that they can.

THE CORE FALLACY

The central error is treating labor substitution as a temporary mismatch between workers and new jobs. The article’s own evidence points to structural displacement: mid-level coding is described as replaceable, translator pay has fallen by more than half, companies are cutting staff, content production is collapsing, and robots threaten delivery work.

It also assumes that AI creates compensating opportunity merely because it enables independent ventures. When AI makes specialized output available to almost anyone, supply expands, prices fall, and bargaining power moves toward whoever owns the models, platforms, capital, energy, and distribution. The newly independent worker may simply be a cheaper, less protected servitor—not a restored participant in the old wage-consumption circuit.

The aging-population argument is a category error. A shrinking workforce may increase the need for automation in selected sectors, but it does not restore mass labor income or productive necessity. Demographic scarcity in some occupations can coexist with a surplus of displaced workers elsewhere. P1 and P3 are not reversed by population aging.

HIDDEN ASSUMPTIONS

• AI-adjacent businesses will generate enough income and employment to absorb displaced workers.
• Human judgment will remain scarce even when generation, comparison, and verification become cheap.
• AI errors will continue protecting professions instead of becoming cheaper to detect and correct.
• Fewer workers will automatically translate into stronger worker bargaining power.
• Individuals can retrain, access capital, and build businesses during synchronized displacement.
• Government-led AI diffusion is merely productivity policy rather than an accelerator of labor’s loss of bargaining power.
• Transfers or institutional responses can preserve consumption without resolving the collapse of productive participation.

SOCIAL FUNCTION

Partial truth functioning as transition management and ideological anesthetic, with a layer of copium. The article reports enough real damage to appear serious, then redirects the conclusion toward personal adaptation and demographic necessity. It is not pure propaganda; it is more useful than that. It normalizes the early casualties while implying that the system can absorb them through individual improvisation.

THE VERDICT

This is a credible early-warning report wrapped in a labor-market lullaby. Its own facts describe AI diffusing faster than institutions can preserve human economic roles: productivity rises, labor income and occupational scarcity weaken, and fear suppresses consumption. The proposed exits—use AI, become independent, or wait for demographic demand—are individual gambits, not systemic solutions.

Under P1, P2, and P3, the article’s reassuring tail is its least defensible component. It documents the machine severing the employment-to-consumption circuit, then calls the sparks adaptation.