Technology adoption patterns diverge by sector based on demand satiation rather than task capability

str 8 4/25/2026 · 1 article

structural · economic · AI, labor markets · US, UK

Analysis

The article demonstrates that whether automation reduces or increases employment depends primarily on whether pent-up demand exists for the service, not on whether the technology can perform the task. This inverts the conventional 'jobpocalypse' narrative and suggests AI's labor impact will be highly heterogeneous across sectors.

Key actors

knowledge workersmanufacturing sectorprofessional services

Source article

What the AI ‘jobpocalypse’ narrative misses

Financial Times — Work, Skills and Society · 4/25/2026 · extracted in run pdf-import-2026-04-25-1779256502138-89 · 5/20/2026, 7:14:36 AM

"Large increases in software productivity since the 1990s were accompanied by rising, not falling, employment in web development" [1990s]

This concrete historical example directly supports the structural claim that productivity gains do not automatically reduce employment when demand for the service is pent-up and growing faster than efficiency gains.

Reasoning from this article

The article uses multiple historical cases (software, accounting, manufacturing, retail, logistics, spreadsheets, ATMs, newspapers) to establish a general principle: task capability is decoupled from employment outcomes. The mechanism is demand elasticity—sectors with unsatiated demand (healthcare, professional services, tech) see employment rise despite productivity gains, while saturated sectors (manufacturing, retail) see employment fall. This pattern is presented as predictive for AI's labor impact across different occupations.