Why AI adoption hasn't yet moved the US labour market

Summary

A detailed X thread arguing that despite significant AI adoption in the US, unemployment has remained stable and AI has not yet caused measurable labour displacement. McCrory's case rests on two observations. First, the aggregate US labour market shows no signs of AI disruption: unemployment remains stable near full employment, job openings exceed jobseekers, and workers are not switching occupations faster than before. Second, AI so far amplifies rather than replaces skilled workers—it complements domain expertise rather than eliminating it wholesale. He qualifies this with uncertainty: future capabilities may be more substitutive, and weak signals of hiring weakness for young workers in AI-exposed roles warrant monitoring.

Key Claims

• The US unemployment rate in June 2026 stood at 4.2%, consistent with full employment; layoffs remain at pre-pandemic levels; no material increase in unemployment for workers in high-AI-exposure roles has occurred to date.
• Quality-adjusted AI output grew over 2,000% per year in 2024 and 2025, yet no O*NET job taxonomy entry lists all associated tasks as systematically handled by Claude, meaning even high-exposure roles retain unauto­mated components.
• AI currently complements domain expertise, requires human direction, and expands what workers can do rather than replacing roles. The gap between weak hiring for young workers in AI-exposed occupations and broader economic headwinds (post-pandemic slowdown, tightened money supply, remote work) is difficult to separate.
• Labour productivity growth reached 2.0% per year from early 2022 to early 2026, up from 1.6% pre-pandemic, and sectors with higher AI adoption show faster productivity growth; total factor productivity gains are more modest, signalling mixed evidence for a broad acceleration.
• Jobs are not fixed task bundles: new technologies historically reshape work within roles and create new combinations of human expertise and technical capability; users report perceived productivity gains from scope—doing more, more proficiently—rather than task elimination.
• Recursive self-improvement in AI could speed up innovation itself, potentially creating economic singularities. But if some essential tasks prove difficult or impossible to automate—in R&D or software production, for instance—they may cap the pace of displacement even as capabilities improve.

Quotes

• "Even as AI automates some aspects of work, complementary human expertise amplifies what AI (or humans) can achieve alone."
• "It relies on humans in the loop to direct and evaluate the most complex work. And it rewards AI proficiency. Model capabilities are improving fast, but remain stubbornly jagged."
• "Even after just six months of use, people are more likely to interact with Claude as a thought partner, and have more successful interactions with Claude. If AI was good enough on its own, we wouldn't expect to see this effect."
• "People make planning decisions, and delegate implementation to Claude. People with more domain expertise succeed in their tasks more often, and recover more consistently when Claude makes an error."