AGI Won't Create New Jobs

This is a coherent capability argument, but it quietly slides from AGI generalization to labor market immobility. That's the weak link: job creation is not just an intelligence problem, it also depends on institutions, demand shifts, regulation, and human preference.

• –The low-rank and g-factor analogy is useful for thinking about transfer, but it does not prove that all economically valuable work compresses the same way.
• –Benchmark gains show generalization, yet real jobs also depend on reliability, latency, compliance, coordination, and trust.
• –The thesis is strongest for productivity-measured roles; work built around care, status, taste, or politics is much harder to reduce to a single accuracy metric.
• –Even if some human niches survive, they may be too small or too temporary to offset displacement at scale.