What Everyone Is Getting Wrong About AI And Jobs

YC’s Garry Tan argues AI follows Jevons Paradox — cheaper work creates more demand for human judgment, not mass unemployment.

• Geoffrey Hinton predicted in 2016 that radiologists would be obsolete in 5 years; demand for radiologists is now at an all-time high.
• Jevons Paradox: efficiency gains lower cost, reveal latent demand, and expand total consumption — true for coal, shipping, cloud, and now AI inference.
• Containerization cut shipping costs 90% in the 1960s; laid off some dock workers but spawned billion-dollar freight and logistics empires.
• Cloud computing made infrastructure 10x cheaper; server admins became DevOps and cloud architects managing previously impossible scale.
• Andrej Karpathy argues AI first automates rote, low-context, error-forgiving tasks like data entry and customer service, refactoring them into supervisor roles rather than eliminating them.
• YC portfolio co AOKA (AI sales agent for plumbing/HVAC) and Tenor (healthcare paperwork automation) are shifting humans from rote tasks to higher-judgment work.
• Tan’s core bet: as AI cuts the cost of analysis, drafting, and coding, demand for radiologists, lawyers, and engineers will broadly increase, not decrease.

2025-10-14 · Watch on YouTube