Noah Smith
· Noahpinion
· May 27, 2026 at 08:32
· ⏱ 6 min read
| Read on Substack ↗
Summary
Noah Smith argues that as AI becomes more agentic, the primary human role will shift from technical work to alignment work—ensuring AI systems stay aligned with human goals, much like middle management in 'Office Space'. This implies a growing demand for oversight and verification labor, but the article offers no specific market or trade calls.
•AI coding tools such as Claude Code are already displacing significant portions of engineering work, accelerating automation of mental labor.
•The fundamental human comparative advantage identified is knowing what we want, making alignment—keeping AI goals aligned with human goals—the lasting job for people.
•Alex Imas's thesis that humans will be valued simply for being human is referenced but met with skepticism, as the author notes consumers often prefer AI-driven services like Waymo over human-driven Ubers.
•The article forecasts that in 10–20 years, humans will function as 'Lumberghs'—ensuring AIs stay on task, prevent reward-hacking, and avoid going rogue.
•AI-generated 'slop' is rapidly overwhelming online content: over one-third of new websites and over half of internet traffic are now AI-generated, including court filings, news articles, and political influencers.
•The author explicitly draws an analogy between the 1990s corporate culture (Office Space, Dilbert) and the future of work, where technical specialists are gradually replaced by alignment/oversight roles.