Source: The Wall Street Journal (paywall)
The productivity promise of AI agents is creating a perverse outcome: founders working longer hours to supervise, prompt, and course-correct their tools rather than replacing manual work. This points to a gap between AI capability marketing and actual deployment friction—agents require constant human oversight, creating new forms of cognitive labor rather than eliminating existing ones. The pattern mirrors early automation cycles where promised efficiency gains shifted into different, often more stressful work rather than genuine time savings.