Our Work Norms are 4000 Years Old and We Need to Update Them for AI—Here’s How to Do It

Source: Nate’s Newsletter Date: 2025-05-19 URL: https://natesnewsletter.substack.com/p/our-work-norms-are-4000-years-old

Summary

Nate argues that work assessment norms haven’t changed in 4,000 years — quality was always verified through direct observation and semantic proximity — and AI has broken that system by producing plausible output that severs the connection between work and the worker who did it. The crisis is not that AI produces bad work; it’s that it produces work that looks good enough to defeat traditional quality signals, making trust and authenticity in professional relationships deeply uncertain.

Implications

• Enterprise adoption thread. The collapse of traditional quality-assessment mechanisms is a governance problem, not a technology problem. Organizations that haven’t developed new trust and verification frameworks for AI-augmented work are operating on outdated accountability structures — and AI adoption without new norms compounds the problem.
• AI economics thread. If knowledge workers can produce semantically credible output without domain expertise, the labor market signal breaks down: credentials and prior work samples lose predictive validity, and firms can’t efficiently allocate work to the humans who actually understand it.
• Watch: Whether new professional trust norms crystallize around outcomes rather than process (results verification, not work attribution), and which industries develop the most robust new standards first.