Does code quality still matter?

Mark Seemann explores whether code quality will remain relevant in a future where LLMs write all code without human oversight. He argues that current code quality principles are rooted in human cognitive constraints, and questions whether LLMs would accumulate their own form of technical debt in hypothetical LLM-optimized languages. He concludes that fully autonomous LLM-driven development is more complicated than AI optimists suggest, and that human oversight removal poses risks we may not be equipped to troubleshoot.