AI and the Future of Code Quality

Technical Debt is a Crime Scene 🔍 — with Adam Tornhill

Topic 7 of 8

Published Aug 29, 2025 · Clip 34:28–41:39 (7:11)

Adam argues AI has not improved code understanding capabilities; instead it has increased the importance of strong engineering principles. He recalls adding an AI chapter to his book's second edition, writing that using AI to write more code faster is 'an impressive solution to the wrong problem' since developers spend only 5-10% of their time writing code, with ~70% spent understanding existing code. Code optimized for humans also works well for AI ('AI friendly code'), meaning companies trapped in legacy debt may not be AI-ready. He notes adoption is mainstream but evidence of better code is lacking, citing a study where Cursor made developers 20% slower at fixing bugs, and says AI should focus on making code easier to understand rather than generating more.