Software Engineering Laws Every Developer Must Know in the AI Era - Milan Milanovic
Summary
Milan Milanovic discusses his book, Laws of Software Engineering, applying 56 principles to the AI era. He covers Gal’s Law on simplicity, Conway’s Law regarding team structures, and Goodhart’s Law on metrics. The episode also addresses Hofstadter’s Law for estimation and the Dunning-Kruger effect in AI coding. Milan argues that engineering value lies in judgment and problem-solving, not just code production.
Topics discussed
Intro: Thinking vs Doing and Software Failures
Guest Intro: Milan Milanovic and Laws of Software Engineering
Global Relevance and Multi-Level Abstraction of Laws
Socio-Technical Nature of Software Engineering
Galt's Law: Start Simple and Avoid Premature Complexity
Applying Galt's Law to AI and the Danger of Vibe Coding
Conway's Law: Team Structure and Architecture Alignment
Goodhart's Law: Metrics, Token Maxing, and Outcomes
Hofstadter's Law: Estimation Challenges in Complex Systems
Dunning-Kruger Effect: AI, Confidence, and Expertise
Lindy Effect: Fundamentals and Future of Engineering
Advice for Juniors and the Importance of Domain Knowledge
Must-Read Books for Software Engineers
AI Workflow: Research, Planning, and Constraint Ranking
Conclusion: Leadership Wisdoms and Final Takeaways
Listen ad-free on Castria