From Stonemasons to Carpenters

The piece argues that AI-driven code generation is changing software work from stonemasonry - crafting systems line by line - to carpentry - building the forms that shape generated code. Using the concrete-bridge metaphor, it contrasts builders who place each stone with carpenters who build wooden molds for poured concrete; similarly, modern developers should spend more time creating tests, documentation, interfaces and guardrails so AI can reliably produce the implementation. The shift reframes the craft: instead of focusing on low-level coding, engineers define the desired external behavior and constraints that will drive correct code generation.

That shift has concrete implications: every developer must think more like a product manager, prioritizing what should be built and the system boundaries rather than only how to implement internals. Software engineering becomes more about system design, validation and verification as immediate code generation shortens the feedback loop and makes design consequences visible instantly. Engineers will need stronger design skills and an outside-in mindset, while still taking advantage of software’s ability to iterate (unlike poured concrete). The result is wider adoption of engineering practices that were once too costly, enabling faster, higher-level system building enabled by AI.