Conceptual integrity and counting lines of code
- Source
- simonwillison.net
- Date
Context
Simon Willison argues that increased code output can indicate improved productivity only when quality stays comparable: the code must still be tested, maintainable and ready for production. The post, built from podcast excerpts, pairs that conditional argument with a second constraint: an engineer's ability to understand and oversee the growing system.
Cheap implementation also makes loosely related features easier to justify. Willison calls the resulting design problem a loss of conceptual integrity, meaning that the software's parts stop fitting together coherently. The argument preserves a role for engineering teams even when agents write quickly: people must share responsibility for understanding the code and deciding which additions belong.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
Checking sign-in…
Loading comments…


