Generating Code Faster Is Only Valuable If You Can Validate Every Change With Confidence

This article argues that generating code faster is only valuable if every change can be validated with confidence. It addresses the widening gap between code production and testing as tools accelerate development. The author outlines a practical testing strategy: three levels of automated testing at pull request time (unit tests, service-level functional tests, and platform-level functional tests) to ensure changes don't break the service. Additionally, nightly testing with end-to-end customer journey, performance, and chaos tests catches issues like memory leaks and performance degradation. The piece emphasizes defining what 'good' testing means and building pipelines that scale with the system and organization, making it highly relevant to software engineering and IT best practices.