A developer explains why they reject AI-generated code even when it functions correctly. The author argues that working code is not the same as good code, citing issues like maintainability, security, and alignment with project architecture. The post reflects a growing tension in software development between productivity and craftsmanship.
We've been told AI will free us from drudgery. But what if the drudgery was the point? Writing code isn't just about output. It's about understanding, ownership, and building something that lasts. The developer who rejects working AI code isn't a Luddite. They're a craftsman.
AI-generated code is a black box. It works today, but tomorrow? When something breaks, you'll have to reverse-engineer a ghost. That's not progress. That's technical debt with a smile. We need tools that augment our thinking, not replace it. The best code isn't the fastest to write. It's the clearest to read.