Cornelius Renken keeps writing code that AI will make obsolete within two years. As Product Lead at Kombo, specializing in applied AI and browser automation, he's not ignorant of the tools reshaping his field. He's made a deliberate choice.
The argument: hands-on coding builds system understanding that prompt engineering doesn't. When you're debugging a production issue at 2am, knowing how the pieces actually fit together beats knowing which AI tool to ask.
This tracks with what we're seeing in the enterprise. AI coding tools are accelerating development, yes. But they're also accelerating technical debt. Fast ships with shallow understanding create maintenance nightmares. The code works until it doesn't, and then nobody knows why.
Renken's position reflects a broader tension in enterprise tech. CTOs are under pressure to adopt AI tools for velocity. But they're also responsible for systems that need to run reliably for years, not sprints. Writing code the old way might seem inefficient until you need to refactor a critical module and nobody understands its dependencies.
The enterprise context matters here. Startups can rewrite fast. Government departments and banks can't. The code you ship today will likely still be running in 2028, AI revolution or not. Understanding how it works isn't optional.
Worth noting: Renken isn't arguing against AI tools. He uses them. But he's pushing back against the idea that prompt engineering replaces systems thinking. In fast-evolving fields like AI automation, that deep knowledge becomes a competitive advantage, not a liability.
The real question for tech leaders: are you optimizing for shipping this quarter or maintaining next year? The answer probably isn't binary, but it's worth asking before your entire team stops understanding the codebase.