Vibe coding is manufacturing tech debt at machine speed

Vibe coding is moving fast, but is it moving software forward or just piling up a new kind of debt. In this episode of Pop Goes the Stack, F5's Lori MacVittie and Joel Moses are joined by Narayan Veeramani to dig into what happens when AI-generated code lands in real production environments, long after the demo passes and the pull request looks “clean.”They break down how today’s coding assistants can produce working code at unprecedented speed, while humans still have to read, review, and maintain it at human speed. The result is what the team calls comprehension debt: code that appears correct, ships quickly, and then becomes a 3 a.m. incident because no one truly understands its assumptions, its architecture, or its operational impact.The conversation also highlights why green tests can be misleading when AI generates tests from its own code instead of from the spec or requirements, reinforcing incorrect assumptions like a US-only interpretation of “Thanksgiving promotions.” Joel adds an architectural cautionary tale where “beautiful” async logic passed mocks but hammered a production database with inefficient query patterns.The takeaway is practical and direct: vibe coding doesn’t eliminate engineering rigor, it changes where the rigor must happen. Bring context into prompts, generate tests from specifications, review continuously instead of in a big-bang PR, and ensure someone accountable can explain the code in an audit or an outage. AI can handle the mechanical work, but teams still own what they ship.