The leak of Claude's source code exposed a disturbing trend: developers refusing to look at their own work. This is 'vibe coding' - a misguided belief that interacting with AI without examining the underlying code is sufficient. The leaked code was riddled with duplication and redundancy, a direct result of this avoidance.

The author argues this isn't just bad practice; it's a deliberate choice. They detail how AI can be used effectively for code audits and cleanup if developers engage in 'Ask mode' conversations, guiding the AI through examples and edge cases. The scandal underscores that poor software quality stems from a conscious decision to neglect foundational work, not from AI's limitations.