AI coding tools are real productivity multipliers — and they are also accelerating a form of technical debt that most organizations aren't measuring. The core problem isn't code quality; it's the widening gap between the volume of code that exists and the volume that anyone actually understands. AI optimizes for local correctness.