The problem is not the AI code, but nobody knows anything anymore

The article argues that the real risk of AI-assisted software development is not code quality alone, but the loss of institutional knowledge, architectural intent, and system ownership as teams increasingly rely on LLMs to generate specs, tests, and code. For CIOs and technology leaders, the strategic implication is clear: AI can accelerate delivery, but without strong human oversight and engineering discipline it can also create fragile systems, higher maintenance costs, and a workforce that no longer understands what it ships.