When a lower layer fails, your ability to remedy a situation depends on your ability to understand that layer of abstraction. Now: if an LLM produces wrong code, how do you know why it did that?