Managers are often really impressed with trivial code changes or assume that complex impossible changes are trivial.
This makes sense. Being a manager (as a standpoint) puts you far away from the code which limits your ability to understand it.
LLMs turning coders "into managers" then makes it difficult for them to evaluate coding agents' effectiveness. The longer you spend further away from coding, the more difficult it is to tell what is trivial and what isn't.
β




