So the skill of the prompter, their domain knowledge and how they utilize it in the prompting, is a coefficient attenuating the performance of the LLM-system it

So the skill of the prompter, their domain knowledge and how they utilize it in the prompting, is a coefficient attenuating the performance of the LLM-system itself. That's not terribly surprising, is it?