No, I didn't. I claimed that if my coding ability atrophies, I expect that my ability to code through LLMs will decrease. And this is an outcome I want to avoid for the stated reasons and others.

That said, I get your point: maybe "the thing writing the code" will be as abstract as "the thing moving the electricity through transistors" at some point. And maybe I will be wrong about coding skill translating to good agent-herding skill. But I don't think I'm wrong now.