That's your prerogative, but be aware you'll continue to remain confused about LLMs. Anthropomorphizing them is what gives you the best high-level intuition about where and how to employ them, and where and how not to.

If there was anything that made sense to anthropomorphise it would be a machine meant to mimic talking, thinking and answering like a human, one that even passes the Turing test.

When we built the idea that anthropomorphising is wrong, we meant when doing it for rocks or trees or thunders or deer or some such.