Why would I want an AI assistant to have agency? I want them to help me, not to further their personal goals. In fact, I don't want them to have personal goals other than helping people.
I didn't say it should have one, I'm saying that LLMs statistically finding the next bit of information aren't really making decisions, which is to counter-argue the fact that it's not different from how we reason.
Your comment doesn't convince me you're making decisions either. "Wurd" could've just what you considered the best token to get your point across in the same way that LLMs choose the best token
