I suspect there are much more than two kinds. there is a varying degree of understanding what these tools are capable of, and that multiplied by what people need and how much they care about outcomes and consequences is the number of "kinds" of AI users.
Let's take the group of developers (to keep it simple) that have a deep understanding of LLMs and how they work. Even then, some don't care if it generates entire codebases for them, some know there will be bugs in it, they just don't care. Some care, but they know their job is to make their project managers happy. Others don't have apathy or pressure like that, but they'll still use it in the same way, because for one reason or the other it saves them time. I'm probably missing more examples, but it is the same usage, but different motivations, people, and environments.
Let's take the group of developers (to keep it simple) that have a deep understanding of LLMs and how they work. Even then, some don't care if it generates entire codebases for them, some know there will be bugs in it, they just don't care. Some care, but they know their job is to make their project managers happy. Others don't have apathy or pressure like that, but they'll still use it in the same way, because for one reason or the other it saves them time. I'm probably missing more examples, but it is the same usage, but different motivations, people, and environments.