The article argues that AI, specifically Large Language Models (LLMs), should be viewed as tools that require skill to use effectively, rather than as automation that replaces human judgment. The author contends that using LLMs well depends on the operator's ability to direct them, interpret feedback, and make corrections. This perspective is presented as a counterpoint to the idea that LLMs are not tools because they cannot be wielded. The article emphasizes the importance of human expertise and intention in using LLMs.