AI Is a Tool. Using It Well Takes Skill. (yyhh.org)

In a recent article, Rich Hickey argues that large language models (LLMs) should not be considered tools but rather forms of automation that lack human oversight. While he raises valid points about the risks of outsourcing cognitive tasks, the counterargument asserts that LLMs, by their very design, can indeed be wielded effectively with skill. The effective use of LLMs requires the same level of human judgment and expertise needed when using traditional tools, such as saws or chisels, which can amplify both good and bad decisions based on operator skill. The main takeaway highlights a crucial distinction: the ability to direct and refine the output of LLMs is essential for meaningful engagement, illustrating that these models can be effective tools when utilized thoughtfully. Despite concerns about unpredictability and reliability, the fundamental qualities that define a tool should include the operator's capacity to engage with the technology critically. Moreover, dismissing LLMs as mere automation overlooks the nuanced feedback loops and intentionality that skilled users bring to bear, reinforcing the importance of human oversight in leveraging AI technologies effectively.