'AI' Is Not a Tool by Rich Hickey challenges the notion that large language models (LLMs) are tools. He argues that tools are hand-held implements requiring intentional control, feedback, and expertise to master. In contrast, LLMs are automation systems that operate without direct human control, continuous direction, or sensory feedback.
Hickey distinguishes between wielding a tool—exerting power and learning through practice—and using automation, which replaces human effort with passive requests. He notes that tools build expertise, reveal nuances, and foster deep understanding of tasks and the world, while LLMs lack predictability, deterministic function, and the intimate engagement that defines true tool use. The author emphasizes that relying on LLMs for tasks like coding or research means abdicating active learning, critical thinking, and the development of expertise.
He warns that treating automation as a tool risks reducing complex learning to superficial interactions, such as prompting a model to summarize papers without truly engaging with the material. Ultimately, Hickey asserts that the essence of tool use—control, mastery, and insight—is fundamentally absent from LLM interactions, making them fundamentally different from genuine tools.
स्रोत: Hacker News · HeadlinesBriefing द्वारा सारांशित