Author argues writing is essential for true thinking in LLM era
A blogger argues that writing is a critical tool for processing information and that Large Language Models lack the human intuition required for creative leaps.
The author began taking notes in middle school to help understand long passages, eventually realizing that creating notes aids in processing information rather than merely storing it. After spending ten years exploring why writing helps thinking, the author notes that thoughts not written down often fail to stick because they were never fully formed. This perspective aligns with advice from Manuel Blum, who told graduate students that Turing machines are more powerful than finite automata because they have paper and pencil. Blum claimed that with writing, a person has the power of a Turing machine. Leslie Lamport similarly stated that if you think without writing, you only think you're thinking.
The author suspects that Large Language Models inherit intelligence because human intelligence is deposited in language. However, they argue that LLMs cannot take the creative leap to make new connections because they lack half-formed human intuition. The author notes that Amazon's six-pager memo is no longer proof of thinking in the age of LLMs. They advise writers to produce text in bulk without restraints during the first phase and to keep their writing simple, plain, and succinct. The author has been writing and sharing on their blog since 2011.
Sources
Every sentence of this story was checked against facts quoted from these articles. How we make the news
Spotted an error? Tell us
