TL;DR
Get the little things that make your day delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
A November 2026 blog post argues that writing is not just a way to record ideas: putting thoughts on the page can help people develop, test and remember them. The author says AI-generated text may look like evidence of careful thought without doing that work for its user, and recommends starting with an unrestricted draft.
A researcher writing on the Murat Buffalo blog argues that writing can function as a tool for thinking, not merely as a record of ideas, and says the distinction matters as people increasingly use large language models to generate text. The post draws on the author’s experience with note-taking and research writing, but presents a personal argument rather than new experimental evidence.
The author traces the idea to middle school, when reading long passages did not seem to result in lasting understanding. Taking notes, including elaborate mind maps, helped make sense of material by forcing the author to decide how each idea related to others. The notes, the writer later concluded, were useful less as storage than as a way to process information.
The essay connects that experience to advice attributed to computer scientist Manuel Blum. Blum compares finite automata, which have limited computational abilities, with Turing machines, which can compute any computable function and use a tape. The post treats the tape as an analogy for paper: writing gives a person a place to hold intermediate thoughts, making it possible to work through problems that would be difficult to manage mentally alone. It likens this to external memory and “out-of-core” computing.
The author says the effect becomes clear when reviewing research papers: a paper may seem understood while reading, but writing a review can reveal gaps in comprehension and help place the work in a wider context. The post also recounts a 2018 blog entry about similarities between Paxos and Nakamoto consensus. The author says the connection became clearer during the writing process and later helped seed a research paper. These examples are personal accounts, not evidence that writing will produce the same outcome for every reader.
How Writing Changes the Thinking Process
The essay’s practical point is that drafting can be part of reasoning. Writing makes ideas visible and revisable, allowing a person to notice contradictions, supply missing steps and connect concepts that might otherwise remain vague. That can matter to students, researchers and professionals whose work depends on explaining complex material, not just recalling it.
The argument also distinguishes between using a tool to support thought and using it to replace thought. The author says language models may be able to summarize literature and reproduce familiar frameworks, while still lacking access to a user’s unformed intuitions. The essay does not establish a general limit on what AI systems can do; it warns that accepting generated prose without developing or checking the underlying ideas can create a misleading appearance of understanding.
Products Worth Considering
eufy Security 4G LTE Cam S330 4K Solar Outdoor Camera with Pan and Tilt, AI Tracking, SIM Card and 32GB SD Card
Philips Hue White and Color Ambiance Smart Retrofit Recessed LED Downlights, 5/6-Inch, 4-Pack
Roborock Q7 M5+ Robot Vacuum and Mop with Self-Emptying Dock
As an affiliate, we earn on qualifying purchases.
From Mind Maps to AI Drafting
The post begins with a specific learning problem: the author could read a page without retaining its contents. In the late 1980s, after encountering Tony Buzan’s mind maps in Turkey, the author began making extensive branching notes. The method helped organize reading by requiring decisions about where each idea belonged.
Later, during a period of PhD procrastination, the author encountered Blum’s comparison of finite automata and Turing machines. The essay uses the computer science analogy to explain why paper can extend working thought: a writer can preserve intermediate results and return to them rather than holding every detail in mind. It also invokes freewriting, in which a person produces an unpolished draft before trying to shape it.
The post brings that argument into the current debate over AI-assisted writing. It references Amazon’s six-page narrative memo as an example associated with writing for clarity and quotes Jeff Bezos’s view that producing such a memo requires clear thinking. The author argues that language models weaken the memo’s value as a signal of thought because polished pages can now be generated without the writer doing comparable intellectual work.
“Without writing, you are reduced to a finite automaton. With writing you have the extraordinary power of a Turing machine.”
— Murat Buffalo blog author, quoting computer scientist Manuel Blum
What the Essay Does Not Establish
The post is a personal essay and argument, not a controlled study. It does not measure how much writing improves comprehension or memory, compare writing with other learning methods, or establish how broadly the author’s experiences apply. Its account of AI is also a judgment about current systems and their use, not a demonstrated finding that models cannot make creative connections.
The essay does not give data on how often people use AI to draft, whether generated writing reduces users’ independent reasoning, or which forms of AI assistance might support rather than displace thought. Those questions remain open in the source material.
A Practical Case for Rough Drafts
The author’s immediate advice is to begin with an unrestricted draft: write in bulk, tolerate errors and fragments, and postpone editing. The aim is to create material that can be examined and improved, rather than trying to produce polished prose at the outset. The post recommends keeping a working notes file during research, using it as a lab notebook for questions, tangents and emerging connections.
That recommendation is not a reported policy change or a scheduled research milestone. The next step for readers is practical: write through an idea before asking a model to frame it, then use AI assistance, if desired, in ways that do not obscure who did the reasoning or whether the claims hold up. The essay leaves the broader effects of AI-assisted writing unresolved.
Key Questions
What is the essay’s main argument?
It argues that writing can help generate and test ideas, rather than simply preserve thoughts already formed. The author says putting ideas on a page makes them easier to organize, revisit and connect.
The analogy is that paper stores intermediate thoughts outside the mind. That external record can help a person work through more steps than they could comfortably hold in memory at once.
Does the post prove that writing improves thinking for everyone?
No. It offers the author’s experiences and interpretation, including examples from reading and research, but does not report a controlled study or quantify the effect.
What does the author recommend doing before editing?
The author recommends producing a rough draft freely, without stopping to correct errors or make every sentence polished. The draft then gives the writer something concrete to assess and revise.
Does the essay say people should not use AI for writing?
No. It warns against relying on generated text in place of working through one’s own ideas. It does not set out a rule against AI assistance or evaluate particular tools.
Source: muratbuffalo.blogspot.com
