For many researchers and engineers, the act of writing is not a final step to summarize results but a primary mechanism for discovering what they are actually researching. This shift in perspective—from viewing writing as a 'boring epilogue' to a core cognitive tool—highlights how externalizing thoughts helps clarify the mind and structure complex systems.
Writing as a Discovery Mechanism
One commenter, writing as billypilgrim, shared that during a PhD, they discovered that the scientific process relied heavily on the ability to convey ideas to others. They noted that many research questions were only discovered during the writing phase, as the need for clarity of mind forced them to work through multiple drafts to reach a point efficiently.
The discussion emphasized that writing is a tool for deep thinking. billypilgrim argued that AI-assisted writing can be problematic because it may lead people to believe that writing is just a matter of producing text, rather than a method for thinking deeply about a topic.
External Storage and System Design
Several participants highlighted the importance of physical media in the thinking process. One user, hoppp, described how most projects begin as scribbles on paper or system diagrams with boxes and arrows to describe a thought process before a single line of code is written.
Another commenter, noir_lord, noted that despite trying various tools, they consistently returned to a simple A4 pad and ballpoint pens. They observed that the 'complete free form nature' of pen and paper provided a cognitive benefit that digital tools like an iPad could not fully replicate.
Technical Analogies and Cognitive Risks
The discussion included a technical comparison between Finite Automata (FA) and Turing Machines (TM). One commenter, yubblegum, suggested that if a human aspires to 'TM-hood,' the guidance is to write, re-read, and re-evaluate thoughts, noting that this process is subject to the halting problem.
Conversely, baxtr warned of the pitfalls of this approach, noting that for highly analytical people, writing can become counterproductive. They argued that it can lead to 'pseudo-rationalizing' if the writer focuses on continuously elaborating thoughts instead of using the writing to challenge assumptions or test hypotheses.
The Role of AI and Audience Tailoring
The distinction between human-authored drafts and AI output was a recurring theme. One commenter, jbs789, found it difficult to read AI-generated words because they lacked the 'weight' of human thought, which can lead to giving the text more credit than it deserves.
Additionally, sublinear pointed out that human writing involves tailoring a message to an audience's 'pain points'—a nuance that AI may struggle with because it lacks the ability to perform a 'vibe check' of a room.