The Dangers of ChatGPT and Excessive Note-taking
In recent years, I have gotten into the habit of using Obsidian for taking notes. Whilst it has been immensely useful (at least that’s how it feels when I use it), there is arguably very little output from all the incessant documentation I do in that app, and whilst the concept of a ‘second brain’ is very attractive to me, somehow I feel it is atrophying the first (brain).
Note-taking itself, whilst not inherently evil, certainly can give the feeble-minded of us a false sense of accomplishment - we feel that, by writing down our idea in the most basic of formats, we have already completed 90% of the project. The note is then filed away in our folder of ‘ground-breaking discoveries’, where it will be connected to our infinitely expanding knowledge graph of unfulfilled fantasies, giving us a little kick of dopamine as we do so, never manifesting into anything concrete, and befuddling our tiny vestigial rat brains the next time we happen upon it on one of our endless note-organisation quests.
The addition of ChatGPT and other language models to this workflow is even more nefarious. Prior to 2022, whenever I had to write something, I would actually sit down and type out full sentences, like I am doing right now. The fact that this concept seems so alien to me now (and I suspect many others too), is very telling. Nowadays I will just jot down the basic concept or outline of the thing I want to communicate, and get AI to do the rest. However, I have come to realise that by doing this, we lose something very important.
This is the first time I have written something out in longhand in who knows how long, and I am finding it very strange! I fear that the delegation of our thought formulation processes to external forces deprives us of many of the benefits of doing it ourselves. Not only do we reduce our ability to think and communicate coherently, paying attention to what we’re doing for long periods of time, we also deprive ourselves of the actual process of writing. This process can give rise to many ideas and thoughts which otherwise would have not come about, and by using a model that is trained on the collective online drivel of mankind, as well as potentially losing the accuracy, I think we may also be allowing our internal thought processes themselves to become corrupt.
That being said, I think there is huge value to modern ways of note-taking, like Obsidian. These workflows allow us to connect thoughts in a much more immediate way, making the type of lateral thinking that is the essence of creativity (which would have otherwise taken us years to develop by sifting through dusty handwritten tomes), much more accessible for the average person. At the same time, the use of AI for summarisation and contextual ‘grokking’ of large amounts of text data, is incredibly useful for research purposes; not to mention having a free consultant to discuss and stress test any of your ideas against in your pocket at any given moment - something that business people would have previously paid hundreds, if not thousands for, to consult with. Whether this ability actually produces any new work, at least in my case, remains to be seen.
I will make an effort to write longhand again this year, and see where the process of posting on this blog takes me. Inspired by my good friend Jack.
This article was generated by ChatGPT.