In Search of the Perfect Note-Taking System: A Decade of Trial and Error
Reading time: 15 minutes
Summary: Throughout my 10-year journey with note-taking, I realized the importance of balancing efficiency with memory retention. Starting from not taking notes at all, I evolved to trying various methods, each with its own set of challenges, ultimately leading to a need for a more efficient and organized approach.
The Beginning of My Note-Taking Journey
I've always struggled with note-taking, a challenge that began about 10 years ago when I started my reading journey. Initially, I didn't take notes at all, believing they were unnecessary. I wanted to enjoy reading without extra burdens. However, I soon realized this approach was a mistake. Without notes, retaining much of anything became difficult. After all, what's the point of reading if you don't remember the content?
I then explored various note-taking strategies. The common advice was to create summaries in my own words, draw connections between the content, and use mind maps. These techniques were genuinely effective. Active engagement with the material helped me understand and remember it better, and I ended up with valuable learning materials for future reference. However, this method was incredibly time-consuming and, frankly, I despised the process. It felt like a return to school, with the obligation to study material. Although effective in theory, it made me reluctant to read, so I soon abandoned this approach and resumed reading without taking notes, leading to the inevitable consequence of forgetting almost everything.
Finding Simplicity
Eventually, I realized this couldn't continue and decided to give note-taking another shot, this time aiming for a simpler process. I was willing to accept a trade-off in memorization quality. My new method involved tracking passages I found important. Whenever I encountered a crucial argument or piece of information, I noted the page and paragraph in a Google Keep file. At the time, writing in my books was an unthinkable act, as I regarded them as precious and sacred, to be kept in pristine condition. Thankfully, I've since moved beyond such snobbery.
This method initially seemed effective for my purposes; it allowed me to save the most important aspects of a book. However, I soon realized that revisiting this material was challenging. To find something specific, I had to open my file, locate the notes, and then read through everything until I found what I was looking for. The number of times I actually went through this process is almost zero. This experience led me to understand that simply tracking important passages wasn't sufficient—they needed to be searchable. But how do you make notes from paperbacks searchable?
Searchability And Digitalization
In theory, there are methods like using physical markers or stickers, implementing titles and color coding, etc. But this brought me back to my initial problem: the process is time-consuming and tedious. The easiest way to make something searchable is to have a digital version, where you can simply type in the word you're looking for. For instance, if I want to revisit something about Finland, I just search for 'Finland'; it's straightforward. However, the issue is that I exclusively read paperbacks. I find it hard to focus on ebooks and lose concentration more quickly. So, I found myself at an impasse.
To resolve my dilemma, I resorted to the only feasible solution I could think of: digitizing all the passages I deemed important. Initially, I continued to simply track the locations of these passages in a file, as before. However, after finishing a book, I would go to each marked passage and start typing it out. This process resulted in a separate file for each book, filled with all the crucial information I had noted. This compilation of the most valuable insights from each book felt like a breakthrough in my note-taking journey. It seemed like I had finally mastered the art of note-taking.
Yet, this method proved to be quite tedious and time-consuming. While it wasn't as demanding as creating my own summaries or mind maps, it was far from enjoyable. Each time I finished a book, I dreaded the task of typing out everything I had saved. Despite this, it became my most-used system. For many years, I applied this method to hundreds of books, and I still retain all these files. It's astonishing to reflect on the wealth of valuable information accumulated in them.
The main drawback, however, was the significant amount of time it required. Yet, this was a testament to how important it was for me to preserve this information, and I was willing to make that sacrifice. Despite monotonous, at least typing out the existing paragraphs was less mentally taxing compared to creating my own notes in a more proper sense. Creating summaries and mind maps required considerably more mental effort and were far more challenging. The reason for typing the text manually was that the optical word recognition technology wasn't as advanced back then. I found correcting its errors so frustrating that I preferred manual typing, and being a fast typer helps. While OCR technology has improved significantly since, the process remains a little tedious.
Improvements
Eventually, I devised a more efficient approach to streamline this process. Instead of typing out passages manually, I began copying them from the ebook version. Although I only owned the physical copies, I could find the ebook versions online through less-than-legal means. When noting the location of a passage in my physical book, I started recording the first and last few words of the passage. This made it easier to search for the text in the PDF or EPUB file, allowing me to copy and paste it directly into my document.
This method was significantly faster than manual typing, though not as quick as one might expect. Locating the book, finding each passage, and copying it still consumed time and was quite tedious. Additionally, copying from PDFs often messed up the formatting, as each new line was interpreted as a new paragraph, requiring extra time to correct.
Remaining Problems
Even with these documents being invaluable sources of information, consulting them posed its own set of problems. For instance, I recently opened my file on "Becoming Human: A Theory of Ontogeny" by Michael Tomasello, a fascinating but dense book on human development and the establishment of distinctly human traits in early childhood. The file contains my 23 notes highlighting the most important and worthwhile passages. Even though I read this book nearly four years ago, having this file is incredible. Reading through it feels almost as beneficial as re-reading the entire book, but far more efficient.
The issue is that these 23 notes, despite being so carefully curated and representing only a fraction of the whole book, still amount to a lot of content—around 6000 words. With an average reading speed, it takes me about half an hour to go through them. This is much shorter than the 10-12 hours it would take to read the entire book, but half an hour is still substantial if I just want to quickly revisit a specific topic.
Searchability Challenge
The additional challenge I faced was ensuring the searchability of my digital notes. A key advantage I sought by digitizing everything was the ability to perform typical CTRL+F style searches. However, this wasn't as straightforward as I had anticipated. First, I had to remember the exact keyword for the search to be effective. Variations or synonyms didn't yield results, but it was often challenging to recall the specific term used in the text.
For instance, in the "Becoming Human" book I mentioned earlier, it extensively discussed human collaboration. However, I couldn't just search for phrases like "working together" or "collaboration" because the book used the term "cooperation." Unless I remembered that exact word, my search would be fruitless.
Expanding the Search Scope
Another issue was that even if I managed to use the right keyword, it would only search within that specific book. If I wanted to explore the topic of human cooperation more broadly, ideally, I should be able to search across all the books I've ever read. Many other books I've read also touch on this subject, and accessing them all would be beneficial. A potential solution was to combine all my documents into one file, creating a comprehensive resource.
However, this approach also had its problems. The resulting file was enormous, and broader searches often yielded too many matches to be useful. Most terms, unless highly specialized or technical, would return hundreds of results, making it difficult to sift through. While there were minor workarounds and tweaks, they invariably circled back to the same problem: they required more effort and time.
Despite the challenges, I continued to use this system for years. I experimented with other methods and variations of my own approach, but none proved superior. I consistently encountered an insurmountable problem: either the system generated thorough notes but was too time-consuming to justify, or it was quick and efficient but the notes produced were of little use.
The Cost of Time
I've come to realize that time is a factor that's often undervalued. Many people are willing to adopt time-intensive systems for the benefits they offer, but not all of these systems are worth the cost. What's crucial is a cost-benefit analysis of each alternative. In my case, devoting a significant amount of time to note-taking was a steep price, and I approached this decision with extreme caution.
Firstly, reading is paramount for me, and if a note-taking system makes me dread reading due to its time-consuming nature, it's clearly not the right choice. Secondly, the purpose of note-taking is to enhance learning and memorization. However, another effective way to better learn and remember information is through extra reading.
Revisting and using notes from a book on evolutionary psychology can solidify understanding and retention, but so can just re-reading the book. Spacing out re-readings can be particularly beneficial and has been shown to be highly effective for learning and memorization.
This argument becomes even stronger when considering reading a different book on the same topic. Not only do you benefit from improved learning and memorization, but you also gain exposure to a different perspective. This approach keeps the material interesting, provides a more comprehensive understanding of the topic, and minimizes the errors and biases inherent in a single source. In short, the issue of time is not trivial and involves a multitude of trade-offs and considerations.
This system, while creating a massive, personally curated database of knowledge, proved frustrating due to its impracticality. I used these documents mainly when writing book reviews, helping me to revisit and better understand the material. But once the review was published, I rarely returned to the files. I often wanted to, but because of the difficulty in using it in any meaningful sense, I dreaded it and avoided it. I knew the time and effort it would take to locate the specific information I wanted.
The Birth of Modern LLMs
In late 2022 I started hearing about ChatGPT. Initially, I didn’t think much of it. Despite my long-standing interest in artificial intelligence, nothing I had seen in language processing models had impressed me, and I wrongly thought it was the wrong approach. However, after encountering some Twitter posts showcasing its capabilities, my curiosity was piqued, and I began experimenting with it. Even before the advent of GPT-4, I was amazed at its power.
Since then, I've used it daily, accumulating hundreds of hours of experience. I was particularly struck by its ability to summarize content effectively and its pragmatic understanding of language, if even if in a strict philosophical sense that claim is dubious. I started to see how Large Language Models (LLMs) could be the solution to my long-standing frustrations with traditional note-taking methods.
Previously, I had tried optimizing my notes by assigning titles and tags to passages, which made the documents more searchable. However, this required substantial additional time and effort, which I didn’t find worth it. But LLMs seemed almost magical in their ability to generate accurate summaries and context-relevant titles and tags. I began experimenting with feeding notes to ChatGPT to create titles, summaries, and tags for each note.
While this wasn’t very efficient and still time-consuming for each note, I realized this process could be automated. I envisioned some kind of platform built upon this concept, where one could read just the titles, click to expand summaries, and further into the original notes. I started to think of an app designed to create the most efficient note-taking system possible.
The Power of AI
This part of my journey marks a significant shift, where the emergence of LLMs opened up new horizons. Here's a glimpse into how integrating AI technology aligns with and enhances the note-taking solutions I've been seeking all along:
Digitalization: A perennial challenge was converting handwritten or printed notes into digital format. Integrating OCR technology into an app could vastly simplify this process. This app would not just convert text, but also address common formatting issues, perhaps with the aid of algorithms or ChatGPT itself.
Efficient Search: Another issue was the cumbersome task of locating specific notes among many. A system where the primary content I read were summaries or titles would greatly expedite this process. These elements could serve as quick identifiers, allowing me to locate and delve into the original text as needed. For instance to locate a specific note of interest in book previously mentioned, instead of having to read 26 passages of text, I now only have to read 26 titles. The search time is reduced from 30 mins to 46 seconds.
Contextual Understanding: The introduction of LLMs brought the possibility of meaning-based search. This means a search for a term like 'collaboration' could also bring up results for related terms like 'cooperation', thanks to the model’s ability to understand and relate concepts.
Right-Sized Searchable Database: Finding an efficient way to search through my notes was always a challenge. A single book provided too narrow a scope, while searching across all books was too broad. However, categorizing information by topic, which was previously a time-consuming task, could now be automated.
Building the Idea
I was incredibly excited about this idea, but two major obstacles stood in my way. Firstly, I didn't know how to code, so how would I build this app? Secondly, I had some experience with startups and understood the immense challenges involved in launching a successful business. It's a significant commitment, often without any guarantee of success. Having a good idea is just the starting point; building a team, developing the product, raising funds, marketing it, and generating revenue are incredibly difficult tasks. Despite recognizing the potential of my idea, these challenges led me to initially set it aside. I didn’t want to risk thousands of hours of my life for something that may not work out.
However, everything changed when I started working on my thesis for my Master's degree in Neuroscience a few months later. The process of reading numerous papers and making highlights for essential information brought me back to my struggles with book note-taking, but applied in academia. I had already experienced this with my undergraduate degree in Philosophy and Psychology, but the amount of information required for a thesis was another level entirely. This was when I realized the broader application of my idea. It wasn't just about book notes; it was about organizing information, which is a widespread need. It applies not only to readers, but to academics, writers, journalists, and much more.
Writing my thesis ended up being a very frustrating process. Not only because of the usual difficulty in writing it, common to every student, but in my case it was extra frustrating because I knew that if the app I had envisioned existed, it would reduce my thesis research and reading time by manifold. This realization was the catalyst for me to embrace the risks and challenges of starting what is now known as Raven.
Lacking coding skills, I reached out to a developer friend who resonated with my idea and agreed to join as the technical co-founder. Together, we embarked on a journey to transform this vision into a tangible reality.
Conclusion
I've written this post to illuminate my personal journey with note-taking, and while introducing Raven is a part of my story, promoting it isn't my primary intention here.
I recognize that my approach to note-taking and the emphasis I place on time efficiency may not resonate with everyone. Yet, in my discussions with numerous individuals about Raven's concept and the broader challenges of traditional note-taking, I've found that many share my frustrations and are curious about trying the app. If you're intrigued, feel free to join the waitlist.
Lastly, for the sake of transparency, I must note that this narrative is a simplified version of my experiences. I experimented with various other note-taking methods that aren't detailed here, and there are numerous other factors that played a role in my journey and in Raven. I've condensed my story to maintain brevity in this post.
Thank you for taking the time to read, and I sincerely hope you discover a note-taking system that suits you well. Ideally one that doesn't take a decade to perfect…
