Table of Contents
Note-taking apps have gained immense popularity across a wide spectrum of people and professionals, transcending age and occupational boundaries. Whether you’re a university student, a journalist, or a software developer, the allure of these digital tools is undeniable. The promise they hold is exciting: a more organized and stress-free life, increased productivity, and a smoother journey through the complexities of everyday existence.
Simultaneously, note-taking, once a deeply personal endeavor, has undergone a transformation. It’s now a collaborative practice. The advent of team workspace platforms like Confluence, Notion, and Coda has fostered shared knowledge management among diverse entities, ranging from next-generation startups to large enterprises.
Despite the promises, modern note-taking apps have fallen short of fully realizing them. Irregardless of specific workflows or favoring any particular note-taking app, in this blog, we try to answer the question: what are some of the problems of modern note-taking apps, and how can we build them differently to overcome their shortcomings?
We begin by defining three levels of note-taking goals, ranging from the fundamental task of gathering information to the loftier aspiration of enhancing productivity and fostering creativity. Next, we present three problems associated with each of these goals, examining the shortcomings of current note-taking apps and the hurdles they face in fully realizing these objectives. Finally, we wrap up this post by imagining the future of note taking.
Three levels of note-taking goals
Level 1: Capture information
At its most basic level, note taking captures information. It’s a fundamental process and it can be as simple as grabbing a pen and jotting down your thoughts on any medium from a napkin, a takeout box to a digital notebook etc. These informal, short-term notes are usually categorized into separate buckets both mentally and physically. You have sticky notes all over the place on your desk, you jot down quick notes on your iPhone, you add short comments on YouTube videos, Substack blogs – the information is scattered and as it accumulates, it becomes clear that merely recording information is no longer sufficient; you must also be able to retrieve specific information when you need it.
Level 2: Facilitate information retrieval
The second level of note-taking aims to make it easier for users to retrieve valuable information quickly, tailored to specific contexts. Whether you’re studying for an exam, crafting an essay, or tracking events in your personal life, efficient retrieval of relevant information is crucial. In this context, the vast expanse of the internet can be seen as a behemoth notebook, with a wealth of “notes” in the form of blogs, news, wikis, videos, or audio transcriptions etc.
Currently, two approaches are widely adopted to achieve this goal. The first is through organization (“indexing”), which involves structuring (sometimes called indexing) information for easy access. Google indexes more than 130 trillion web pages in its database and has been continuously capturing (crawling) and indexing new content on the Internet. Its search engine can retrieve relevant information in milliseconds, whereas an average knowledge worker would take 1-3 hours daily just trying to find information in a particular document, about a million times slower. Google has indeed done an amazing feat. The second approach is through knowledge synthesizing & distillation (Note that this differs from distillation in the context of generative AI), rather, a process of us internalizing information captured from diverse sources.
Before RLHF (Reinforcement Learning from Human Feedback) made LLMs (Large Language Models) more aligned to human values, knowledge distillation mainly involved humans spending hours sifting through webpages, Reddit threads, and other sources of information. LLMs have the ability to ingest vast amounts of human knowledge from the entire internet, summarizing and synthesizing it into well-rounded opinions and concise knowledge for consumption. Google’s newly-released Generative AI Search Experience is a good example on how AI can synthesize information for us and provide more useful context.
On a personal level, note-taking apps should expedite the process of interacting with our knowledge. If we can’t efficiently retrieve or search through our notes, the value of taking notes diminishes significantly. After all, what’s the point of capturing information if you can’t find what you’re looking for? It’s akin to jotting something down and immediately tossing it into a pile of trash. Our memory is the bottleneck that limits us from the abundance of information in both the real and digital world.
Level 3: Promote productivity and creativity
The third and most ambitious level of goals is to boost productivity and creativity. Note-taking apps focusing on achieving this goal are often referred to as Personal Knowledge Management (PKM) tools or Second Brains. Our cognitive abilities have struggled to keep pace with the rapid growth of digital content. How much can a note-taking app elevate our thinking, enhance our productivity, and accelerate our learning compared to traditional paper notebooks? In theory, it can take us far.
Drawing on the assimilation theory of learning, developed by David Ausubel in the 1960s, we understand that new knowledge is most effectively integrated when it can be connected to existing mental structures. In other words, we learn most efficiently when we can link new information to what’s already understood. Creativity, too, revolves around connecting our existing knowledge with new ones.
Imagine a note-taking app that can help you connect the dots, becoming your thinking partner. As you accumulate more notes in your digital notebook, it evolves into a tool that comprehends the context in which you write and read notes, ultimately synthesizing new insights from your past notes.
The problems of note-taking apps
Problem 1: Context loss
Still remember the first level of goals about capturing information? It seems to be a trivial task at first sight. Afterall, you don’t even need a digital note-taking app to do that. A problem with it is capturing the context, additional information that surrounds a focal event.
Our ability to remember is greatly enhanced when we can place our thinking within a meaningful context. It acts as a mental anchor, allowing our brains to focus on what’s relevant and retrieve memories more efficiently. While creativity often demands thinking outside-the-box, context helps us recollect information at a local level, making it an invaluable tool for memory recall.
However, the nature of information is multimodal; it encompasses not just text and visuals but also sensory experiences like touch (haptic sense) and smell. While textual and visual information are ubiquitous on digital platforms, other modes of information are challenging to replicate. Imagine capturing a hiking trip on a video – much of the context gets lost in translation, leading to “context loss.” Similarly, when documenting a recipe, the associated context could include the time and place of preparation, the method used, and how it tasted. Although most information can be condensed into text or visuals, certain aspects, like the aroma and taste of a meal, or the ambient environment in which you cooked the dish, remain challenging to recreate using language.
Note-taking’s primary objective is to capture important information. Without adequate context, notes become meaningless strings of characters, making it challenging to relate the information to our past experiences. Conversely, an excess of context does not necessarily improve the quality of what information provides.
Problem 2: Paradox of choice
As mentioned in the second level of note-taking goals, information retrieval builds upon the foundation of organizing and distilling captured information. Many note-taking apps offer web clipper feature, allowing users to save web content as notes. While this feature is empowering, it can also be overwhelming and distracting. Digital hoarding becomes a real concern as individuals save content while fearing they might miss out on valuable information.
We’ve all been there – browsing social media platforms like Instagram or LinkedIn, coming across seemingly great content that we can’t process immediately. It’s so tempting to click that save icon and think you will come back to it later. But then we forget it. Today, we average about half the productivity growth rate today that we saw in the 1950s and ’60s. Meanwhile, generative AI has made it simpler than ever to fabricate misinformation. Ezra Klein, the founder of Vox and the author of the book “Why We’re Polarized”, wrote about his opinions on potential risks posed by AI on human cognition and productivity where he talked about the increasing overhead cost for humans to parse through AI-generated content in important areas like our justice system.
The convenience of saving freely accessible content in digital note-taking apps has led to indiscriminate content consumption, resulting in feelings of unhappiness, stress, and powerlessness. This is what we call the Paradox of Choice.
Problem 3: The unautomatable mind
Thinking is an active pursuit — one that often happens when you are spending long stretches of time staring into space, then writing a bit, and then staring into space a bit more. It’s here here that the connections are made and the insights are formed. And it is a process that stubbornly resists automation. — Casey Newton, Why note-taking apps don’t make us smarter
The concept of building a Second Brain has gained popularity, but I find the terminology somewhat misleading. A brain doesn’t just capture, organize, and distill information; it excels at making connections, turning information into knowledge, and ultimately expressing it in a creative manner. Andy Matuschak points out that note-taking apps emphasize the display and manipulation of notes but often miss the mark when it comes to making meaningful connections between them. While many note-taking apps promise increased creativity through AI-powered editors and customizable features, they sometimes lose sight of their true purpose – to serve the user.
Instead of obsessing over questions like “How should I structure my notes?” or “What layout should I use for my Notion homepage?”, the more pertinent questions should revolve around discovering workflows that foster the development of insights over time. Note-taking apps cannot provide a one-size-fits-all solution, and your workflow will largely depend on your individual preferences. Some apps make it easier to align with your style, but the true power lies in what the workflow produces.
In fact, most effective thinkers don’t write about how they organize journals, nor do they dazzle you with some fancy-looking templates and a massive knowledge base.
The future
In this part, we will discuss some potential features of a future note-taking app. Generally, a future note-taking app should have four fundamental features:
- Autonomous information capture
- Self organization
- Knowledge fusion
- “Connecting the dots”
Autonomous information capture
Capturing information can be a huge overhead. A future note-taking app would automate this process and capture vast amounts of information from your everyday experience: every piece of content you consumed online, every conversation with your co-workers during happy hours, every sport game with your friends, etc.
Self organization
We spend a lot of time structuring and organizing our notes using a top-down approach. However, ideas and knowledge emerge from atomic concepts in a bottom-up fashion. The new generation of note-taking apps should be able to structure the vast amount of information captured earlier in a meaningful way with an ever-evolving ontology.
Knowledge fusion
Knowledge fusion is about subtraction. If you have captured redundant information, a future note-taking app should consolidate different information using an external knowledge graph and delete knowledge deemed irrelevant.
Here we show an example of using external knowledge source to “fuse” multiple notes about Taylor Swift into one note. Ideally, this will not be limited to factual knowledge. More abstract concepts can be fused as well.
“Connecting the dots”
A future note-taking app should seamlessly connect your existing knowledge with the unknown, functioning as a thinking partner that can generate fresh, inspirational ideas. Furthermore, it should adeptly identify your knowledge gaps and offer learning recommendations to fortify your existing knowledge network. For instance, if a user has taken notes on topics like ChatGPT and Python, the knowledge assistant could bridge these two concepts by suggesting a learning path from Python to NLP, then progressing to neural language models (GloVe, Word2Vec, RNN, LSTM, ELMo), training for text embeddings, transformer architecture, GPT, and finally reaching ChatGPT. As a user, you will gain a deeper understanding on how Python and ChatGPT can be connected. There are usually other ways to connect entities. The figure below depicts another path: Python → NLP → LLM → ChatGPT, which is much shorter and might be more suitable for users don’t want to dig too deep.
Another compelling use case involves nurturing creativity, where the note-taking app predicts what new, unfamiliar knowledge could spark the most connections with your existing knowledge base. Consider the story of Spencer Silver, a scientist at 3M in 1968, who aimed to create a super-strong adhesive but serendipitously crafted a weak yet reusable one. Years later, his colleague Art Fry found a practical application for this weak adhesive by inventing sticky bookmarks, leading to the birth of Post-it Notes. Similarly, in the late summer of 1666, Isaac Newton had his first insight into the theory of gravitation while observing an apple fall from a tree.
Creativity is a beautiful process, often born from seemingly unrelated events. The ultimate question facing the next-generation note-taking apps is this: How can it foresee the significance of an event akin to the “falling apple” to Newton based on our unique experiences?
Acknowledgement
I’d like to thank the amazing early reviewers who help make this post better: Julien Zhu and Ian Webster.