In this article, I lay out a vision of the dream tools to help content creators manage and apply their cumulative learnings. I outline the desirable properties of such tools and how we can build them.

Let's get started.



Knowledge Synthesis vs Passive Consumption

I use the term "content creators" to refer to a variety of knowledge workers, including bloggers, video producers, researchers, journalists, and more. The many types of content creators are characterized by a common basic workflow: the process of consuming and synthesizing knowledge. Or more generally, learning and teaching.

This is distinct from the mere passive consumption of information. It involves digesting information, trying to apply it, condensing it, sharing it, and making it your own. In the process, you understand the knowledge better yourself. And the output you create helps others learn.



Ease of Capture

As you read, learn, or just live life, you come across many tidbits and ideas you want to remember. These are inspirations and materials that you can later make use of.

But if you don't capture these ideas, you tend to lose them forever. And it is hard to generate relevant ideas on demand when you need them. So capturing is an always-on process, and your collection of ideas grows over time.

The challenge with capturing ideas is that it is hard to know in advance when you will need them. If the friction to capture is too high, it simply won't happen. The capturing process must be effortless.



Save from Anywhere

Ideas can come to you in many different contexts: conversations with friends, newsletters, books, podcasts, documentaries, walking a dog in the park, etc. The best ways to capture ideas are different for each scenario. You want to go with whichever method is the most convenient in the moment so you can resume whatever activity you were doing.

As technology evolves, we would have an increasing variety of platforms or media we can use to save ideas. Currently, you can write it down on a piece of paper, take notes on your phone, leave a bookmark, highlight the passage, take a screenshot, and more.

But in the future it seems inevitable that we would incorporate additional modes of inputs such as wearables like glasses with cameras, virtual reality recordings, and brain-computer interfaces.



Integrated and Queryable Knowledge

As you use multiple tools, a problem that arises over time is that your captured knowledge becomes scattered across different places. Using a fragmented set of tools to manage your knowledge is costly. It makes it harder to retrieve any information you need.

Currently, your book notes, product specs, and notes for online courses likely live in disparate platforms. We can do better. Your information should not be siloed. It should be searchable in one place.



Customizable and Extensible

It is unlikely that any single tool will be the best knowledge management solution for all scenarios and workflows. No matter how well-designed a product is, there are use cases that are not accounted for, or are de-prioritized due to resource constraints. The best knowledge management tools you use should be extensible to fit your nuanced needs. Each tool you use needs to work well with the other components in your workflow, over a shared standard, open source code, or API.



Free-form Digital Drawing

If I ask you to explain an abstract concept, one of the first things you would reach for is probably a pen, then paper or a whiteboard. This free-form drawing medium is expressive. It makes use of our spatial intuition. It is free-form. It helps you see otherwise non-obvious connections. But it is inconvenient to store, retrieve, and connect to existing knowledge.

In contrast, a digital interface is easy to manipulate across platforms, but tends to be limited in expression. It tends to rely on text representation, which is often insufficient in conveying an idea, especially a complicated or abstract one.

An ideal knowledge management system would marry the two to get us the best of both worlds: a digital interface that affords frictionless free-form drawing and is easy to maintain. Products such as Miro and OneNote are a good step in this direction.



Structured Knowledge & Personal Knowledge Graph

The most common digital form of an idea is a note , typically stored in plaintext and organized in notebooks or folders. This is usually sufficient for the purpose of jotting down your thoughts.

But these generic "Note" objects can become more useful when they are augmented with properties, relationships with other notes, and other metadata. These properties can be automatically extracted in the capturing process (e.g. "page number" and "book title" in a e-Book highlight), or custom-defined by the user (e.g. "years of experience required" on a "Job Posting" page).

The interconnections between these individual notes can be structured in a knowledge graph , where an entity "Google" is connected to a collection of "Job" entities via the relationship "is hiring". A knowledge graph can be valuable in retrieving knowledge and answering queries. And search engines make extensive use of this to structure the world's knowledge. And on an individual level, personal knowledge graphs can have many interesting use cases for retrieval and automation. Tools like Notion , Coda , and Roam Research offer more options for structuring knowledge for individual users. But this is just the beginning.



Autosuggest for Thoughts

With richer representations of knowledge, machines can better "understand" our thoughts and compute on them. This can help streamline (or even automate) much of the research and idea generation workflows.

Imagine this: when you are writing about a topic, your knowledge base can suggest semantically relevant content, either from your existing data, your team, or the collective knowledge on the web. You would no longer have to switch contexts to look up simple facts. Rather, thoughts can flow frictionlessly from inside your head to external digital artifacts that you can edit and share.

And with generative language models (such as the GPT ) maturing, you can provide a skeletal structure of your ideas, and have AI models complete your thoughts. Instead of guessing and generating random tokens, the model's outputs would be based on your thoughts that have already been digitized in your system.

With this close collaboration with machines in your knowledge work, information becomes truly at your fingertips.



Prefer simple over complex

Complexity can creep in any system. Technology tools tend to get bloated over time. An initially focused feature set can turn into a maze of confusing and loosely related features. A constant challenge for feature-rich tools is discoverability of functionalities without cluttering the UI. These tools are supposed to save you time. It would defeat the purpose of using the tool if it takes longer to make it work the way you want than to work without it.

To this end, Notion has raised the bar for simplicity for knowledge management tools. Figma also does a good job for hiding its rich feature set behind contextual actions.



Shared, Collaborative Knowledge Structures

Knowledge synthesis is collaborative by nature. You consume works created by others and in turn create yours by mixing in your knowledge and experience. Knowledge management systems can accelerate this collaboration and empower every individual to leverage and build on each other's work as much as possible.

Instead of knowledge being siloed within disciplines and locked inside individuals' minds, there is tremendous potential for tools to enable individuals to contribute to a collective knowledge structure, while saving time from having to build their own from scratch. The successes of platforms such as Wikipedia and open source software development are encouraging. We can build towards a future where individuals expose and attach parts of their private knowledge base to a public topic entity for others to fork for their own use. And this process can even be automated at some point.



We live in an exciting time for content creators and for innovation in the knowledge management space. The above are some of the promising directions for development, to work towards accelerating learning, making new discoveries, and making progress towards solving big problems.

To contribute to this vision, I recently open sourced an extensible web clipper browser extension , and I am working towards a few of the directions outlined above with Rumin . If you would like to chat about this further, feel free to get in touch on Twitter .