Evergreen Notes Aren't Enough: The Value of Connections and the Case for Public Knowledge Graphs

"If you had to set one metric to use as a leading indicator for yourself as a knowledge worker, the best I know might be the number of Evergreen notes written per day," writes former KhanAcademy R&D chief Andy Matuschak.

Most people take transient notes to get down thoughts or ideas in a specific context. These notes are useless after that context passes by. More importantly, this kind of notetaking encourages transient thinking rather than . Evergreen notes, on the other hand, aren't bound to a specific moment or project: they store pieces of knowledge to be shared and referenced across all work and thinking. Consequently, taking evergreen notes encourages the development of a much larger and more interconnected personal knowledge graph.

These are the same conclusions I arrived at through the philosophy of learning in public. "...among all notetaking strategies, the only one that actually simplifies knowledge management and makes it more effective is to publish your learning, experiences, and insights in public," I wrote in January, a month when I published 20 posts on my blog with the guiding rule that "I haven't learned something until I've published writing about it."

Evergreen notes are great for capturing ideas, but on their own they're not enough. Just as crucial as evergreen ideas are the connections between them. This is fundamental to how human intelligence operates[^nosource]: innovation and creativity aren't contained in experiences and facts themselves, but rather in the connections between them. Out of repeated observations emerge patterns, which we call insights.

"Your brain is like a solar system," Ideaflow engineer Linus Lee puts it. "When you first hear about an idea it's like an asteroid that gets thrown into orbit somewhere. Over time, as the idea finds other related ideas, their gravitational pull pulls them together into something bigger; over time it accretes mass and falls into orbit somewhere close to the "star" which are the things you think about every day. [Knowledge capture] is about increasing the number of asteroids you retain. [Knowledge organization] is about expanding the gravitational field of your star."

This is nowhere near overlooked by knowledge management thought leaders, of course. Highly backlinked knowledge graphs are worshipped, fundamental to products like Roam, Obsidian, and Andy Matuschak's website.

What I think is overlooked is the potential of a global knowledge graph, rather than a purely personal one. Tools like Roam and Obsidian claim to provide you a "second brain" to consolidate all your thoughts in, and if you're diligent, to build connections between. The extent of your graph is purely within the bounds of your own head. Yet, if you think about the connections that you make, I'll bet that almost all of them come from external sources and interactions: a conversation; an article you read; an event that occurs in the world. When you have such an external encounter, you create a node in your personal knowledge graph, then link your previous internal notes to it.

To me, this seems like a massive and unnecessary bottleneck in the biulding of knowledge graphs, both for individuals and collectively. What if, instead of having millions of silo'd individually maintained knowledge graphs, we had one global graph that millions of people contributed to? The pace of knowledge work, and the development of new ideas and connections, would skyrocket.

Such a global knowledge graph already exists, of course. There are many forms of it: published papers, books, pamphlets, available to the entire public for consumption and contribution (what segments of the public are not able to consume or contribute to this knowledge graph is also important to analyze, but I digress). The internet in particular has been a huge facilitator of this type of knowledge graph. This very blog is a node directly linked to several of Andy Matuschak's public nodes, as well as a previous public node of my own, and a private-made-public one of my conversation with Linus.

My thesis in my learning in public blog post, refined now, is that contributing to the public knowledge graph is what all private knowledge graphs should strive towards. This is beneficial to the advancement of all human knowledge, of course, but I believe it's also the optimal method of developing your own knowledge graph.

The global knowledge graph contains many, many more nodes than any private knowledge graph; but more importantly, it also shoots off connections in a million different directions at all times. Frameworks seek applications, solutions seek validation. In a private knowledge graph, you are the only agent building connections. In the global knowledge graph, there are millions of agents building connections, strenghtening existing ones or pushing them in new directions. The more of a channel you create between your private knowledge graph and the public human knowledge graph, the more you will be able to make use of the connections shooting out of the public one to build your own connections.

What does this look like concretely? For me, it means publishing everything I learn on my blog. Notes kept private, even evergreen ones, are useless to me, except for their facilitation of the creation of public notes. Others curating more sensitive knowledge and connections may desire to keep their notes private for longer, or share them at a semi-public level, i.e. within a team or company.

Beyond an individual scale, the fundamental value of the public knowledge graph is the driving thesis behind my development of Postulate, the site on which you're reading this post. Postulate is a notetaking tool, designed from public-down rather than private-up, reducing the friction of turning private notes into public evergreen ones as much as possible. I'm working on a separate write-up about Postulate, but for now I'll allow two screenshots to demonstrate my vision.

Here's Roam Research's landing page screenshot:



It's a nicely crafted place to write snippets of notes, linking them to each other to form a tightly interconnected personal knowledge graph.

The project interface of Postulate is somewhat similar, containing discrete notes connected temporally and through tags:



But you, a public visitor to my Postulate project, do not see these snippets. You see this post, a public evergreen note rather than a snippet. The entire point of Postulate snippets is for them to be turned into posts, unifying notetaking and publishing into streamlined thinking workflow rather than being disjoint pieces. Here's what the post editor interface looks like:



Post on left, referenced snippets on the right.

For all my emphasis on connections, Postulate doesn't even currently support backlinking of any sort (it's on the roadmap!). Yet, I argue that it's a better facilitator of connection-building than Notion, Obsidian, Roam, Ideaflow, or any other private notetaking tool for all but the most diligent of users. This is because of the implicit connection-building power of simply publishing your work, and making use of the millions of agents in the public knowledge graph of humanity.

Revisiting Linus' solar system analogy: publishing your writing is like taking asteroids in your system, positioning them not in tight close orbits around your star, but in the farthest out orbits you can without losing them, maximizing impacts with asteroids outside of your solar system so that new materials and trajectories may be captured.

[^nosource]: ...based off of intuition and not any knowledge of neuroscience or psychology, though I'm guessing rigorous explanations exist for the same principle