This year my life unexpectedly moved almost 100% online. Both my job and classes have been fully remote for months. So my working and learning processes have changed. In this article, I will cover the latter.

It all happened at a turning point for me. I am going to finish a master's in Machine Learning & AI. I was already preparing to reinvest that study time more independently, but this unexpected online migration has opened up entirely new perspectives. Here are some personal considerations.

Knowledge has changed

Online information is fast, cheap, asynchronous, heterogeneous, fragmented. The opposite of the offline information I have been used to for nearly 20 years in school. I have realized that this difference has implications for knowledge itself.

For centuries, knowledge has been developed, preserved, and communicated primarily with books. The properties of the paper influenced the properties of the information transmitted within it:

• First of all, the paper is bulky and expensive. Then only the best works pass the cut and are published. Everything else is very unlikely to reach an audience.
• Because of it, information is curated by a group of experts who filter it for us. So we are guaranteed to have access to the best content available.
• In a paper book, information has to be presented as a sequence of facts in a finite space. This means having some stop-points for the mind to progressively gain understanding.
• Content is static. Once printed, it is not possible to change it.
• It is also not possible to have a different point of view. The reader reads what the writer wrote. No more, no less.

In books, information has a shape and space. Over the past 20 years, the Internet has changed them with major consequences for knowledge:

• In the digital world, the publishing process is fast and cheap. Low-quality content can be put a few clicks further from us at most, but still available.
• Information is made by everyone. No credential is needed to say something. There is almost no curation process in place. The quality of what we read depends on us.
• Information is networked. Each piece of content links to another one increasing rather than reducing the number of facts to be processed. So it is much harder to set some stop-points to gain understanding.
• Content is dynamic. It's easier to change or misunderstand its original meaning.
• Everyone can add an opinion, making it difficult to reach stable and shared truths.

Having access to millions of networked dynamic sources in a split second is something relatively new to the human mind. Books have always worked as filters. Their properties have prevented information from causing us knowledge overload. The Internet removed these filters and online content now exceeds the brain's limits. Everything is too much and too fast. We have to constantly face information overload barehanded.

I think that networked knowledge brings with it completely new challenges. The Internet architecture had profound consequences on my learning and creative processes. So I felt the need for something more suited to its properties.

A new personal knowledge space

If the very nature of knowledge has changed, so must do the tools to deal with it. I started looking for ways to capture facts and ideas on the fly, which is typical when you browse on the Internet and the sources of information are diverse and change fast. But collect information doesn't provide much value by itself. So I was also interested in how to engage with it and generate lasting learning that would enrich my overall understanding of phenomena.

After some research, social networks led me to the figure of Niklas Luhmann. He was one of the most important social theorists of the 20th century also known for being very prolific with his publications. The secret was in the way he took and organized his notes. Luhmann's note-taking system was based on two simple principles that are curiously central to AI and computer programming as well: modularity and links.

In his system called Zettelkasten, ideas become like building blocks. They can be connected in endless ways to generate multiple flows of thoughts, find contradictions, and see new patterns. These building blocks are called Indexes and Notes. Indexes, typically labeled with a keyword, are the entry points to a knowledge base made of Notes linked together. There are no hierarchies nor notebooks. Each Index starts a flow. Each Note refers to a single concept and can be part of multiple flows.

The knowledge base is not intended to be an archive, rather an artificial extension of the mind. This is a critical difference compared to traditional systems. For almost 20 years, my question has been: “Where do I put this information?”. Now it is “How could this information enrich my existing flows?”. It's a shift from topic-based to concept-based thinking. To make notes reusable in multiple flows, they need to express concepts instead of topics.

Again, the shape and space you give to information have consequences on knowledge. Take the summary of a book for example. Inside a notebook, you would probably have a folder for the book itself and subfolders for chapter summaries. However, books tend to repeat the same concepts across chapters with different words. The knowledge base structure makes it more convenient to find and link them. This maximizes the value extracted from single readings and encourages connections with previous ones (which is how effective learning works).

It's now like playing LEGO with my thoughts. I consume content through new lenses. I actually interact with it to take notes. Then I combine the bricks to explore and find unexpected insights. The goal is not to remember more things, which is often the main expectation when taking notes. To me, the purpose is to have a tool that turns information into ready-to-use knowledge to learn and create better content faster. Notebooks are not equally effective because the information is locked in silos and ideas don't talk to each other.

I am now building a dedicated app to support my learning and creative efforts for years to come. This same article was written in a few hours combining some notes together.

Some information is based on “Too Big to Know” by David Weinberger which I have read and recommend (no affiliation).