Second brain in practice: fewer files, more connections

I went to a talk about second brain and left with some ideas about how to organize my notes better. I have been using Obsidian for some time, I already have my folder structure and a few routines that work well for me, so I thought I knew most of the subject. It was more interesting than I expected. It made me think about what happens after some information is saved, a part I did not think so much about before.

When we talk about a second brain, I think the first image that comes to mind is an app full of folders, tags, links and automations. We want to save everything, meetings, videos, PDFs, things we study, ideas that appear during the day and anything else that looks interesting. It feels like nothing will be lost if everything is saved, but in practice it doesn’t work like that. I have saved many links, transcripts, PDFs and book highlights that I never opened again.

Sometimes I find one of these things months later and I don’t even remember why it looked important at the time. You can have thousands of well-organized notes and still not find what you need when you really need it. The organization is there, the folders look nice and the files have proper names, but they help less than they should. One question from the talk stayed in my head: when I need this six months from now, will I be able to use it or will I only remember that I saved it somewhere?

When organization becomes only accumulation

For a long time I organized my files thinking mainly about where each thing should go. I created a folder, then a subfolder and when a new subject appeared I created another category. Everything looked organized, but I often had the same problem: I could not remember where I had put something, or I found the note and no longer understood why it had been important.

A folder is good at showing where the information is, but it doesn’t always explain why I saved it, what it is related to, what I learned and if it still makes sense today. Without this context the note loses value very fast. When I write down the reason and make some connections, it usually keeps making sense for longer. The problem is that this means going back to the note, changing it, deleting things, filtering and organizing. This was the part I left for later most of the time.

The flow I started using

Capturing information needs to be easy, because if I have to choose a folder, tag, template and format every time an idea appears, I will probably end up not saving anything. This is why I still like having an Inbox, where I put meeting notes, book excerpts, questions, links and thoughts that are not complete yet. For me it is a temporary place, not the final destination.

The Inbox solves this first part. Then comes the work that really matters, going back there and deciding what to do with each thing. First I try to understand why it matters and where it fits. If it really has some value, I write the main point with my own words and try to connect it with something that already exists. When I do this, the note starts being useful because I can find it later and use it in a decision, meeting, article or anything else.

I also need to go back from time to time and review what is there. Some things still make sense, others need an update and many things can simply be deleted or archived. When I don’t do this second part, the Inbox starts growing and after some time I even avoid looking at it because I know it will be a lot of work. It is like that pile of clean clothes you washed but keep delaying to put away :-)

A real example

This article became an example of this. The source was the talk I attended, and later I generated a transcript that was quite bad, with several strange parts, but I could still recover the main ideas. If I had simply dropped that file inside a folder, it would probably stay there for months and I would never open it again.

So I took the transcript and started separating what really got my attention. There was a lot about capture, context, connections, queries and review, so I started connecting this with the structure I already use in Obsidian and mainly with the problems I had when trying to organize my notes. Then I wrote what I understood with my own words and mixed it with my experience until I got to this article.

In the end a talk became a transcript, the transcript became a few insights and these insights were connected with things I was already doing. I will keep the transcript as a source because I may have understood something wrong or may want to check a part in the future, but I know I will probably not start with it. I will most likely come back to this article first, because it already has the part that interested me and the connections I made at that moment.

Projects, areas, references and archive

The talk also mentioned separating projects, areas, references and archive. Projects are things with a delivery or expected result, while areas are responsibilities that continue to exist. References are materials that I may need to check again, and the archive is where I put what is not active for now. This idea is very similar to the structure I already use in Obsidian and explained in How I organize my thoughts, studies, and notes with Obsidian.

But I did not copy this structure exactly because I think it is hard to have one organization that works the same for everyone. Someone studying languages has very different needs from someone working with commercial proposals, software development or academic research. For me the structure needs to follow the way I work and also change when new needs appear. I don’t see much sense in planning twenty folders before I even have the notes, I created mine as the problems appeared and I still move things around sometimes.

How this appears in my daily life

In my English studies, for example, every class is saved in chronological order with the date, teacher and what we studied, but this is only the capture part. When I review a class, the words I really need to practice go to the Vocabulary note and the mistakes I keep making go to Mistakes Patterns. Later, these are the notes I check before a conversation, use to create exercises or even bring back to my teacher.

The classes are still saved, but I don’t need to read all of them again to find out that I have been making the same mistake for three weeks. The original information is still available if I need it, but there is a more useful layer on top of it. I think this was something I did not do so well before. I saved many things, but I didn’t always turn them into information that I could actually check later.

Something similar happens at work. In a commercial proposal, the meeting transcript can be a source, but later the decisions need to go into the scope, the questions into the Q&A and the risks need to appear in the assumptions or reach the responsible people. If I do this well, I don’t need to read several meetings again looking for something I remember someone saying, but have no idea in which meeting it was.

I can ask a more specific question, like which decisions made in the last meetings changed the proposal scope, and which risks still don’t have an owner? This question connects things that are probably spread across different files, and the answer may become a pending items list, a proposal update or even the agenda for the next meeting. English classes and a commercial proposal are very different situations, but the logic is similar: the source feeds something that already exists and normally ends with a next action.

Connections between notes

I still use directories normally and I think they work very well for navigation. Links come in when I find a relation that a folder cannot represent very well. A note about pricing, for example, can be related to a competitor, a proposal, a business rule and a decision made in a meeting. Each one will probably be in a different place, and the link helps me reach them without remembering where every file is.

These relations also appear with time and I don’t think they need to be planned from the beginning. A note can start connected to one project and a few weeks later be useful in another one. Sometimes I also find an old conclusion that is already outdated or that I would write in a completely different way today. When I summarize a source I almost always find some relation with something I had already saved, and many times this is where a new idea appears.

What is not worth keeping

Another thing I started accepting is that many things simply don’t deserve to become a permanent note. Some information may be useful only until the end of a project, a meeting note may completely lose its value after the decisions are saved, and a link I saved without thinking may never deserve to leave the Inbox. Before, this was harder for me because I always had the feeling that I could need it later.

Keeping everything forever also has a cost. It increases the amount of things I need to organize and may even make my searches worse, so I am trying to choose better what really deserves to stay there. Some things I simply delete, some I keep as a source because I may need the original, and others have a lesson that is worth summarizing or connecting with a project, decision or article.

Archiving is still something I am learning to do better because I tend to leave everything close, afraid that I may need it later. But after some time this excess also gets in the way and it becomes harder to separate what is still relevant from what is there only because I never had the courage to delete it. I need to let go a little more and accept that not every piece of information needs to be available forever.

Where artificial intelligence enters

Language models help a lot with all of this, mainly when we need to search through a lot of text, summarize information or find relations between different content. We can put together documents, transcripts, business rules, analyses and many other sources, then start asking questions about this material. My first idea was exactly this, put everything inside one place and start asking.

After testing it I noticed that it doesn’t work so well. If a source is outdated, duplicated or has no context, it is still bad after going into the AI. And if the question is too generic, the answer usually becomes generic too. It looks correct and many times it is correct, but it is not very useful when I need to make a decision or do something with that information.

One of the ideas I liked most from the talk was to think more about the questions. I spent a lot of time worried about how to add new information to my second brain and little time thinking about what I would really like to ask it later. If I ask what do I have about this client?, I will probably get a summary of many things and it may even be useful, but it is too open.

If I ask which scope changes happened in the last meetings and which risks still don’t have an owner, the answer can already be used in a meeting or proposal review. Today I try to imagine this type of question before simply saving one more source. This already changes what I choose to save and the context I add with it.

A simple routine to keep the system alive

If I create a huge routine to take care of my notes, I know I will abandon it at some point, so I am trying to do a simple weekly review. I look at the Inbox and try to reduce what has accumulated, delete captures that have no value anymore, and I try to summarize the really important sources or connect them with something that already exists.

I also look at active projects and see if something needs an update, while what has finished its cycle goes to the archive. Another thing I started doing is writing down questions that my own material still cannot answer, because this also helps me see where context is missing. I don’t do all of this perfectly every week. Some weeks I do a lot and some weeks almost nothing, but it already helps.

How I notice it is working

For me it works when something I saved becomes useful again, maybe helping in a meeting, stopping me from searching for the same information again or saving those minutes I spend trying to remember where I saw something. Sometimes a note I had almost forgotten appears when I ask a specific question. In other cases, the history of a project avoids a discussion we already had or a query finds a risk that was lost somewhere inside a meeting.

This is where I see value in all of this, when it saves me time or helps me solve something faster. There are also weeks when the Inbox is full and I review absolutely nothing, and sometimes it shows quite well how messy my own week was. I think this is part of it. The important thing for me is being able to come back later without turning note organization into another job.

After the talk I wrote down five verbs that I liked as a reminder of the process: capture, understand, connect, query and review. I don’t use all of them every time and I don’t think every note needs to go through everything, it is only a reference I use when I notice that I am starting to collect too much. My structure will probably be different in less than one year, but if it keeps helping me recover an idea, make a decision or write something, for me it is already great.