The idea of a “digital brain” is appealing. Put your notes, documents, links and half-formed ideas into one system, connect them all together, then expect your accumulated knowledge to become easier to use.
There seems to be enough truth in this concept to warrant further investigation. But there is also a great deal of hype.
My interest in Obsidian started with an investigation into knowledge graphs: the visual networks of linked notes that promise to reveal relationships between ideas.
My initial investigation into Obsidian does not persuade me that a “digital brain” is the right model for a consultant’s working knowledge.
What Obsidian does well
Obsidian is built around local Markdown files. A vault is simply a folder containing notes, attachments and supporting metadata. The notes remain readable outside the application, which is a major advantage over a proprietary archive.
Its linking model is also clear. A note can link explicitly to another note using a simple reference such as [[Semantic Kernel]]. Obsidian then calculates backlinks, identifies unlinked mentions and can display those relationships visually in graph form.
This is an interesting way to curate active ideas, but it does need you to create these links intentionally.
If a team has notes about industrial AI, automation, software strategy and knowledge management, deliberate links can help maintain continuity between related work. A note about “AI governance” might sensibly link to other notes on engineering judgement, information quality or the limits of automation. But this is a multidimensional complex relationship and it is not evident to me that a tool can help represent these connections.
Migrating from Evernote
I used to capture all my notes in Evernote. My library is well over 12,000 notes collected over 7 years. Evernote has unfortunately become an expensive subscription that unfortunately still relies on a proprietary format – to get value out of Evernote you need to be fully committed to using its premium features. My problem is that my digital world and that of the teams I work with has now far more complex than a simple note taking app. Evernote was no longer cutting it, I had already decided months ago to cancel.
Obsidian has an official importer for Evernote .enex exports. If it converts the archive cleanly enough, it offers a practical route from Evernote to a portable Markdown vault. The resulting files can then become a stable source archive for apLabs. In my own case this move away from proprietary formats seems to be the only reason my continued exploration of Obsidian might be worthwhile.
Where the digital-brain idea loses me
Evernote aside, my exploration so far has not revealed any “light bulb” moment when it comes to the concept of a digital brain. The problem begins when a note graph linked to other notes is treated as strong evidence of inherent complex relationships. It is not.
A graph can visually show that notes are linked. It cannot tell you whether the links are useful, accurate or important. A dense visual network can easily become a diagram of filing habits rather than a representation of knowledge. The person who linked two notes together may or may not have represented the relationship correctly. Busy consultants working in teams do not have time to design and figure out complex relationships in the off chance they may want to unravel them later. One consultant may create links between notes that another consultant fails to understand, the result in a team setting is chaos.
The same limitation applies to automatic linking. AI tools can suggest related notes using keywords, embeddings or language-model analysis. Some suggestions will be useful. Others will be plausible but weak, based on superficial similarity rather than a relationship worth preserving. The idea of AI agents working overnight to construct your digital brain for you is interesting, but in practice full of flawed assumptions about the way consultants and engineers actually do work.
A note system should reduce the effort required to prepare for a meeting, find a past decision, retrieve an example, develop an article or reconnect with work done for a client. It does not need to imitate the structure of a human brain. It needs to help a human be more productive.
As I see it, the “digital brain” label risks encouraging the wrong kind of effort: extensive time spent tagging, link maintenance and graph grooming without a clear return.
How apLabs approaches the same problem
apLabs had evolved a mature knowledge management approach before I encountered Obsidian. The apLabs purpose was never to create a perfect personal knowledge graph. It should be a productivity and knowledge-management toolkit for working with material that already exists: conversations, notes, documents, links, projects and emerging ideas. It is designed with teams in mind.
The practical difference lies in the way information is retrieved.
apLabs can use semantic search to find relevant material by meaning rather than requiring an exact folder name, tag or phrase. Semantic search represents text as numerical embeddings, allowing related material to be found even where the same language has not been used. As I see it with AI and RAG, semantic search can far more accurately simulate the way the human brain retrieves and processes information than a knowledge graph can.
Semantic search with AI is more useful than retroactively creating thousands of links across my enormous Evernote archive.
A note about “operational technology decision-making” may still be relevant to a search for “industrial digital strategy”. A manually curated graph may never capture that connection. Semantic retrieval can surface it when required, while leaving the human to judge whether it is useful.
Semantic search is not without its limitations. I will cover this in a future post. But for now, the way semantic search with embeddings works in apLabs continues to meet my needs exactly and reliably.
Avoiding graph links does not remove the need for structure. Projects, folders, tags, dates, source information and explicit links still matter. They provide context, separation, accountability and a way to navigate material intentionally.
Closing comments
My early investigation of Obsidian is still at the conceptual stage. I am sure the tool holds promise for specific use cases. I believe there is a role for knowledge graphs in certain tightly managed situations – for example to represent the complex relationship between equipment and activities in a chemical processing plant. But as a substitute for the “human brain”, especially when used in teams, it falls far short of what I was expecting.
References
- Obsidian Help — Obsidian. Accessed July 2026. https://obsidian.md/help/
- Link notes — Obsidian Help. Accessed July 2026. https://help.obsidian.md/links
- Import notes — Obsidian Help. Accessed July 2026. https://help.obsidian.md/import
