Hello!
Long time lurker, first time poster.
I’m Aseel and I’m new-ish to Logseq. I’m a career switching new Software Developer who needs to take notes to process, and, for my current hobby, I’ve also got a creative writing bug.
I was first introduced to Logseq and Personal Knowledge Management / Second Brain through a course I took at work. I was already looking for my own Wikipedia style note taking application - I have bounced around OneNote (I know…), Notion, Obsidian, Campfire etc. They all have their quirks, but when I found Logseq, I was instantly hooked.
I started on Logseq MD, and it was going fine. When I heard about Logseq DB, I definitely hit the change bump like many of you. I’m fortunate, my old graph was relatively small and mostly work-related notes and… well… they laid me off recently… so that’s you know, life.
On the plus side, it’s given me time to digest Logseq DB and review the docs properly. It took a while, but it finally clicked and I am 100% on board. I can see the parallels with a Software Engineering Object/Class/Properties model and how it can scale in future.
I’m comfortable with AI tools, through work and personal coding. It’s also great for writing research as it summarizes topics without me having to read through a hundred links. Before my layoff, I was developing a RAG model to provide extra context to an LLM for a business use case.
I started to think, “How can I connect my Logseq DB knowledge graph to an LLM, so it has my own specific context in its answers, and I can chat with, basically, the knowledge I curated into it?” It turns out people are doing this in the real world (I listened to this really cool podcast the other day where this company ingested all service manuals and tickets - all proprietary - into a GraphRAG model so its field technicians could have citable context for issue resolution and could see the parallels), so the knowledge is out there - albeit in sometimes technical YouTube videos. Also on the plus side: AI can summarize all of that, once you get it to play nice. As I understand it, Logseq DB is well suited for this purpose.
In order to do this, I learned about ontology, POLE models, Subject-Predicate-Object patterns - adding the right properties / tags to pages to provide context for an AI - which is what Logseq DB supports well with native features.
I set myself up with a starter knowledge graph, with basic in-built ontology, that fits industry norm LLM knowledge graph and ontology principles. I used AI to help sanity check for correctness, fitting Logseq DB’s unique quirks.
I’m hoping to use this to store notes on Software Engineering, Cooking, Writing Theory, Literary Projects and more. Writing, especially, has a web of interconnected concepts which Logseq DB, with backlinks, properties, and tagging, is a great fit for (Robert Jordan’s Wheel of Time series, talking about “Threads” and “The Pattern” makes sense!).
I also love that Logseq is open source, doesn’t share my data or train an LLM, is stored on my hard drive (with backups!), responsive, and that I can eventually host my own local LLM to connect up to it - eventually I hope to write a Logseq DB LLM chat-window plugin using Logseq’s MCP, equivalent to an IDE chat window, to help automate the maintenance of my knowledge graph and allow for direct interface, so I don’t have to copy/paste.
It’s been a great learning, one that is not over, and it’s certainly building real world skills that I hope to add to my resume/CV.
Outside of this, I love to go on bike rides, watch hummingbirds perform aerial dogfights, cook, travel, and disappear into an album.
Hello, all, again and thank you for reading!
-Aseel