Logseq for digital hoarders

General
1

I plead guilty. I am a digital hoarder. I collect songs, papers, books, articles, places, videos, etc, etc. Because of this I use several systems to manage those things: Google Maps favorites, bookmarks, read-it-later apps, Calibre, Zotero, playlists, and lists everywhere. Those things generate thoughts, which I collect in Logseq as they come and link to other thoughts. I constantly try to keep those things out of Logseq because those are too many, and I prefer to keep them isolated until they earn their place on Logseq, near my thoughts.
However, I’m a bit tired of having so many boxes in my life, some of them are paid services. Some others take my lists and make inferences about my life to send me ads or lock me in their subscriptions forever. In general the work of those apps is futile; they just have an appropriate metadata structure for the thing they store and a database to collect many of those things in lists linked to my profile.

I have come to think that having those apps is unnecessary because Logseq has the capability to do both, create metadata and collect stuff, which I’ll own in my graph , and manage with my workflow or AI.

I’m hesitant to do this because the many things I collect , will not be useful, and I want to keep my graph lean.

I want to know if some of you feel a similar problem, and if you have used Logseq to solve it. How? Did it work for you? How long are your collections?, How do you capture, on your desk, on the go, or on paper? What would you do with your very appreciated collections if those were not tied to your “collection provider”? Would you still be using your “collection provider”? Are you a bit tired of the workflow you need to maintain for those things? Are you tired of thinking, "Where do I put this thing when something 'collectible” comes into your digital life?

I would like to have one system that collects them all and syncs seamlessly with the service that best reproduces the collection, but ultimately I want my collections under my sovereignty.
Do you?

2

I think collecting everything in Logseq would be terrible for building a knowledge graph, to say nothing of the overhead of syncing of that stuff. What is locking you into subscriptions? I use open-source tools for most of this stuff. Calibre and Zotero are open source. I use Obsidian as a read-it-later app since the demise of Omnivore, but there are other more elegant open-source and self-hosted solutions. I use Floccus for bookmarks. You can find free storage solutions for all of these and none of them have ads. I also use Joplin for some of what you describe because it does not require all attachments to be synced to every instance, such as mobile. I think the instinct to keep the Logseq graph lean is a smart one. I do not think the program would handle all types of collections very well. Better to use the tool best suited for the task.

Master Foo asked the novice: “If you have a text file, what tool would you use to produce a copy with a few words in it replaced by strings of your choosing?”

The novice frowned and said: “Perl’s regexps would be excessive for so simple a task. I do not know awk, and I have been writing sed scripts in the last few weeks. As I have some experience with sed, at the moment I would prefer it. But if the job only needed to be done once rather than repeatedly, a text editor would suffice.”

Master Foo nodded and replied: “When you are hungry, eat; when you are thirsty, drink; when you are tired, sleep.”

Upon hearing this, the novice was enlightened.

1 Like
3

Yes, mostly agree. My current setup is similar, though Spotify, YouTube, and GMaps are not that easily substituted for me. I like to share and analyze my tastes based on those collections as they evolved over time , and also build better lists using other users tastes. What I don’t like is being tied to the proprietary network just because they “own” their users.
I was considering using a separate Logseq graph for collecting “things,” isolated from my Logseq PKM, because I agree with you; having both in the same place would be a terrible idea. Eventually have a curation workflow for playlists, places, and links all within the Collecting Graph and sync back to the appropiate app of your like. For instance I have Apple Music and Spotify, and would like to have my playlist always in sync, obviously I need an streaming app to reproduce the playlist, but I was wishing there would be some way to manage and maintain the playlist outside the audio streming provider. I am trying to reduce my apps, and the time I put on them. You will tell me Just choose one!! However this is not always as simple as it sound, at least not for me.

In short, in my mind that would not be a knowledge graph, but a collecting graph isolated from the knowledge graph, to be curated for later.

PS: Lovely Master foo story, thks for sharing.

4

My current setup for my collections is:
Apple music, spotify, soundhound for music, youtube for videos, Gmaps for places, Raindrop and firefox for bookmarks and read it later, Calibre, Kindle cloud for books, Zotero for papers, logseq Everbook and pocketmods for thoughts.
Plus one desktop, laptop, tablet and ereaders to sync.
It becomes exponential.
I wish everything would stay in sync without my intervention, and be exportable in case I want to share, import, or just go away.

5

Yes, some things are hard to replicate outside proprietary services, but that’s just the nature of the digital service economy. That said, I’ve found OpenStreetMap to be useful in a lot of contexts, in some cases better than Google Maps, and it’s completely offline. For YouTube, I’ve used FreeTube or Invidious to save lists of subscriptions as json files. You don’t get recommended videos this way, which can be a positive and negative (honestly, probably more of a positive these days). I use YouTube music for music discovery, but I also curate my own collection that I host on a NAS and access via Jellyfin. I know not everyone is able to do this, but I think having your own collection is probably the most reliable way of managing custom playlists over the long term.

All of these things can be saved in files and folders, the synced to the cloud storage of your choice. Then the issue isn’t where to find the collections, but which app to use to access the information. Like you’re considering with Logseq, I also use Obsidian vaults for different kinds of collections, but mostly things that primarily text-based like read-it-later collections. As an outliner, Logseq works best just as a note-taking app for me.

1 Like
6

It is nice to see how some parts of your collection process resemble my setup (Invidous, Open Street maps), some others were on my plan too.
As you pointed out, ”Some things are difficult to replicate outside proprietary services, but that’s just the nature of the digital service economy”, which is true and it is getting harder over time for the very same reason you mention.
There’s this concept in economics Search cost, which is one of the milestones of the digital economy mediated by algorithms; those algorithms decrease the cost to nearly zero.
However, when that happens, we lose the pleasure of searching and the creative mental process that comes with this, “the joy of the road,” because every so often you learn from the search process and the goal is less important.

For instance, if I get my playlist in Logseq, I can add my reflections in blocks, connect to ideas, and ask for recommendations based on my playlist and my comments, using any API. I feel more control this way. This could be applied to other collections as well, and I could have more influence on the recommendation algorithm than just relying on Spotify or YouTube, adding, of course, more search cost and perhaps more pleasure in the process.

I like to add blocks tagged as “LLM:Some intruction for the LLM”, and sharing those blocks with a model work nicely if the prompt has the instructions to focus on those commments, or ingest the whole graph. This kind of control is what I’m looking for, though collecting everything in logseq might be too much.
That said, I think that I want to have full ownership of my collections because I want to get more control of the AI inferring things about me using that collection. I also want to add some search friction to the process, which is something I enjoy.

7 
This is an example of how I use my notes to feed the LLM:**Prompt to consume My logseq notes and generate meaningful output:**
	Read this
	## PARAMETERS:
		Personality*:
		  Matter*:  
		  Destinatary:  
		  Purpose:  
		And follow this
		## Instructions:
		You will analyze my notes regarding the [MATTER] acting as described in the [PERSONALITY] and if found write to the [DESTINATARY] for the [PURPOSE] described. 
		    
		  Please follow these  
		## Rules:
			Lines that begin with **`LLM:`** at my notes, contain specific **instructions for you**. You must act on each one.
			The content is formatted in an **outliner structure**, where:
				Indented lines belong to the parent line directly above them.
				You must respect these parent-child relationships for accurate context.
			You will use the **surrounding text (non-LLM lines)** as **context** to inform your writing or actions.
			Once I paste the notes, repeat these instructions back to me for confirmation. Do nothing else until I confirm.
	After I confirm, execute all `LLM:` tasks as requested.
	    
	  I will now paste my notes.