Lego - the ultimate plugin system (source)
I've recently been getting back into using Obsidian after a years long hiatus mostly using Logseq and -- gulp -- Apple Notes.
While Obsidian comes with some great features and "core plugins" out-of-the-box, a big part of my re-onboarding experience has been customizing the environment with community plugins to fit my specific needs: exporting flashcards to anki with custom templates, converting kindle highlights to obsidian notes, setting up kanban boards, directory navigation enhancements, etc.
This process is trial-and-error: I search for existing plugins and test them out to see how they work. If I like them, I'll continue to practice using them (tools are best paired with habits). If not, I remove it and look for another. If I really can't find exactly what I'm looking for it's not a deal-breaker for Obsidian. I could always try building my own obsidian plugin or, since all the data in stored locally and formatted as (obsidian-flavoured) markdown, I could build/use any kind of program that can access local, easy-to-read/manipulate files.
The key point is that much of the functionality that makes Obsidian so valuable doesn't come from the core Obsidian team.
I expect that if the atproto ecosystem continues to mature, we'll find ourselves in an analogous situation as the Obsidian ecosystem does -- except the community plugins won't be for a single app, but the ecosystem at large.
We're already catching glimpses of this future.
Emerging tools like allow anyone to create atproto automations -- protocol plugins, really. For example, automatically converting stars of Tangled repos to Semble cards (by ) or taking any liked bsky post and creating Semble cards for any links in the post (by ).
These kinds of conversions enable customized forms of interoperability without apps needing to share lexicons or coordinate on standards, which can be a time consuming and difficult process (especially if the domain is not yet well established -- being a great example of when lexicon standardization makes sense).
On top of tools like airglow, there are emerging tools around interactive embeddings and components like webtil.es and . Both of which can empower community members to not only build protocol plugins at the data-layer, but at the interface-layer -- whether completely separate apps or embedded widgets in existing apps.
It's not hard to imagine a future where new users of the open social web can be greeted with countless plugins to customize their own information environments, just like with Obsidian. Except they aren't limited to a single app but instead an entire protocol-based ecosystem.
This is a post-app paradigm.