PKM apps need to get better at resurfacing information

I’m a happy user of a number of apps that can be classified under the nebulous category of Personal Knowledge Management (PKM) software:

  • Obsidian (note taking)
  • Things (task management)
  • Drafts (quick capture)
  • Readwise Reader (RSS and read-later)
  • Raindrop (bookmarking and archiving)

These apps allow me to work with vast amounts of digital information. They let me:

  • Quickly capture information from my computer or physical environment
  • Organize captured data using categories and/or tags
  • Connect related information using bidirectional linking
  • Annotate information (by e.g. highlighting parts of web pages, PDFs, or EPUBs)
  • Quickly capture my own fleeting thoughts as I engage with the information (or the world in general)
  • Synthesize what I’ve learned by writing my own notes, potentially creating new knowledge in the process

But something crucial is missing from modern PKM apps: they do a poor job of helping me re-engage with information that I’ve already captured but forgotten about. None of them can automatically resurface relevant material from my system based on my current context.

For example, if I’m using Obsidian to write a journal entry about my exercise routine, I want the app to show me every other journal entry about the same topic. I could pull them up manually, but that would require searching for multiple keywords and wading through more than a decade of writing. In 2025, even a tiny on-device embeddings model could find them instantly.

Resurfacing information doesn’t even have to rely on complex AI algorithms. I want my todo list app to show me every task that was created more than six months ago, which can be easily accomplished using a simple database query. This feature would save me a ton of time and energy by allowing me to prune outdated tasks from my system without having to manually read through each one.

With their current abilities, most PKM apps feel like inert stores of data, no different from the old paper notebooks I’ve stashed away in my desk drawer. There’s a ton of useful information in there, but there’s no way to find it unless I arduously flip through the pages.

Why is this useful?

I personally want to re-engage with the information stored in my PKM apps so that I can:

  • Remember what already exists in my system: PKM tools remove friction from the process of capturing information, which means I sometimes forget what I captured. It’s good to be reminded.
  • Prune stale data: sometimes I need to remove outdated or irrelevant items from the system. Manually sifting through them can be an overwhelming task, but it’s easy to progressively delete them as the app resurfaces them over time.
  • Reduce paralysis of choice: looking at a list of hundreds of documents or articles can be paralyzing. Having an algorithm automatically select a subset of them eases the process of choosing what to engage with.
  • Observe patterns: if I’m constantly reacquainted with information I captured at different points in my life, I can begin to see patterns in what I read and write.
  • Create serendipity: sometimes the mere act of placing two unrelated notes together can spark new ideas.

A possible solution

I love Spotify’s Home tab. It does a great job of resurfacing music I’ve already listened to or saved in my library. Right now, it shows me:

  • The eight most recent albums, playlists, or artists I listened to
  • A selection of mixes created from music in my library
  • A set of recommended albums based on my recent listening activity
  • A set of albums featuring songs I’ve added to my playlists
  • A list my favorite artists, based on recent listening
  • A list of new releases from artists I like
  • Recommended playlists based on genre and mood

If I ignore the algorithmic recommendations—as I usually do—there’s a lot on this list that nudges me towards re-engaging with music I already love dearly. This is exactly what I want my PKM apps to do for me.

A magical far-future version of Obsidian could start with a homepage containing:

  • My most recently created and edited notes
  • A “This Day in History” section, listing notes created on today’s date a year ago, two years ago, and so on
  • Notes similar to the notes I’ve recently created and edited
  • Notes I haven’t opened in a long time
  • My most frequently accessed notes
  • Notes grouped by sentiment or some other algorithmic criteria

Of course, this is only one possible solution that applies to one specific kind of app. I’m including it here in the hopes that it can serve as a starting point that can be further iterated upon.

Plugging the gap myself

I’ve built a few (admittedly very simple) tools to surface data automatically from the apps I use regularly:

  • I use a function on Val Town to take me to a random bookmark stored in my Raindrop account. I call it StumbleDrop, an homage to the beloved Web 1.0 service StumbleUpon.
  • Another function on Val Town takes me to a random article in my Readwise Reader account. Of course, I call it StumbleWise.
  • I’ve created custom views1 in Readwise Reader to show me stale RSS feed items, stale read-later items, and items with a certain combination of tags.
  • Every few days, I Command + click the Open random note button in Obsidian eleventy billion times and read whatever comes up.
  • I use the Similar Notes plugin in Obsidian to show me notes similar to the one I’m editing in a sidebar2.

These tools have improved my experience of using my PKM apps, but there’s only so much functionality I can hack together with scripting and extensions3. I want these features built into my software, not haphazardly pasted-on using third-party extensions and services.

Whoever solves this problem will have my heart forever

If I’m not careful, my PKM apps become black holes where information goes to die. I end up with thousands of notes, tasks, bookmarks, and read-later items that just sit there, completely inert and useless, making me anxious.

I could be more disciplined about what I collect, but why do difficult thing when have computer? What’s the point of using a computer for knowledge work if it can’t amplify my cognitive abilities?

If you’re building a PKM app today, I urge you to consider building features that allow your users to periodically re-engage with the information they collect. Your users will appreciate it, and I will love you forever.

Footnotes

  1. I don’t understand why more apps don’t support custom views. Some macOS apps have Smart Folders, but what you can do with them is often limited.

  2. It mostly works well, until it mysteriously breaks.

  3. Allow me to take this moment to complain about the quality of extension and automation APIs in most software. Many macOS apps still don’t support Shortcuts (looking at you, Chrome), forcing me to write AppleScript. Some apps expose more functionality via AppleScript than shortcuts, or vice versa. Some apps omit basic functionality from their automation suites, such as navigating to a list item. And don’t even get me started about REST and GraphQL APIs exposed by webapps. For many SaaS apps, it’s easier to write a Puppeteer script and extract data from the DOM than use the official API.