Order from Chaos

I’ve been interested in personal knowledge management for well over a decade now. Like many others, I started with OneNote - adding sections, creating pages for different subjects and categories in my life. It worked. For a while.

Then it became unwieldy. So I’d reorganise. Create a new structure. Feel in control again. And then, inevitably, watch it spiral back into chaos. Rinse and repeat… for years.

If you’ve ever felt that cycle of organisation-followed-by-overwhelm, this post is for you.

What Even Is PKM?

Personal Knowledge Management (PKM) is exactly what it sounds like: a system for capturing, organising, and retrieving the information that matters to you. Some call it a “Second Brain”: an external system that extends your memory and thinking capacity.

The terminology varies. The goal doesn’t: stop losing ideas, find what you need when you need it, and actually use what you’ve learned.

The Patterns I Explored

When my notes became perpetually messier, I started investigating systems that others had built:

  • Zettelkasten: Atomic notes, heavily interlinked. Fascinating in theory, but the numbering systems felt dated in an age of search and custom sorting.
  • Johnny Decimal: Numeric categorisation. Elegant, but I wanted folders that said what they were, not what number they’d been assigned.
  • ACE (Atlas, Calendar, Efforts): Nick Milo’s approach from Linking Your Thinking. Interesting philosophy around emergence.
  • PARA: Projects, Areas, Resources, Archive. Tiago Forte’s system.

And PARA? This was revolutionary for me.

Why PARA Changed Everything

The categorisation of Projects, Areas, Resources, and Archive was exactly what I needed. But after years of practice, honing, and learning, here’s what I feel most people miss when they first encounter PARA: it’s not really about the folders.

The underlying philosophy is what makes it work: higher actionability sits higher in the stack.

PARA Actionability Stack Actionability is the hidden power behind PARA that most people miss.

  • Projects are active, time-bound work. Most actionable.
  • Areas are ongoing responsibilities with no end date.
  • Resources are reference material—useful but not urgent.
  • Archive is inactive. Least actionable, but preserved.

This spectrum of actionability lets me concentrate on what matters while still having a place for everything else. Nearly a decade later, I’m still using PARA as my foundation. But I’ve made adjustments.

My Twist: Promoting the Inbox and Journal

I’ve pulled two concepts out and given them their own space:

The Inbox: Capture first, decide later. Zero friction. Anything can go here and the triage (sorting into PARA+ folders) happens later. This separation is crucial: it means I never hesitate to write something down because I’m not sure where it belongs.

The Journal: This is where things get interesting. Every note that is triaged out of the inbox gets logged in my daily journal with a link to wherever that note is moved to. Think of it like Zettelkasten’s linking philosophy, but with time as the spine.

Workflow Diagram Random notes get captured and funnelled into the vault through triage.

Why does this matter? If I remember I had a project idea on a specific day, I can find the journal entry and follow the link. If I’m in the project file, I can trace back to the moment that idea was born, what I was doing, other events or things that happened around that time, what I was thinking, how I was feeling. Context that would otherwise be lost.

The Tools Along the Way

I spent years in OneNote. It genuinely has great features: stylus support, audio notes, handwriting recognition. It’s served me well.

But when AI CLI tools started maturing, I realised something: AI could help manage my notebook, but only if my notes were in a format AI could easily read. OneNote’s proprietary format wasn’t it.

So I migrated to Obsidian; local-first, markdown-based, and tons of useful plugins. With ChatGPT’s help, I exported my OneNote sections to PDF and converted them to markdown. Painful, but worth it.

Other tools exist, good ones, better ones even. And I presume they’ll have AI implementations in no time, but I want something more: customisation (and I want it now).

Where AI Fits In

And so this leads me to my present journey.

After discovering Claude CLI, an AI that could read and modify documents on my computer, I realised a new kind of augmentation was possible.

I can have AI look over my journal and tell me about myself. Could I improve some aspect of my processes? Is there a long forgotten project I could revisit? Am I overanalysing certain things? Are there patterns I haven’t consciously noticed? These are questions I never thought I’d be able to ask my notes.

With a promoted Inbox, AI helps me triage. I drop notes in and depending on whether it’s a research note or a quick log entry, the AI picks it up, decides whether to help research, whether to create or update projects etc. Over the last few months, a select few AI models have become remarkably accurate, partly because of improvements, partly because I’ve also improved at describing my processes.

Right now, these steps are manual. I type into my Inbox, then ask AI to triage it. But I’m working toward automation: notes that process themselves, voice recordings from my phone that transcribe and flow into my vault without me touching a keyboard.

What’s Next

This post is the beginning of a series. I’ll be diving deeper into:

  • The Journal system and atomic rollup
  • How I’m building AI automation into my PKM workflow
  • The friction points I’m still solving

If you’re drowning in notes, or you’ve never quite found a system that sticks, I hope my journey gives you some ideas—or at least the reassurance that everyone struggles with this.

The perfect system doesn’t exist. But a system that evolves with you? That’s worth building.