The Year Everyone Started Building
Notes from PKM Summit 2026 on AI, friction, and what's worth protecting
I spoke at the European PKM Summit this year. I spoke there last year too. Two years in a row, same event, very different room. (The only constant is my fashion choices, I guess...)
If you’ve ever built a system to capture your thinking, convinced yourself this one would finally stick, and quietly abandoned it three weeks later - this one might resonate.
Last year, the tech geeks were quietly talking about AI tools in the dark corners of the conference. Still niche, still early adopter energy.
This year, that was everyone.
Old, young, no coding background whatsoever. Building read-it-later apps, book catalogues, study assistants, home automation systems, mindfulness applications. Someone even built a personalised music teacher for their kid. All of it inside Obsidian, all with Claude Code.
You can now build something like that in 10 minutes. And the best part? It’s yours. No subscription. No company that owns your data. Yours.
The age of traditional apps is dead. The age of personalised, self-built apps is here.
Clive Thomson is a tech journalist who has spent decades covering the intersection of people and technology. He attended the summit and said it gave him a glimpse of what the future of PKM will look like. And he’s right. This is where we’re heading.
But Clive also said this:
“Every time new technology arrives, we’re excited and overwhelmed in equal measure. 20 years ago, what tools did we have? A word processor. Today we have too many to choose from. How do we manage this?”
My take-away question became: how do we stay essentialists now that almost anything is possible?
Most people never figure it out. Most hop from one app to the next promising one that promises to relieve all the tension that has been building up in their disorganised minds. But sooner or later, they all collapse under their own maintenance.
What finally fixes it isn’t discipline or a better app. It’s removing the friction entirely.
PKM systems don’t fail because people aren’t disciplined. They fail because the friction of filling things in is just enough to stop you from reaping the benefits. In the past it almost felt like an unteachable lesson - that you need to start and stop 10 different systems and methodologies, that you need to grind your teeth and build up the mental resilience to input the data in the right way, to use the correct template, to diligently sit and fill in your journal, to give that meeting a summary. End of a long day, you tell yourself you’ll process it later. You don’t. Three weeks later you’re piecing together fragments of memories you should know by heart.
I can finally say: it’s no longer a discipline problem. It’s a design problem. And it’s fixable.
And the scale of what’s now possible is bigger than most people realise. You can have multiple Claude instances running at the same time. One is processing your meeting. Another is doing your week review. You’re making coffee. This isn’t one assistant. It’s a department.
But with that power came a question nobody had a clean answer to. Another speaker at the conference, Thijs Pepping, said it best: “The more you measure, the more you behave like a robot.”
With data about ourselves so easily acquirable, how can we remain in the driver’s seat of our own lives? Will we quietly hand over the thinking to an algorithm?
We’re building systems to track sleep scores, day ratings, step counts, productivity metrics. At some point the measurement starts to erode the thing you’re measuring. Worth sitting with, especially for those of us building quantified-self systems in Obsidian.
What I’m trying to hold onto is this: AI wraps around your methodology, it doesn’t replace it. The structure of your life - your life areas, your projects, your daily rhythm - stays yours. Designed by you, for you. AI handles the admin layer around it. The logging, the summarising, the linking, the processing. The thinking stays human.
I don’t know if that’s the answer. But it’s the line I’m trying to protect.
I’ve never been more excited about where this is going. And I’ve never thought harder about what to protect from it.
We’re launching a 4-week live cohort in April to help you build this yourself.
Interested in building your own system?
Sign up and you’ll get a free email series + a free live webinar before the cohort opens - and first access when sign-up goes live.
https://lifestack-os.com/
Even one workflow that removes one piece of friction is worth it. That’s how the avalanche starts.
Much love,
Dee
Construct by DeeWebsite | YouTube | X / Twitter | Newsletter
Thank you for coming back for a second year, Dee — and for writing this. You
captured something I noticed too: the room was fundamentally different this
year. Last year it was curiosity. This year it was all about construction.
Your line "it's not a discipline problem, it's a design problem" is one I want
to frame. That reframe alone is worth more than most productivity advice out
there. The moment you remove friction from capture, everything changes. Not
because people become more disciplined — because design finally stopped
punishing them for being human.
And I'm glad you included what Thijs said about measurement turning you into a
robot. That question deserves to sit at the centre of this movement, not at
the edges. The easier it becomes to build these systems, the more important it
becomes to ask what we're protecting from them.
Thanks for being part of this community, and for putting into words what so
many in that room were feeling. See you next year??
Martijn