The app is mine. The data is ours.
Personal interfaces, shared data, and what we should actually be paying for.
I’ve been swapping note-taking and planning apps my often than I care to admit. Not because I love shiny new things, but because none of them work the way my brain does. Small example: I have trouble with calendars that have date gaps. Most weekly views skip weekends, or monthly views break up sequences in ways that make me lose the thread. It sounds minor, but it shapes whether I can actually use a tool effectively. We’ve all got our quirks. And if you’ve ever felt a tool was almost right but kept snagging on the one thing you couldn’t change, you know the frustration. The answer has always been: accept, or move on. I’ve moved on many times. And every time, I’ve lost data.
So I decided to see if I could prototype my own.
If we can build personal software, should we?
I started vibe coding a note-taking tool that fits the way I think. It’s not a big change from what you know, it’s just shaped around my own quirks instead of someone else’s defaults. The process of building it surfaced a question I wasn’t expecting.
If I can build my own interface, why can’t everyone? But if everyone did, how would we collaborate? How would we create shared truths?
Things work better when they fit your mental model, and we’ve had preferences in apps for years; themes, font sizes, light and dark mode. Those are merely cosmetic. What’s becoming possible now is customizing apps. Which features are front and center and which ones can you hide out of sight? Whether your calendar shows continuous time or paginated months. Whether you have a calendar view at all versus just a timeline. Not preferences about how a tool looks, but about how it works, how it matches the way you organize, prioritize, and think.
The data underneath: notes, tags, relationships, timestamps, they shouldn’t be too personal. They should be can be shareable, synced, and collaborative: speak a common language.
Infrastructure, not apps
Right now, we pay for apps. We subscribe to tools that bundle the interface, the data, and the syncing into one package. The package however is designed to keep you inside it. When you leave, your notes come with you as a stripped-down export and your years of organization mostly evaporate.
There are tools that do it better. Obsidian stores your notes as plain Markdown files on your own machine. That’s real data sovereignty. But if you want multi-device editing that’s seamless, you still have to pay for syncing. And that’s actually fine. Infrastructure is worth paying for. Servers, sync, uptime, storage, those are real services. What’s harder to justify is paying for lock-in disguised as a product.
We’re already seeing this unbundling in other places. With most AI tools what you pay for has moved from the tool to the usage itself. What if personal software worked the same way? You build or customize the interface that fits your brain. You pay for the infrastructure that stores, syncs, and connects your data. The interface is yours. The data is interoperable. The infrastructure is a service.
The trajectory of personal AI
Chief designer Mark Rolston has been talking for years about ephemeral applications; software generated at runtime, spun up to serve a momentary and personal need, then discarded. It’s a compelling idea, and we’re starting to see early versions of it.
Here’s the catch: if everyone’s tools are personal, how do we still create shared truths? Ephemeral apps are just for you, just for now, with no shared surface at all. But even less radical personalization raises the question. If your view of a project and mine look completely different, how do we stay aligned? How do we point at the same thing and know we’re seeing the same facts?
This is the hard part. Keeping people connected while letting them work differently. It means the data layer has to do real work: shared objects, shared state, shared history. The interface can be personal, but the truth underneath can be common ground.
Personalization without isolation
Data sovereignty has become a familiar idea: you should own your data. But what I’m really talking about is interface sovereignty: the freedom to shape not just what your tool looks like, but how it works. All without giving up the ability to collaborate with people who work differently.
People have always organized things differently in their heads. But through language and shared conventions, we find ways to work together anyway. Software could work the same way. Personal interfaces, common infrastructure. Your note-taking system and mine might look nothing alike, but the underlying objects speak the same language. Each of us thinking in our own view, all of it connected underneath.
We’re not there yet, but the foundations are starting to show, the building blocks exist. The question is whether we’ll use them to keep making the same ten apps for everyone, or to let people shape their own tools while still finding common ground.
Update March 18 2026
Actually I might have been a bit naive with my call to action. Since writing the draft of this article, Apple has blocked several vibe coding apps in the App Store. Turns out Apple is not interested in allowing people to install apps that can change themselves, due to App Store quality processes. Let’s see who will win.