Fly.io departure: author bets on a phone with no fixed-function apps
The author of a blog post titled "What Even Is An OS Now?" announces he is leaving Fly.io to work with a collaborator named Kurt on a new project: a phone designed around software that users generate for themselves with AI, rather than fixed-function apps built by professional developers and installed from an app store. He argues that AI is dissolving the boundary between programmers and users: he has spent months building his own native macOS apps by describing what he wants in English, and expects ordinary power users will soon do the same. His central claim is that most future software will not come from strangers shipping finished apps, but from people assembling small, idiosyncratic programs for themselves, reusing only a fraction of the code that goes into something like a full word processor, which he puts at roughly "1/7th of the guts of a word processor". He gives everyday examples of the kind of narrow, personal problems these programs would solve: getting an accurate weather forecast for a specific Chicago neighborhood, deciding which street to take home instead of following a map, or checking whether a meeting is happening right now. He argues that the core purpose of a modern operating system, which he defines as partitioning applications from each other and controlling how they communicate, makes far less sense once most software is generated by the user rather than imported from strangers. He contrasts this with today's phones: every mainstream phone on the market in 2026 was designed back in 2023, when computers were still assumed to run fixed-function apps built by a "priesthood of professional programmers", and he says the underlying design goes back to the 1970s and machines like the DEC PDP-11. The new phone he and Kurt are building is explicitly not meant to run prefab apps, AAA games or short-form video; instead it is meant to build apps for the user on demand, whenever asked. He does not name the phone, give a price or release date, or explain any technical detail of how it would work, saying he is deliberately withholding those details to avoid sounding like a typical startup pitch. He frames the whole piece as a personal bet: "Every limit AI runs up against collapses within a month," and he expects that what frontier models do today will be done instantly and for free within a few years.
Key facts
- The author is leaving Fly.io to work with a collaborator named Kurt on a new phone project.
- The phone is designed to run AI-generated, personal software rather than fixed-function apps from professional developers.
- He estimates such personal apps will reuse something like 1/7th of the functionality of a full application such as a word processor.
- He argues mainstream 2026 phones were designed back in 2023 around a decades-old assumption that software comes from professional developers.
- No name, price, release date or technical details of the phone are given; the author says he is withholding them deliberately.
Why it matters
The essay argues that AI is erasing the line between programmers and users, letting ordinary people generate small, personal programs instead of buying apps built by strangers. If that holds, it undercuts the reason phones and operating systems are built the way they are today, since their core job of walling off untrusted third-party apps from each other matters much less when most software is created by the user for themselves.
Who it affects
It concerns anyone who builds or relies on phone operating systems and app ecosystems, since the author's claim is that the entire model of prefabricated, professionally built fixed-function apps is starting to wash away, similar to how boxed software and CD-ROM distribution did before it.
How to use it
There is nothing to use yet: the author names no product, price, release date or launch timeline for the phone he and Kurt are building, and he explicitly declines to share technical details, saying he wants to avoid the feel of a typical startup announcement.
How solid is it
This is a first-person blog post and personal announcement, not a report backed by data, a product demo or independent verification; its claims about how software and phones will change rest entirely on the author's own predictions and experience building native apps for himself.
Risks and caveats
The piece is explicitly framed by its own author as a founder making sweeping claims about the future right before unveiling a new venture, a pattern he acknowledges tends to age badly; the phone itself has no announced name, timeline or working details, so the core promise is unverified.
“the core purpose of a modern operating system is to partition different applications off from each other, and carefully control how they can communicate”
— the author of the essay