On September 19, the KDE Community posted Scott Jenson’s Akademy 2026 talk to YouTube under a title that doubles as the question: Are we really going to use the same Desktop UX forever? Jenson has spent a career on interfaces at Apple and Google, and he used KDE’s annual developer conference to argue that open-source projects should stop refining windows, icons, menus and pointers — the WIMP model — and start inventing past it. LWN’s jzb wrote it up on September 23. The Hacker News thread ran 307 points and 354 comments, and the forums it landed on were not warm. The recurring objection was precise: every attempt at a new paradigm quietly deletes the standard functionality people came for.
That objection is correct. It is also diagnosed backwards, and the misdiagnosis is why the freeze persists.
The Forums Are Right for the Wrong Reason
The claim is that users reject change. What users actually reject is unrecoverable change. Watch what got adopted without a fight over the last fifteen years: fuzzy search, the command palette, tiling workspaces, tabbed everything. All additive. All reversible. Now watch what fought: GNOME 3, Unity, KDE 4. Every one of those was a rewrite of the frame rather than an addition to it, and every one of them turned into a years-long referendum.
The difference isn’t novelty. It’s whether a user can back out. Backing out requires someone with the authority to run a change as an experiment, watch it fail, and undo it quietly. Commercial vendors have that person. Volunteer projects have a mailing list. So design changes arrive as fait accomplis, get litigated in a comment thread, and the next generation of maintainers learns the only durable lesson available: don’t touch the frame.
The Same Week’s Front Page Was All Plumbing
Here is the part that ought to embarrass everyone involved. In the same days LWN was publishing Jenson’s talk, its front page carried Git 2.56, gccrs, io_uring work, NetBSD’s compat_linux, GNOME 51, and systemd v262. That is a shipping machine operating at the top of its field. The plumbing is genuinely world-class.
The surface is the one layer nobody is structurally rewarded for changing. Fixing io_uring has an owner, a patch series, and a merge window. Redesigning the window frame has a founder, a philosophy, and a schism. The problem was never capability. It was that the desktop’s user-facing layer is the only part of the stack where open source governance has no mechanism for a controlled experiment — which means it has no mechanism for anything but revolution, and revolutions lose.
Compatibility Is a Go-to-Market, Not a Design Philosophy
Open desktops acquire users by being the thing that runs what you already know. That is not a philosophical commitment to tradition; it is a customer acquisition strategy, and historically an effective one. But it has a cost that rarely gets stated: the people you win are the people who switched because it looked like home. They are the most change-averse constituency in computing, and they are now your veto.
A distributor’s desktop packaging lead, typing into a Slack DM while the Akademy stream was still running, put it more bluntly than I can: “We don’t ship a desktop. We ship an escape hatch. Nobody escapes to a place that looks unfamiliar.”
So Jenson’s ask isn’t really a design proposal. It’s a request that KDE abandon its only working growth channel on a bet about users who don’t exist yet. That is the kind of request that gets a standing ovation at a conference and a polite nothing in the commit log.
There Is No Commercial Desktop Lab to Hand This To
Here is where the comfortable answers on both sides fail.
The reflex on my side of the aisle is that markets handle this and the commons can’t. But the market already answered. The best interface people at the largest firms were routed off the desktop years ago — first to phones, then to assistants, and now to chat. The desktop became a legacy surface inside the very companies that invented it. There is no commercial desktop UX lab to outsource this to, because desktop users are the least monetizable users in computing and nobody funds a research group for a customer they’ve stopped courting.
Which leaves a nonprofit. KDE is now the last institution with any institutional reason to care about what a desktop should become — and simultaneously the institution least capable of running the experiment Jenson wants. The commons is the only place left that will do this work, and the commons is structurally the worst at it. If you believe markets solve this, you need to explain why the market’s answer was to walk away.
And if you were in that comment thread defending stasis, your defense isn’t wisdom. It’s a bill someone else pays — everyone who will meet a computer for the first time on a surface whose design was frozen by a company that lost interest in 2012.
Jenson is probably right, and probably won’t be listened to. Not because KDE lacks imagination, but because the only institution left that could act on his talk is the one that cannot survive acting on it.