KDE turns 30 and someone's brought an AI-native desktop proposal

> A desktop that loves you back has to know you. Personal AI has crossed the threshold where the desktop itself can be compiled per user, per device, per moment, from a small portable model of the user – what we are sketching as Kadai: an encrypted, vendor-agnostic personal kernel from which Plasma assembles each Activity.

Sounds like a gimmick that will be annoying after a week. Users who aren’t KDE developers don’t want a desktop to love them back, in fact they don’t care or want to think about their desktop beyond launching their apps and window management.

The desktop is a solved problem. Now we are moving into the era of agents using the mouse and keyboard and interacting with the windows manager.

When I see stuff like this, I think it's clear that Linux is not going to be part of that era outside of the CLI. The Linux desktop is never going to recover from this.

Gnome and KDE just keep messing. That's why I stuck to openbox/labwc.

Because, these are not solving problems they are inventing them. Everything that makes Linux great, simply hasn't translated to the desktop GUI. And I can see that will not change.

We are done with desktops. They will be handed off to ai. Humans will use them less and less. Because they will have better things to do. And likely better interactions. After all a mouse and keyboard while I still prefer it over voice, is still from an era where that was the best we had.

Now, I actually think I prefer tmux in a terminal on my phone on a software keyboard. Over a desktop with mouse and keyboard. As clunky as it is. Because I can use that anywhere, anytime to talk to agents.

It’s going to happen in windows and most likely Mac too. In both cases, all your data will be shipped off to the model providers and then out into the panopticon.

The open source desktop environments don’t have the profit motivator that drives the commercial alternatives. They’re driven to deliver their interpretation of great UI. We can each debate whether we agree with the vision or execution but that’s the motive.

If they can employ AI to help remove desktop friction - and it’s local and not shipping my data off to the surveillance machine - I’d be open to it.

In fact, if it could sort out KDE wallet’s amnesia for my wifi password, I’d give it a roll now…

Although Xfce 1.0 appeared slightly earlier, its initial releases were proprietary, making what was originally known as the Kool Desktop Environment one of the earliest entirely FOSS desktop environments for Linux. Version 1.0 was released in July 1998.

What a weird opening.

Also, both projects were started late 1996. A lot of people were already running KDE alphas/betas before the 1.0 release, because as a desktop it was vastly better than any of the window managers at the time.

I remember installing beta 1 in 1997 was a big relief for me. It felt like the last barrier was broken for Linux taking over the desktop, now that it finally had a desktop that could compete with Windows 95. Of course, I was terribly naive, but I blame it on only being 15yo.

I don't recall Xfce being used much at all. It become more popular until 5-10 years later because people were complaining that KDE/GNOME became too heavy and they were looking for more lightweight alternatives.

I love making kustom KDE widgets with LLMs[0][1]. If you like customizing your desktop, I highly recommend you take advantage of this new era.

[0] https://github.com/dginovker/KDE-Mouse-Click-Counter-Widget [1] https://github.com/dginovker/KDE-AI-Usage-Tracking-Widget

The best way that my desktop can show me love is by being a glorified application launcher and nothing more. Also, the less meta/data that it sends back to its vendor the more love I feel that I am receiving.

News piece with unnecessary annoying editorializing.

I think the idea of plasma widgets and rails fully being customized per user giving them material control on ones desktop is a good idea.

We need good ideas to counter and take on the omarchy nonsense.

Guess they never used QtQuick with LLM. LLMs are really bad at QML/QtQuick, they often use older APIs/syntaxes and generate total garbage.

Is it going to just be an Omarchy + KDE iteration? The article was very light on content but DHH is definitely driving more results and new contributions which is great for free software.

This article has no content whatsoever... Just sending the links of relevant materials to AI would produce a better-organized result than this.