What would deeper AI integration on the Ubuntu desktop look like? Well, you don’t need to think too hard to find out.

A Canonical engineer has published a prototype launcher that simulates how on-device AI could be used to enhance and aid users of the Ubuntu desktop.

But the warty-prototype1 snap by Didier Roche-Tolomelli (aka didrocks) is not a real launcher. It’s a semi-interactive, automated mockup. It simulates things a user could do, and that includes interacting with a (theoretical) on-device AI model to complete tasks.

The Snap Store listing refers to it as a ‘design-discussion prototype of the next Ubuntu launcher’.

An intriguing choice of words – that’s as far as I’d read into it for now, though. A hint of a possible future feature perhaps, but not quite a commitment to shipping or actually building this as a real thing.

A launcher and a prompt box

I installed the prototype and ran through the simulated tasks to get an idea of what (there’s a menu full of them). The launcher itself is triggered by a small Ubuntu logo in the top panel and is only ostensibly interactive.

One simulated task saw the launcher used to find a local image file (using the file filter). The launcher’s AI box is used, and the model is asked to set the image as the desktop wallpaper and adjust the system accent colour to match.

In another, the “Followup with AI…” is used to find a suitable app the user can install to open an ePub. Matches from the Snap Store filter back inside of the launcher, complete with install buttons, which then trigger an installation. Nice.

And another, and to my mind more impressive use-case demo tasks AI with creating an add-able calendar itinerary from a handful of items selected from the clipboard history. Press play on the video embed below to watch that:

It’s intended to function as a regular launcher too, e.g., opening apps, findings files, picking emoji, clipboard history, finding open browser tabs and reviewing recent ‘tasks’.

Throughout the simulation results appear with all kinds of icons and actions to manipulate or manage them inside of the launcher, making this more of an active tool than a passive one.

Much like Apple’s (more capable) Spotlight in recent versions of macOS, in that regard. It’s not hard to imagine how the launcher could be extended further, to handle more than simple find-and-retrieve tasks.

We do know Canonical is keen to ‘ramp up’ AI features in Ubuntu. It released the Myna AI dictation tool in tech preview this month, and its engineers are using LLMs and agentic coding tools for various efforts, infra management, etc.

But the company has also said it won’t be bolting AI on for its own sake. Any integrations it does explore will, even in this ‘simulation’, be predicated around practical use-cases and making use of on-device AI models.

Try the demo (remember it’s a demo)

You can install the warty-prototype launcher demo from the Snap Store. It will run on most supported Ubuntu releases:

sudo snap install warty-prototype --edge

Once installed, launch it from the command-line. Click the Ubuntu icon in the top panel to reveal the launcher (and the menu of simulations to see in action).

Just keep in mind that you can’t use this to do anything. It’s a launcher mockup, not a real tool. It won’t integrate with your system, find your apps or double-up as your therapist-cum-psychosis-inducer.

For those wondering why ‘warty’, that was the codename of the first Ubuntu release – so named as the first version was likely to be a bit of a unrefined one. ↩︎

all 2 comments

sorted by: hot top controversial new old
[–] 8 points 2 weeks ago

Every day, I'm so thankful I use Guix.

  • source
  • [–] 7 points 2 weeks ago

    ubuntu is a misnomer

  • source
  • 7 comments in c/ubuntu@lemmy.ml view that post
    [–] 2 points 2 weeks ago (6 children)

    Given how open the traditional Linux desktop APIs are, how much functionality is exposed in CLI form, a "desktop" OpenCode-like system could transform a Linux desktop into something other OSes coild achieve with a lot more difficulty. For example, reducing the size of a video file is just a cryptic ffmpeg command on Ubuntu. If ffmpeg isn't installed, it's an apt command to get it. A "make this big video 4GB" prompt is trivial to execute by a local model. And yes, an "AI desktop" would only work for most of us if it can do it all on-device. Not throwing our trusted data to OpenAI's servers. A system like that would also require a strong harness to prevent hallucinations from rm -rf-ing the wrong thing.

  • source
  • hideshow 6 child comments
  • [–] 3 points 2 weeks ago (5 children)

    It's going to train new people to blindly type in their admin passwords and risk hosing the system

  • source
  • parent
  • hideshow 5 child comments
  • [–] 1 point 2 weeks ago* (last edited 2 weeks ago) (4 children)

    Quite possibly. But also possibly not. I've see plenty hosed systems by colleagues at work as they start using Ubuntu workstations with no experience.

    Some years ago: "Why can't do I get permission error, let me fix it good chmod -R 777 /. Hey Amoeba, my system doesn't boot anymore."

    One could in principle create a model that knows the best practices, the no-nos, the dos and don'ts of Linux and try to stand in the way of the user holding the hose.

  • source
  • parent
  • hideshow 4 child comments
  • [–] 2 points 2 weeks ago (3 children)

    Any system like this is going to be limited by the fundamental inability of genAI to be hallucination-free. Being programmed with best practices will help but eventually it will suggest something that will be bad for the user.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 1 point 2 weeks ago (2 children)

    True but a system like that wouldn't just rely on the model training. It'd also rely on the harness around it. The program that takes the text from the clanker's output and executes it. It can have, as existing tools do, a hefty static ruleset that allows / disallows / asks for user confirmation before executing. The ruleset would have to be significant and well designed to catch most disasters. Still wouldn't be foolproof, but neither are my own CLI commands, or my brain at all times of the day. A not-so-reliable operator can be further mitigated by using a snapshotting filesystem.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 2 points 2 weeks ago (1 child)

    Yes the harness is what i was talking about.

    You can get footguns also from your own cli commands, misunderstanding the documentation, or by copy/pasting from a guide without knowing what you're doing. But in both those cases it is clearer that it is your responsibility for understanding the context. And eventually that's how you learn.

    A generative, agentic system that does things for you without you needing to understand, and gets it right a lot of the time, but trains the user to enter the sudo password wihout understanding is not the UX I would encourage in an operating system.

  • source
  • parent
  • hideshow 1 child comment