Linux doesn’t deserve this
Imagine taking the lightweight, privacy-focused appeal of Linux and completely ruining it by hardwiring AI models straight into the OS. Nobody is switching from Windows for more forced AI features.
Noteworthy that this is being done by the guy behind “Cyanogen” and he’s doing it with Linux for the same reason as Valve; so that they’re not dependent on a Commercial OS that they cannot control. Their “Cortex” engine has real possibility since it leverages local resources before shedding work to the cloud. If it works as promised it’s likely that Microsoft will soon be making them an offer they cannot refuse.
Here’s the original techcrunch article: https://techcrunch.com/2026/11/10/after-challenging-googles-android-with-cyanogen-its-co-founder-takes-aim-at-windows-for-ai-era/
Here’s a link to Kora’s Website: https://kora.fun/
My god what an obnoxious website.
If it works as promised it’s likely that Microsoft will soon be making them an offer they cannot refuse.
Yet another reason to avoid it
I’d object to this if it were open source and forkable, but looks like it is not and so your OS becoming corpo-fied is an especially salient and unavoidable risk:
Cortex is KORA’s own technology and is not open source
I will probably give KORA a try when it’s available. I’m not on the “fuck ai” train and I see no reason to avoid experimenting with something like this.
To your point Microsoft has already started on something similar anyway.
Fuck it, how do I make my computer even stupider?
Buy a Mac
I’d sooner buy one from McDonald’s
Apple is preoccupied with making everything smaller; McDonald’s, meanwhile, went the complete opposite direction with the Big Mac.
Apparently there’s supposed to be a link to an image hosted on lemmy.ml. My instance not federating with that one may be why I don’t see it. If anyone else was wondering, it’s just some random startup advertising their new “AI-focused” OS.
Beehaw defederated from .ml and I’m still seeing the image, oddly.
Fedia defeds .ml? That’s awesome LMAO
But federation generally doesn’t affect links like image links since they’re direct URLs instead of federation.
But I also don’t know fedia’s platform all that well so maybe it does on fedia lol
The image also doesn’t load from kbin.earth, which does federate with lemmy.ml. I see this occasionally, and I’ve considered filing a bug with the mbin folks. It’s interesting that it also happens with piefed.
EDIT: I’m dumb, fedia is also mbin.
I have two accounts on pure lemmy servers. The one that doesn’t federate with lemmy.ml doesn’t display the thumbnail for this post, but coughs up the image if I click where the thumbnail is supposed to be. The one that does federate with lemmy.ml shows both the thumbnail and the image. Someone probably has to put a bit more thought into image propagation and storage for corner cases like this.
Who wouldnt prefer their desktop to guess what they want it to do?
I feel that the AI desktop is really seductive and I’m not surprised a company is using Linux this way, even though I feel it is misguided because LLMs are not the AI we were promised.
Really, it appears like the fulfillment of the scifi future depicted in the pre-home-computing era for automated computing devices using “natural” interfaces. Think say the Jetsons or something. We are free from the keyboard and mouse drudgery of organizing and finding things ourselves,just tell the computer to do things.
A lot of people really really do hate using their computer and this is their dream, for better or worse.
The majority of people would probably use a voice UI (VUI) over a GUI or CLI, but even more so with NLP involved. Hey I want it just for accessibility reasons.
Its just not worth it to rent from cult, facist, or megacorps willing to destroy the world to do it. Local LLMs though are cool, and hopefully we set up for better distributed training again.
Its just not worth it to rent from cult, facist, or megacorps willing to destroy the world to do it. Local LLMs though are cool, and hopefully we set up for better distributed training again.
KORA is built around that idea. Their “Cortex” engine runs workloads on local LLMs first and will use resources on other computers on the network before it sends work to the cloud.
It’s a very good idea and if it works as advertised I suspect Microsoft will offer them a LOT of money for it because it’s almost exactly what Microsoft’s “Hybrid Intelligence” is trying to do.
Local LLMs can be really useful if your hardware can handle them.
Combine it with that OS that wipes itself when you enter an invalid terminal command.
