What does everyone think about the comparison of title at this point in 2024? We’ve seen only devs copying Windows 11 like “accent colors” (an embarassment to be copying the thing everyone in the community hated and a competitor all the time, but that could be just me as I see no controversy).
I told penguin about ai but I see not even an ant worth of action on it after 2-3 years of warning but M$ brought it in instantly without breaking a sweat. Not a single pip about it on linux. Again the, pouring glue speed of linux is haunting the platform…
I am confused about this post… What’s the actual question/argument of this post?
The title asks for a comparison of Gnome of Windows. First of all: Such a comparison is difficult, as one is a desktop, the other a system. Also, the UX of Gnome and Windows is quite different. So, in order to make a good comparison, you’d need a use-case as reference.
Then you talk first about “Gnome copying accent colors from Windows”. Accent colors has been a concept on Linux for a long time, its just recently that it was standardized into one system most desktops agree on. Before that, it did exists if one wanted to.
I believe the last paragraph is your actual point, as in “Gnome does not do enough AI”?
While I can’t speak for everyone, here is my view:
AI, as in Large Language Models (LLM), has a few issues, from extensive use of resources, the blatant disrespect for the works of others (and most likely massive copyright infringement), its tendency to hallucinate or questions about privacy when using online AI services.
For this reasons, I don’t believe AI should be integrated tightly into Gnome itself.
That being said, there are good ways to use AI on Gnome, following the idea “one app for one task”. If you want to use an LLM, you could use apps like Alpaca or Bavader. There are also extensions if you want the interface in the shell.
The great thing about open-source is that everyone can use code to build the environment suited for them.
If you think AI is something that should be core to Gnome, you would need to provide a good, precise use-case that can be best solved with AI, that most user would actually benefit from being in core (instead of an optional app or extension), and maintains good performance and privacy.
Like I said in that sentence, I told penguin, which is the kernel developer. It wasn’t for gnome that suggestion but everything. The AI would be one that did not steal/retain data like current ones obviously.
The comparison is about the desktop since that is shared between both, and intuitive to select to compare.
The accent color mention was because it is in the latest release of gnome that iI briefly checked before making this topic. My task with this topic is to see if there’s remaining dissatisfaction with linux desktop pace.
Well, first of all: The Linux kernel and Gnome are separate projects. Linux is often used without Gnome, and Gnome can be used outside of Linux (with BSD for example).
I could respond to every point in detail, but there is one sentence in your reply I want to specifically point out:
This is the problem with your question:
You’re not proposing that AI should be used to solve or aid a specific use-case. You want AI for everything. Such a sentence basically says: “I’m looking for a problem I can solve with my solution”. That’s just the wrong way around.
Just adding AI for the sake of adding AI is not sensible. A better idea is to have a specific use-case, where AI would be a good solution, and then implement it there.
As an example: If you’re browsing the web, you might have websites in foreign languages. Here, you might want to translate those sites for the user.
So, you implement an LLM which can locally translate the content of a website. This is what the Firefox translation feature does.
This is an example on how AI is implemented to solve a use-case.
If you think Linux should have more AI, you should identity such use-cases which would benefit from it. The advantage would also be that you don’t need to create a “Do-it-all” model, reducing the need for resources and copyright infringement.
Otherwise, the LLM apps can already be used.
As a final note:
You mentioned the “slow pace of development”. First of all, in the last release there were more than just accent colors. Though not all of it is immediately visible, it still had a lot of effort put into it.
And, a lot of the developers do this work in their spare time. There is no “Gnome LLC” with employees and a CEO making decisions. Its the Gnome community, working on a shared idea. Please keep that in mind.
Of course I have case uses and can see the purpose in many areas. Discrete cases are noisey and spammy in the overall interface, so you would want it to do everything needed. from one input. I should be able to type, “screenshot of ui bar please, and put it in …, reducing size 50% or attach to a new email.”
Blockquote
There is no “Gnome LLC” with employees and a CEO making decisions. Its the Gnome community, working on a shared idea. Please keep that in mind.
I thought there was a huge amount put into linux now, with people being able to attend linux conferences etc. Seems crazy there isn’t someone on this to destroy competition. It makes sense for other countries to do such a thing and have such a goal, rather than change a “UIX” label here or there as part of an update.
I didn’t request to be a mirror of windows, but simply to stop copying it like “accent color” and many other incidents of copying (another was the ubuntu thing of trying to put linux desktop on all devices with a touch interface after the windows 8 release…) and dumbing the interface then down to a phone interface. I remember all that and more. You will see it yourselves with some self-honesty. I requested an aggresive competitor mentality of professionals.
Now it could be the case, countries that have commercial OSes will have outfielders constantly pouring negativity on those who want something in linux desktops, or dissuading developers or others away from such paths, as a strategic, commercial objective. That would explain the hundreds of different linux variants that all are the same more or less, without anything competitive.
What you’re “requesting” here is the improbable, if not impossible.
An “everything AI”, if it is possible, would be an enormous effort, probably requiring completely new systems for interaction between AI and UI, etc.
First of all, that would be an really expensive endeavor.
And for what benefit?
Yes, you can write “screenshot of ui bar please, and put it in …” to an AI and let it work for a time.
You could also just press the Screenshot key, select the area you want, and then paste it into your document/mail/folder. The latter is fully implemented in Gnome and works, most likely faster than you could ever write that prompt in a way the AI understands without error.
The idea of Gnome is to create a UI in desktop and apps which gets out of the way of the user and let’s him do his tasks as efficiently as possible. That’s also the reason for the simple “phone interface”, which has been in the making long before Windows 8 was a thing.
And use-cases are important. As an actual developer, you need to understand what it is you want to solve, and use-cases are vital for that. Especially if you don’t just want to waste resources on unhelpful waste.
And for the other point:
We don’t aim to “destroy our competition”. We just want to create and use our desktop environment with our goals of privacy and simplicity. If someone is fine with KDE? Sure, have fun! Actually, why don’t work together on a standard or two, so apps from one desktop work better on the other.
This is not a profit-orientated company, this is a community of like-minded users and developers. Sure, we all wouldn’t mind if there were more funds available to work on it more or more people using Gnome, but not with the goal of yours in mind.
What you are envisioning is very different from the goals here.
What’s been curiously comical to me using Linux over the last 20 years, it this notion of competition between Linux and Windows, I don’t think there is/was. Linux for me is/has been simply an “alternative” to Apple/Microsoft and their ‘proprietary’ colony.
I do constantly witness Linux ‘users’ attempting to inject competition using the DE’s to convince an audience to create/join their NotWindows colony. Personally, I don’t believe any of the current platforms will exist on any endpoint in another 7-10 years.
You would simply need to expose many/all of the app functionality commands asynchronously to the open ai, not all too much different to the web apis at present have been trying to do. Then it can call whatever part it needs and attach things like shortcuts on the fly, or not, with other program apis, where the output is piped to. Maybe something like that where anything in an app is open to the system. Then you could say i want 5 screenshots and on shortcut key 1-5, too, and open each one in an editing app to custom crop and “i will select the area to screenshot each time”, then “ask me what to do when each is closed”. Or, “tell me if the weather is going to change to rain an hour before it does on the desktop” as a permanent state, with the app having various display methods to give out. things like that.
Blockquote
This is not a profit-orientated company, this is a community of like-minded users and developers. Sure, we all wouldn’t mind if there were more funds available to work on it more or more people using Gnome, but not with the goal of yours in mind.
What you are envisioning is very different from the goals here.
And this is my point about gatekeepers who want to strategically limit linux. I disagree with it strongly and condemn the mentality. It’s about the strategic limitation and diminishment of linux desktop on purpose. Already penguin is on a leash to a US company as a retainer to keep him from doing anything big.
“I want to painstakingly open a screenshot app, select the function, save to a location i browse to each time, open the file explorer to find it, choose a file name, open an app to crop and resize each and go through the save interface all over again only to then go to another app to finally search for the files in file dialog box to attach to an email”
is NOT “out of the way” and is as annoying, cumbersome and painstaking as the commandline.
You would “simply” need to rework the entire system. Well, if its that “simple”, well, get to it then!
Gnome is a big ecosystem full of apps, the shell, and many components beneath them.
Your idea may seem “simple” at first, but only until you try to actually do it.
More importantly though, you still don’t have any use-cases. A use-case describes a problem or task a user want to accomplish, without an solution being predefined. The whole idea is to find the best solution to the use-case.
You still just try to find problems to your solution.
This makes me wonder if you actually tried Gnome before… If not, the newest release of Fedora Workstation provides it in an unmodified way.
I did mentioned the actual behavior before, and I would say it really is quite “out of the way”:
And now we got a conspiracy myth against Linux as well! How fun!
What’s been curiously comical to me using Linux over the last 20 years, it this notion of competition between Linux and Windows, I don’t think there is/was.
Gates: Linux is no threat to Windows (1999)
March 24, 1999
Microsoft CEO Bill Gates, on tour to promote his new book, “Business @ the Speed of Thought,” said Wednesday that the popular open source operating system Linux is no threat to Windows. Gates says that Windows offers far more functionality
