Sigh, right, I forget that we live in the era of censorship, and words are more powerful/scary then atomic bombs!!
Thank you for your help anyway.
1 Like
Please don’t do that.
Words matter and there’s a standard of communication that is expected from all us community members. You are, of course, free to write whatever you wish on your personal space. So you are not being censored, you are simply being asked to follow the communication standards that the community has agreed on.
I have edited your post.
5 Likes
You’re missing the point entirely, but fine, as you wish. I got better things to do with my time then try to argue for a seemingly lost cause anyway. Humanity will reap exactly what it sows…
2 Likes
LLMs, or what is being called AI, is as I understand it, a large lookup table utilising a prediction algorithm. LLMs are also not really an AI.
I present a link by “professors of data science and biology at the University of Washington” https://thebullshitmachines.com/
I hope the title is not too provocative given the context, it is a university publication 
2 Likes
AI is just a conglomerate of selected algorithms to achieve the desired goal. Even spell checking is an algorithm and is/will be a part of the more advanced AI features.
AI is nothing more than a collected group of previously defined algorithms (and probably many new ones as well) that are put together to work in the desired manner.
1 Like
Ok @computersavvy and @theprogram maybe I didn’t express myself properly, my bad.
So let me rephrase, what I was trying to say is that what regular google search has been using/doing for decades is not generative. Whereas the main…(attraction? defining feature?) that has people so hyped over the latest stuff that they call “ai” is that it’s generative.
Does that make sense?
Pffff, looooolz, of course none of this is real ai, whenever I see these so called “experts” holding conferences saying stuff like “AI will be as powerful as the human brain in just 5 years.” I literally wet myself!! Well, almost, a couple of drops come out at least… 
P.S. Same as the current headsets are not actually VR, but if you wanna see how far I got when I asked reddit if I was the only one who hates calling it that you can do so here.
Spoiler it didn’t go very well for me… My takeaway from that was that we can’t seem to figure out some catchy industry leading/defining terms for what these things actually are so we’ve decided we’re good with just pilfering sci-fi for these awesome futuristic(very, very far in the future) terms. 
2 Likes
My beef is more with people tolerating 50ms+ latency with Virtual Desktop just because of wireless 
(Rift CV1 ftw and I had it at 15ms SPUD disabled).
As long as the immersion’s there I’m fine with it being called VR, but I can get more immersion from a low-latency tethered PCVR headset vs cheap single-LCD panel Quests with higher latency.
Then I guess your beef is with me 
, as I love wireless, so much so that when I first got into VR with the Vive Cosmos(yeah, yeah, I know, shut up, stop laughing! it wasn’t that bad a year or so after release…
) I also bought the wireless adapter for it which cost almost as much as the headset itself aka another thousand bucks or so…
Anyway at this point I would never even consider going back to anything wired, and if that means I have to put up with higher latency so be it. As far as the actual numbers I had with Virtual Desktop back when I was using windows, I don’t remember them exactly.
If the 50ms+ you mentioned is when you get stuttering and hickups then yeah I would consider that unacceptable and try my best to bring it down some. But as long as the latency(even if a bit too high) doesn’t cause such problems I’m fine with it.
1 Like
I’m not entirely against AI as long as it’s all optional and not baked in the core system and impossible to remove. I tinkered with Stable Diffusion but I don’t want AIs help to run my systems, do search for me or anything else. It’s a terrible idea.
The AI insanity is one of the main reasons I want to leave Windows as Microsoft will for sure bake AI in into core Windows 11 25H2 (or Windows 12) within the next two years. I build my own Windows 11 ISOs with all unwanted components removed but it gets more difficult with every yearly Windows 11 update. I’m still on 23H2 and I hope to find a Linux alternative soon.
Fedora 41 KDE is my fav distro so far. I had it installed on my second PC since early December, but if any form of forced, non-optional AI is in the Fedora’s future, then I’m back to testing other distros.
1 Like
Interestingly, I was searching for the AI model with the properties and licensing terms that Granite does offer. As a former IBM employee I’m quite optimistic and will check it out myself.
Not sure however what will be the shape of the final solution offered with the next releases of Fedora/GNOME, but imo, all concerns of users who have negative sentiment towards AI must be diligently addressed, added some opt-in, very good manual, kill switch, easy access to audit log, configurations that allow to limit resources for the local providers, GUI etc.
2 Likes
Very nice website, loving it.
1 Like
These people, myself included, are worried, more like “certain” actually, that AI tech will be exploited and abused by big corpos. So having AI as an integral part of the OS, that’s always there and that can’t be removed is a big NO for me and many others. Unfortunately, vast majority of the population doesn’t understand and doesn’t care, they just want new, shiny stuff.
Big corpos lie, they count on general public ignorance and lack of attention span, you can’t believe a word they say today, as they say something else today and will do something else tomorrow.
They want more money, more control and all of your data. Today Microsoft says Recall stores data only locally, on your computer. That’s great. Tomorrow we’ll find that the data is somewhere on Microsoft’s servers, because oops they “changed their minds”, because it’s all about “user experiences in the cloud” they would say. And if the govt comes knocking on Microsoft’s door asking for your Recall data, Microsoft will not refuse.
The AI tech is too powerful and corpos are too greedy and they won’t be able to resist the gobs of personal data captured and stored by AI tech. They just won’t.
I hope this won’t be a problem with Linux being opened and all that, but even now, some distros are adding questionable, polarizing features that can’t be removed or disabled. And seeing Fedora’s ties to a billion dollar corporation make me wonder if I am making the right choice. Perhaps I should have a longer look at the upcoming Debian 13.
One thing I like about Fedora’s ties to a billion dollar company is the surety and the engineers it brings in to my favorite free desktop.
As a self-confessed Paranoid Panda, I’m always on the lookout for corporate abuses. Yet if we consider the role Red Hat plays in large global infrastructure deployments, I can see little reason to think they are going to backdoor all their systems with an evil AI.
I suspect that within weeks we will see a clear and unambiguous statement from Fedora about AI, and I think you are going to like it.
1 Like
Oh, that is absolutely true. There is some good software (for all OSs) written by talented individuals, but there are certainly advantages to an organized group of professional developers working as a team (well, some exceptions do exists…). Mainly, this ensures constant improvement and longevity of software. My main issue with Linux is low quality and lack of polish of many desktop applications. Outside of Mozilla, KDE and GNOME, Linux desktop software leaves a lot to be desired (again, exceptions do exist). Which is a reason I ended up with Fedora KDE.
But I don’t know to what extent Red Hat is pulling Fedora’s strings and this news got me worried a bit. This is not the same Red Hat from 20 years ago. Though I don’t know enough to form my own opinion, but I saw some negativity directed towards them.
Red Hat will certainly not backdoor their enterprise warez with evil AI but their customers might be interested in some form of AI as most large corpo CEOs these days are gung-ho about AI. Also, Fedora is a “consumer product”, so who knows?
I hope you’re right and we will get some clarification from Fedora.
1 Like
I don’t think Ai is ready to be integrated into our operating systems yet. It makes too many mistakes and has stability issues. Maybe down the road, but not yet.
1 Like
Microsoft thinks otherwise. It’s fully expected that Recall and other AI invasive technologies will become integral part of Windows 11 25H2 or Windows 12, without any means to remove them, unless you run Windows Enterprise or Education edition.
I hope Linux devs will have more sense and keep the AI stuff optional.
1 Like
This!!! I just came here wanting to add another thing I just thought of, just in case anyone with power of decision is actually gonna end up reading this thread… And that is that something I value a lot(and I think should be among the things Linux focuses on) is consistency!!
Whether something works or fails, I at least want it to do so in a consistent way. But with an ai system that one second gives you the correct answer to your querry and the next second tells you you should eat at least one small rock per day…
I don’t know about you, but to me, the definition of consistency that ain’t! 
LOL, to be honest, a true AI (not what we have now) would be more like a human mind, no?
So I would absolutely expect it to be inconsistent, with moods affecting its “mental processes”, reliability and consistency. It would learn like humans, often using incomplete, inaccurate, ambiguous data, coming to wrong conclusions. At least its memory retention would be superior. But I think the general expectation and understanding is that AI should be vastly superior to humans, but what if it is a moody, irritable genius kind of thing?
Many people are concerned based on the experience they have with LLM, and as far as I am aware of this is not the thing that Fedora is going to implement.
Take a look at this more as a set of very specialized small tools and not an overwhelmingly large AI being.
For example, I’ve been experimenting with Goose, that is a client for AI language models that does not aim to answer every question possible but performs certain specific tasks instead, like designing automation scripts, doing small translations, or validating the correctness of playbooks, explaining manuals, or fixing config files for our projects. Narrow, but specific usage.
1 Like
As long as all this is optional, installable/removable apps, I’m fine. None of this needs to be an integral part of the OS.