There are several ways to counter that sort of thing, but let’s start from the beginning. LLMs (what people call AI) is VERY computational heavy, you need a powerful GPU to run a model locally, and it occupies lots of power and memory. The idea that we’re even remotely close to something like that being embed into hardware without people realizing it is just absurd.

But let’s imagine someone is able to make it, and magically prevents hackers from breaking it and using it as extra free power. This will have to live in the CPU as anywhere else wouldn’t have authority to “delete files”, and even the CPU would have a hard time doing that. Now this LLM needs to distinguish stuff I’m writing with stuff I’m reading, otherwise it would also delete files when someone is observing me. It also needs to reply in sub millisecond otherwise the computer will lag absurdly. It also can’t update it’s local model because it doesn’t have network access, so just use tokens it hasn’t heard of.

In short if someone managed to add a piece of hardware capable of doing that it would have to be significantly more powerful than the piece of hardware it’s embed in, and it would only work until someone breaks it and gives everyone a free hardware upgrade.

You can relax, nothing like that is even remotely close of being theoretically possible.

That being said, Windows doing this or similar is a possibility, your best bet is to use an open source system.

I feel like you might be young and not have had to actually use analog stuff. Your whole family trip photos could be gone in an instant because you burnt the film accidentally, or you could lose the film before getting it revealed, and although rare (probably as rare as a bit flip destroying data nowadays) it happened that pictures were destroyed during the revealing process, and even if all of that worked the pictures could have been over/under exposed, out of focus, or any other variety of issues.

Typewriters? I wrote stuff in them when I was a kid, granted computers were around back then, but I liked the sound. They would jam the hammers, run out of ink, or just annoyingly one letter would not work. If you’ve made a typo or wanted to edit something the entire page had to be thrown out and rewritten, and if it didn’t fit now the next page would have to be rewritten as well. And now that you’ve finished writing and left the pages on the table a spill could destroy your day’s work, or your dog could eat your homework.

Film? I don’t think anyone here has actually dealt with film unless you work in the industry. For home users we used to use videocassette, which is a digital medium, and a very flimsy at that, dropped soda on it? Gone, it got stuck in your player? Gone, you put a magnet near it? Gone.

On the other hand digital pictures, text and movies you can have multiple backups effortlessly and completely avoid any possible single disaster scenario.

Maybe not yours, but I’ve had a fire in my life that deleted lots of pictures and stuff, whereas all of the digital media we have is still with us because we copied it to several places so no single event could destroy it. If you only had one copy of an important digital file, you’re doing it wrong.

That would make Mint unstable. That is exactly what unstable means in Linux context. There are debian based rolling-release distros, including Debian Sid. This is one of the reasons people choose Arch, because it’s a rolling release you never have to worry about version.

There’s a good chance you might break stuff by upgrading major version like you fear, and that’s why it doesn’t happen automatically. That being said it should be safe, but good on you to prepare backups.

They do when it makes sense. Hosting small Giles is not a big deal, and small files would be a bad experience on torrent, since by the time you get peers to start the download you would have already finished it from a normal server, not to mention that you need to host the torrent file anyways. Also things that have lots of releases/patches are a bad experience, since people might be seeding an old version.

However, large files that only get sporadic releases, such as distros iso, can definitely benefit from it and you can usually find torrent links for them.

Please research the meaning of stability when applied to Linux before parroting stuff. Also, who mentioned Arch?

Warning, I’ve had an issue in the past where I couldn’t play a game (Deus Ex Mankind divided) because it needed a specific instruction set on the CPU (SSSE3). While not your specific case since the FX-8350 supports SSSE3 (I should know, that was the exact CPU I switched to to be able to play the game) there might be newer instructions sets that this old CPU does not support.

Also that GPU is older than what people like to remember, from a time where AMD was the worst GPU option on Linux. It’s very likely that the open source driver is good enough for that card by now, but there’s a good chance you might need to wrestle with the AMD proprietary GPU driver (fglrx) which is worse than the Nvidia one in some aspects and some distros don’t even package it anymore.

If you plug your Nvidia GPU that rig would be usable for gaming, I’m not sure what fps you would get as games keep getting updates and old hardware remains the same so old benchmarks might not be reliable, but I suppose it should run plenty of stuff.

You used Linux like Windows and got bad results, OP treated Windows like Linux and got bad results. The problem is neither OS but how familiar you are with it and their peculiarities.

That being said, GPU drivers are not a rough edge on Linux, only Nvidia drivers are. And even then it’s usually a single click/command to install the proprietary drivers if you need them, otherwise the open source ones work like a charm. This used to be more of a problem a few years back, when both manufacturers used proprietary drivers, but AMD ones are open now and therefore integrated into the mainstream kernel, so they just work.

Yes, but you would need to know to run that command, so it’s the same situation as the windows case where he didn’t know which drivers to get. So the argument is disingenuous in that it either ignores the case or he has knowledge on one OS that he doesn’t on the other. On the other side of the coin someone could be making a similar post saying in windows they just switched hardware, installed drivers and done, in Linux they spent hours trying to figure out how to install the drivers.

I’m not saying it’s hard (on any OS) but it requires previous knowledge on both (although to a much lesser extent on Linux since this only happens when switching GPUs and only under specific conditions).

Can you give me an example of which distro/hardware change gave you a black screen? Because unless it was Gentoo or something you built the kernel yourself a black screen is extremely unlikely. Unlike Windows which requires drivers for everything, in Linux the drivers are baked into the kernel, so any hardware change should just work out of the box (there are some caveats to get the best possible driver, but even the included driver should be more than enough for almost anything except heavy use on Nvidia GPUs).

I agree that on average the Linux user has more technical expertise than the average windows user, but that’s mainly because the average user doesn’t choose their OS. If you take into consideration only people who actually chose their OS, I think it’s very similar.

And OP talked about his experience doing that, the default windows driver gave him a crappy resolution, and he had lots of issues getting the right driver and making it work. You skipped all of those issues because you knew beforehand which was the correct driver, and pre-downloaded it.

Didn’t knew about the mods, that might be a rabbit hole to dig into.

This is disingenuous, if you had an AMD GPU on Linux and switched to an Nvidia card you would be using the nouveau drivers so you would need to install the proprietary drivers to get the best performance.

And lots of the same issues that are listed on the windows side could happen on Linux as well since they relate to connectivity.

I agree that the original post is dishonest, but your solution is exactly the same as what they said with the exception that you knew it would be a problem so you downloaded the driver beforehand. Had you not known that would happen the series of misfortunes could have happened to you too.

If you know that that’s the right driver to install, what is your question?

I usually give detailed responses, but honestly the correct response here is RTFM. The short answer is to install nvidia-580xx-dkms.

Arch wiki is such a great place that has the answer to most technical questions you might want to ask. I strongly dislike the idea that Arch is for advanced users, but it does expect you to read the documentation (which is why I dislike stuff like Manjaro that try to make Arch “accessible”, but end up leaving people in similar situations without even knowing where to look for the solution to their issues).

Because you brought up a non-proton related issue. It’s highly unlikely this is proton related, this is an issue on how the steam virtual keyboard and Skyrim interact, and since the keyboard doesn’t cause issues on any other game I think it’s very likely the culprit is Skyrim, and if that’s the case I want to know if the issue also happens on Windows.

Is it possible that the issue is within proton? Yes, but the problem you presented can be in multiple other places that proton can’t touch. For example if the game crashes on windows too then proton is working correctly, and even if it doesn’t it can still be something else besides proton, e.g. steam closing the keyboard in a different manner, Skyrim has had a famous bug where it crashes when you alt+tab away from it and come back on Windows for a long time, and opening a virtual keyboard on top of the game and going back is essentially alt+tabing, so I wouldn’t be surprised to find out that this is the same bug in Skyrim showing up with a different face.

PS: out of curiosity I went and loaded my old prisoner save, and named the character which is how many people reported the bug and it just worked, so even if this used to be an issue, doesn’t seem like it is anymore.

What happens when you do that? And does doing that on windows work? Because it sounds like an incompatibility between the steam virtual keyboard and Skyrim.

The solution I’m talking about should already be the standard by most devs (especially small studios), even before LLM was a thing. See, small teams can’t afford QA, at least not to the same extent as big studis, so they need to add checks to stuff in a way that catches large problems, and a placeholder making it into the final game is a big problem. Even before generated images were a thing devs would just use any random image they had that more or less worked, and those images could have copyright or be problematic in any other way, so ensuring none of that made it into the final release has always been important.

Dude, naming the textures placeholder_<name> doesn’t take any more time and ensures you won’t ship a game with a placeholder. This is, or at least should be, common practice even without using LLMs, and only takes a couple of seconds, not enough to cause any inconvenience.

That’s not what a concept artist does, concept artists (if they had one) did the work before, game artists are still doing the work while the generated placeholders are in place, no person’s job was compromised by using generated placeholders. That being said, if any placeholder made it into the final game then fuck them.