Nvidia pilfers Youtube for AI, GNOME HDR, Immutable Manjaro: Linux & Open Source News

[@TheLinuxEXP]

Get the free report on the increase in Linux CVEs, and how they can affect your organization: tuxcare.com/downloadables/a-look-at-3-months-of-linux-kernel-cves/?The%20Linux%20Experiment&

[@Riv_Nexus]

two hours ago??

[@TLink9]

I don’t think video game companies should be forced to keep live service games up and running. However they should be required to release tools for the community to run their own servers after discontinuing support.

[@TheLinuxEXP]

Yep, totally agree

[@softwarelivre2389]

I wouldn't say release tools, but qt least enough documentation and change API keys to a public one.

[@askeladden450]

They should have support for self hosted servers from day 1 so people can self mod their own servers to get rid of cheaters without needing rootkit anticheats, as well as running modded servers and custom game modes.

[@Sherolox]

And for games with cosmetics or unlockable weapons/characters, they should be made available free of charge, or exclusively through being earned in game, after the game is no longer continued officially.

[@TheFrontmanRocker]

That's exactly what the initiative is about

[@toxiccan175]

The “proposed law” is actually an initiative, or loose proposal, for a law to be written. It is less about continued maintenance and more about permanent software ownership and the facilitation of passing off games to their communities rather than dropping support and destroying playability.

[@Patryk128pl]

16:00 That's completely wrong. The idea is when the publisher/developer stops the support for a game, they should've left it in a playable state. That would imply that if a live service game stops getting support and the main server is shut down the publisher or developer would have to release a patch that would allow for offline game, or to share tools to host private servers. That doesn't mean to maintain servers indefinitely.

[@adambyte256]

That's exactly what he says from 17:35 to 18:00. Watch the damn video.

[@CuteSkyler]

The little guy using copyrighted music in the background? Gets taken down immediately. A gigantic company downloading 130 million hours worth of KZbin videos for commercial purposes? Damn, sounds like a gray area.

[@askeladden450]

Funny how all these corporations are so vocal when it comes to piracy (which is copyright violation) but have no issue doing copyright violations on a scale a million times larger. Seems like its only an issue if consumers do it. If what Nvidia is doing is considered legal, then piracy should be considered legal too.

[@techpriest4787]

They did not violate copyright law nor moral standards. TY offers a download button in the first place. I am also a trained concept artist and not only software engineer for game logic and game engines. As artists we ARE instructed to use other artist's work in other words to copy or learn from them. AI does not do any different. Both are neuromorphic system even. It is about time that your type quit spreading misinformation on this. I am tired having to explain it to you know-nothing people: you are not a victim

[@askeladden450]

​@@techpriest4787 this is the funniest stuff I've seen all day. You apparently have no idea how the copyright law works, how AI works, or even how artists work. 1. There are limitations on how much of the source material you can use under fair use laws without the permission of the copyright holders. This is usually limited to portions of the source material. These neural networks use the entire source material (every single pixel/byte of it) without any permission. 2. When artists 'copy' other people's stuff, they are morally trained to try and make it not look too similar to the source material. That doesn't exist with neural networks. On specific runs, they can make pretty close copies of the materials they were used to train on.

[@monkyyy0]

@@techpriest4787 this is probaly a paid shrill but.... anyway > neuromorphic NN'S ARE NOT BRAINS, human brains are not doing perfect calculus, human brains have loops, etc etc etc

[@MrDiarukia]

@@askeladden450 "There are limitations on how much of the source material you can use under fair use laws without the permission of the copyright holders." - Not knowing the exact wording, is using not in terms of displaying the copyrighted work, which AI doesn't "When artists 'copy' other people's stuff, they are morally trained to try and make it not look too similar to the source material." - Well, fanartists exist, and they often technically violate copyright, though given, no company in their right mind goes after it. The thing with AI is, unless you try to recreate something specific, you stay pretty far away from a copy/imitation. Collecting training data is the grey part, and the morally questionable one. I personally, do not have a problem with general purpose models, i.e. learning from everyone. But making LoRas etc. (specializations on certain subsets) of an artist style and/or original characters should not be happening without consent/compensation of the rights holder, since your stated goal is to copy in that case.

[@lenOwOo]

For real, the guy must be from another world timeline ​@@askeladden450

[@willi1978]

copyright only helps big companies, current law favours them

[@MiukuMac]

Finland, the only country that has surpassed the 100% so far in the initiative. Heeeyyoooo!

[@erufailon4723]

Though we're also one of the countries that doesn't require authentication to sign it and will only verify the signatures after the campaign ends, so I'm a little wary of those numbers. But it does show that the initiative is popular here, because those signatures can't all be invalid. Still, I'd only treat the ones from countries that require authentication as 100% confirmed.

[@rafalg87]

"Secure session" feels kind of misleading to me. I know Wayland is more secure, but it's not like Xorg users get hacked left and right. I think I'd go with "Default (Wayland)" and "Legacy (Xorg)".

[@connivingkhajiit]

I thought the same thing and I agree

[@kokop1107]

It's completely wrong and a very misguided view on the threadmodel of desktop sessions (which isn't even considered a real threat, like at all in the cybersecurity world). Very anoying they try to steer users with this loaded language that is based on misinformation

[@MyReviews_karkan]

"Stop killing games" was started by a man named "Ross Scott" and he took it to the EU. He has a KZbin cannel called "Accursed Farms".

[@OraOraOra]

Thanks for the News! Wtf.. the Nvidia thing is crazy! And thanks for covering "Stop Killing Games" based take!

[@theirs7351]

nvidia is genuinely an evil company

[@negirno]

They became filthy rich on the AI hype.

[@hotrodjones74]

@@negirno Part of the recent stock market crash was caused by AI hype and American investors abusing the the Japanese Yen in the carry trade. Also Intel laying off 15% of their workforce didn't help. Corporate big tech is hot trash. The massive bullshit company AI bubble will crash very soon and it'll be a hot mess.

[@cyberturkey77]

I don't think what NVIDIA is doing here is wrong tbh. Most of whats on KZbin violates copyright anyway. They're not reusing the videos (uploading them). They can do this and its fine imo

[@wojtek-33]

​@@cyberturkey77 So i should be able to download a bunch of tutorial videos from particular creator, package them into a course, and sell that course to others?

[@ytechnology]

NVIDA is pulling the “It's easier to ask forgiveness than to beg for permission” routine. Individuals and companies will have to press NVDIA for compensation in due time.

[@toxiccan175]

Hoping XFCE Wayland support comes sooner rather than later…

[@skelebro9999]

Finally, Nixjaro is real.

[@DigitalEntity]

How about making the games to work offline after the eol from the developer/publisher?

[@KhakiCap]

I thought that was one of the proposals in the initiative.

[@SuperTort0ise]

​@@KhakiCapit's an option, but not required so long as you can play the game, which might mean server software, patching to make it entirely offline, or just release the code, whatever works.

[@ran_red]

How long until tech companies train AI on private cloud storage. I honestly fear it's probably already happening, I wouldn't put it past *any* of these big tech companies to start scraping your private S3 buckets, Google photos, etc

[@33KK]

>Google Photos Oh you sweet summer child

[@JE-MG]

They already are. Why do you think MAcrohard forced everyone’s local one drive into the cloud.

[@ran_red]

@@33KK lol I have no clue what's what in the Google Shit Suite, I've never used it

[@igordasunddas3377]

​@@ran_redit's everyone. That's why I am so puzzled when companies and individuals want to use a cloud service... Are they too dumb to recognize, that it's basically a computer of someone else, possibly your enemy and a thief even?... Jeez :/

[@TheVincentKyle]

Does anybody else watch the piece on GNOME and ISO files and think "yay, GNOME is now on feature parity with [checks notes] Windows XP!" I mean it's of course welcome, but doesn't really feel like anything to talk about.

[@oimateusaqui]

I'm just commenting to say that I love your videos ❤️

[@tgheretford]

The idea that NVidia's defence being that humans do it too and therefore it's fine falls flat when you realise that humans also suffer consequences for doing the same thing that they're doing with AI at a much smaller scale - it's called plagiarism. If I watched your video and then made my own video using my own words but your ideas and passed it off as my own work, you may not be happy about that. It's not legally classed as copyright infringement but you'll suffer serious consequences for doing it. If people want plagiarism to be classed as copyright infringement, they need to write to their democratically elected representatives.

[@TheLinuxEXP]

Absolutely!

[@godminnette2]

They are not doing that though - they are training a model on many videos, one whose output does not directly reference those videos and has randomness as a factor when creating output. It is not plagiarizing - however, models do not learn exactly as humans do (we learn a bit by data training, but we learn more by logic training, which these models don't do), and if Nvidia tried to train someone on these videos and also claim perpetual ownership of that individual and anything they ever produce... Well, slavery is illegal. A company owning an asset they can profit on is different from a free individual.

[@temari2860]

@@godminnette2 What do you define as logic training? That's just data training but internal, which reuses some data for more diverse data. Advanced ML training also has that if that's what you're revering to, otherwise I've never heard of any "logic training" happening in human brain.

[@temari2860]

Nah if you're taking it to human terms, that would be an equivalent of you creating content based on your entire experience with everything you've seen and interacted with in your life, so just based on your personality and ideas. Whether people want to make special separate rules for AI is another question, but as it stands NVIDIA's argument is valid and that's exactly why AI training on data is a gray legal area, because what's it's doing is learning and legally it's no different from you looking at pictures which impacts you as a person and your future works even if a little and in an unpredictable way.

[@cyberneticsquid]

@@temari2860I think legally there's the rather notable distinction that humans are people and AI isn't

[@torzk7966]

I agree with your points regarding Cosmic. I'm probably not gonna try it out until release, but I am very excited for it. Hopefully, Nvidia cards will be more stable on Wayland by then as well (especially on hybrid GPU laptops).

[@NiffirgkcaJ]

I really hope that the EU will do this! It's really absurd for these publishers to not allow people to extend the life of these online games the moment they decide that it's time to make money again.

[@zocker1600]

Hey Nick, this is everyone! :)

[@BelarusUnite]

hehe, i usually write the same to Tim Cain

[@drewo.127]

15:49 my thoughts exactly!!! I’ll keep repeating it until the heat death of the universe: AI NEEDS TO BE USED TO HELP HUMANS, NOT REPLACE HUMANS! ESPECIALLY IN THE CREATIVE FIELDS!!!

[@smashed_penguin]

better-fs vs 🧈-fs

[@TheRealStus]

Don't underestimate humans. I can ignore every single license too.

[@savagepro9060]

Butter-FS? Isn't that how TUX gets . . . High CORE-lesterol!

[@DrathVader]

Really hyped for HDR on Gnome. I got an HDR monitor a couple of weeks ago and it looks amazing when it works properly.

[@YermakTheTrainEngineer]

Nick this one was awesome, thanks for the update man

[@pixiedev]

3:50 yes you're right. the perfection in the ui they have is much appreciated. But I stop using it because of slow boot time (3 minutes) 😅

[@quickkennedy3226]

hopefully all the EU friends sign the initiative!!!!

[@infinitebeast5517]

Making companies maintain games is not realistic but what they should absolutely do is take out any drm that won't let you play the game without connecting to their servers and also provide tools to allow the community to run servers or the game

[@SuperTort0ise]

So you agree with the initiative, It is absolutely *not* asking for companies to maintain a game forever. it is however asking for exactly what you said!!

[@CaracuSC]

I've never been this early to a video. HDR on GNOME would be great!

[@tutacat]

The new thing for ISOs on GNOME is just the popup. You could already use all those features either from open with, or through the disks app.

[@tutacat]

I think the petiton means they just have to not be hostile.

[@juxuanu]

As a dev, there is no way I am trusting a bash script to manage my whole system images.

[@Winnetou17]

Not even if it was written by you ?

[@juxuanu]

Even less

[@gus3000spam]

​@@Winnetou17 especially if it was written by me lmao

[@underratedphonker6395]

Devs shouldn't use such a mess like manjaro anyways

[@Slugbunny]

DRM and always-online have been such a disservice to customers. About time there was legal pushback for a return to owning what you paid for. It would be very simple to achieve with software. Speaking of property, hope Ngreedia faces consequences sooner rather than later for yoinking IP to only contribute to worse versions of it - and their bottom line, of course.

[@fuseteam]

The hardware Enablement Kernel is a newer version of the kernel; 22.04 currently has 6.x instead of 5.15 which it shipped with

[@ronm6585]

Thanks Nick.

[@linuxstreamer8910]

let's hope elementary os for the update does not need a fresh install. it is not like it is impossible because every distro can do it

[@GXShade]

I have a feeling that Immutable distro will become the norm when it comes to desktop Linux.

[@pradeepmalar327]

Louis Rossmann made a great video on the "Stop killing Video Games" petition.

[@nittani.]

Stay tuned for axylos2

[@BruceCarbonLakeriver]

I'm full on your side on the AI crap, besides the fact you're doing great with your Linux vids. Greetz to France from Austria :)

[@TrevorCachin]

Live service games are like investing in a shell company knowing it's just going to disappear and everything you put into it will be gone. Games just aren't made to play LAN or host P2P sessions anymore...

[@DaveRyder-po4gq]

Always informative.. 👍 👍

[@dand337]

Deepin 23 is planned to release next week!

[@danfg7215]

I'm trying to like immutable distros, tried out Fedora Kinoite for a while, and now Vanilla OS Orchid, but I can't seem to get anything basic working that would usually require root access. Right now I'm stumbling with mounting external drives, easy to do on regular Linux, not so much in Vanilla OS.

[@Techmagus76]

In theory robust, but very experimental and unstable fits perfectly how i remember Manjaro.

[@a.randomjack6661]

There's a bunch of AI generated content on YT. Trying to understand this world, I like to be informed about topics like geopolitics and human behavior types, to name 2, I come across AI generated content curiously ignoring important facts, which make them as useless as mass medias. Mass medias. "If you don't read the news, you are uninformed. If you read the news, you are misinformed" -Not Mark Twain: That's why I'm here

[@barkhascherp7746]

I don't trust nearly half of what is said in random videos on YT, so why would I trust an AI that has been trained on them? Even if it was ethical for them to do, I wouldn't want to touch their AI with a 10 foot poll considering what it was trained on.

[@a.randomjack6661]

@@barkhascherp7746 But it's touching you. Like any tool, it can be used in more than one way. It's not only videos, it's social media comments, likes/dislikes... And there's AI made algorithms, like those used in some elaborate financial tradings.

[@jd31068]

The ability to run our own server for games like Call of Duty, coop with the Juggernauts. Would be so great.

[@rothn2]

I don't believe training on any public information is immoral. It's public!

[@godminnette2]

There can be issues with game developers publishing server binaries - there are risks of security and issues of code also used in other projects being entangled. It does not necessarily cost a company nothing to do this - however, a company that plans to do this in advance can take simple steps to make it cost less when the time comes. Something all companies should have to do.

[@vikiai4241]

Yes, the summary presented above muddies things a bit, which is likely unavoidable when summarising something this complex down to a few minutes. Accursed Farms released a long-form QA recently that goes into far more detail than possible in a rapid-fire magazine format video: kzbin.info/www/bejne/qXa5c5yEarh-q6M

[@ettoreatalan8303]

The chapter about Ubuntu already starts at 10:57.

[@TheLinuxEXP]

I changed the time code, but as always, KZbin takes a while to update

[@KuruGDI]

If we complain about nvidia scraping videos from KZbin for training their AI let's not forget that there are also videos on this platform that are licenced under a free Creative Commons licence. Highly unlikely that nvidia really cares about this or provide the appropriate credit to the original author, but not impossible until we know otherwise.

[@NoCults4Moi]

FreeBSD ftw

[@Number2Vaderfan]

how is nvidia able to do is? is the youtube-scraping ai baked into their video drivers? are we viewers contributing to it simply by watching videos right now?

[@netkv]

if the eu initiative was actually implemented, what's stopping the game publishers simply saying they "lease" games instead of "sell" games, ig it'd be more honest but won't change much

[@-1_void]

Well that would make sales go down

[@m4sterred853]

Leasing requires knowledge of when the term ends. Currently, live service games shut down at arbitrary dates, often times way earlier than expected when talking about the smaller ones.

[@kokop1107]

This has already been the standard for years. You're already not buying the game but the licence to use it

[@PC4USE1]

The viewing of the videos by the AI is not illegal(unethical,yes) but the downloading of said videos is. If it were not,why would they need 'rotating IP addresses to prevent being banned"?

[@Gigusx]

Look... if I had Nvidia's resources and wanted to train my AI model(s) on something KZbin's videos would be perfect for, I'd to the exact same thing. I would also add movies, tv shows, and documentaries to the list, lol.

[@abdullahzafar4401]

Hi, Nick. system 76 currently doesn't have an amd graphics card laptop, Will there be good laptop battery optimization on all amd laptops ?

[@needsLITHIUM]

If the Atomic/Immutable Manjaro can get stable, this might make the distro useable lmao.

[@raute2687]

great video, but I have one minor criticism. BTRFS stands for b-tree filesystem, not "better filesystem".

[@SuperTort0ise]

Um it's actually butter FS

[@gbjbaanb]

So does this Manjaro still include Gnome after the controversary around the maintainer for gnome on Manjaro?! That was fun. Bizarre, i mean.

[@agr-tech]

What is the difference between ai training and search indexing? Google has been training their search engine forever but if AI crawls this data there is now a problem? I don’t get it.

[@the-answer-is-42]

Guess: Googles search engine indexing is meant to find stuff, it's not generative and thus it won't use that stuff to create new stuff. In other words, it's how the crawled data gets used that's the difference. Maybe, just guessing here. Also iirc, search engines are at least on paper not supposed to crawl through everything. Websites can say "don't do this". Now if crawlers follow that or not in practice is another thing, but it's safe to say NVIDIA didn't.

[@iferca]

When a review of the tuxedo convertible?❤

[@hugo-g5r]

no timestamps? :(

[@caspersmith7112]

i agree with the ai statement you made

[@Verssales]

Elementary, for me at least, died after one of the founders quit... The development slowed down and the features aren't so attractive anymore.

[@weab]

I'm not a big fan of Nvidia, but I'm in agreement with their reasoning on this matter. Aggregating and utilizing publicly available information isn't theft, piracy or immoral in any way. If the AI is utilized in a way that results in copyright infringement or plagiarism then those instances can be handled with existing legal doctrine on a per case basis.

[@gbjbaanb]

publically available doesn't mean tyou can still use it for your commercial practices. the football is shown on TV, but that doesn't mean I can put a big TV in the park and broadcast it to everyone as an event. I assume YT has a terms and conditions of viewing that says its for personal purposes only.

[@I-USELINUX]

Linux user's rule no. 1: Never forget your first distro!

[@Weezerfier]

Why is HDR so important for Gnormies when there are LOTS of other very important and serious things they should focus on?

[@Can_You_Hear_Me]

so immutable distros are getting mainstream...

[@realchoodle]

I feel like all of Gnomes recent updates have just been groundwork/foundation stuff and it’s getting really frustrating.

[@kodemasterx]

HDR on Gnome, click click click 😅

[@Clickworker101]

Elementary

[@akinne_]

It's so frustrating that I can't install Cosmic and try it out. I tried to install Cosmic on Arch, but I failed. Technically, it works, but for example, I don't have icons, and instead, I see black squares everywhere. I saw one post on Reddit about that problem, but no one responded in the comments.

[@stwenty5758]

I tried ArkaneOS - or at least I wanted to. I couldn't get it running on my laptop. I guess I could've tried running in a VM, but meh. And while I will probably try again at some point, I'm just not using Manjaro. I cannot - and will not - trust a distro that thought that the solution to them screwing up was for me to change my system clock. Also, DDOSing the AUR was a kinda of deal breaker.

[@fernandoperez8587]

Ai are not humans so I don't get why Nvidia tries to agree with that line of reasoning.

[@danielkemmet2594]

I don't understand why people think you need permission to access publically available information that's not even protected by a login mechanism. Image going to the library and learning engineering by the books, going to get a job, and then a competing engineering company sues your for using books to learn a skill that benefits their competitors. There's some cognitive dissonance at play here. If you don't want your videos scraped. Don't put them where you KNOW they are likely to get scraped, otherwise, if you're putting something out there publicly don't be surprised if the public uses it.

[@33KK]

You really don't see anything immoral with a company scraping millions of videos without authors consent? There is a very clear difference, video authors very much know that humans will watch their videos, they never were asked if they want AI trained on it though, most would sure decline. Educational books are quite obviously published with humans learning from them in mind, and you totally could get sued if you copied parts of the book too closely. >Don't put them where you KNOW they are likely to get scraped Yeah just don't use the biggest and pretty much only video hosting platform! Peak aibro brainrot

[@SuperTort0ise]

​@@33KKFr, but they're right that "there's some cognitive dissonance here" it's just not who they think it is. Like you don't want your bike stolen, then don't leave it where it could be stolen! Even if this is true it doesn't excuse the thieves.

[@def1nt]

Regarding Nvidia. It's not even about capability comparison. I as a human watch a lot of KZbin videos and decide to create my own commercial product. Nvidia's neural network watches a lot of KZbin videos an it IS THE commercial product.

[@safsaf99]

I'm out of the loop, can someone explain what are the benefits of Immutable distros, besides less chances of users messing up their os

[@zacharydacus6720]

The current initiative is to vague and could be applied retroactively to even older games (which it shouldn't imo) overall I agree with sentiment but this could hurt current smaller studios if done incorrectly, now if it was this law applies after a certain date (kind of like safety laws for vehicles) then id agree 100% because that would allow game makers to have it programmed into the game proactively

[@SuperTort0ise]

It's just an initiative, it's supposed to be vague, and no it won't be retroactive no law in this field works like that. Also ofc it have a date far in advance for when it'd take effect.

[@thescrewfly]

Corporations be like: humans, machines, what's the difference? Exploit the humans, exploit the machines, exploit the humans by using the machines. Nothing to see here, just move along...

[@dewdgiplay]

don't think learning ai with all sources possible is bad

[@Software-sb1gx]

I thought the entire point of having a community driven project was to build only things that the community wants. the companies I know do what's best for the company not for the customer. I mean it's literally the flaw of our entire societal and economic system and why products and services can be so shit

[@ehabsamir6195]

🔥👏

[@dritzzka6177]

Why isn't Google suing Nvidia then?

[@TheLinuxEXP]

Likely because they did the exact same thing for their own AI 😂

[@dritzzka6177]

@@TheLinuxEXP thanks for the info! Didn't think of it that way. thought for a second if Google did bite Nvidia they wouldn't be getting GPUs.

[@savagepro9060]

Question, can you use Elementary OS in High School?

[@pocketglove]

wdym?

[@savagepro9060]

@@pocketglove it's a joke, 'elementary' meaning primary school, as apposed to high school

[@pocketglove]

@@savagepro9060 ahhh, mb mb, I thought you were asking if it's a good choice for someone in high school

[@_a_x_s_]

Wait Ubuntu 32.04? Is it Ubuntu 26.04? And I think the release is too fast for LTS? Let me check what’s going on.

[@brandonchappell1535]

The numbered versions relate to the year and month of release so 32.04 is april 2032, he was just using it as example but i too thought it was odd to go so far ahead!

[@temari2860]

There is a funny situation in the Linux desktop space, in theory it's a good idea to get away from random technical terminology like "wayland" and "x11" in user facing interfaces, but in reality for 99% of current users random terms like "secure session" and "compatibility session" can be more confusing

[@JamesRogersProgrammer]

Does KZbin own even a fraction of the content on its site?

[@justahumanwithamask4089]

Eventually we'll get an immutable distro that's basically android where you can't do anything

[@JacksonNick-j6i]

Comparing Linux distros to Android is like comparing Android to iOS.

[@tablettablete186]

​@@JacksonNick-j6iYeah, the name immutable is so misleading

[@tridens6708]

Ai shit being forced on people who don't want it 😢 Sounds like Ciberdine Systems

[@Little-bird-told-me]

i am all linux MINT

[@Beryesa.]

Wait, elementaryOS is still keeping up? owo

[@ElisLawel]

pretty much all ai models are stealing. chatgpt and deviantart and all

[@savagepro9060]

Envy-dia pilfers KZbin for AI

[@A_Retired_MSgt]

All AI steals, that's how it works

[@Kaptime]

Nvidia owes KZbin billions then and owes the State billions too for breaking the law.

[@Legion-495]

Welcome back indeed.