Computers mostly don’t make mistake, software makes mistakes.

edit: Added mostly because I do suppose there are occasions where hardware level mistakes can happen…

It isn’t fair use, See most of faq @ fairuse faq.

“Fair Use” is often the subject of discussion when talking about online copyright with regards to online video content or music sampling, but it’s notably a flawed defense as it generally has no legal definition for how much of certain content can be used or referenced. The very first line of that faq has the following note:

How do I get permission to use somebody else’s work?
You can ask for it. If you know who the copyright owner is, you may contact the owner directly. If you are not certain about the ownership or have other related questions, you may wish to request that the Copyright Office conduct a search of its records or you may search yourself. See the next question for more details.

All artists / writers and others are asking LLM model producers to do is a) Ask for permission or B) Attribute the artists work in some kind of ledger, respecting the copyright of their work. Every work you make (write/play/draw/whatever) has a copyright that should be respected by companies and are not waived by EULA or TOS (ever) and must be respected in order for author attribution as a concept to work at all. There is plenty of free, permissive copyrighted content on the internet that can be used instead to train an LLM, but simply asking for permission or giving attribution would at least be a step in the right direction for these companies and for the industry as a whole.

Defenders of AI will note that the “use” of art in LLM is limited and thus protected by fair use, but that is debatable based on the content of the above listed FAQ.

How much of someone else’s work can I use without getting permission?
Under the fair use doctrine of the U.S. copyright statute, it is permissible to use limited portions of a work including quotes, for purposes such as commentary, criticism, news reporting, and scholarly reports. There are no legal rules permitting the use of a specific number of words, a certain number of musical notes, or percentage of a work. Whether a particular use qualifies as fair use depends on all the circumstances. See, Fair Use Index, and Circular 21, Reproductions of Copyrighted Works by Educators and Librarians.

You can see that the use cases above (commentary, criticism, news reporting and scholarly reports) does not qualify LLM companies to use or train their models with copyrighted data for privatized industry. Additionally, you’ll note that “market disruptive” uses cannot be protected by fair use in it’s definition, meaning that displacing artists with AI automatically makes LLM use of copyrighted material an infraction of copyright that is not protected by the fair use clause.

Regardless, this will need to be proved in court and even if it passes certain criteria, it will not apply to all infractions. Fair use is a defense, not a protection, and thus LLM producers will have to spend time in court in order to defend individual infractions. There’s no way for them to catch all copyright infringement with one ruling, it needs to be proved on a case-by-case basis.

IANAL but this is my 2 cents on the matter.

Ah fair enough then. :) It is a pretty funny thing to watch 13 times in a row.

Consider using a USB3 SSD as your boot drive if you want long term usage from your pi. The SD card is prone to failure relatively quickly on Raspian and is even worse on OSes that aren’t optimized for the PI directly.

Consider the fact they may be using a 3rd party application, like ErsatzTV or similar, to create a “virtual” tv station that just plays videos in sequence after being added to a channel.

You can always reach out to the user in question if you’re concerned though. I generally wouldn’t open my plex up to users unless I had regular communication with them anyway, but that’s a personal thing I think…

Edit: Another thing to note is that, depending on the user’s tech savvyness, they might not be aware that you can see what content they’re watching. If you aren’t close friends, it might turn them off the service to have an admin approach them about the content that you yourself are hosting. Consider this and really weigh whether it’s your own business or not. And obviously, to be clear, you definitely shouldn’t be charging for this service so make sure that you’re not putting yourself in potential legal problems.

I mean, I sort of get why the developers say it’s Discord’s policy even if it’s a bit misleading.

Game developers don’t really want to moderate their own discord server and simply want to use the strictest automated filtering system available and this just happens to include phone number linking. The operators of the servers themselves do not have access to these phone numbers and they are only stored by discord directly to prevent spam.

I would personally prefer games to not have their communities tied to discord, akin to how forums were big deal for games back in the day, but even then they do need some kind of automated way to filter out all the crap. This is a problem with moderating any community, including a lemmy/kbin/mastodon, and I don’t blame them for simply picking the strictest option to ease the burden on the 1 or 2 people who are charged with managing these servers (especially if they are unpaid or volunteers, which is a whole other can of worms that shouldn’t happen…)

Until there’s evidence there’s basically no reason to treat this as news.

The ai is trained on recordings of his voice which they have not secured the rights to though. You can’t simply use any data you find on the street and use it professionally in any field.

An impression is a very different context. You’re vastly overestimating the independence of an AI model to equate it to human performance or impersonators.

Insanely deep rpgs are a bit of an issue for me as well. And I generally do love rpg games, but I feel like the good ones should ease you into decisions a bit better than dropping you into a character creator.

So it’s a mixed bag for me.

Does anyone know if this affects mods like CS?

Edit: Idk if it’s just me on linux, but it seems like CS1.6 no longer launches on the new update. A bummer, but I had a lot of fun play HLDM just now so I think they’ll fix it eventually.

signing into cloud services and downloading apps is just so much easier to do!

This is actually true, but it doesn’t speak to why self hosting is “impossible” and more to how the lack of education around computers have reached an inflection point.

There’s no reason why self hosting should be some bizarre concept; In another reality, we would all have local servers and firewalls that then push our content into the wider internet and perhaps even intranet based notes. Society as a whole would be better if we chose to structure the internet that way instead of handing the keys to the biggest companies on the stock market.

I’ll give this podcast a listen to though, as it might be interesting. I think the reality is that some more docker frontends might help casual users jump into the realm of self hosting – especially be setting up proxy managers and homepage sites (like homarr) that work intuitively that never requires you to enter ports and IPs (though fearing that is also an education problem, not a problem with the concept itself.)

I’ll probably stick with kbin and hope you guys join a threadiverse standard again at some point.

Is there any reason why you can’t start by improving KBin for moderation needs? It seems like they could use the extra help.

Some people on the fediverse do not want their views challenged and react with hostility. It’s not like that all over, but it can certainly be frustrating if some users treat this as a walled garden of thought and lash out at others for challenging their viewpoints.

Crazy to me that they’d shut down instead of going open source and integrating with the fediverse. Doesn’t even seem like a good business move as offering hosting for other companies and professional groups seems like a good market opportunity in a world where businesses even dislike Twitter.

Edit: for example, offer gitlab like service but for social media.

If an intern gives you some song lyrics on demand, do they sue the parents?

Uh— what? That analogy makes no sense. AI is trained off actual lyrics, which is why companies who create these models are at risk (they don’t own the data they’re feeding into the model.)

Also your comment is completely mixing Trademark and Copyright examples. It has nothing to do with brand names and everything to do with intellectual property.

deleted by creator

I have a kind of complicated system for organizing my music files – some of which is admittedly way too much maintenance but it might be of interest to some.

For my general “commercial” music collection, the folder structure is roughly
Music/%Release Artist | Band%/%Album%[%Year%]/%Track No.% - %Title%.%Format%

This is simple to maintain. I basically just use MusicBrainz Picard and set up appropriate paths.

For my soundtrack collection, it gets a bit more complicated. For Anime/Film/Whatever, I have it sorted basically the same way but in a different root folder. So something like:
Music/Anime/%Release Artist | Band%/%Album%[%Year%]/%Track No.% - %Title%.%Format%

Which is also easy to maintain since most of these also have commercial releases.

But games are sorted more strangely. To put it simply, I have a folder structure that puts the console or platform first, followed by the game name and then the loose files. Since some of these files are emulated formats (.vgm, .nsf, .spc), I generally don’t bother renaming them and keep them as is and trust that the music program in question has tagging support. It also means that having them sorted by console is mostly beneficial to quickly find emulated file formats, but YMMV and I have regretted the choice on occasion.

Obviously game soundtracks are spotty when it comes to releases. Some companies have reliable metadata you can get from MusicBrainz Picard, like SquareEnix, but others have no tagging at all or very incorrect tag values. Because of this, I generally use something like VGMDB, which is usually higher quality but not always. I do have to resort to manually correcting files on occasion.

If anyone has a nice automated way to sort this stuff out, it would be a real benefit to me as well.

Spotify serves mp3s because it uses less bandwidth and most people can’t tell the difference on their 30€ Bluetooth headset.

I think this highlights a bigger issue when it comes to this discussion.

The issue isn’t the mp3 format – for the most part, the format of any lossy encoder can sound good with the right settings. The problem is that, unlike flac, all encoded lossy files are essentially untrustworthy audio formats. So when people say mp3 sounds bad, it’s only a half truth in the same way that it’s a half truth to say that people cannot tell a difference. You are putting trust in the person who encoded the audio to make the right choice and the encoder is putting trust in the idea that the person consuming the media can’t tell the difference.

When it comes to being cheap on bandwidth since most users can’t hear it, that’s a huge cop-out being made for a company that can do better. While Apple is pretty notorious for making terrible decisions for arbitrary reasons, even they respect the user enough to allow you to opt into higher audio format quality. It’s decisions like these that cement Apple as the kings of the creative computer user.

You might have to do the pen and paper method.

I agree. It’s a mistake for Valve to reduce platform support as it becomes a justification for dropping other platforms for other developers as well.

I know that MacOS is a bit of a pain to support right now with how steam and third party applications are treated, but it would be a bigger issue for Apple to drop support if Valve maintained a strong presence on the platform via steam. With the way things are now, Apple might rip the bandaid off and just remove the ability to have third party stores on their computers. This is already what they do with Iphones/Ipads