Think about compositions, samples, performance rights, and so on. There is a lot more at stake.
Like software patents - when you’re not a normie.
We'll only get the art that artists really wanted to make? Great!
What's the benefit of protecting monetary IP rights to software?
What's the benefit of consolidating all meaningful access to computing services to a few trillion-dollar gate-keeping corpos?
I’m guessing you’d want to restrict lyrics to encourage more plays of the song by people who are motivated to understand them. Along with the artist’s appreciation of that experience of extracting what you’re fascinated by. Burdensome processes generate love and connection to things.
Not everything is a functional commodity to be used and discarded at whim.
The only thing funnier than that is when they do have the lyrics to a song that probably doesn't need them, like Hocus Pocus by Focus: https://open.spotify.com/track/2uzyiRdvfNI5WxUiItv1y9?si=7a7...
The fact that they don't do anything with that information is unrelated.
Why would people invest in destroying what they love?
While exclusively-controlled LLMs would be mildly useful to them, the technology existing is dangerous to them, and they already have a surplus supply of content at low cost that they monetize by controlling discovery, gatekeeping, and promotion, so I don't think it makes sense for them to put energy into LLMs even if they had the technical acumen to recognize the possibilities (much the same way that Google, despite leading in developing the underlying technology, had vvery little incentive to productize it since it was disruptive to their established business, until someone else already did and the choice was to compete on that or lose entirely.)
1) RIAA is evil for enforcing copyrights on lyrics?
2) OpenAI is evil for training on lyrics?
RAII is a different kind of (necessary) evil
*There are some funny "play from sheet music without knowing the song" type videos out there, with funny results. YouTube/google search is no longe usable, so I can't find any.
Considering this is hacker news, it seems to be such an odd dichotomy. Sometimes it feels like anti-hacker news. The halcyon days of 2010 after long gone. Now we need to apparently be angry at all tech.
LLMs are amazing and I wish they could train on anything and everything. LLMs are the smartphone to the fax machines of Google search.
Sorry this such a (purposefully?) naive take. In reality the thoughts are much more nuanced. For one open source/free software doesn't exist without copyright. Then there is the whole issue that these companies use vast amount of copyrighted material to train their models, arguing that all this is fair use. But on the other hand they lock their models behind walls, disallow training on them, keep the training methods and data selection secret...
This tends to be what people disagree with. It feels very much different rules for thee and me. Just imagine how outraged Sam Altman would act if someone leaked the code for Gpt5 and all the training scripts.
If we agree that copyright does not apply to llms, then it should also not apply to llms and they should be required to release all their models and the way of training them.
As a side note i am definitely not strong into IP rights, but I can see the benefits of copyright much more clearly than patents.
That aside, I think there's a lot more complexity than you're presenting. The issue is who gets to benefit from what work.
As hackers, we build cool things. And our ability to build cool things comes in large part from standing on the shoulders of giants. Free and open sharing of ideas is a powerful force for human progress.
But people also have to eat. Which means even as hackers focused on building cool things, we need to get paid. We need to capture for ourselves some of the economic value of what we produce. There's nothing wrong with wanting to get paid for what you create.
Right now, there is a great deal of hacker output the economic value of which is being captured almost exclusively by LLM vendors. And sure, the LLM is more amazing than whatever code or post or book or lyric it was trained on. And sure, the LLM value comes from the sum of the parts of its source material instead of the value of any individual source. But fundamentally the LLM couldn't exist without the source material, and yet the LLM vendor is the one who gets to eat.
The balance between free and open exchange of ideas and paying value creators a portion of the value they create is not an easy question, and it's not anti-hacker to raise it. There are places where patents and other forms of exclusive rights seem to be criminally mismanaged, stifling progress. But there's also "some random person in Nebraska" who has produced billions of dollars in value and will never see a penny of it. Choosing progress alone as the goal will systematically deprive and ultimately drive away the very people whose contributions are enabling the progress. (And of course choosing "fair" repayment alone as the goal will shut down progress and allow less "fair" players to take over... that's why this isn't easy.)
Is it simply "we like college kids" and "we hate OpenAI"? that dictates this?
I'm ready, hit me with the nuance.
... Alright?
A corporation who pirates, is more likely to pirate en masse everything that they can get their hands on, in an ongoing manner, and throw everything they can at contesting their right to do so in court.
Training on Wikipedia? Cool! Training on pirated copies of books? Not cool! Training on lyrics? IMO that's on the "cool" side of the line, because the "product" is not the words, it's the composition and mastered song.
Plus, from TFA,
"GEMA hoped discussions could now take place with OpenAI on how copyright holders can be remunerated."
Getting something back is better than nothing
I don't respect this opinion but it is unfortunately infesting tech spaces right now.
While I do not agree with your take, FWIW I found your comment substantive and constructive.
You seem to be making two points that are both controversial:
The first is that generative AI makes the availability of lyrics more problematic, given new kinds of reuse and transformation it enables. The second is that AI companies owe something (legally or morally) to lyric rights holders, and that it is better to have some mechanism for compensation, even if the details are not ideal.
I personally do not believe that AI training is meaningfully different from traditional data analysis, which has long been accepted and rarely problematized.
While I understand that reproducing original lyrics raises copyright issues, this should only be a concern in terms of reproduction, not analysis. Example: Even if you do no data analysis at all and your random character generator publishes the lyrics of a famous Beatles song (or other forbidden numbers) by sheer coincidence, it would still be a copyright issue.
I also do not believe in selective compensation schemes driven by legal events. If a legitimate mechanism for rights holders cannot be constructed in general, it is poor policy craftsmanship to privilege the music industry specifically.
Doing so relieves the pressure to find a universal solution once powerful stakeholders are satisfied. While this might be seen as setting a useful precedent by small-scale creators, I doubt it will help them.
But anyway it seems I misinterpreted the issue and record companies have always been against reproduction of lyrics whether an AI or human is doing it
Record companies have in fact, for decades, been going after sites for showing lyrics. If you play guitar, for example, it's almost impossible to find chords/tabs that include the lyrics because sites get shut down for doing that.
> I think having the lyrics reproducible in text form isn't the problem. Many sites have been doing that for decades and as far as I know record companies haven't gone after them.
Reproducing lyrics in text form is, in fact, a problem, independent of AI. The music industry has historically been aggressively litigious in going after websites which post unlicensed song lyrics[0]. There are many arcane and bizarre copyright rules around lyrics. e.g. If you've ever watched a TV show with subtitles where there's a musical number but none of the lyrics are subtitled, you might think it was just laziness, but it's more likely the subtitlers didn't have permission to translate&subtitle the lyrics. And many songs on Spotify which you'd assume would have lyrics available, just don't, because they don't have the rights to publish them.
[0] https://www.billboard.com/music/music-news/nmpa-targets-unli...