cross-posted from: https://lemmy.world/post/48284303
A sum that would bankrupt these companies and set an important precedent.
Yesterday, there was news of publishers doing the same. Let’s see what other industries follow.
Thats it? When DMCA first hit,it was millions per torrented song
They should be charged like an 18 year old kid in 2004
By their logic, that’s a whole song, not “just” pieces of it.* Also I’m sure they somehow calculate how many people then listen to that and base their “math” on that. But wasn’t that logic proven to be ludicrous later on, i.e. their numbers way too high? Not all torrented music reflects profits lost 1:1.
TBH I don’t know the last bit but that’s how people like me see it.
* the headline should read “per affected song” - the original Atlantic article makes this even clearer
OP, you’re ironically posting content mill slop that’s poorly regurgitating this article from The Atlantic. It seems like a good idea to swap them out.
I don’t think so. The Atlantic article doesn’t say $150k per song.
I think this article is written by a human, not AI.
It references Engadget and Reuters, in addition to The Atlantic. Among others too.
I think this article is more informative and better than The Atlantic piece.
- The headline reads “now there’s proof”, but they leave the proof as an exercise to the reader to go read The Atlantic’s article where the proof actually is.
- The fact it’s “$150,000 per song” is functionally meaningless. Anyone can slap any amount onto a lawsuit (in this case, over $3 trillion total), and that’s two-year-old news anyway.
- So much of the Gadget Review article is tangents; for example, it randomly devotes a paragraph to a tool developed at U. Tennessee two years ago that had no relevance to this lawsuit or to The Atlantic’s investigation – just that it can prevent this kind of plagiarism. Which, “cool”, but this article is allegedly about how this already happened and “now there’s proof”. This is trying to be a secondary and a tertiary source at the same time (respectively, summarizing this specific issue and the AI industry’s theft writ large), and it fails at being either.
- Its two throwaway references to Engadget once again don’t mean the overall article isn’t just a worse, surface-level version of the one from The Atlantic; it just means there are tangents thrown in as filler.
- There’s a paragraph: ““Trained on copyrighted recordings without permission” — that’s how label plaintiffs have characterized the practice in filings, as summarized in industry commentary.”
- This citation has zero reason to exist. You do not now or ever need to cite “industry commentary” for that absolutely trivial analysis. No fucking shit that’s what they claim; that’s the whole point that we’ve been talking about. A professional writer wanting to write informatively would not write like this. If written by a human, it’s begging you to think it’s done more than five minutes of surface-level research. This alone shows an utter disregard for quality.
- It’s blatant that this site is a slop content mill. The thumbnail image here is generated, but human authors have done that too, so let’s look at their[1] series[2] of[3] “resource[4] articles[5]” that are among the sloppiest shit I’ve ever seen attributed to a human writer. This mill evidently has no hangups with publishing LLM slop and saying a human did it.
- Gadget Review is just another in the sea of no-name tech content mills that have sprung up recently; I’ve literally never heard of it.
- And the coup de grace: “Whether you’re an indie artist wondering if your EP got scraped, or someone who generated a birthday jingle on Suno last week, The Atlantic’s databases aren’t just journalism. They’re evidence.”
I think this article is more informative and better than The Atlantic piece.
I genuinely don’t know what to say except that The Atlantic has a proper investigation from Alex Reisner as part of a series he’s been doing for three years, whereas you apparently prefer obviously generated slop. I write encyclopedically as a hobby – reasonably adjacent to this style of “and this happened[1] and this happened[2]” from this article (and constantly reading news articles to write what I do) – and this is barely focused crap with little you couldn’t get by asking ChatGPT about The Atlantic’s article. Except I think it’d somehow do a better job than this LLM-generated shit.
Ty for the heads-up. I still care about quality research and reporting.
The fact that an ai assistant pop immediately takes over the article page when you land on it is so hilariously infuriating.
I’m not optimistic any of this will matter in the end, but I’m hoping I’m wrong an these AI billionaires get a massive kick in the dick.
I’m not optimistic. In these fights, the deeper pockets usually win.
The AI industry is throwing around such incredible amounts of money right now, it would be possible to solve this problem by just buying the record labels outright. The big AI companies could do the following:
-
Form a consortium, call it the “AI Industry IP License Alliance” or similar.
-
Have the leading AI companies offer massive loans to the organization, assembling a war chest of say, $50 billion.
-
The consortium then buys up rights holders one at a time. The largest record label in the US, Universal Music group, has a market cap of about $34 billion. It would cost $17 billion to buy a controlling stake of Universal.
-
After a record label, a book publisher, etc. is bought up, have the label sign extremely generous very long-term contracts with the big AI companies. Offer the same cheap access equally to all the AI companies to avoid antitrust concerns.
-
After a long-term contract is signed, slowly sell of the shares of the company and give up ownership of it. The AI companies might have to sell their shares for slightly less than they bought them for, but that would likely be an acceptable loss.
They could just churn through entire industries like this. Just buy up companies that own IP, lock the companies into 20 year contracts licensing their IP for a pittance to the AI industry. Then dump them back on the market. The same war chest could be used again and again and again. Eventually rights holders would just start offering generous terms at the mere threat of such a hostile takeover.
The real methods used would likely be more subtle than this direct buy-out approach. But in these fights the deeper pockets tend to win. Or alternately, maybe the AI companies only do this if they start getting massive judgments against them. If Universal Music gets a $50 billion judgment against OpenAI, rather than paying it, it will be cheaper for OpenAI to simply buy Universal, agree to settle the settlement for a pittance, dump it back on the market, and take whatever loss they have to.
-
Looking forward to open source projects suing copilot for the same.
Finally Microsoft will really help Foss.
I hope they end up paying up, even if it’s just cost of business being multi billion companies and all, and then everyone starts listening to other kinds of music, because AI destroyed the point in listening to slop.
The pattern is record companies are only upset because they want to AIfy their own catalogs to profit off making music worse. Ain’t no friends there.
Fuck yeah
I have a friend who is a corporate lawyer in Manhattan. SDNY has so many lawsuits for stolen IP by AI companies they just call it too many to count, but certainly in the thousands.
Oh yes, let them both eat each other alive.
deleted by creator
