We should get rid of ramps and handicapped parking at grocery stores too. If it’s such a large market someone can start their own accessible grocery store…
A flagship LLM today is trained on tens of trillions of tokens, the equivalent of hundreds of millions of 100k word books. No human has that kind of appetite.
As an aside, the entire Google Books corpus is generally estimated at tens of millions of books.
Exactly. It’s the funniest thing. These books are only being bought by AI companies. It seems pretty clear they only have value to AI companies. Otherwise all these people bemoaning the loss of rare books would be… buying them.
“How dare these companies buy rare and valuable books that nobody else values enough to buy” is a self-canceling argument.
It’s been established in the DC circuit, where basically any challenges to copyright office decisions would need to be brought. For all intents and purposes, until the Supreme Court or Congress decide to weigh in, it’s pretty well established that AI generated content cannot receive copyright. There will probably be some decisions around the edges regarding when human editorial input might result in a copyright being awarded to the editor, but it would be very surprising to see the precedent overturned completely, particularly given its grounding in previous caselaw.
Authorship is also required in the Constitution's section on copyright, and is likely entangled in international treaties. Any change that foundational would probably require a constitutional amendment.
Did anyone expect it to go the other way? Europe is more hardcore about copyright than the US, and that's saying something. Germany is practically...well I was going to say they're nazis about it but perhaps that's in poor taste.
That’s not how that would work at all. The copyright of a derivative work has no effect on the existing copyright of an underlying work.
If you record yourself reading a book, you own the audio recording copyright but it would be a copyright violation to reproduce that copy without a license for the underlying rights.
In this situation:
If the AI generates the audio recording of a book, no one owns the copyright of the audio recording but it would still be a copyright violation to reproduce that copy without the underlying rights.
I probably wouldn't try to adjust for both of those at once. You could ask whether inflation-adjusted revenue is up, you could ask whether per capita ticket sales are up, but I don't think the combined concept, whatever it might be, is very meaningful.
My first thought reading this was I wonder if Sony is planning to wind down or spin off Sony DADC, their commercial optical disc manufacturing subsidiary they’ve run since the dawn of the CD.
Sony DADC presses every PlayStation games but also has spent decades manufacturing CDs, DVDs, Blu-rays, UHD Blu-rays, software, and other optical media for 3rd party customers across the entertainment and computer industry.
If PlayStation discs disappear in 2028, that removes one of the last massive, guaranteed sources of demand for optical disc production inside Sony. Even if the plants continue serving movies, music, and third-party clients for a while, it’s hard not to wonder if this announcement is also the beginning of the end for one of the last major end-to-end optical media manufacturers in the supply chain.
The certificate was issued before the Nightmare Eclipse zero day thing started but I suppose it’s possible there are other certificates expiring around the same time that could be connected to the Nightmare deadline. Probably a coincidence though
Yeah if you have an actual really good idea for a certain kind of under analyzed data that could give you an edge but it would take too much effort or time to compile or analyze the data, an LLM has potential to make a viable strategy out of an otherwise correct but unuseful insight
