Protect the artists and their work. Fuck AI and the tech bros.
Significant_Sale6172 on
Good.
AcanthaceaeEqual4286 on

HopefulTangerine5913 on
I fucking hate AI but I can’t help thinking he’s just panicking that it’s going to finish The Winds of Winter before him and doesn’t want to feel pressure to beat it to the punch
cactus_jilly on
Burn it down, George.
Transitsystem on
ChatGPT came up with nothing. It’s an aggregate of all the human knowledge put into it, it made nothing.
HonkyTonkPianola on

dremolus on
Good.
Now finish the goddamn books, George!
augustfolk on
Show up the AI by finishing the goddamn book, George
Sadboi395 on
Would this set the legal precedent to sue anyone who writes a fan-made story? I know he’s not doing that, but am curious if this will extend copyright protections to where people could sue for fan writing, art, etc.
CheckMateFluff on
TLDR for anyone paywalled:
A federal judge in Manhattan let authors’ copyright claims against OpenAI and Microsoft move forward, pointing to a ChatGPT response that reads a lot like a GRRM sequel outline. The court said a jury could find those outputs “substantially similar,” so the output-infringement claims stay in. The judge did not decide whether training on books is fair use yet.
So, for context, In a separate case, a California judge said training on lawfully obtained books can be fair use, but storing pirated book “corpora” is not, which is what pushed Anthropic into a roughly $1.5B settlement. So the heat right now is on outputs and any tainted data sources. Expect discovery and pressure to settle.
If you want odds? I mean, courtroom win for the authors is about 20%; a settlement, authors would call a win about 60%, but thats just a guess, obviously.
AdamOfIzalith on
They have two options:
1. Pay out and set the precedent that they are responsible for the things their AI does.
2. Fight the case and argue that their own marketing of the product is bullshit that aggregates stolen work.
targaryeh on
thank god
ANALOGPHENOMENA on
There really is a bubble.
Zanzibardragonlion on
Not the point, but also wanted to mention that ChatGPT’s sequel ideas were really fucking stupid.
15 Comments
Protect the artists and their work. Fuck AI and the tech bros.
Good.

I fucking hate AI but I can’t help thinking he’s just panicking that it’s going to finish The Winds of Winter before him and doesn’t want to feel pressure to beat it to the punch
Burn it down, George.
ChatGPT came up with nothing. It’s an aggregate of all the human knowledge put into it, it made nothing.

Good.
Now finish the goddamn books, George!
Show up the AI by finishing the goddamn book, George
Would this set the legal precedent to sue anyone who writes a fan-made story? I know he’s not doing that, but am curious if this will extend copyright protections to where people could sue for fan writing, art, etc.
TLDR for anyone paywalled:
A federal judge in Manhattan let authors’ copyright claims against OpenAI and Microsoft move forward, pointing to a ChatGPT response that reads a lot like a GRRM sequel outline. The court said a jury could find those outputs “substantially similar,” so the output-infringement claims stay in. The judge did not decide whether training on books is fair use yet.
So, for context, In a separate case, a California judge said training on lawfully obtained books can be fair use, but storing pirated book “corpora” is not, which is what pushed Anthropic into a roughly $1.5B settlement. So the heat right now is on outputs and any tainted data sources. Expect discovery and pressure to settle.
If you want odds? I mean, courtroom win for the authors is about 20%; a settlement, authors would call a win about 60%, but thats just a guess, obviously.
They have two options:
1. Pay out and set the precedent that they are responsible for the things their AI does.
2. Fight the case and argue that their own marketing of the product is bullshit that aggregates stolen work.
thank god
There really is a bubble.
Not the point, but also wanted to mention that ChatGPT’s sequel ideas were really fucking stupid.