Less Evil Hank Scorpio said:

---

Logos Stick said:

---

Less Evil Hank Scorpio said:

---

Logos Stick said:

---

Less Evil Hank Scorpio said:

---

All he said was "stole her work". Doesn't matter what they did with it, piracy is a form of theft. So I think your semantic at best argument is moot.

---

All the words matter my friend.

He said this: "LLMs steal work from others.", followed by... "stole her work"


Do I need to explain it to you?

---

Actually, yes please do.

Anthropic bought and scanned books and used them for training. This is fair use.
Anthropic also pirated books and used them for training. The piracy and storage of these works is not fair use. The training part is moot.

LLMs steal work from others - were there works stolen used to build LLMs? Yes.

"stole her work" - here you left out some important words. He said "Anthropic is about to cut my wife a check because they stole her work" - here it is much more clear that he's referring to the actual company and not the LLM itself. That is what the ruling is about, did the company steal/pirate works? Yes.

I'm failing to see how your distinction disproves anything he said.

---

Wow!!!

He said that LLMs steal work from others! That is completely baseless! That is not what the case was about and not what was ruled on!

Let me see if I can give you something to help you comprehend (doubtful)...

Had Anthropic not fed the books to the LLM, did they still commit a civil offense by storing copies of the pirated books?

Lets start simple here.

---

So you're just ignoring the additional context provided by his last sentence?

Like I said, the training is not the issue, it was the piracy. I have never claimed that training a model based on legally acquired books is stealing. I understand that it is fair use.

Perhaps I am being too generous in my interpretation that when he says "LLMs steal works" he is referring to the companies that develop the models rather than the actual Large Language Model. I understand that an LLM isn't truly regurgitating words stolen from books and reddit.

The model isn't stealing. The company that built the model stole to be able to acquire more material to train their clanker. I still don't think he said anything wrong, but I guess it would have been more correct for him to say "LLM companies steal work from others" rather than "LLMs steal work from others". But again, his reference to Anthropic later in the same post makes it clear to me that he is commingling the terms.

---

LLMs have ZERO to do with anything in this case!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

LLMS DO NOT STEAL WORK FROM OTHERS. PERIOD! I can take every book in the country that I buy and train Claude with it! Legal! I can buy, download and scan them into a digital library! Legal!

Anyone who keeps pirated copies of books is GUILTY!!! Anyone!

Good grief!

I now have stars. IGNORE this user.

Less Evil Hank Scorpio said:

---

Logos Stick said:

---

Less Evil Hank Scorpio said:

---

Logos Stick said:

---

Less Evil Hank Scorpio said:

---

All he said was "stole her work". Doesn't matter what they did with it, piracy is a form of theft. So I think your semantic at best argument is moot.

---

All the words matter my friend.

He said this: "LLMs steal work from others.", followed by... "stole her work"


Do I need to explain it to you?

---

Actually, yes please do.

Anthropic bought and scanned books and used them for training. This is fair use.
Anthropic also pirated books and used them for training. The piracy and storage of these works is not fair use. The training part is moot.

LLMs steal work from others - were there works stolen used to build LLMs? Yes.

"stole her work" - here you left out some important words. He said "Anthropic is about to cut my wife a check because they stole her work" - here it is much more clear that he's referring to the actual company and not the LLM itself. That is what the ruling is about, did the company steal/pirate works? Yes.

I'm failing to see how your distinction disproves anything he said.

---

Wow!!!

He said that LLMs steal work from others! That is completely baseless! That is not what the case was about and not what was ruled on!

Let me see if I can give you something to help you comprehend (doubtful)...

Had Anthropic not fed the books to the LLM, did they still commit a civil offense by storing copies of the pirated books?

Lets start simple here.

---

So you're just ignoring the additional context provided by his last sentence?

Like I said, the training is not the issue, it was the piracy. I have never claimed that training a model based on legally acquired books is stealing. I understand that it is fair use.

Perhaps I am being too generous in my interpretation that when he says "LLMs steal works" he is referring to the companies that develop the models rather than the actual Large Language Model. I understand that an LLM isn't truly regurgitating words stolen from books and reddit.

The model isn't stealing. The company that built the model stole to be able to acquire more material to train their clanker. I still don't think he said anything wrong, but I guess it would have been more correct for him to say "LLM companies steal work from others" rather than "LLMs steal work from others". But again, his reference to Anthropic later in the same post makes it clear to me that he is commingling the terms.

None of this changes my opinion that your correction of him is far more semantic than substantive.

---

Except that is definitely happening. LLMs can and do regurgitate entire sections of copyrighted work as part of the new work they produce. Any regular person doing that and trying to claim it as their own work product would be guilty of plagiarism.

That isn't why your wife is getting paid though. As far as I know there hasn't been a case about an LLM using pieces of copied text in a response.

Logos Stick said:

---

Less Evil Hank Scorpio said:

---

Logos Stick said:

---

Less Evil Hank Scorpio said:

---

Logos Stick said:

---

Less Evil Hank Scorpio said:

---

All he said was "stole her work". Doesn't matter what they did with it, piracy is a form of theft. So I think your semantic at best argument is moot.

---

All the words matter my friend.

He said this: "LLMs steal work from others.", followed by... "stole her work"


Do I need to explain it to you?

---

Actually, yes please do.

Anthropic bought and scanned books and used them for training. This is fair use.
Anthropic also pirated books and used them for training. The piracy and storage of these works is not fair use. The training part is moot.

LLMs steal work from others - were there works stolen used to build LLMs? Yes.

"stole her work" - here you left out some important words. He said "Anthropic is about to cut my wife a check because they stole her work" - here it is much more clear that he's referring to the actual company and not the LLM itself. That is what the ruling is about, did the company steal/pirate works? Yes.

I'm failing to see how your distinction disproves anything he said.

---

Wow!!!

He said that LLMs steal work from others! That is completely baseless! That is not what the case was about and not what was ruled on!

Let me see if I can give you something to help you comprehend (doubtful)...

Had Anthropic not fed the books to the LLM, did they still commit a civil offense by storing copies of the pirated books?

Lets start simple here.

---

So you're just ignoring the additional context provided by his last sentence?

Like I said, the training is not the issue, it was the piracy. I have never claimed that training a model based on legally acquired books is stealing. I understand that it is fair use.

Perhaps I am being too generous in my interpretation that when he says "LLMs steal works" he is referring to the companies that develop the models rather than the actual Large Language Model. I understand that an LLM isn't truly regurgitating words stolen from books and reddit.

The model isn't stealing. The company that built the model stole to be able to acquire more material to train their clanker. I still don't think he said anything wrong, but I guess it would have been more correct for him to say "LLM companies steal work from others" rather than "LLMs steal work from others". But again, his reference to Anthropic later in the same post makes it clear to me that he is commingling the terms.

---

LLMs have ZERO to do with anything in this case!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

LLMS DO NOT STEAL WORK FROM OTHERS. PERIOD! I can take every book in the country that I buy and train Claude with it! Legal! I can buy, download and scan them into a digital library! Legal!

Anyone who keeps pirated copies of books is GUILTY!!! Anyone!

Good grief!

I now have stars. IGNORE this user.

---

Bro, chill out. We aren't even disagreeing, you just seem to take major issue with someone criticizing "LLMs" instead of "Anthropic"/"LLM Companies"

nai06 said:

---

infinity ag said:

---

Logos Stick said:

---

This is a good lesson on making sure you don't allow the LLMs to use your info/data for training. In this case, the info was uploaded by someone. The blame lies there.

Quote:

---

NYU math professor Tristan Buckmaster says OpenAI's Tuesday drop of 722 AI-generated math papers wiped out entire research programs:

"This recent dump of papers on Tuesday destroyed the careers of early career mathematicians."

---

NYU math professor Tristan Buckmaster says OpenAI’s Tuesday drop of 722 AI-generated math papers wiped out entire research programs:

“This recent dump of papers on Tuesday destroyed the careers of early career mathematicians.” pic.twitter.com/0yiz7G9wwr

— Big Brain AI (@realBigBrainAI) October 9, 2026

---

One can argue that LLMs "steal" information from other sources.
In effect, these companies have created a machine that goes around and take (for free) all information that others have spent time and money to create/generate/build. In other words, it is IP theft.

Someone is going to complain about that leading to lawsuits and shutdowns.

That is coming next.

---

There is no arguing about it. LLMs steal work from others. They cannot exist without using human created work as a basis.

Hell Anthropic is about to cut my wife a check because they stole her work.

---

I was just being nice so as to not trigger the AI CEO boot-lickers here.
Yes, it is stealing. As you say, they wouldn't exist without work that others did and which they used for free.

100%. Now what? How can they quantify the value? Is your wife getting the value of what she created that Anthro is using? What if she said it is not worth $100k, it is worth $1 Million, then what? Why should AI companies decide what the value is?


Edit: Too late. I see someone triggered, right after my post.

nai06 said:

---

Less Evil Hank Scorpio said:

---

Logos Stick said:

---

Less Evil Hank Scorpio said:

---

Logos Stick said:

---

Less Evil Hank Scorpio said:

---

All he said was "stole her work". Doesn't matter what they did with it, piracy is a form of theft. So I think your semantic at best argument is moot.

---

All the words matter my friend.

He said this: "LLMs steal work from others.", followed by... "stole her work"


Do I need to explain it to you?

---

Actually, yes please do.

Anthropic bought and scanned books and used them for training. This is fair use.
Anthropic also pirated books and used them for training. The piracy and storage of these works is not fair use. The training part is moot.

LLMs steal work from others - were there works stolen used to build LLMs? Yes.

"stole her work" - here you left out some important words. He said "Anthropic is about to cut my wife a check because they stole her work" - here it is much more clear that he's referring to the actual company and not the LLM itself. That is what the ruling is about, did the company steal/pirate works? Yes.

I'm failing to see how your distinction disproves anything he said.

---

Wow!!!

He said that LLMs steal work from others! That is completely baseless! That is not what the case was about and not what was ruled on!

Let me see if I can give you something to help you comprehend (doubtful)...

Had Anthropic not fed the books to the LLM, did they still commit a civil offense by storing copies of the pirated books?

Lets start simple here.

---

So you're just ignoring the additional context provided by his last sentence?

Like I said, the training is not the issue, it was the piracy. I have never claimed that training a model based on legally acquired books is stealing. I understand that it is fair use.

Perhaps I am being too generous in my interpretation that when he says "LLMs steal works" he is referring to the companies that develop the models rather than the actual Large Language Model. I understand that an LLM isn't truly regurgitating words stolen from books and reddit.

The model isn't stealing. The company that built the model stole to be able to acquire more material to train their clanker. I still don't think he said anything wrong, but I guess it would have been more correct for him to say "LLM companies steal work from others" rather than "LLMs steal work from others". But again, his reference to Anthropic later in the same post makes it clear to me that he is commingling the terms.

None of this changes my opinion that your correction of him is far more semantic than substantive.

---

Except that is definitely happening. LLMs can and do regurgitate entire sections of copyrighted work as part of the new work they produce. Any regular person doing that and trying to claim it as their own work product would be guilty of plagiarism.

---

That has nothing to do with whatever pap your wife produced that she is getting paid for. Not a single thing.

And more importantly, the LLM is not stealing anything!

Less Evil Hank Scorpio said:

---

That isn't why your wife is getting paid though. As far as I know there hasn't been a case about an LLM using pieces of copied text in a response.

---

Two more recent ongoing cases:

The New York Times currently has a pending case against OpenAI for a variety of things including reproducing their articles verbatim. They also claim that (and appear to have proof) that OpenAI found a way to get arround NYT paywalls to scrape even more articles. There are examples of ChatGPT reproducing entire articles verbatim.

George R.R. Martin and several other authors also have a pending case against OpenAI (separate from the Bartz case)
It has three major prongs.

1. The act of training their models on copyrighted books is plagiarism/a violation of their copyright
2. Training on pirated books constitutes and infringement
3. ChatGPT is producing results that are substantially similar and don't fall under fair use.

When ruling to allow the claims to move forward, the judge found that a reader could "easily conclude that this detailed summary is substantially similar" to the book because it "conveys the overall tone and feel of the original work by parroting the plot, characters and themes of the original."

It's happening even on the small scale. Except people either don't have the money or wherewithal to defend their copyrights. This person gives and example of Google Bard copying their own work scraped from their website in it's response.
https://blog.coredump.cx/p/large-language-models-and-plagiarism



I am obviously more invested than a lot of people on this board. My wife is an author and I work in publishing. I'm also subscribed to various industry publications and get updates on AI in publishing almost daily. These things are happening even if they don't get a ton of press.


Links
https://www.hollywoodreporter.com/business/business-news/george-rr-martin-openai-court-legal-1236413461/
https://www.nytimes.com/2026/09/17/technology/microsoft-openai-publishing-industry.html
https://gizmodo.com/george-r-r-martin-can-sue-openai-over-game-of-thrones-rip-offs-2000679057
https://en.wikipedia.org/wiki/The_New_York_Times_v._Microsoft_and_OpenAI

infinity ag said:

---

nai06 said:

---

infinity ag said:

---

Logos Stick said:

---

This is a good lesson on making sure you don't allow the LLMs to use your info/data for training. In this case, the info was uploaded by someone. The blame lies there.

Quote:

---

NYU math professor Tristan Buckmaster says OpenAI's Tuesday drop of 722 AI-generated math papers wiped out entire research programs:

"This recent dump of papers on Tuesday destroyed the careers of early career mathematicians."

---

NYU math professor Tristan Buckmaster says OpenAI’s Tuesday drop of 722 AI-generated math papers wiped out entire research programs:

“This recent dump of papers on Tuesday destroyed the careers of early career mathematicians.” pic.twitter.com/0yiz7G9wwr

— Big Brain AI (@realBigBrainAI) October 9, 2026

---

One can argue that LLMs "steal" information from other sources.
In effect, these companies have created a machine that goes around and take (for free) all information that others have spent time and money to create/generate/build. In other words, it is IP theft.

Someone is going to complain about that leading to lawsuits and shutdowns.

That is coming next.

---

There is no arguing about it. LLMs steal work from others. They cannot exist without using human created work as a basis.

Hell Anthropic is about to cut my wife a check because they stole her work.

---

I was just being nice so as to not trigger the AI CEO boot-lickers here.
Yes, it is stealing. As you say, they wouldn't exist without work that others did and which they used for free.

100%. Now what? How can they quantify the value? Is your wife getting the value of what she created that Anthro is using? What if she said it is not worth $100k, it is worth $1 Million, then what? Why should AI companies decide what the value is?


Edit: Too late. I see someone triggered, right after my post.

---

As far as the current anthropic settlement goes, it's about $3K per book. If it's a traditionally published book, half will go to the publisher (assuming the rights haven't expired). The rights for my wife's very first book sold for about 10 times that amount.

But such is the nature of class action lawsuits.

nai06 said:

---

Less Evil Hank Scorpio said:

---

That isn't why your wife is getting paid though. As far as I know there hasn't been a case about an LLM using pieces of copied text in a response.

---

Two more recent ongoing cases:

The New York Times currently has a pending case against OpenAI for a variety of things including reproducing their articles verbatim. They also claim that (and appear to have proof) that OpenAI found a way to get arround NYT paywalls to scrape even more articles. There are examples of ChatGPT reproducing entire articles verbatim.

George R.R. Martin and several other authors also have a pending case against OpenAI (separate from the Bartz case)
It has three major prongs.

1. The act of training their models on copyrighted books is plagiarism/a violation of their copyright
2. Training on pirated books constitutes and infringement
3. ChatGPT is producing results that are substantially similar and don't fall under fair use.

When ruling to allow the claims to move forward, the judge found that a reader could "easily conclude that this detailed summary is substantially similar" to the book because it "conveys the overall tone and feel of the original work by parroting the plot, characters and themes of the original."

It's happening even on the small scale. Except people either don't have the money or wherewithal to defend their copyrights. This person gives and example of Google Bard copying their own work scraped from their website in it's response.
https://blog.coredump.cx/p/large-language-models-and-plagiarism



I am obviously more invested than a lot of people on this board. My wife is an author and I work in publishing. I'm also subscribed to various industry publications and get updates on AI in publishing almost daily. These things are happening even if they don't get a ton of press.


Links
https://www.hollywoodreporter.com/business/business-news/george-rr-martin-openai-court-legal-1236413461/
https://www.nytimes.com/2026/09/17/technology/microsoft-openai-publishing-industry.html
https://gizmodo.com/george-r-r-martin-can-sue-openai-over-game-of-thrones-rip-offs-2000679057
https://en.wikipedia.org/wiki/The_New_York_Times_v._Microsoft_and_OpenAI

---

Thanks for the update. Will be interesting to see where it goes.

For these scammers, it is the "cost of doing business". They will do it again and pay the fee.

Only imprisonment deters them.