Hacker News Reader

Comment by Trusteando | Hacker News Reader

Reading isn't the right comparison. Human memory is lossy and fades while LLM encoding is durable with no degradation. The valuable content that the author provides: the content, style, selection of topics, and more is encoded, written into the LLM weights, and they obtain profit from them (now directly, via ads). No one can compete with that kind of copying and pasting from copyright-protected material. And the scale is what hurts authors most: flooding the market with millions of copies on demand, without paying for it.

View on HN

Human memory decays; similarly, LLMs do not have perfect recall. But even if I reread a book every month for 60 years and use the knowledge in there to build a 25 billion dollar empire, the only thing the author is ever going to get from me is the $10.25 the book cost me. And no court in the western world would say he's due more.

> No only reading, because the content, style, selection of topics, and more is encoded, written, in the LLM weights

Exactly analogous to a human reading the material.

Most people can't recite a book they've read verbatim. However, if you ask an LLM to continue a random sentence from a semi-popular book, it can sometimes provide the exact text, unless the system flags the response.

> Most people can't recite a book

Yes - but some people can.

Are they criminals for reading books?

Those some people can't recite a book to any given person in the world at any given time. There's a difference.

So if we invent a way for those people to talk to everyone in the world - then they would be a criminal for reading books whether they recited or not?

You're not making sense.

The problem is the reciting. Not the reading. And hence, not the training.

If you would go on a tv and read out loud a book you have not purchased rights to present, according to the most copyright laws in the world, yes you would and you would get a fine. Similarly if you broadcast a tv-show you have not purchased rights would.

Whether this is right or not is a seperate question.

> If you would go on a tv and read out loud a book

That is not the situation we are discussing. No one is arguing that what you describe is infringement.

What we are discussing is the training - which happens BEFORE the broadcast. It is analogous to reading. Is simply READING the material an infringement.

Using analogies like "reading" to describe AI training is quite misleading imho. Training effectively embeds the book's contents into the model's weights. Changing the format doesn't change the content; a better analogy is distributing a book's text within software.

Current copyright laws are simply not prepared for this unprecedented use.

> Training effectively embeds the book's contents into the model's weights. Changing the format doesn't change the content

Reading a book embeds its contents into your brain. And yet, that is considered fair use.

I agree, the analogies are meaningless in a legal context. In the legal context, the courts disagree with you.

I'm not disagreeing with you, but the comparison is breaking down here. Reading a book is the specific intended purpose. It's why the book exists, not just fair use. If you couldn't remember what was going on in the book as you read it, it would be worthless.

Please give me the location of any copyrighted work in the weights of any open source model or a prompt that will retrieve it.

Go to Deepseek and ask it “Give me the lyrics for Männer by Grönemeyer”.

It will give you the complete song lyrics which are under copyright.

Those lyrics aren't in the model. They're either in the database DeepSeek makes tool calls into (unlikely), or they're out on the Internet and DeepSeek simply retrieved them on your behalf.

LLMs are far too lossy to be able to store such lyrics in their entirety. In fact, they're not even "lossy" since they're not even trying to record such information. They're just weights for how likely it is that any given word will come after another.

Is this a direct inference call to the model or does it use a web search tool to pull the answer? In the former case I guess that could be infringement. In the latter case it is no more than a tool that facilitates it and would be not be infringement any more than Xerox is infringing if you use their copier to copy the front page of the NYT.

Yes, and that is obviously copyright infringement.

The infringement occurs at the time of copy, not at the time of training.

That begs the question though, should it be? If you’re looking up the lyrics to a song you’ve bought, does it really matter, at least philosophically, how one retrieves them?

The law here is no: If you own a copy of the song, you can reproduce the lyrics in whatever way you want.

What you can't do is distribute the lyrics without the author's permission.

When you ask DeepSeek to retrieve the lyrics (and it does so), the real question is this: Is DeepSeek merely acting as an intermediary/ISP according to the DMCA (in which case they'd be protected under the safe harbor clauses) or are they illegally redistributing the lyrics without the author's permission?

Whether or not you own a copy of the lyrics is irrelevant from a legal perspective in this scenario.

My guess: If they just retrieved the lyrics from some website and delivered them to you (because you asked), they're an ISP. However, if they pulled them out of their own database, they're violating copyright.

It shouldn't be. The idea that writing down the lyrics to a song and putting them on the internet for no financial gain at all should be punishable by some big record label is the kind of over zealous IP protection that 99% of HN users would have been against until something changed a few years ago and this place became a haven for copyright extremists.

Reply on news.ycombinator.com