Hacker News Reader

Comment by xlayn | Hacker News Reader

Training on the book being fair use for something whose whole deal is to take those ideas and exercise them is the steal of our time...

So you come with this amazing idea regarding anything... how to build a chip, how to write software... and you write a book on how to...

Then a model eat it... spit it for free (they could get an advantage from being able to exercise that logic) and it's fair use...

Then... why torrent it, can't they sent someone to buy the book and scan it or buy it in pdf?

View on HN

Torrenting just scales better, I think it's simple as that.

Torrenting doesn’t make anyone money. So no way to pay off senators and judges.

Neural network is a method of compression. Lossy compression. What the AI companies want to push constantly is misconception:

  This is not algorithm

And

  They think

They are juggling between a personification and de-personification of their models according to what they need at the moment. So when you're hacking using fable, fable says you hacked.

When anthropic wants to gain money they say "our model hacked and we do not know why (damn it's so good)"

When using publications, "they have been used for learning" (not algorithmic compression)

Etc

I don't want to have to pay licensing costs for books I read since I "trained" on it.

I'm deeply concerned my children will have to pay rent for the rest of their lives for the schooling they got.

I imagine it's the same reason a lot of people pirate content instead of buying it legally. It's cheaper and faster.

I've pirated several books, movies, games, and music albums because there was literally no way to give someone my money for them. I can't imagine trying to acquire things legitimately, at scale.

I wonder why would anyone write a book if an LLM company can just buy 1 copy and reproduce it forever?

This is such an amazing take... if "training" an LLM with a book is fine, if you rewrite the book with slightly different wording... are you plagiarizing it? if you are then every application of AI over trained data of that trained data is plagiarizing?

Say you have this Math book and you rewrite it from third to first person and then replace all the sine examples with cosine and modify the exercises to have different numbers... is it stealing? you can buy 1 book, "train on it" and then generate your "personal original approved copy"

Reply on news.ycombinator.com