Hacker News Reader

Comment by runarberg | Hacker News Reader

I am thinking about a long essay Icelandic author Þórbergur Þórðarson wrote to his pals abroad, and were left abroad. I am thinking about a photo book by an Indonesian naturalist who is famous on Bali, and took amazing photos of wildlife on Sulawesi in 1926, and colored in, and somehow ended up in New Jersey in the 1980s. I am thinking about a collection of essays written by a teenage J.D. Salinger who he left unsigned at a café thinking nobody would want to read them but just leaving it up to chance. Or maybe a Jackson Pollock sketchbook he lost at a party which ended in the host’s bookshelf, and finally at an estate sale.

Plenty of such unknown unknown exist, and the AI machine will inevitably destroy a bunch of them at this scale.

View on HN

You're trying to imagine rare valuable books and then fantasizing about AI companies destroying them, but what's actually happening here is that AI companies are acquiring, digitizing, and then pulping the instruction manuals to 1983-vintage copy machines.

This is all such a special-pleading argument. You know what other institution snatches up books and destroys them at huge scale? Public library systems. People clean out their attics and basements and drop off huge boxes full of books at libraries; libraries take the things they know will circulate, and destroy the rest. Take a guess as to how Þórbergur Þórðarson fares at the Newark Public Library. Wait, bad example, they stopped accepting book donations because nobody wants your old books. They tell you to give the books to thrift stores instead. Guess what the thrift stores do with them?

You know how many times I've read stories about the grave damage libraries are doing to human culture? Zero, zero times.

Source? You state that in a tone that implies you have verifiable knowlege of this. All the information I found says that the exact number, titles and authors are under NDA.

It's in the 404 story.

It's not. The 404 article states (paraphrase): we're not going to disclose any titles, so where does your information come from. Please provide a link.

https://www.npr.org/2026/08/19/nx-s1-5936438/why-tech-compan...

But we don't necessarily know what we might find valuable in the future.

1. This is not the 404 media story you quoted as source

2. Maiberg's answer is most likely a book that has no emotional value. Translation: they do not know

You're right. It's merely the writer of the 404 story, discussing their story.

Given no information, would you assume companies that are buying "huge quantities" of "rare" books are getting first edition Jules Verne by the crate, or 1989 Highlights magazines recovered from dentist offices?

Both. I think they are just buying whatever they can get their hands on most of it is going to be junk, but one in a million is going to be a rare and valuable book which we didn’t even know existed.

After scanning and destroying 10 million books, AI companies will have destroyed 10 such books.

(I am obviously assuming a Poisson distribution here where I pulled the parameter p = 1/1000000 out of thin air; point is p is non-zero; and at this scale the undesirable event is bound to happen a bunch of times).

I assume public libraries know what they are doing and know which book they are destroying, that they keep an accurate inventory and hire professionals maintaining said inventory and marking which books can safely be destroyed.

I assume no such things of AI companies.

Obviously that's not true. Newark is literally telling people to talk all their old rare books to thrift stores, which will immediately throw them in dumpsters.

This is a bad comparison. Public libraries are trying their best to preserve rare copies, despite being overwhelmed and underfunded. If they come across a rare copy chances are it will be preserved. When an AI company comes across a rare book, it will not recognize it as such.

There is also a fundamental difference in intention. AI companies are seeking out old books and will destroy them. Public libraries are trying their best but simply don‘t have the resources to find rare books in every collection.

And finally there is a fundamental difference in scale. Public libraries are not buying millions of copies to destroy, completely eliminating the chance they will ever be discovered.

---

PS. I don‘t like the anti-academic tone of your post. Librarians know what they are doing, their expertise is valuable, and when they act according to their specialized knowledge it does have a positive effect on the world.

What comparison? I'm talking about libraries as they actually exist and you're talking about what you think libraries should be doing. I'm arguing positively, and you're arguing normatively.

I'm not "anti-academic". I'm just very involved in my local community and I read the annual reports from our library system. We accept book donations, like a lot of suburban library systems do (from our extremely book-y community), and we are open about the fact that most of those books get trashed. We do not have the personnel on staff you claim libraries generally do.

I would say that between the two of us, I'm the one arguing more respectfully about libraries. I see them as real institutions with real pressures and constraints that do an important public service, and you see them as an instrumentality in your argument against AI.

You were the one who brought public libraries into this conversation, in an attempt to legitimize this practice by AI companies. I assumed this attempt was you saying the two behaviors are comparable.

I pushed back on that comparison. These behaviors are in fact not comparable. What the AI companies are doing is bad actually, and what libraries are doing is, while unfortunate, acceptable, given the limited resources they have.

Yes, I brought public libraries into this conversation, because they daily destroy hundreds of thousands of books, most of them without any professional evaluation whatsoever. I don't have a problem with this, because I understand a little how the book trade works, and whatever romantic idea people have about the preservation of books, it has little to do with what books actually are.

What I don't understand is your response to this. An AI company digitizes a book before destroying it: "bad actually". A public library takes a cartload of books to a dumpster without so much as opening the front cover of any of the books: just fine.

Can you explain why AI companies destroying these is worse than a library or bookseller destroying these?

Public libraries staff professional archivist, they take careful inventory and know exactly which book they are destroying.

AI companies buy boxes and boxes of books with unknown content staff anybody who can operate a scanner and have no idea which books they are destroying.

If a public library comes across a book they didn’t know they had, it is very likely that somebody will notice and know how to continue, to find out if this book is worth saving etc. AI companies will treat this book exactly like any other and destroy it to feed the plagiarist machine.

At this point I believe you're just sort of making up a fantasy of what you think a public library should do, not what they actually do.

This is not what most public libraries do with donation books. They mostly just recycle them.

The very act of cataloging donation books takes a huge amount of resources that the collection managers don’t have.

Many public libraries don’t have a collection management department at all! They outsource that function entirely to vendors and those vendors mostly source new books and spend resources preparing them for the hard use of a library (changing bindings, uploading and cleaning metadata, tagging, etc).

Decommissioning is done by volunteers who just look at the condition of books and chuck the bad looking ones in a bin for recycling.

My wife worked in this industry on the vendor side, your model of public libraries is closer to a small subset of certain big city libraries narrowed to their rare and research departments. The median book bought by a library is a Danielle Steele romance novel packaged for library consumption and sent to a small town library system that will be lucky to have a single professional librarian for the whole system.

Rare and valuable books are not going to be in the giant pallets AI companies are buying. Most public libraries do not have professional archivists or anything like that, not so charity shops, or booksellers.

No, but they are going to be intermixed in boxes from an old estate, or a used books store which is closing up, etc.

> Most public libraries do not have professional archivists

Most reasonably sized library systems do in fact. Even small ones have trained librarians with an some degree in library science who took at least an introductory class in archiving as a part of their degree, and very likely has some idea how to prevent rare books from being destroyed.

Right, and what happens to those boxes from old estates are that they get put into dumpsters. The only difference here are:

* It's happening on a much smaller scale

* They're actually digitizing the books before they destroy them

This obviously isn't about the books. It's about people want reasons not to like AI companies.

The scale matters. It is disingenuous to just remove it from the equation as if it doesn’t.

As does the intention matter. The AI companies are doing this because they want to profit off of it. They don’t have to do this, and the fact that they do is part of what makes them bad for humanity.

Right: the scale matters, which is why it's very weird to take AI companies to task for something the public library system does on a far greater scale. And, in fact, the AI companies do have to do this: they're required by law to destroy these books.

AI companies are doing this at a far lesser scale than libraries are, are you implying doing it less is worse?

I am going to have to see the numbers involved before I believe this. But even if it is true, it is still not a comparable behavior as I have argued above.

You didn't know it was happening at all just a few hours ago; you believed libraries had archivists on staff trained to preserve all the donations. You're in kind of an odd rhetorical position to be saying [citation needed].

Libraries do have archivists on staff. And I never believed libraries preserve all donations. I didn’t even mention donation, you did that.

I honestly don‘t understand what kind of a mission you are on here, and why you feel the need to defend AI companies by whatabouting libraries into this conversation. But whatever you are doing here, I think you have succeeded. This thread was supposed to be about the unambiguously bad behavior of AI companies, but now it is about libraries.

But OK. The numbers don’t add up. The 404 article talk about millions of copies, you mentioned hundreds of thousand. If we take these (extremely vague and unworkable) numbers at face value (which I am willing to do with the 404 article but not you because I trust journalists more than strangers on the internet), the scale of AI companies is one order of magnitude larger then libraries.

I still think this is a bad comparison regardless of the numbers. Seeking out books to destroy is worse behavior then not accepting donation. There is no equivalence in behavior here to compare.

Reply on news.ycombinator.com