That is of course not totally true, the copies are in private digital stores.
But for general public the knowledge is gone. Someone researching let’s say the history of a little village won’t be able to catch it.
This happens as well in current private collections. The physical book gets stored in a private book collection. But many also get out by the years when people are moving or die.
The amount of companies doing this is the real harm. There will be many companies running this process. It is absolutely not comparable to a single book collection getting burned in a house fire or being thrown away instead of donated to a second hand store.
It might feel for some as a not really big issue but actually it is in time.
That is not a realistic approach.
What's stopping say next year's billion or trillion dollar company from going into x market for used y, or vacuums up z limited natural resource? Should people have to suffer for the whims of a board of directors?
There's little difference and I would suggest there's a pity in knowledge being hoarded by collectors.
If we were getting into the differences, though, it would be that is now happening en masse in order for a handful of corporations to monopolise knowledge, history and ideas from books that were until recently available somewhere to the public.
I also doubt that many collectors are buying rare books with the intention of destroying them.
And it's irreversible when the books are really destroyed.
That combination makes it an urgent and important issue.
Or do they keep it on their private, warded off storage?
Better yet, why not raise prices by 10x for these online AI orders? Then again, I suppose you don’t really run a book shop in 2026 because you want to make money.
I guess I am criminal.
/s
Legally they can't, they would immediately land in copyright claim hell.
For example, require the digitized copies to be put in a public trust where for example libraries can provide these books as digital loans. (Per author adjustment for copyright, if it’s not possible now, store them nonpublically in the trust until later.)
And I bet it’s more likely that in 100 years you can read a book printed nowadays than you can read data digitized nowadays.
Canada is one of the few “western” countries without annoying customs duty/tariff requirements to ship to by post these days. US is a mess, and UK/Europe are going in the same direction.
That’s a good built-in way to ensure public preservation of knowledge IMO
If these AI companies were creating public repositories where researchers could access digital copies then I would not have such a problem with this, but that is not what they are doing. The ability of researchers to cite these books as sources is being snuffed out.
1) It's easier to scan.
2) US copyright law (currently, sorta, kinda) requires that only one copy of the book you purchased exist. If you scan it, you must destroy the other copy.
What's the "sorta, kinda" as you understand it? Has any judge, or law, created a requirement to destroy copies?
From what I understand, depending on the purpose, making digital copies may be fair use if it's "transformative". In a decision involving Google's book search feature, the court's decision highlighted that the amount of text displayed in the product was limited and not a "significant market substitute" [2]. The AI firms aren't trying to show you a digital copy of the books, and in a bunch of contexts already, if their product regurgitates too much verbatim text, that's going to be a liability (e.g. educational uses and plagiarism). I've also tried prompting LLMs to finish a paragraph from inside a novel from a well-known author, and they have not been able to get the next sentence typically. You can't easily and with confidence just extract back a book. I have not tried the experiment of seeing how many tokens you need to spend before a books is reproduced and what the cost is relative to the original work -- and you would already need the digital text to recognize when you had succeeded.
What's transformative? Previously the courts have said that scanning to offer a search product was transformative, given some limitations on how much of the matching book is shown to the user. I think there's a decent argument that training an AI model is at least as transformative as search indexing.
https://en.wikipedia.org/wiki/Authors_Guild,_Inc._v._HathiTr...
https://en.wikipedia.org/wiki/Authors_Guild,_Inc._v._Google,...
Ah yes, the old "if you can't beat the ruthless type A sociopathic American capitalists who literally don't give a shit about anything other than money and power, then you should join them and make a quick buck while the world burns". A tale as old as time really; we must mirror the behaviour of people who'd tell their self-driving car to drive over a gaggle of baby geese in order to continue working on the important task of privatizing the world's knowledge. \s
Hell no. Small town bookstore owners probably have the best job in the world; if they're smart, they'd continue doing exactly as they're doing.
This website for example may look like a user-to-user book sales platform but most books there are supplied by used bookstores: https://www.bokborsen.se/
I can morn paper but it's not like the contents are being lost.
Rumors of rare books being destroyed would be the bad part of this but if "rare" is the tourist guide to a local cave why should we care?
Feels alot like the water argument
Apparently not valuable enough that they could be arsed to save a copy from Anthropic's greedy grubby hands even after DECADES of head start.
How dare they!
Unless they're sharing the scans publicly or with the Internet Archive or equivalent, they are being lost.
The same effect would be if a hoarder purchased the book and never sold it or showed it to another soul. Same effect if a toddler purchased it, dropped it in their spaghetti bowl, and tore it apart.
But a copy of a book is a copy of a book! Books were designed for multiple copies being printed. Perhaps we recall Gutenberg and that whole controversy of easy printing from a press. So, the thing getting people all in a tizzy is that a single copy of a book that has, perhaps, become rare in 2026, a single copy that languished on a shelf somewhere, is being purchased and consumed.
Honestly, folks, books are really in no danger here. If you know any rare books that are irreplaceable, and with only one extant copy, just buy it yourself. For God's sake! Not rocket science.
Buyers and sellers come together of their own free will, unencumbered by interfering third parties and exchange money for goods and services.
There are a lot of comments in this thread proposing some kind of interfering in these transactions.
Don't like Amazon or Anthropic digitizing rare books? Then buy them yourself and stop with the kvetching.
The actual reason to get angry about this is copyright law, which would rather have all surviving copies of a (niche, old, rare but still covered by copyright) book in private hands and inaccessible to the public, rather than make it public domain.
Moreover, by scanning the books and putting them into AI models they make the knowledge actually publicly available. There's just no fucking way you'd ever read these books, while there is non-zero chance you might ask an LLM about it.
IMO the correct solution here is:
1. Things should become public domain after much shorter period of time, like 10 years or so.
2. Government should maintain a freely accessible archive of immaterial goods in public domain - music, movies, books.
The effect would be twofold:
1. Open models would have access to similar training data as private models, which would be a huge game changer.
2. Companies would be stimulated to come up with new, exciting media, rather than sell us same shit over and over again.
Or she just gets cold. My daughter wears those kind of gloves all winter regardless of how warm we heat the house.
The gloves are a special kind made of thermal material for people with Raynaud's disease, where narrowing of the blood vessels reduces circulation in cold temperatures.
People here don't/can't really heat their houses properly.
https://en.wikipedia.org/wiki/List_of_countries_by_wealth_pe...
Of course they have very low wealth inequality so they don’t have trillionaires at all which some people may see as a problem. In fact remarkably despite the median wealth of NZ being so much higher than the USA they have a lower average wealth per capita.
They just spread the wealth around making the middle class 3x richer.
I wonder if there is a per adult vs per person vs per working person difference in the measurement or something like that?
Edit: I've dug into this more, and the closest thing I can come up with for an explanation is that UBS changed its model for 2026 to "normalize household debt" in order to make it more comparable across countries. There doesn't seem to be sufficient information to explain how that change caused such a radical change.
[1]https://web.archive.org/web/20230815103944/https://www.ubs.c...
Emphasis mine. US median household is 2.53 persons [1]. 193/2.53 = 76k, which is within spitting distance of the 69k per person listed. The wikipedia is median wealth per capita
[1] https://www.census.gov/quickfacts/fact/table/US/HCN010222*