>VERBOSE Maximum verbosity.

Confirmed.

"training AI on a book without paying for it is stealing" -> "fine, we'll buy the book"

"if you pay for one copy you can only keep one copy" -> "okay, so we'll shred the physical copy and keep the digital one"

"now nobody can read it!" -> "you couldn't read our copy anyway, because we're not interested in selling it, and you've set up a legal framework where you can't read books without paying for them"

Point 3 has been a problem as long as books and markets have existed. Point 2 is the most common legal solution for digital lending libraries. The only thing that's new is point 1, and that's a direct response to lawsuits that people on this website have long cheered.

There is no good reason to require this. There are several bad reasons to require this. But it is required.

This is why digital libraries keep these physical copies, and can't legally sell them, or give them away, or lend them (libraries forbidden to lend!), or destroy them. The books just stand there uselessly, taking storage space. A complete waste, and yet another monument to the stupidity of copyright law.

If you're not lending digital copies though, you are not obliged to keep the originals. You could sell them or give them away. In the US this would be perfectly legal, since the first-sale doctrine summarily overrides copyright. You bought it, you own it, and you don't need the rightsholder's permission to sell it again. Or, you can shred it to pieces.

So let's talk about how book-scanning on a mass scale actually works.

Google Books and the Internet Archive pioneered book-scanning methods that do the job without harming the physical copy. The last part was necessary when the books were on loan from libraries: the Google Books M.O. for the biggest copyright heist in history was going to university libraries and telling them "hey, lend me your collection and I'll digitise it for free!", which of course they would be crazy to refuse. When the books were borrowed by the Internet Archive (typically public domain), or purchased with the required ratio of 1 physical / 1 digital copy, it was also necessary to keep the book intact. (Also, unlike everyone else discussed here, people in the Internet Archive actually love books.) But these scanning methods require skilled workers, and some very specialised and expensive machines.

[x]

More importantly, these methods take a lot of extra time. They're manual.

On the other hand, the fastest way to scan a book is to unceremoniously break off the spine, and aggressively feed the pages to a high-speed scanner. Google Books did that first, for the books it purchased or otherwise collected (not all of it is from libraries). Because a company that scans on a mass scale its own, not on-loan books, and is NOT obliged by law to keep a copy, has simply no incentive to wait who knows how long, and pay extra cash, just to keep the books intact. All it wants is to do this as fast as possible.

So basically, AI companies are using the same scanning method Google Books used for the books it didn't borrow. (Intellectual Property lovers, remember to give credit to the folks at Google for inventing the method!) By design, the books are destroyed in the process of scanning.

Speaking of Intellectual Property lovers, I don't know what their problem is, they got what they wanted! The rightsholders attempted to make scraping/analysing data a matter of copyright infringement (it's not) and they failed (thankfully). Then they noticed that in some cases the training corpus was acquired via good old piracy, and went oh no, you can't do that! It's copyright infringement! And here they succeeded (because let's be real, it was), and the AI companies went "alright then, we won't pirate books any more, we'll legally buy them". And they did.

The tiny violins we're hearing right now are not really about "rare" (pre-2022, huh?) books getting destroyed. They're about books bought in bulk second-hand. And hey, remember that US first-sale doctrine? You know what it means when you buy a book second-hand? No royalties for the rightsholders. The book was already sold, and they've already got their cut. They're not getting it again. Sorry guys. If it makes you feel better, when you buy second-hand furniture, you don't pay royalties to the carpenter either.

P.S. Free tip, never say "authors" or "artists" or "creators" in matters of copyright. Instead, train yourself to always say "rightsholders". Because that's where copyright applies to, NOT authors. (Of course sometimes they're the same person, but for most works in circulation, they're not, there's a publisher or a platform or an IP hoarding corporation, heirs and "estates" and trolls and middlemen of all sorts.) It will help you understand the situation and the stakes.

A photo of Dr. Nancy Grace Roman holding a model of the Hubble Space Telescope.

You have to be a pretty big deal for NASA to name a space telescope after you, and that is certainly the case for Dr. Nancy Grace Roman — our Roman Space Telescope’s namesake. She’s sometimes called the “mother of Hubble,” but she’s really the mother of all space-based observatories and NASA science in general. As NASA’s first chief astronomer, Dr. Roman's vision and leadership supercharged space science and helped make NASA the world-class scientific institution it is today.

Other useful AI Removal/Detection links

How to spot AI Hallucinations like a Reference Librarian (originally on Substack)

this will kill people and they know it, because they prefer us dead and out of sight. the government, the NHS and even the opposition have made this clear again and again. a twenty-five year old can have sex, get married have children, join the military, earn a living, be halfway up the corporate ladder or highly successful in their chosen career field by that age, but they can’t transition under NHS rules. this is a death sentence for trans teenagers and they are going ahead with it. it’s trans genocide, same as across the pond.

But I know. I think it was me.

In my defense I never intended to start a rumor. Many years ago I worked as a cleaner at the school and one evening I had to transport one of those big floor washing machines from the basement to the second level via the elevator. When the doors opened a very confused looking man stood inside. He was one of those slicked back gym-bro IT guys and made no movement to get out. The elevator wouldn’t fit him, me and the machine so I asked “Where are you going? Up or down?”

He gave me a smug shit-eating grin and said “Down?” in a mocking tone.

It took me a second to realize that of course he wasn’t going down, we were in the basement, but his look and tone annoyed me so much I refused to admit I misspoke and instead said “Yeah, down. I don’t know if you’re going to the second basement”

His smile disappeared “There’s a second basement?”

“Yeah but it sounds like you don’t have access to it so I guess you’re going up? I’ll just wait”

I never thought of it as anything other than a funny story to tell about that time I got so annoyed with a guy that I invented an entire second basement, but it turns out he probably refused to believe a cleaner fooled him and the story spread.

We already saw during the primaries that the GOP is cutting as many polling places as possible so people can't make it in time or aren't able to stand in line long enough to vote.

Tarrant County is about to have a vote to cut about half of their polling sites. Expect more of that in your area, and expect that you might not hear about it.

Also, check REPEATEDLY that you're registered to vote, especially as we get closer to the election! Just because you were registered in March doesn't necessarily mean you'll still be registered in November.

The last day to register to vote in the midterms is October 5.

Oh, the party's nominee wasn't even in your primary? Too bad, the GOP was worried their candidate might lose so they redid the voting maps. You get next door's candidate instead.