Nice article. Just today we had an order for a signed limited edition with a drawing by R. Crumb.
I happened to have the trade edition of the same book, so I gave them a partial refund and sent them the trade edition. I'll tell you what happens in a week or so. Our fulfillment rated is crashing
AI just tried to order a copy of your edition of Futuria Fantasia inscribed by Ray Bradbury to Tom Garner. There was an unsigned copy for $60 less that I tried to have drop shipped but the seller said he wouldn't send that book to the chipper.
These buyers only care about ISBNs, so they might not accept the trade edition (unless it had both ISBNs on the copyright page - I doubt the people unpacking the boxes would be hip to that nuance. And never give a refund! They have all the money!
Some books need to be preserved, and other don't. This is sad to say, but it is important in the long run.
Let's face it, we don't need 100 million mass market paperback copies of Sydney Sheldon's Master Of The Game, even if it is a classic 1970's novel. They have no real value on the used book market, and are fairly easy to find. Let 10 million or so of them go into the chopper and nothing is lost. It's different if it's a 1st edition hardback, especially if it's signed. But there's a lot of mass market paperback books just gathering dust on the shelves of used bookstores and private libraries around the world. Using them to train AI might not be a bad idea.
The same is true of Reader's Digest Condensed Books, Harlequin romance books, and similar publications; back issues of most magazines and newspapers, old text books, out of date encyclopedias and dictionaries, Bibles and religious books, and other works that overload bookshelves without being touched for years at a time.
I think that most booksellers and the market is intelligent enough to know what works are worth preserving and which ones are not. At least all of those old Green Leaf adult books from the 1970's have a practical use now instead of hiding in someone's closet. Good bye to most copies of the Sweet Valley High series, Flowers in the Attic, The One Minute Manager, The Peter Principle, and The Late Great Planet Earth.
I have no idea how an AI bot will react when trained on a lot of some very strange books that have been published over the decades, but that's not my concern. I'm sure that some of the results will be interesting. But I don't expect anything that is rare or valuable will be sacrificed to the great AI bot in Google or Meta. Plus, most of these titles are already available either as ebooks or on the internet.
What no one seems to be thinking about are the number of records, CDs, and DVDs that are feeding the circuits of AI bots. But recordings seldom get destroyed in the process.
But I do look forward to hearing Humphrey Bogart read The Valley Of The Dolls to me as I surf the internet. It is a brave new world in which we live, indeed.
Fantastic overview, the best I've read on the topic. Selfishly, you’ve saved me from having to attempt to write anything similar, which I wouldn’t have been able to do half as well. You do a great job with the numbers, which are illuminating.
The thought I have been wrestling with from the beginning, and still holding, is that these companies aren't singularly responsible for the destruction of "rare books." First, as discussed, the great majority aren't rare books to begin with (even if scarce). But more importantly, books are already being lost to time and apathy at a much higher rate. The biggest threat to books remains disinterest. A text can die at the hands of Anthropic, etc, or it can succumb to slow neglect stuffed away on a shelf or buried in some pile, after which it will meet the same end anyway; at least the former offers something of an honourable death, for there being the prospect of some use. I'm not saying I'm happy about either.
I think it may be worthwhile to write a new novel in the vein of Ray Bradbury - the dystopian future that may arise from Fahrenheit 451 or rather Nanometer 451 - the broadband white light spectrum used by a dual-LED illumination system. How is society shaped from here?
Thanks, Ray. I don't know... between F451, Player Piano, The Neuromancer, and heaven forbid, The Parable of the Sower, I think novelists already have it covered.
Outstanding overview of the current situation. One point that you did not push quite so hard is that many, even most of these books had large printings (they are, after all, post ISBN titles). Thus, even with this destruction by a few AI companies, a great many copies remain in circulation of the vast majority of these titles. It was duly noted that many are books that may have already been headed for recycle bins and/or forgotten on bookseller shelves. I think the concerns are overblown. I agree with MadBlue's comment...keeping obscure information preserved digitally may just be superior to there existing one, two, even five or 10 more copies of an obscure book gathering dust on someones shelves. I seriously doubt if AI is often chewing up the last existing copy of anything. And if there was a fire, flood or catastrophe...the information is preserved. But yes, preserving the intact copy IS the best solution.
Thanks, Bud. I generally agree that most of the books scanned (and pirated) were common. Those mostly came from Better World Books, World of Books, etc. The material being purchased from small sellers has tended to be much more obscure and or specialized. Almost all of the books destructively scanned will be preserved in libraries. But taking the four I sold. One is still in print (AI just bought a used copy at half price); one was a somewhat scarce collectible book of which there will be one less copy now; and two are not available for purchase - I had the only copies on the market. I also keep in mind that we're not talking about one AI company. Many companies are following the same path, including companies in other countries. So AI could easily exhaust the supply of tens of thousands (or more) scarce books.
Thank you for this thoughtful and enlightening essay. I am saddened at the thought of the destruction of any book, but your point about allowing some obscure information to live on was helpful. What interesting times we live in.
Nice article. Just today we had an order for a signed limited edition with a drawing by R. Crumb.
I happened to have the trade edition of the same book, so I gave them a partial refund and sent them the trade edition. I'll tell you what happens in a week or so. Our fulfillment rated is crashing
AI just tried to order a copy of your edition of Futuria Fantasia inscribed by Ray Bradbury to Tom Garner. There was an unsigned copy for $60 less that I tried to have drop shipped but the seller said he wouldn't send that book to the chipper.
These buyers only care about ISBNs, so they might not accept the trade edition (unless it had both ISBNs on the copyright page - I doubt the people unpacking the boxes would be hip to that nuance. And never give a refund! They have all the money!
Some books need to be preserved, and other don't. This is sad to say, but it is important in the long run.
Let's face it, we don't need 100 million mass market paperback copies of Sydney Sheldon's Master Of The Game, even if it is a classic 1970's novel. They have no real value on the used book market, and are fairly easy to find. Let 10 million or so of them go into the chopper and nothing is lost. It's different if it's a 1st edition hardback, especially if it's signed. But there's a lot of mass market paperback books just gathering dust on the shelves of used bookstores and private libraries around the world. Using them to train AI might not be a bad idea.
The same is true of Reader's Digest Condensed Books, Harlequin romance books, and similar publications; back issues of most magazines and newspapers, old text books, out of date encyclopedias and dictionaries, Bibles and religious books, and other works that overload bookshelves without being touched for years at a time.
I think that most booksellers and the market is intelligent enough to know what works are worth preserving and which ones are not. At least all of those old Green Leaf adult books from the 1970's have a practical use now instead of hiding in someone's closet. Good bye to most copies of the Sweet Valley High series, Flowers in the Attic, The One Minute Manager, The Peter Principle, and The Late Great Planet Earth.
I have no idea how an AI bot will react when trained on a lot of some very strange books that have been published over the decades, but that's not my concern. I'm sure that some of the results will be interesting. But I don't expect anything that is rare or valuable will be sacrificed to the great AI bot in Google or Meta. Plus, most of these titles are already available either as ebooks or on the internet.
What no one seems to be thinking about are the number of records, CDs, and DVDs that are feeding the circuits of AI bots. But recordings seldom get destroyed in the process.
But I do look forward to hearing Humphrey Bogart read The Valley Of The Dolls to me as I surf the internet. It is a brave new world in which we live, indeed.
Scott,
Fantastic overview, the best I've read on the topic. Selfishly, you’ve saved me from having to attempt to write anything similar, which I wouldn’t have been able to do half as well. You do a great job with the numbers, which are illuminating.
The thought I have been wrestling with from the beginning, and still holding, is that these companies aren't singularly responsible for the destruction of "rare books." First, as discussed, the great majority aren't rare books to begin with (even if scarce). But more importantly, books are already being lost to time and apathy at a much higher rate. The biggest threat to books remains disinterest. A text can die at the hands of Anthropic, etc, or it can succumb to slow neglect stuffed away on a shelf or buried in some pile, after which it will meet the same end anyway; at least the former offers something of an honourable death, for there being the prospect of some use. I'm not saying I'm happy about either.
Scott,
Truly insightful - as usual.
I think it may be worthwhile to write a new novel in the vein of Ray Bradbury - the dystopian future that may arise from Fahrenheit 451 or rather Nanometer 451 - the broadband white light spectrum used by a dual-LED illumination system. How is society shaped from here?
Ray
Thanks, Ray. I don't know... between F451, Player Piano, The Neuromancer, and heaven forbid, The Parable of the Sower, I think novelists already have it covered.
Outstanding overview of the current situation. One point that you did not push quite so hard is that many, even most of these books had large printings (they are, after all, post ISBN titles). Thus, even with this destruction by a few AI companies, a great many copies remain in circulation of the vast majority of these titles. It was duly noted that many are books that may have already been headed for recycle bins and/or forgotten on bookseller shelves. I think the concerns are overblown. I agree with MadBlue's comment...keeping obscure information preserved digitally may just be superior to there existing one, two, even five or 10 more copies of an obscure book gathering dust on someones shelves. I seriously doubt if AI is often chewing up the last existing copy of anything. And if there was a fire, flood or catastrophe...the information is preserved. But yes, preserving the intact copy IS the best solution.
Thanks, Bud. I generally agree that most of the books scanned (and pirated) were common. Those mostly came from Better World Books, World of Books, etc. The material being purchased from small sellers has tended to be much more obscure and or specialized. Almost all of the books destructively scanned will be preserved in libraries. But taking the four I sold. One is still in print (AI just bought a used copy at half price); one was a somewhat scarce collectible book of which there will be one less copy now; and two are not available for purchase - I had the only copies on the market. I also keep in mind that we're not talking about one AI company. Many companies are following the same path, including companies in other countries. So AI could easily exhaust the supply of tens of thousands (or more) scarce books.
Thank you for this thoughtful and enlightening essay. I am saddened at the thought of the destruction of any book, but your point about allowing some obscure information to live on was helpful. What interesting times we live in.
It's a very muddy situation.