Readers of fictional novels don't expect (or even want) such informational accuracy. Disruptive technologies like LLMs only need to produce books of good enough quality for the vast majority of such readers. If they're happy with what they're reading in the books that they purchased then so will the publishers and booksellers making more…
Readers of fictional novels don't expect (or even want) such informational accuracy. Disruptive technologies like LLMs only need to produce books of good enough quality for the vast majority of such readers. If they're happy with what they're reading in the books that they purchased then so will the publishers and booksellers making more money from them by adopting such innovations. What the authors think (and how they're financially impacted) won't come into it.
The book market is very heavily slated towards a small number of popular authors; most books sell fewer than 5,000 copies. Most books do not make money, and certainly not significant amounts of it.
You need to write something like JK Rowling or George RR Martin. Otherwise, you aren't actually meaningful competition in that market.
The quality of writing in the AI written works is extremely low, which means you need to compete in a market which is more focused on disposable, lower end content. The most vulnerable market would probably be romance novels, as being able to write a dirty story to yourself with an AI is appealing to a lot of people and the fact that it might not be of the best quality may not be as important, as a lot of romance novels aren't the highest quality to begin with, and being able to personalize it might make up for the other quality issues, as sure, it may not be the best writing, but it is specifically appealing to you in particular.
That's an actually valuable market that might be targetable. But the present quality of these AIs is not even to this level (no one wants a romantic partner to suddenly turn cold on them, or radically change personality or backstory) and a lot of these AIs aren't even trained to produce such things.
You're missing the bigger picture. You aren't appreciating the impact of scalability. Authors don't compete with each other, their books do. But you've identified the low-hanging fruit of women's romantic fiction as the largest sales segment.
Amazon (and the publishing industry in general) don't need human authors to write books. In fact, such authors are a cost for them. They 'only' need content to publish, with an appealing marketable front-end. Amazon has the infrastructure capable of generating an army of bot instances churning out cheap content, plus deriving training feedback advantage from their advertising, sales and Goodreads site. It's only the readership which will decide if LLM generated content is 'good enough', and so far they've proven that cheaper, more easily consumable content is a market winner.
There is no scalability because most books don't make significant money. The market size is heavily constrained for low-end books; producing 100x as many of them won't increase the market by 100x, and may actually decrease the size of the market by flooding the market with low-quality stuff that causes it to crash as people lose interest and go do other things (as seen in the video game crash of 1983).
The low end, low print run book market is simply not where the money in publishing lies. It's in the books that sell millions of copies.
Harry Potter sold about 80,000,000 copies per book in the series, or 600,000,000 for the seven books.
The typical self-published book sells about 250 copies. by comparison. And even the typical commercially published book sells only 3,000 copies.
Filling the market with more dross won't increase sales, it will decrease them. Indeed, we saw that with Steam, even without AI content - opening the floodgates lowered per-game sales.
I meant scalability of infrastructure, of which Amazon in particular has little shortage of. They also have 60% of the eBook market. Book electrons don't take up much space and the data overheads don't mean margin erosion at high volumes, unlike with printing and warehousing. When you don't pay for editing, design and printing then total high volume is how you offset low runs per author. 1 book for $10 @ 10% net is equivalent to 10 books at $1. Or just run the whole operation at cost to compete, because books are far from the only items on their marketplace shelves...
Authors are a small cost to publishers. They literally work for free to produce content and afterwards work free again to promote it. There is no better deal in business. The high costs of publishing have little to do with authors and replacing them with LLM content risks a public backlash against the cost of books. Will the public accept a mere %10 reduction in the price of books (the authors royalties) that were written by AI? Not likely. They'll want them practically free. Note I am talking about the products of large publishing houses. Independent operations run by one or two people and producing only e-books could get away with it. But if you are printing books, getting rid of authors is a bad idea.
Millions of books are already sold via subscription (eg KU) or for less than a dollar. Amazon eBook royalties are 35%/70%. Their paperback royalties are 60%. That's a noticeable amount at volume. For just one large publisher, a $250m chunk of like for like sales has been sliced from their genre market over the last decade, mostly by eBook self publishing, picked up mostly by KDP. Any opportunity for unit margin increases would be welcomed by a company's shareholders.
As with processed/frozen vs chef-cooked/organic food, readers will pay a premium price for a premium human-written book product. But most won't.
Right now authors bear the cost of formatting e-books and also printing when they sell on Amazon. It's very hard to make money through print books if you use Amazon's service though as you need to price your book high enough to cover costs. Most self-published authors work with printers in China, for example, to lower costs. Amazon shuffles off as much of the cost of producing books to others. Publishers have been complaining about this for years. Anyway, I'm not sure we're disagreeing in much. Amazon might be able to grow it's share of cheap genre fiction ebooks and super low quality paperbacks. It's just not clear LLM's are viable as authors since human fees are a tiny part of publishing costs.
Good points. But traditional publishing remains stuffed full of human-based process friction. Automation of 'manuscript workflow management' may be slow to come, but it will shake things up, as sure as DTP dumped the typist profession, no matter how bad writers still are at typing (or tech).
If LLMs will be able to create passable fiction, then I could just cut out the middleman altogether. Instead of paying a human author to create fiction using LLMs, I could just use an LLM to create my own custom fiction.
But, that's a hypothetical problem at the moment. LLMs can't produce compelling fiction because they don't have a well-developed theory of mind.
That's my point. But, similar to the Turing test, creating fiction doesn't require sentience, consciousness or the practical application of psychology or philosophy. We already happily anthropomorphise animals and robots. It just requires a machine to win a place at the 'imitation game' table. And 'compelling' will continue to remain in the eye of the beholder-cum-buyer.
". . . creating fiction doesn't require sentience, consciousness or the practical application of psychology or philosophy."
Well, technically you're right. LLMs *can* create fiction of some kind. The question is how many people will enjoy reading it. An entertaining novel involves suspense, psychological insight, and unexpected twists and turns. The author has to anticipate the reactions of the readers to some extent, which requires a sophisticated theory of mind. An A.I. program of some kind might be able to do this someday, but it probably won't be an LLM.
I've tried using ChatGPT-4 to write fiction and the results weren't that interesting. ChatGPT is very good at mimicking prose style (e.g., it can produce stuff that sounds like D.H. Lawrence or Edgar Allen Poe), but the writing had no real direction; it just cranked out descriptive paragraphs that didn't tell a story or lead anywhere.
I'm sure somebody will find a way to monetize LLM fiction and it might even be lucrative, but I doubt that LLM-produced novels will rise to the top of the best-seller list any time soon.
But their potential use in automating the publishing pipeline to replace human copy editors, proofreaders etc is no doubt being explored by companies looking to cut costs.
Readers of fictional novels don't expect (or even want) such informational accuracy. Disruptive technologies like LLMs only need to produce books of good enough quality for the vast majority of such readers. If they're happy with what they're reading in the books that they purchased then so will the publishers and booksellers making more money from them by adopting such innovations. What the authors think (and how they're financially impacted) won't come into it.
The book market is very heavily slated towards a small number of popular authors; most books sell fewer than 5,000 copies. Most books do not make money, and certainly not significant amounts of it.
You need to write something like JK Rowling or George RR Martin. Otherwise, you aren't actually meaningful competition in that market.
The quality of writing in the AI written works is extremely low, which means you need to compete in a market which is more focused on disposable, lower end content. The most vulnerable market would probably be romance novels, as being able to write a dirty story to yourself with an AI is appealing to a lot of people and the fact that it might not be of the best quality may not be as important, as a lot of romance novels aren't the highest quality to begin with, and being able to personalize it might make up for the other quality issues, as sure, it may not be the best writing, but it is specifically appealing to you in particular.
That's an actually valuable market that might be targetable. But the present quality of these AIs is not even to this level (no one wants a romantic partner to suddenly turn cold on them, or radically change personality or backstory) and a lot of these AIs aren't even trained to produce such things.
You're missing the bigger picture. You aren't appreciating the impact of scalability. Authors don't compete with each other, their books do. But you've identified the low-hanging fruit of women's romantic fiction as the largest sales segment.
Amazon (and the publishing industry in general) don't need human authors to write books. In fact, such authors are a cost for them. They 'only' need content to publish, with an appealing marketable front-end. Amazon has the infrastructure capable of generating an army of bot instances churning out cheap content, plus deriving training feedback advantage from their advertising, sales and Goodreads site. It's only the readership which will decide if LLM generated content is 'good enough', and so far they've proven that cheaper, more easily consumable content is a market winner.
There is no scalability because most books don't make significant money. The market size is heavily constrained for low-end books; producing 100x as many of them won't increase the market by 100x, and may actually decrease the size of the market by flooding the market with low-quality stuff that causes it to crash as people lose interest and go do other things (as seen in the video game crash of 1983).
The low end, low print run book market is simply not where the money in publishing lies. It's in the books that sell millions of copies.
Harry Potter sold about 80,000,000 copies per book in the series, or 600,000,000 for the seven books.
The typical self-published book sells about 250 copies. by comparison. And even the typical commercially published book sells only 3,000 copies.
Filling the market with more dross won't increase sales, it will decrease them. Indeed, we saw that with Steam, even without AI content - opening the floodgates lowered per-game sales.
*clap clap*
I meant scalability of infrastructure, of which Amazon in particular has little shortage of. They also have 60% of the eBook market. Book electrons don't take up much space and the data overheads don't mean margin erosion at high volumes, unlike with printing and warehousing. When you don't pay for editing, design and printing then total high volume is how you offset low runs per author. 1 book for $10 @ 10% net is equivalent to 10 books at $1. Or just run the whole operation at cost to compete, because books are far from the only items on their marketplace shelves...
Authors are a small cost to publishers. They literally work for free to produce content and afterwards work free again to promote it. There is no better deal in business. The high costs of publishing have little to do with authors and replacing them with LLM content risks a public backlash against the cost of books. Will the public accept a mere %10 reduction in the price of books (the authors royalties) that were written by AI? Not likely. They'll want them practically free. Note I am talking about the products of large publishing houses. Independent operations run by one or two people and producing only e-books could get away with it. But if you are printing books, getting rid of authors is a bad idea.
Millions of books are already sold via subscription (eg KU) or for less than a dollar. Amazon eBook royalties are 35%/70%. Their paperback royalties are 60%. That's a noticeable amount at volume. For just one large publisher, a $250m chunk of like for like sales has been sliced from their genre market over the last decade, mostly by eBook self publishing, picked up mostly by KDP. Any opportunity for unit margin increases would be welcomed by a company's shareholders.
As with processed/frozen vs chef-cooked/organic food, readers will pay a premium price for a premium human-written book product. But most won't.
Right now authors bear the cost of formatting e-books and also printing when they sell on Amazon. It's very hard to make money through print books if you use Amazon's service though as you need to price your book high enough to cover costs. Most self-published authors work with printers in China, for example, to lower costs. Amazon shuffles off as much of the cost of producing books to others. Publishers have been complaining about this for years. Anyway, I'm not sure we're disagreeing in much. Amazon might be able to grow it's share of cheap genre fiction ebooks and super low quality paperbacks. It's just not clear LLM's are viable as authors since human fees are a tiny part of publishing costs.
Good points. But traditional publishing remains stuffed full of human-based process friction. Automation of 'manuscript workflow management' may be slow to come, but it will shake things up, as sure as DTP dumped the typist profession, no matter how bad writers still are at typing (or tech).
If LLMs will be able to create passable fiction, then I could just cut out the middleman altogether. Instead of paying a human author to create fiction using LLMs, I could just use an LLM to create my own custom fiction.
But, that's a hypothetical problem at the moment. LLMs can't produce compelling fiction because they don't have a well-developed theory of mind.
That's my point. But, similar to the Turing test, creating fiction doesn't require sentience, consciousness or the practical application of psychology or philosophy. We already happily anthropomorphise animals and robots. It just requires a machine to win a place at the 'imitation game' table. And 'compelling' will continue to remain in the eye of the beholder-cum-buyer.
". . . creating fiction doesn't require sentience, consciousness or the practical application of psychology or philosophy."
Well, technically you're right. LLMs *can* create fiction of some kind. The question is how many people will enjoy reading it. An entertaining novel involves suspense, psychological insight, and unexpected twists and turns. The author has to anticipate the reactions of the readers to some extent, which requires a sophisticated theory of mind. An A.I. program of some kind might be able to do this someday, but it probably won't be an LLM.
I've tried using ChatGPT-4 to write fiction and the results weren't that interesting. ChatGPT is very good at mimicking prose style (e.g., it can produce stuff that sounds like D.H. Lawrence or Edgar Allen Poe), but the writing had no real direction; it just cranked out descriptive paragraphs that didn't tell a story or lead anywhere.
I'm sure somebody will find a way to monetize LLM fiction and it might even be lucrative, but I doubt that LLM-produced novels will rise to the top of the best-seller list any time soon.
True, they're more of a text generation accelerator/author's assistant at the moment, as with this account of Sudowrite usage: https://www.thecreativepenn.com/2023/02/08/how-i-used-generative-ai-tools-for-my-short-story-with-a-demons-eye/
But their potential use in automating the publishing pipeline to replace human copy editors, proofreaders etc is no doubt being explored by companies looking to cut costs.