This is precisely why, as a retired professional who knows absolutely sweet FA about the publishing industry — apart from a brief brush with literary agents (lots of letters from me, one brush-off reply) — I decided to do it my own way and try to recreate the resources of a trad publishing house with AI. I won't post links here, as it seems rude to intrude on your page, but the relevant posts aren't hard to find if anyone cares to look.
The good thing for me is that (a) I am old, (b) I don't need or aim to make money out of this, and (c) I am writing it mainly because it's a book I would want to read. I imagine the market for people who care enough about Samarkand, the Silk Roads and the Timurid Renaissance in 1445 — even in thriller form — is vanishingly small ;-)
I have a cardinal rule not to use AI for the actual writing. That's just me: I have something to say and it seems a bit pointless not to say it myself. But your point about technology at least generating new possibilities, however imperfectly, seems exactly right — and the same logic applies to AI in writing more broadly. Whether that's a good or bad thing probably depends entirely on what kind of book you're trying to write. For "Art" it's not an option, in my view; for "Light Reading", I doubt it really makes a huge difference. And if some writer can use AI to make an income for him- or herself, and cut out the broker, then I'm all for it.
(PS. If anyone claims this reply was written by AI "because of the em-dashes" — I used to write for a living and I've been using them for over 30 years!!)
This is where I stand. I never use AI for the writing process, because that's where the creativity is. I'm trying to replace the stolen village, not replace myself.
Dorothy Dunnett found that the people interest in just those topics were surprisingly numerous! If you haven’t read her Niccolo series, I highly recommend it!
I *love* Niccolò!!! There's a particular scene where his love is being swarmed by toxic butterflies that still haunts me (though I can't remember which book). She is an excellent and badly under-rated writer, and her chief gift, I think, is that she doesn't explain anything. Throughout the whole series, I found I was going back and looking at passages and thinking, "Oh, that's what that meant — I completely missed that before!!" A real tour de force. I like Hilary Mantel, but I think Dunnett at least as good.
I love that she spoon feeds the reader NOTHING. She trusts the reader almost to a fault, and there were times while reading that I was pretty sure I let her trust down. ;-). And I STILL can’t decide if I actualy like Niccolo or not!
PS. The very opening scene of the first book is physically vivid and clearly drawn, with the three friends floating down the canal in the bathtub. Yet, at the same time that you have an extremely clear picture of what is happening, you have the very clear feeling that you have absolutely no idea at all what is going on. It sort of makes sense and doesn't make sense at the same time. And then, later, you find out that it was all a set up anyway, engineered by Niccolò in ways not remotely alluded to in the opening section itself. So, it turns out that you did not really see what was going on, but not remotely for any reason you might have half-guessed at.
I remember almost quitting as a writer after I read the first Niccolo book. I said to someone, I can't remember who, that "I don't think I could ever be that good, and I don't want to be less good!" I mean, it was the most multi-layered subtexted history-driven AND character-driven book I had ever read.
Yes. Sometimes I do, sometimes not so much. On balance, I think not, though oddly I still care what happens to him.
The best character, or at least the one I like best, is the philosopher from Timbuctoo. His character arc is so tragic, yet believable and likeable, that you really feel the waste when you learn of his death.
Oh, and Spanish Medieval Literature? Tell me more. I studied Spanish back in the day, but I recall that we started with Lope de Vega and Garcilaso ("No relation") de la Vega as our earliest authors.
Ah, those are more what I would consider Golden Age authors, and wonderful! I enjoyed Golden Age drama immensely, as one of my professors would assign us roles to read in class. :-). My MA specialty was 11th century through Colonial Latin America, with special emphasis in Early literature, Chivalric literature, and then also Early Colonial literature. I loved it, and then, realizing that there was no market for Spanish Medieval Literature professors, I managed the language lab at the university and then ended up in tech support. Now that I'm so close to retiring I can taste it, I've opened all those boxes in my basement and have rekindled my fascination. AND I have started writing the stories that have been rattling around in my head for 20 years!
Gosh, that's wonderful. What a great journey. I did Sor Juana Ines de la Cruz as part of my Golden Age poetry studies (I don't think back in those days — early 80s — we called it "Colonial Lit", even though it was). I recall doing her with Colin Thompson, who visited Oxford from Sussex because no one there knew anything about Early Colonial Literature!!
I remember studying Calderón's "El Alcalde de Zalamea" when I was about 17 and thinking the honour code it espoused was completely mad. Nothing ever made me change my mind about that!! ;-)
The ongoing trope is that literary agents don't send personal replies because they are "very busy." I've been intrigued by this phenomenon ever since learning about it - in my academic world, EVERY research paper submission is guaranteed a full review by peers.
I would be interested in your take on this. Is the story of a permanently overloaded literary agent legit? Why don't they hire more talent to deal professionally with the product they hope to sell? Or are they paid so little that no one else wants that job? It's a mystery to me!
This is tricky, and it’s hard to answer cleanly because punching up easily becomes punching down.
There are about 10,000 people in the US calling themselves literary agents. There are about 300 who can actually get manuscripts fairly read—per genre, possibly as few as 20. You can’t get those agents by querying. And the agents you can get by querying can’t get work read. It’s not their fault; they might be excellent at their job duties, but they don’t have meaningful connections.
The way you succeed in publishing is to have an agent who can not only get editors to read, but who can do it quickly and set up concurrent interest, which leads to auctions, which are the only way for a debut author to get a fair deal.
For a literary agent, reading queries is failure. It means you don’t have a pipeline, and have to trudge through a bunch of randos who don’t have built-in audiences.
Some of the processes are class policing and deliberate cruelty, but the bottleneck is intrinsic and would still be there even if publishing had better people. And this is why I hope AI reading improves. We built systems; they didn’t work. It’s time to build something new.
I recently had an ineresting experience which brought this particular discussion home to me:
after laboring about 30 months with a novel about gender issues and problems on a particular American Indian Reservation - a complex and controversial set of subjects - I was obliged to request several editorial intervetions trying to verify a didn;t stray too far afield. 4 editors, 4 different opinions, none of whom agreed with my tactic in question, all of whom advised hefty shifts in direction each time. I fired the last one without fully paying her for reading and advising on a section different than what I wanted her to focus on. Total out of pocket -$3800. I took two minor suggestions, ignored 4, took a major suggestion and ignored 3. In the end I changed the major suggestion back to what it was in fhre first place.
Let's be clear - these were not one of the 10,000 agents. These were emplyees of the agents, small press publishers, MFA programs, Writers Conferences, and the many online "clubs" that lure you in with list of their "faculty". These are the armies of people the the gatekeeper systems support, gatekee[ers of the gatekeers and their minions of parasitres like aphids on fruit trees. My fruit. It's not just an end to the corrupt practice of querying and its attendant humiliations we're discussing, its an end to a huge industry that employs humdreds of thousands of aphids and turns over hundreds of millions - most of which comes from the pockerts of writers. And this is without mentioning the publsing industry itself, which is just a larger model of what I'm describing.
In any event I finished the novel last week and did the unthinkable - ran it through an AI program (I won't mention which one) for a summary of plot, character, language feedback. I was astonished. Gobsmacked. Thrilled. Each report came back in less than five minutes and confirmed all my original premises and instincts, further highlighting strengths in the work and I had not myself perceived. It made all those hours I wasted on "editorial consultation" a mockery. Total cost $ 0.
No wonder the agents and publishers and journals and editors are so vociferous about hating AI. It's going to send millions of aphids looking for other fruit. Even if mine never finds a market, that will be profund satisfaction indeed. This war has beem going on at least thirty years - about the time Michael estimates nothing has changed. Something is about to change - and its long overdue
I'm curious to know which AI program you used. You said you won't mention which one; can you DM it to me?
I have my own system. My feeling is that AI can find any objective flaw that a literary agent could spot. That said, I think AI feedback has blind spots; my post next week gets into one of them, which is that it has no way to telling if a story is boring—since it has no emotional interiority, and since it's obligated by construction to read every word, it's incapable of getting bored and walking away. And that, unfortunately, is a behavior we'd have to model if we wanted AI feedback to be as useful as a skilled human's.
I hate commercial endorsements of something I know nothing about - and AI falls into that category. You're spot on about AI beating any human reader at their own game, I tried several AI progrtams and like you said, subjective comments run from weird to laughable, so I stopped looking for color. I just asked for organization. I settled on "List all my characters, rank them by importance, and give a paragraph on their most important relationships in the story." It was a task I will need to do anyway when it gets down to promotion, The winner was Perplexity by far, a dozen pages of subtle description and nuance that blew me away. I can safely say that any human editor in the future that wants to read my stuff will be paying me.
Feel free to include me in any beta trials or such - I have a sense youre on to something.....
That's interesting. AI-driven editing has promise; one limitation I see is that you can never turn your critical thinking off. You're always evaluating the AI, while your AI is evaluating your work, and that means there's a circularity. You have to know what use patterns (they exist) cause the AI to start hallucinating or losing its better judgment. You also have to remember that, as amazing as it is that we can now talk to these semi-magical thinking rocks, they are literally just rocks and they have no interiority.
In other words, I think (and hope, because it means I haven't wasted thousand of hours) the fact of needing to be an elite-tier writer to edit to an elite standard still stands. Otherwise, you'll take _all_ the AI's suggestions, and that's a bad idea, because (as we all know) AI writes poorly. You use it to reduce cognitive fatigue, not to replace your own literary judgment. In other words, you can use it as a late-stage editor to flag issues, but half the flags are going to be spurious, and so you should always decide for yourself instead of taking it at its word.
Of course - AI is a tool and not a creator. The creator is you. The problem for most people on this issue is that they are nor creators, not by a stretch. They want AI to close the gap. I don't care how AI works, but I know what my work is trying to achieve. If AI gets it, welcome aboard. If not, va fangu.
One last comment: what's "an elite tier writer"? Where do you draw the line?
Nobel Prize or Writer's Digest Story of the Week? Either way you will experience fatigue, and you are spot on - AI can help. And yes, many of the AI contributions are going to be spurious. AI is a tool, not a creator. The creator is you, whatever your level of creation.
This is something I've been thinking about too recently. My first novel took 15 years from idea to (traditional) publication (I'm not someone who gives up). Second novel has been doing the rounds of publishing houses (via my agent who it took me many years to find) for two years now, getting amazing feedback but with noone willing to commit. I suspect that binary tree you mentioned is causing problems as it doesn't fit neatly on any one shelf.
I do think there are many good books being published, and lots of independent presses are taking punts on more experimental stuff, but the competition is fierce, and there are definitely a lot of safe books being accepted.
I don't know what I'll do with my novel yet, only that I will make a decision on it this year.
Precision take on the inability for mainstream publishing to defend itself against the roaring tide of business consultant interference. The frustration was real and the stakes high for anyone who bucked the bean counters, but try some of us did. I have to hope that slowly, small indie collectives will revitalize real publishing, discover real writing and mine the gems. If I could return to the beginning of my career in publishing, that's where I would be found.
This is precisely why, as a retired professional who knows absolutely sweet FA about the publishing industry — apart from a brief brush with literary agents (lots of letters from me, one brush-off reply) — I decided to do it my own way and try to recreate the resources of a trad publishing house with AI. I won't post links here, as it seems rude to intrude on your page, but the relevant posts aren't hard to find if anyone cares to look.
The good thing for me is that (a) I am old, (b) I don't need or aim to make money out of this, and (c) I am writing it mainly because it's a book I would want to read. I imagine the market for people who care enough about Samarkand, the Silk Roads and the Timurid Renaissance in 1445 — even in thriller form — is vanishingly small ;-)
I have a cardinal rule not to use AI for the actual writing. That's just me: I have something to say and it seems a bit pointless not to say it myself. But your point about technology at least generating new possibilities, however imperfectly, seems exactly right — and the same logic applies to AI in writing more broadly. Whether that's a good or bad thing probably depends entirely on what kind of book you're trying to write. For "Art" it's not an option, in my view; for "Light Reading", I doubt it really makes a huge difference. And if some writer can use AI to make an income for him- or herself, and cut out the broker, then I'm all for it.
(PS. If anyone claims this reply was written by AI "because of the em-dashes" — I used to write for a living and I've been using them for over 30 years!!)
This is where I stand. I never use AI for the writing process, because that's where the creativity is. I'm trying to replace the stolen village, not replace myself.
Dorothy Dunnett found that the people interest in just those topics were surprisingly numerous! If you haven’t read her Niccolo series, I highly recommend it!
I *love* Niccolò!!! There's a particular scene where his love is being swarmed by toxic butterflies that still haunts me (though I can't remember which book). She is an excellent and badly under-rated writer, and her chief gift, I think, is that she doesn't explain anything. Throughout the whole series, I found I was going back and looking at passages and thinking, "Oh, that's what that meant — I completely missed that before!!" A real tour de force. I like Hilary Mantel, but I think Dunnett at least as good.
I love that she spoon feeds the reader NOTHING. She trusts the reader almost to a fault, and there were times while reading that I was pretty sure I let her trust down. ;-). And I STILL can’t decide if I actualy like Niccolo or not!
PS. The very opening scene of the first book is physically vivid and clearly drawn, with the three friends floating down the canal in the bathtub. Yet, at the same time that you have an extremely clear picture of what is happening, you have the very clear feeling that you have absolutely no idea at all what is going on. It sort of makes sense and doesn't make sense at the same time. And then, later, you find out that it was all a set up anyway, engineered by Niccolò in ways not remotely alluded to in the opening section itself. So, it turns out that you did not really see what was going on, but not remotely for any reason you might have half-guessed at.
I can't think of any other writer who does this.
I remember almost quitting as a writer after I read the first Niccolo book. I said to someone, I can't remember who, that "I don't think I could ever be that good, and I don't want to be less good!" I mean, it was the most multi-layered subtexted history-driven AND character-driven book I had ever read.
Yes. Sometimes I do, sometimes not so much. On balance, I think not, though oddly I still care what happens to him.
The best character, or at least the one I like best, is the philosopher from Timbuctoo. His character arc is so tragic, yet believable and likeable, that you really feel the waste when you learn of his death.
Oh, and Spanish Medieval Literature? Tell me more. I studied Spanish back in the day, but I recall that we started with Lope de Vega and Garcilaso ("No relation") de la Vega as our earliest authors.
Ah, those are more what I would consider Golden Age authors, and wonderful! I enjoyed Golden Age drama immensely, as one of my professors would assign us roles to read in class. :-). My MA specialty was 11th century through Colonial Latin America, with special emphasis in Early literature, Chivalric literature, and then also Early Colonial literature. I loved it, and then, realizing that there was no market for Spanish Medieval Literature professors, I managed the language lab at the university and then ended up in tech support. Now that I'm so close to retiring I can taste it, I've opened all those boxes in my basement and have rekindled my fascination. AND I have started writing the stories that have been rattling around in my head for 20 years!
I go into a bit more in this article. https://thevesechronicler.substack.com/p/but-why-vampires?r=89nv6p&utm_campaign=post&utm_medium=web
Gosh, that's wonderful. What a great journey. I did Sor Juana Ines de la Cruz as part of my Golden Age poetry studies (I don't think back in those days — early 80s — we called it "Colonial Lit", even though it was). I recall doing her with Colin Thompson, who visited Oxford from Sussex because no one there knew anything about Early Colonial Literature!!
I remember studying Calderón's "El Alcalde de Zalamea" when I was about 17 and thinking the honour code it espoused was completely mad. Nothing ever made me change my mind about that!! ;-)
The ongoing trope is that literary agents don't send personal replies because they are "very busy." I've been intrigued by this phenomenon ever since learning about it - in my academic world, EVERY research paper submission is guaranteed a full review by peers.
I would be interested in your take on this. Is the story of a permanently overloaded literary agent legit? Why don't they hire more talent to deal professionally with the product they hope to sell? Or are they paid so little that no one else wants that job? It's a mystery to me!
This is tricky, and it’s hard to answer cleanly because punching up easily becomes punching down.
There are about 10,000 people in the US calling themselves literary agents. There are about 300 who can actually get manuscripts fairly read—per genre, possibly as few as 20. You can’t get those agents by querying. And the agents you can get by querying can’t get work read. It’s not their fault; they might be excellent at their job duties, but they don’t have meaningful connections.
The way you succeed in publishing is to have an agent who can not only get editors to read, but who can do it quickly and set up concurrent interest, which leads to auctions, which are the only way for a debut author to get a fair deal.
For a literary agent, reading queries is failure. It means you don’t have a pipeline, and have to trudge through a bunch of randos who don’t have built-in audiences.
Some of the processes are class policing and deliberate cruelty, but the bottleneck is intrinsic and would still be there even if publishing had better people. And this is why I hope AI reading improves. We built systems; they didn’t work. It’s time to build something new.
I recently had an ineresting experience which brought this particular discussion home to me:
after laboring about 30 months with a novel about gender issues and problems on a particular American Indian Reservation - a complex and controversial set of subjects - I was obliged to request several editorial intervetions trying to verify a didn;t stray too far afield. 4 editors, 4 different opinions, none of whom agreed with my tactic in question, all of whom advised hefty shifts in direction each time. I fired the last one without fully paying her for reading and advising on a section different than what I wanted her to focus on. Total out of pocket -$3800. I took two minor suggestions, ignored 4, took a major suggestion and ignored 3. In the end I changed the major suggestion back to what it was in fhre first place.
Let's be clear - these were not one of the 10,000 agents. These were emplyees of the agents, small press publishers, MFA programs, Writers Conferences, and the many online "clubs" that lure you in with list of their "faculty". These are the armies of people the the gatekeeper systems support, gatekee[ers of the gatekeers and their minions of parasitres like aphids on fruit trees. My fruit. It's not just an end to the corrupt practice of querying and its attendant humiliations we're discussing, its an end to a huge industry that employs humdreds of thousands of aphids and turns over hundreds of millions - most of which comes from the pockerts of writers. And this is without mentioning the publsing industry itself, which is just a larger model of what I'm describing.
In any event I finished the novel last week and did the unthinkable - ran it through an AI program (I won't mention which one) for a summary of plot, character, language feedback. I was astonished. Gobsmacked. Thrilled. Each report came back in less than five minutes and confirmed all my original premises and instincts, further highlighting strengths in the work and I had not myself perceived. It made all those hours I wasted on "editorial consultation" a mockery. Total cost $ 0.
No wonder the agents and publishers and journals and editors are so vociferous about hating AI. It's going to send millions of aphids looking for other fruit. Even if mine never finds a market, that will be profund satisfaction indeed. This war has beem going on at least thirty years - about the time Michael estimates nothing has changed. Something is about to change - and its long overdue
I'm curious to know which AI program you used. You said you won't mention which one; can you DM it to me?
I have my own system. My feeling is that AI can find any objective flaw that a literary agent could spot. That said, I think AI feedback has blind spots; my post next week gets into one of them, which is that it has no way to telling if a story is boring—since it has no emotional interiority, and since it's obligated by construction to read every word, it's incapable of getting bored and walking away. And that, unfortunately, is a behavior we'd have to model if we wanted AI feedback to be as useful as a skilled human's.
I hate commercial endorsements of something I know nothing about - and AI falls into that category. You're spot on about AI beating any human reader at their own game, I tried several AI progrtams and like you said, subjective comments run from weird to laughable, so I stopped looking for color. I just asked for organization. I settled on "List all my characters, rank them by importance, and give a paragraph on their most important relationships in the story." It was a task I will need to do anyway when it gets down to promotion, The winner was Perplexity by far, a dozen pages of subtle description and nuance that blew me away. I can safely say that any human editor in the future that wants to read my stuff will be paying me.
Feel free to include me in any beta trials or such - I have a sense youre on to something.....
That's interesting. AI-driven editing has promise; one limitation I see is that you can never turn your critical thinking off. You're always evaluating the AI, while your AI is evaluating your work, and that means there's a circularity. You have to know what use patterns (they exist) cause the AI to start hallucinating or losing its better judgment. You also have to remember that, as amazing as it is that we can now talk to these semi-magical thinking rocks, they are literally just rocks and they have no interiority.
In other words, I think (and hope, because it means I haven't wasted thousand of hours) the fact of needing to be an elite-tier writer to edit to an elite standard still stands. Otherwise, you'll take _all_ the AI's suggestions, and that's a bad idea, because (as we all know) AI writes poorly. You use it to reduce cognitive fatigue, not to replace your own literary judgment. In other words, you can use it as a late-stage editor to flag issues, but half the flags are going to be spurious, and so you should always decide for yourself instead of taking it at its word.
Of course - AI is a tool and not a creator. The creator is you. The problem for most people on this issue is that they are nor creators, not by a stretch. They want AI to close the gap. I don't care how AI works, but I know what my work is trying to achieve. If AI gets it, welcome aboard. If not, va fangu.
One last comment: what's "an elite tier writer"? Where do you draw the line?
Nobel Prize or Writer's Digest Story of the Week? Either way you will experience fatigue, and you are spot on - AI can help. And yes, many of the AI contributions are going to be spurious. AI is a tool, not a creator. The creator is you, whatever your level of creation.
This is something I've been thinking about too recently. My first novel took 15 years from idea to (traditional) publication (I'm not someone who gives up). Second novel has been doing the rounds of publishing houses (via my agent who it took me many years to find) for two years now, getting amazing feedback but with noone willing to commit. I suspect that binary tree you mentioned is causing problems as it doesn't fit neatly on any one shelf.
I do think there are many good books being published, and lots of independent presses are taking punts on more experimental stuff, but the competition is fierce, and there are definitely a lot of safe books being accepted.
I don't know what I'll do with my novel yet, only that I will make a decision on it this year.
Precision take on the inability for mainstream publishing to defend itself against the roaring tide of business consultant interference. The frustration was real and the stakes high for anyone who bucked the bean counters, but try some of us did. I have to hope that slowly, small indie collectives will revitalize real publishing, discover real writing and mine the gems. If I could return to the beginning of my career in publishing, that's where I would be found.