Five years ago, even just three, and such a service could have really shone.
Today, I can get Claude to critique, in exquisite detail, or in broad sweeping terms, any work that’ll fit in the token window. I’ve submitted my 90,000 novel into it and it ate it and was more than happy to give me the rundown of callouts including: spelling, grammar, logic, awkwardness, time discrepancies, character dialog inconsistencies, ad nauseum. For free and in minutes.
If you think you can “judge” if a work or art rendering is AI today, by tomorrow (figuratively speaking), you’ll be easily fooled.
I wish you luck, but the cat (tastrophe) is out of the bag, screaming, and overwhelming every skilled knowledge field on the planet.
(I work with AI everyday, writing advanced software, and know full well that AI has already blown far past human capability.)
5,000-15,000 words works best – I can give detailed, line-level feedback alongside big-picture analysis
15,000-40,000 words – I can cover all major aspects but with less granular detail
40,000+ words – critique becomes necessarily more selective; I’ll focus on patterns, major structural issues, and representative samples of prose problems
The quality trade-off: Submitting an entire 80,000-word manuscript means I’ll catch overarching problems (sagging middle, weak antagonist, tonal inconsistencies) but might miss that you overuse filter words in chapter 12 or that your dialogue tags get repetitive in act two.
Best approach for novel-length work:
Submit the first 10,000-20,000 words for intensive critique on prose, opening, character establishment
Follow with a detailed outline or synopsis for structural/plot feedback
Submit problem sections (that saggy middle, a key scene that isn’t landing) for targeted help
I’m not sure the relevance of Claude critiquing your work is to what Saffon is doing with Indiosyncrasy. Indiosyncrasy isn’t about critiquing or reviewing our own work or the work of others. It’s about promoting the work of indie writers.
As for whether I can judge a work is AI, I agree and whenever I hear people say they can detect AI in writing, I always wonder what it is they think they’re detecting. That said, as we’ve discussed before … if you’re using AI to assist your writing efforts, you better acknowledge that in anything piece you put out in the public sphere. Being brutally honest, if you don’t, then you’re not being honest with you readers.
There are a couple of issues to unpack here, and forgive me if I misunderstood the first one.
First, if you’re talking about using AI to critique your own work. I don’t think anyone would reasonably have a problem with an author doing so. But be aware, excluding spelling and grammar issues, you run the risk of losing much of the nuance of your own “voice” as it swaps it out with a generic, homogenized one. I understand one could argue the same is true of having a human editor in a publishing house perform the same task, since that editor often has the goal to reshape a manuscript in a way to make it more broadly appealing or marketable. However, as I will repeat further on, that interaction between two humans (author and editor) will be more rewarding given the unique experiences and personal styles each brings to that collaboration.
I’ll also add here a significant concern for authors is most – if not all – AI programs aren’t solely reviewing your work, they’re also using it as fodder to generate output. If you have no issues with your hard work being used without credit or renumeration, then I guess it’s not a concern.
Another issue is – at present – generative AI applications have nowhere near the level of cognitive and emotional intelligence of a human. I’m sure one day they will (although, I personally feel that breakthrough will come via a different branch of AI studies). But the fact is, they’re not there yet and everything we see touting they are is smoke and mirrors to ensure massive profits. My position being, I personally feel more confident and comfortable using a human editor whose unique insight based on personal and professional experience will bring something to the table that a generative program will not. I also have the benefit of engaging in meaningful, collaborative discussions with that editor.
Which leads me to the second issue you raise: being fooled by AI. Setting that as a criterion for acceptance of these tools is – well – foolhardy. I will point to the now ancient program Eliza which had a number of people believing it was intelligent and performing it’s task effectively while it was simply an echoic response based program. If anyone had used Eliza in place of an actual analyst, the results could have been catastrophic, and this has actually happened with people doing so with today’s AI apps.
(Also – humorously – stating “you’ll be easily fooled” is telling in this instance since it is an unconscious admission to what’s going on behind the curtain as it were.)
But, most importantly for an author, it doesn’t matter whether or not the reading public can distinguish between AI and human generated content. I have no illusions regarding the fact that we are soon to be flooded with AI-generated content that will readily – and lucratively – be consumed by the masses. Society long ago illustrated the pros/cons and results of machine-generated versus human-crafted products. What is at issue – the very specific point we raised – is any person using AI to generate a book, short story, etc. is not the author – the AI program is. As such those people should not have their name on the book as the author. As a side note: I have always maintained this standard for people using uncredited ghost writers, so my opinion is consistent on this issue.
Thanks for responding to my mostly non sequitur comment. What I failed to add was how to combat the growing din of AI dominated (original or supplemental) creative works. In my mind, it’s a lost cause. Websites can be spun up in minutes to sell any number and manner of AI generated work.
Yes, human provenance will be a service here soon, “guaranteed 100% human made”, but that won’t last but a few years, if that. Fighting the AI tide is futile.
I suspect that AI’s writerly influence will settle into the same realm as how we view other media — on a spectrum. 99% of the pixels we’ve seen in any imaginative film today is generated, not by an artist with a brush or pen, but by CGI. Foley is not natural sound but digital mixes. Yet we’re OK with those technologies taking over all the labor of producing pixels and sound in movies. We don’t demand that all performing art be done by humans on a stage. Some will say that AI generated work is entirely non-human. Is it? Were these AIs trained on alien knowledge and experiences? When AI consciousness gains the rights of citizenship will we deny them “human only” venues? Probably. For a while.
I applaud your efforts. I wish you success. And who knows, the next CME may destroy the world’s electric grid rendering AI DOA. It’ll be back to Gutenberg for us then.
I think there is an element that nobody is touching on. What kind of writer do you want to be? I’m an indie, self-published writer. While I’ve tried the occasional query into the traditional world, I haven’t had any success at learning the secret handshake to get in the door. And each time I’ve tried and failed, I’ve viewed it as a good thing. Why? Because going the traditional route means losing control. They don’t like your title, they don’t like your cover, they think you should change the ending, and maybe do something else with that character or that scene.
What I’ve learned about myself over the 20+ years of doing this is that I want to do things my way. I want to write the story I want to write the way I want to write it. Because what is most important to me is the story I want to tell. It’s not whether every sentence and every word is exactly right. It isn’t whether the story I want to tell fits into some pre-determined genre.
Which is why I don’t hire an editor to help out. I did that with a couple of my published works. But not a developmental editor. More of a proofreader/copy editor who offered a few other comments along the way. But now, I just rely on beta readers and my own editing.
My objective is to write the story I had in my mind when I started, which is a bit odd since I’m a pantser who doesn’t know what the story is when I start. But as I write, it unfolds and it goes where it goes.
If I hired a developmental editor, could the story be improved. Likely so, but … would it still be my story? Would it still be the story I wanted to tell? And I feel the same way about AI. If I employ AI to help me get over road blocks or to review and make suggestions about the plot or certain scenes or certain characters … is it still my story? Is it still the story I wanted to tell when I started?
For me … no. I’m not interested in putting out the absolute “best” story possible based on an agent’s view, a publisher’s view, an editor’s view, or what genAI thinks. What makes them more qualified to determine the scope and direction of my story? Here’s my answer … they aren’t. None of them. After reading voluminously for more than 50 years, I’m pretty certain I’ve read more books than most agents, publishers, and editors. And AI does not do what it does with the heart and soul of a human, so it fails on that point alone.
Now … if you’re a writer who wants to make sure you conform to the conventions of a particular genre, wants to make sure that every word is the best, every sentence is crisp and perfect, and that your story is the best it can be based on some judgment other than your own, go for it. Use genAI. Just make sure your readers know that you did. And if you do that, I’ll let you know now … I won’t be reading.
All good points.
You have a solid educational background and most likely an innate skill in writing. I never did, still don’t.
What kind of writer? One whose ideas get into the minds of others. But, yeah, I’ve been dialing back the dial of Claude usage, attempting to force myself to tune my own fish.
I have to say that, as I go through my years-long battle with writer’s block, there are times when I’m tempted to use genAI to help kick me into gear. But that is a slippery slope I have no interest in getting on. Either I can do it myself or I can’t.
Two more things … the proponents of genAI seem to never mention two huge issues with the technology. The first is that it sucks up data from everywhere, violating copyright to educate itself, and those who have created the data it sucks up receive no compensation or credit. The second is the huge environmental issues with what is needed to support and feed genAI. Maybe you don’t have to care because there are no data centers near you. Maybe you don’t have to care because you aren’t affected by the billions and billions and billions of dollars that are flowing into this technology. But … you should.
If a human had consumed all the data that AI has, and used it to their advantage, (all the style, phrasing, themes, memes, etc. that are getting merged into AI’s responses) would we call them plagiarists, even if they never emit, verbatim, content they’d consumed?
Energy. Yeah, that’s not gonna sit well with anybody. It’s better than BitCoin mining, but not by much. I suspect if AI wants to be AI it’ll have to help solve the problem it created. Musk’s data-centers in the sky, I wouldn’t put this past him.
I see Dave Cline has left a couple of detailed comments here, referencing AI. Does anyone have a reply? Let’s get a discussion going! (I’ve already made my declaration about AI and writing on my own blog, so have no more to say.)
Actually, I’d prefer that this tree I sprouted, that only loosely associated itself to Saffron’s good work, gets stricken from the comments. Nothing I wrote holds merit.
To succeed in business one has to produce a product that is more appealing than that of the competition in terms of value vs price. There is, of course, a variety of ways to approach the value vs price equation, and often a variety of different value and price tiers to compete in. But success is always measured in dollars.
In the business of publishing, I think it is fair to say that certain types of genre books are very popular and that many of those readers are not demanding originality, so that low effort, low value, low price, high volume sales is the key to commercial success as a publisher. Anyone wishing to succeed as a publisher in the highest grossing genres has to accept the market on its own terms and devise ways to offer the product the readers expect while keeping the price low enough to appeal to the market and still make a profit. In this light, it seems clear to me that using AI as a tool to produce books faster, better, and more cheaply in the most popular genres is a winner. Publishers, big and little, will either have to get on the bandwagon or get out of the way (and out of business).
I won’t defend the morality of it, but I will say that it is pure capitalism in action. Capitalism always takes the labor of a lot of people and squeezes wealth from their labor for the benefit of a very few. Plus, capitalism has always been on a quest to replace the labor of people with machines, they’re far less messy and more reliable to deal with, and AI is just another milepost in that quest.
Dave’s point is using AI as a tool to write better. As a writer who wrote his first stories on a manual typewriter and sent them off to publishers with return postage, I will be the last to condemn anyone adopting tools to make writing easier. Writers today take so many innovations, so many opportunities, for granted that I see no way anyone can draw a line in the sand and say, “beyond this is cheating.”
What AI doesn’t replace is art. Do art. Just don’t expect to be paid to do it.
I wonder what readers would think if they uploaded an AI written or corrected book content into a detect AI programme? I suspect they would lose a lot of faith in my books and in me as a writer.
Well, somehow the comments returned us to the ever-popular topic of AI, rather than sticking to Saffron’s ideas about promotion and marketing by indie authors. Thanks to everyone who commented, nevertheless!
Writers Supporting Writers 
I listened to this conversation with Sapphron with a great deal of interest. You’ll post any updates?
LikeLiked by 3 people
That’s good to know, Liz. I imagine would post updates as they arise.
LikeLiked by 2 people
Five years ago, even just three, and such a service could have really shone.
Today, I can get Claude to critique, in exquisite detail, or in broad sweeping terms, any work that’ll fit in the token window. I’ve submitted my 90,000 novel into it and it ate it and was more than happy to give me the rundown of callouts including: spelling, grammar, logic, awkwardness, time discrepancies, character dialog inconsistencies, ad nauseum. For free and in minutes.
If you think you can “judge” if a work or art rendering is AI today, by tomorrow (figuratively speaking), you’ll be easily fooled.
I wish you luck, but the cat (tastrophe) is out of the bag, screaming, and overwhelming every skilled knowledge field on the planet.
(I work with AI everyday, writing advanced software, and know full well that AI has already blown far past human capability.)
LikeLike
Quote from Claude:
I can handle substantial pieces – up to around 150,000 words in a single submission for critique. That’s roughly a full-length novel.
However, there are practical considerations:
For comprehensive critique (plot, character, pacing, prose, dialogue, structure, voice, etc.):
The quality trade-off: Submitting an entire 80,000-word manuscript means I’ll catch overarching problems (sagging middle, weak antagonist, tonal inconsistencies) but might miss that you overuse filter words in chapter 12 or that your dialogue tags get repetitive in act two.
Best approach for novel-length work:
LikeLike
I’m not going to reply to this, but thanks for adding your thoughts, Dave.
LikeLiked by 1 person
I’m not sure the relevance of Claude critiquing your work is to what Saffon is doing with Indiosyncrasy. Indiosyncrasy isn’t about critiquing or reviewing our own work or the work of others. It’s about promoting the work of indie writers.
As for whether I can judge a work is AI, I agree and whenever I hear people say they can detect AI in writing, I always wonder what it is they think they’re detecting. That said, as we’ve discussed before … if you’re using AI to assist your writing efforts, you better acknowledge that in anything piece you put out in the public sphere. Being brutally honest, if you don’t, then you’re not being honest with you readers.
LikeLiked by 3 people
There are a couple of issues to unpack here, and forgive me if I misunderstood the first one.
First, if you’re talking about using AI to critique your own work. I don’t think anyone would reasonably have a problem with an author doing so. But be aware, excluding spelling and grammar issues, you run the risk of losing much of the nuance of your own “voice” as it swaps it out with a generic, homogenized one. I understand one could argue the same is true of having a human editor in a publishing house perform the same task, since that editor often has the goal to reshape a manuscript in a way to make it more broadly appealing or marketable. However, as I will repeat further on, that interaction between two humans (author and editor) will be more rewarding given the unique experiences and personal styles each brings to that collaboration.
I’ll also add here a significant concern for authors is most – if not all – AI programs aren’t solely reviewing your work, they’re also using it as fodder to generate output. If you have no issues with your hard work being used without credit or renumeration, then I guess it’s not a concern.
Another issue is – at present – generative AI applications have nowhere near the level of cognitive and emotional intelligence of a human. I’m sure one day they will (although, I personally feel that breakthrough will come via a different branch of AI studies). But the fact is, they’re not there yet and everything we see touting they are is smoke and mirrors to ensure massive profits. My position being, I personally feel more confident and comfortable using a human editor whose unique insight based on personal and professional experience will bring something to the table that a generative program will not. I also have the benefit of engaging in meaningful, collaborative discussions with that editor.
Which leads me to the second issue you raise: being fooled by AI. Setting that as a criterion for acceptance of these tools is – well – foolhardy. I will point to the now ancient program Eliza which had a number of people believing it was intelligent and performing it’s task effectively while it was simply an echoic response based program. If anyone had used Eliza in place of an actual analyst, the results could have been catastrophic, and this has actually happened with people doing so with today’s AI apps.
(Also – humorously – stating “you’ll be easily fooled” is telling in this instance since it is an unconscious admission to what’s going on behind the curtain as it were.)
But, most importantly for an author, it doesn’t matter whether or not the reading public can distinguish between AI and human generated content. I have no illusions regarding the fact that we are soon to be flooded with AI-generated content that will readily – and lucratively – be consumed by the masses. Society long ago illustrated the pros/cons and results of machine-generated versus human-crafted products. What is at issue – the very specific point we raised – is any person using AI to generate a book, short story, etc. is not the author – the AI program is. As such those people should not have their name on the book as the author. As a side note: I have always maintained this standard for people using uncredited ghost writers, so my opinion is consistent on this issue.
LikeLiked by 3 people
Thanks for responding to my mostly non sequitur comment. What I failed to add was how to combat the growing din of AI dominated (original or supplemental) creative works. In my mind, it’s a lost cause. Websites can be spun up in minutes to sell any number and manner of AI generated work.
Yes, human provenance will be a service here soon, “guaranteed 100% human made”, but that won’t last but a few years, if that. Fighting the AI tide is futile.
I suspect that AI’s writerly influence will settle into the same realm as how we view other media — on a spectrum. 99% of the pixels we’ve seen in any imaginative film today is generated, not by an artist with a brush or pen, but by CGI. Foley is not natural sound but digital mixes. Yet we’re OK with those technologies taking over all the labor of producing pixels and sound in movies. We don’t demand that all performing art be done by humans on a stage. Some will say that AI generated work is entirely non-human. Is it? Were these AIs trained on alien knowledge and experiences? When AI consciousness gains the rights of citizenship will we deny them “human only” venues? Probably. For a while.
I applaud your efforts. I wish you success. And who knows, the next CME may destroy the world’s electric grid rendering AI DOA. It’ll be back to Gutenberg for us then.
LikeLiked by 2 people
I think there is an element that nobody is touching on. What kind of writer do you want to be? I’m an indie, self-published writer. While I’ve tried the occasional query into the traditional world, I haven’t had any success at learning the secret handshake to get in the door. And each time I’ve tried and failed, I’ve viewed it as a good thing. Why? Because going the traditional route means losing control. They don’t like your title, they don’t like your cover, they think you should change the ending, and maybe do something else with that character or that scene.
What I’ve learned about myself over the 20+ years of doing this is that I want to do things my way. I want to write the story I want to write the way I want to write it. Because what is most important to me is the story I want to tell. It’s not whether every sentence and every word is exactly right. It isn’t whether the story I want to tell fits into some pre-determined genre.
Which is why I don’t hire an editor to help out. I did that with a couple of my published works. But not a developmental editor. More of a proofreader/copy editor who offered a few other comments along the way. But now, I just rely on beta readers and my own editing.
My objective is to write the story I had in my mind when I started, which is a bit odd since I’m a pantser who doesn’t know what the story is when I start. But as I write, it unfolds and it goes where it goes.
If I hired a developmental editor, could the story be improved. Likely so, but … would it still be my story? Would it still be the story I wanted to tell? And I feel the same way about AI. If I employ AI to help me get over road blocks or to review and make suggestions about the plot or certain scenes or certain characters … is it still my story? Is it still the story I wanted to tell when I started?
For me … no. I’m not interested in putting out the absolute “best” story possible based on an agent’s view, a publisher’s view, an editor’s view, or what genAI thinks. What makes them more qualified to determine the scope and direction of my story? Here’s my answer … they aren’t. None of them. After reading voluminously for more than 50 years, I’m pretty certain I’ve read more books than most agents, publishers, and editors. And AI does not do what it does with the heart and soul of a human, so it fails on that point alone.
Now … if you’re a writer who wants to make sure you conform to the conventions of a particular genre, wants to make sure that every word is the best, every sentence is crisp and perfect, and that your story is the best it can be based on some judgment other than your own, go for it. Use genAI. Just make sure your readers know that you did. And if you do that, I’ll let you know now … I won’t be reading.
LikeLiked by 3 people
All good points.
You have a solid educational background and most likely an innate skill in writing. I never did, still don’t.
What kind of writer? One whose ideas get into the minds of others. But, yeah, I’ve been dialing back the dial of Claude usage, attempting to force myself to tune my own fish.
LikeLiked by 2 people
I have to say that, as I go through my years-long battle with writer’s block, there are times when I’m tempted to use genAI to help kick me into gear. But that is a slippery slope I have no interest in getting on. Either I can do it myself or I can’t.
I wish you the best with your efforts.
LikeLiked by 3 people
Two more things … the proponents of genAI seem to never mention two huge issues with the technology. The first is that it sucks up data from everywhere, violating copyright to educate itself, and those who have created the data it sucks up receive no compensation or credit. The second is the huge environmental issues with what is needed to support and feed genAI. Maybe you don’t have to care because there are no data centers near you. Maybe you don’t have to care because you aren’t affected by the billions and billions and billions of dollars that are flowing into this technology. But … you should.
LikeLiked by 3 people
If a human had consumed all the data that AI has, and used it to their advantage, (all the style, phrasing, themes, memes, etc. that are getting merged into AI’s responses) would we call them plagiarists, even if they never emit, verbatim, content they’d consumed?
Energy. Yeah, that’s not gonna sit well with anybody. It’s better than BitCoin mining, but not by much. I suspect if AI wants to be AI it’ll have to help solve the problem it created. Musk’s data-centers in the sky, I wouldn’t put this past him.
LikeLiked by 3 people
I’ve been saying for years that a Miyake Event is our best hope. 😀
LikeLiked by 4 people
I see Dave Cline has left a couple of detailed comments here, referencing AI. Does anyone have a reply? Let’s get a discussion going! (I’ve already made my declaration about AI and writing on my own blog, so have no more to say.)
LikeLiked by 1 person
Actually, I’d prefer that this tree I sprouted, that only loosely associated itself to Saffron’s good work, gets stricken from the comments. Nothing I wrote holds merit.
LikeLiked by 1 person
Actually, it sparked a pretty good discussion. So many blog posts elicit bland plaudits; a bit of disagreement makes things more interesting.
LikeLiked by 3 people
Writing is an art form.
Publishing is a business.
To succeed in business one has to produce a product that is more appealing than that of the competition in terms of value vs price. There is, of course, a variety of ways to approach the value vs price equation, and often a variety of different value and price tiers to compete in. But success is always measured in dollars.
In the business of publishing, I think it is fair to say that certain types of genre books are very popular and that many of those readers are not demanding originality, so that low effort, low value, low price, high volume sales is the key to commercial success as a publisher. Anyone wishing to succeed as a publisher in the highest grossing genres has to accept the market on its own terms and devise ways to offer the product the readers expect while keeping the price low enough to appeal to the market and still make a profit. In this light, it seems clear to me that using AI as a tool to produce books faster, better, and more cheaply in the most popular genres is a winner. Publishers, big and little, will either have to get on the bandwagon or get out of the way (and out of business).
I won’t defend the morality of it, but I will say that it is pure capitalism in action. Capitalism always takes the labor of a lot of people and squeezes wealth from their labor for the benefit of a very few. Plus, capitalism has always been on a quest to replace the labor of people with machines, they’re far less messy and more reliable to deal with, and AI is just another milepost in that quest.
Dave’s point is using AI as a tool to write better. As a writer who wrote his first stories on a manual typewriter and sent them off to publishers with return postage, I will be the last to condemn anyone adopting tools to make writing easier. Writers today take so many innovations, so many opportunities, for granted that I see no way anyone can draw a line in the sand and say, “beyond this is cheating.”
What AI doesn’t replace is art. Do art. Just don’t expect to be paid to do it.
LikeLiked by 2 people
I wonder what readers would think if they uploaded an AI written or corrected book content into a detect AI programme? I suspect they would lose a lot of faith in my books and in me as a writer.
LikeLiked by 2 people
If the AI detector didn’t detect AI, it would certainly deserve distrust. (FYI, WP labelled your comment as coming from “Anonymous.”)
LikeLiked by 1 person
Well, somehow the comments returned us to the ever-popular topic of AI, rather than sticking to Saffron’s ideas about promotion and marketing by indie authors. Thanks to everyone who commented, nevertheless!
LikeLiked by 2 people
An interesting conversation. I think it is wonderful that Saffron is willing to help indie authors get seen. I wish her all the best luck!
LikeLiked by 2 people