The snag is that artwork is expensive. That bedevilled the Vulcanverse books, which I thought looked amazing with Mattia Simone's atmospheric filler artwork. Myself, I prefer fillers, like the little vignettes Russ Nicholson does for the Fabled Lands books, but most readers demand drawings that illustrate specific scenes in the book. (It's less bother than reading prose.) Also they'd like a lot more illustrations than the Vulcan Forged company were willing to commission for the gamebooks. So how to pay for all that art?
Perruno suggested turning to AI. I'd been thinking about that, though I'm not sure if it saves a lot of money. You can see from the examples here what Wombo Dream came up with off the top of its artificial head. It could be a lot better if I spent a few months practising. But even if that gave us cover art we could use, there's no way any Fabled Lands book could come out without interior illustrations by Russ; he's an integral part of our creative team. And what about maps? The AI is still a few years off (I'm just guessing) being able to handle those.
Then there's the cost of editing and typesetting. And I haven't even talked about paying the writers. After forty years in this business I'm used to the idea that nobody wants to pay the writer, but I'd love to bring in today's top gamebook talent to work on future Fabled Lands books. Paul Gresty of course, but also people like Jonathan Green or Martin Noutch of Steam Highwayman fame or H L Truslove, author of Alba. I have no idea if they'd even be interested, but I wouldn't insult them by asking until I knew I could write a cheque.
Kickstarter doesn't cut it, as I've explained before. I guess we could wait till AI can do the whole job including the writing, but by then the AIs will be the ones reading the books too.
It's a moral dilemma, isn't it? Use AI and you deprive a cover artist of work. Or do you? If the cost of a cover artist means a book cannot be published, then you are depriving the interior artist (not to mention yourself).
ReplyDeleteIt reminds me a little of the way I used to justify taping my friends' records, back in my impoverished student days. Because of that, I had an interest in music far exceeding that of most of my contemporaries, and because of that I subsequently spent a large sum on music (I have many hundred of CDs, and over 2000 albums in total).
The problem comes with how you deal with the opportunity cost. The IP holders tend to take the view that a copied item is always potentially a missed sale. They could probably argue that even if I didn't buy all of the albums I taped, I would have bought some of them; and that I would have been interested in music even if I had had fewer albums as a student. And they may well be right.
In your case, though, the alternatives are more clear-cut, if no less knotty. And I feel this acutely, as someone who effectively chose to put the stone age equivalent of AI art on the cover of Heart of Ice when I published it, and then proved unable to promote (and therefore sell) the book at all.
Although that edition was actually my favourite Heart of Ice cover...
DeleteI'm more willing to use AI for covers because I don't tend to take much notice of covers, to the extent that when I go to review a book on Goodreads I often have to go back to my copy to check which cover it had. (I can't be alone in this, as it accounts for why publishers have never been too fussed about cover art.)
I mentioned how valuable Russ's creative contributions have been to FL, and I think the interior pictures get more attention than the cover. The filler art in particular helps steep the reader in the world, which is important in open-world books.
So I wouldn't replace Russ with AI -- and it isn't going to be able to handle the maps. But I figure that if I don't have the money for a cover artist in any case, I may as well use AI for the reasons you mention.
I think you could give kickstarter another chance, I know I would back it. I can't imagine the previous problem would repeat itself.
ReplyDeleteIt's going to have to wait till I have a big chunk of spare time on my hands. Kickstarters involve a lot of work and on top of that I already have a Patreon page which is my main priority. But never say never...
DeleteWhy not do a Kickstarter campaign for only a pdf version of the book? This could save printing costs and would also give you some idea about the number of people who would want to buy a printed version of the book, if one came out.
DeleteThat's a good idea. The Walls of Spyte KS showed me that a major part of the work is packaging up and posting out copies, and then dealing with the backers who have moved home and don't know where their copy has got to. A PDF-only campaign would solve that part of the problem. The trouble is, all the backers would end up with only a PDF to show for their support, knowing that the print copies would go on sale the following week on Amazon. So how do we ensure that the backers get something extra?
DeleteThe problem with PDF versions is that they are easy to copy, and therefore more difficult to sell. There actually is a PDF version of Heart of Ice, I remembered when my new edition arrived, which I could have given Dave to use as a reward. But then I thought that the problem with that is that it might actually reduce the demand for the book itself, if released into the wild.
DeleteNeil Gaiman and Amanda Palmer are forever saying that giving your work away only increases demand. But they have the advantage of already being famous, and have no clue what life is like for the majority of writers. If only a few dozen people a year are going to play Heart of Ice, I'd rather those were copies they paid for.
DeleteDoes not having to pay a cover artist change the math even a little bit in terms of being able to publish the books? I just found this Fantasy Map Generator that might be able to help you out: https://azgaar.github.io/Fantasy-Map-Generator/
ReplyDeleteWhat would really change the numbers significantly would be having fewer interior illustrations. Personally I would rather have more fillers, because those set the ambience of the book, rather than illustrations of specific scenes. The fillers also give Russ more freedom to contribute his own ideas, because we don't tell him what to draw.
DeleteThat map generator looks interesting. I need to spend some time with it to see if I can create anything halfway publishable, though.
Then I'd say go with that. Use an AI cover, a map you create and tell Russ to just do filler bits instead of scene illustrations. Anyone who objects can be told that they can either have more Fabled Lands books done that way or they won't get any more books because they're not affordable to produce.
DeleteAfter helping Oliver Johnson create the map for his Yule RPG scenario, I'm thinking it might take me longer than the actual writing.
DeleteThe moral dilemma is less 'depriving a cover artist of work', to me (or artist taking commission for character drawings, or what have you). Most models used by major and minor automatic art generation solutions were made using questionably-sourced data, without necessarily obtaining permission from the creators of the original work. There are ethically-sourced datasets and models out there, but they're capable of much less than models that can't prove they had permission to use all the data they were trained on. This goes for both image generation and text generation.
ReplyDeleteOn top of that, right now it's looking like you won't be able to get copyright on anything generated in this fashion - look at the backtracking over Zarya of the Dawn.
Those are very good points. Human artists have always trained on others' work. I remember as a kid spotting all the poses comics artists swiped from each other (I'm looking at you, Dan Adkins) only now AI can do it faster and hide it better. It opens up the question of what any art would look like if artists only trained on public-domain images... but that's one for the lawyers and philosophers.
DeleteI don't mind about copyright myself. Others are welcome to use any AI images I "create", in fact that's a condition of using Wombo Dream images. I can see how it would be a problem for a graphic novel author, though.
Personally, I never really pay attention to the illustrations inside the books, so I wouldn't mind a book without illustrations if it meant seeing the series getting completed. But that's just me and I've always been an odd duck.
ReplyDeleteI'm with you, Armaan. If I'm reading a comic book then I'm interested in pictures. If it's a novel then I'm happy to use my imagination.
DeleteI'm somewhat torn. Sure I agree that illustrations are secondary for a book, but then again I love looking at Russ' drawings -- especially of the cities and places in fabled lands. They add so much flavor to the descriptions! Also, I started out with the old German editions which have completely different, wavy illustrations. And the new German, Italian and French editions also feature different designs, all of which would be arguments against the necessity of having a certain type of illustrations in the books -- they are great stories regardless.
DeleteMoreover, when I was following the kickstarter campaign, it surprised me how much of the funding went into illustrations.
So all in all, I would say, a few illustrations would be nice. Maybe Russ could make only the map and a black-and-white cover for any upcoming (here's to hoping) book as he did for the Heart of Harkun draft storybook?
Speaking of which: I just googled Heart of Harkun, and there's an amazon audiobook set in the fabled lands in the making!
If I were an FL reader rather than one of the writers, I'd prefer a book with two dozen of Russ's fillers over one with a handful of fillers and lots of illustrations of specific scenes. My favourites are when he depicts a city, as those really set the scene and you come back to them again and again.
DeleteCertainly Russ is a key member of the FL creative team. There must be fans of the art in the 1990s German edition (and I have praised it myself: https://fabledlands.blogspot.com/2011/09/in-lair-of-mantikore.html ) but for me Russ is the definitive visualiser of FL the way Jon Hodgson is the artist for Legend.
An audiobook of The Heart of Hark'un? It's news to me. Now maybe they'll explain what the apostrophe is for!
I rarely if ever leave comments, I'm much more of a lurker, but this discussion and blog post piqued my interest.
ReplyDeleteOn the one hand, I am a trained illustrator and understand the time and effort that goes into creating detailed cover art. But not just that, the time and effort that has gone into learning the craft over many years.
On the other hand, I moved into development and I have a fascination with machine learning. I'm studying machine learning right now, and coding some of the algorithms from scratch and training my own (very small) models based on them.
I think there are a lot of ethical concerns around AI, but it all boils down to change. AI is now inevitable, like nuclear weapons, nobody is going to get rid of it because they don't trust that their "enemies" will get rid of it too. We're already starting to put it into weaponry with the invention of algorithmic warfare.
That inevitability will lead to change, fundamental change to the way our world works. There are some people who are going to be taken advantage of as part of that change. But we want the power that AI brings, it could eventually be like having a creative studio in your pocket. What creator wouldn't find that appealing? But is it okay to have that at the expense of the livelihoods and careers of others? You may not be able to cover art costs, but Netflix can and they have already started trialling anime background creation with AI. However, AI could liberate media from gate-keeping. Anyone eventually may be able to create their own Hollywood blockbuster. Inversely, that would then potentially massively reduce the perceived value of the medium, if anyone can create something in a few minutes that used to take millions of dollars and months or even years to make.
I'm currently playing around with creating AI that can produce retro-style sprites based off of my work only (so I'm training on artwork that I have the copyright for). It's doing okay! I've got a long way to go, it's only just about getting outlining and shading sprites currently. But it's a strange experience watching a system that I created performing a task in seconds, that took me significantly longer.
I think all of this is just going to be a case of coming to terms with the new technology. There are potentially very large problems, such as the pollution of the Internet with AI generated content, and the risk of creating incestuous models as it is hard to filter out that AI generated content completely. The idea is that these models are trained on human output, but if we have to spend time filtering out AI content from the dataset, then it could lead to exponential increases in time and cost for training models.
I haven't come to any conclusions yet, I certainly don't have the answers to these questions. But watching it all pan out will certainly be an interesting experience.
Thanks for weighing in, Adam. I've been thinking along the same lines. I had been in magazine production at college (late 1970s) when we had to get a team of people cutting and pasting for days to put a magazine together. Ten years later I did a Tekumel fanzine and desktop publishing meant I could do the whole job myself in much less time. AI art and writing are to that like a Saturn V to the Kitty Hawk.
DeleteMore than anything else, I would love to complete my Mirabilis comic but I can't possibly afford the art. Could AI help me achieve that? Maybe, and it means that a comic book that otherwise would not have existed can come into the world. But... if it's that easy, who'll want comic books? There will be millions of them. The value may go to zero.
My wife was asking about AI writing last night. She's an author and is interested to know what AI will do to the business. My take is not just that the market could be flooded with AI books, but that people won't even need to buy books. They'll just have the AI create them on the fly. I can say, "I want a story about Sherlock Holmes meeting Mata Hari." Then after a chapter or two, "Mata Hari is boring. Get rid of her and have it turn out to be Irene Adler instead." And so on. Effectively we'll have a storyteller in our pocket, so no need to buy anyone's novel.
Except... AI will be able to replicate any formulaic plot-driven genre. Those formulas specify exactly where plot twists have to occur, so they'll be easy for AI to learn and use. But just as AI artists currently often make a mess of 2D drawings because they have limited experience of the 3D world, AI authors will have no experience of the interior life of human beings. So they can easily emulate Lee Child or Dan Brown, but they won't be able to write anything like Virginia Woolf or William Faulkner. The interiority won't be there.
When creating, AI will spot and use patterns. It's the same approach that makes a lot of Hollywood movies uninteresting. It's the reason DC failed to keep up with Marvel in the 1960s, because they were sticking to the patterns while Stan Lee was happily breaking them. So could this be a moment in history like Delaroche's "painting is dead" pronouncement? Painting didn't die, it just changed to go to places photography couldn't.
The AI genie is out of the bottle now, as you say. We can't just smash the looms and hope the industrial revolution doesn't happen. Writing and art and movie-making will change forever. Maybe amid all that change there will be some true innovations we could never have seen before.
Me and 15 mates of mine decided to have a vote to see which villains from Way of The Tiger we'd like to face reincarnations of in The Isle of A Thousand Spires. Here's the results. What do you think?
ReplyDeleteIn 1st place with 6 votes is Honric
In 2nd place with 4 votes is Mandrake
In 3rd place with 3 votes is Astrograth
In 4th place with 2 votes is Foxglove
I've never played Way of the Tiger, so I don't know those characters, but if Jamie writes FL book 9 then possibly he might include them.
DeleteI've just realized something. This article confirmed that Isle of A Thousand Spires is being written but Isle of A Thousand Spires is Book 9, what about Book 8? which is called The Lone and Level Sands, or have I missed a much earlier article?
DeletePaul Gresty is writing book 8, but there aren't currently any plans for book 9. That could change, and if so you'll read about it right here!
DeleteI realise from the alt-text that the image of Book 9 is "just" a tryout or an idea, but I wanted to highlight the sheer onslaught of emotions I got from just seeing a new possible FL future. I have a ton of positive memories connected with the books and, while I'd love to see it continue in physical, even just knowing it's in your thoughts brings me some happiness.
ReplyDeleteThanks, Arc. As soon as I finish the last Vulcanverse book (which seems to be taking forever) I'm going to check with Paul Gresty how FL8 and the FL Quest version of Castle of Lost Souls are coming along. Hopefully we'll get something published before another year goes by.
DeleteIn my opinion with any other author or authors we'd have no chance of seeing both FL8 and the FL Quest version of Castle of Lost Souls released this year
DeleteBut with people as good as you, Paul and hopefully Jamie we have a chance of seeing both FL8 and the FL Quest version of Castle of Lost Souls released this year
Paul is a superb writer, and he's a lot more prolific than me and Jamie these days. Still, we have to allow for his day job. I hope he is able to finish FL8 soon, and if all goes well we'd like to release it and CoLS as soon as possible.
DeleteMay I suggest that me make a lot of minor changes to Castle of Lost Souls? and the reason I ask is because I know a lot of people that have played the original Castle of Souls enough times that they know it well enough that even with a newly created character it will pose no challenge to them
ReplyDeleteThat's a very good point, JM. I'll mention it to Paul Gresty.
Delete