Tess Rafferty
I have written and produced on over 500 episodes of television. I have also written a series of novels. And as a reader, I devour murder mysteries like they’re pie the day after Thanksgiving.
Two years ago I started learning Ai filmmaking which made me curious about other emerging technologies and how they were going to shape the future of how we enjoy entertainment.
And then year after that, I put on a friend’s igoggles (not their real name) and I saw the future. I didn’t understand why all of our TV wasn’t in VR right now. I saw so much potential, especially for my own book series which had been under option for years but never made it to the flat screen. When I tried Apple Vision Pro I thought, “My books weren’t made for 2D. They were made for this.”
Similarly, when I joined the Future of Text symposium and learned about their work, I saw the great potential there, only this time not as a TV viewer and producer – but as a writer. The features being developed here can give me the support I’m looking for as a creator.
In my Kat Kelly Mystery series, I have written 4 full-length novels and a holiday novella. That’s a lot to keep track of. I try to make beat sheets as I go, and keep track of character names in each book, but it becomes a lot to deal with when you’re also trying to write and publish a murder mystery. And yet, when crafting a murder mystery, sometimes it’s these exact details that you most need to keep track of.
For those of you who have never written a murder mystery series – or any series – you want to make sure you don’t repeat character names. (I go one step further: I don’t like to use names with the same first initial in any individual book. I think it’s too much for the reader to keep straight, especially when they’re holding so many other details in their head. I’m looking at you Stieg Larsson.) A book is a lot of words: you want to be able to remind yourself of the important things that happened – and the order in which they happened – without having to keep rereading the whole manuscript. Plus, mystery novels are like a magic trick. You’re laying out not only clues, but also distractions, and you have to do it in such ratios as to always keep the answer just out of reach from your reader. You want them to feel surprise at the reveal – but also like they could have figured it out if only they had just reached a little further. Like that crossword clue you just gave up on because your coffee was getting cold, but you really should have known the answer. It can’t be obvious – but it also can’t be so ridiculous as to not be believable.
When I sat down to begin my 6th book this year, I realized I needed some help beyond my various ad hoc documents. I needed a way to take the sum of all of my books and both organize and analyze it.
Some of what I needed was basic indexes: character names, page appearances and bios; locations used; murder methodologies and motives. Some of it was more complex: I wanted to analyze the word count of each book and see if there was a basic formula for when the first and subsequent murders happened. How much time was spent introducing characters and establishing the world? When did inciting incidents appear? Which books were slower paced? In essence, what, if any, was the Kat Kelly formula?
Also, security was a concern. While my books are out there on the internet for anyone to purchase and download, I didn’t want to upload these documents and make it that much easier for a company to train on them. I needed a project space that would be four walled.
With all I knew about LLMs and various related apps, it seemed it should be easy enough to upload the PDFs of my books and get answers to these and other questions. It was not. I got some good suggestions and tried a couple of different models, but simple things were an issue. Some apps weren’t made to take in nearly half a million words at a time. And despite Word Count being one of our oldest word processing tools, LLMs have a hell of a time doing an actual word count themselves.
In the Future of Text symposium I see a near-future where I can do all of this and more. In addition to auto-indexing details like characters, generating beat sheets and comparing and contrasting book structures, I see an interface where we can “Seed Map” our plots and mysteries. As I write a mystery, I often “seed” ideas: what I call introducing hints and allusions to plot elements that will be developed throughout the story. Sometimes these elements are motives, sometimes they’re red herrings, sometimes they’re just character motivation. Often, I put a seed in without knowing exactly what it’s going to become; other times I leave a space for a seed to go, and then fill it in as more of the plot develops and I learn what’s needed to build the story.
In an XR space, we can designate these SEEDS, allowing us to easily toggle back to them when needed, while also seeing where in the story they fall in a visual outline or map that can be drawn in a number of different ways to best help the writer view their work as a whole.
And that’s just the beginning. An XR text system could allow writers to interface with any number of generative Ai APIs, creating character prototypes, location floor plans, or even whole worlds. This would allow creators to more accurately see what they’re describing and hone what their vision of it is. Likewise, any XR environment with a browser could give a writer a literal window onto the world to go to the exact location in West Africa or Antarctica they’re writing about – as they’re writing about it. And we’d have the opportunity to be part of virtual writers’ groups, where we can get together to read each others’ work and discuss constructive ideas.
There is similar potential for readers here, too, building on some of the hypertext interface of Kindle or other two dimensional reading experiences. In an XR environment we can HEAR songs that are referenced or the correct pronunciation of foreign phrases. We can SEE locations that are mentioned or LEARN more about the history informing the plot. And we can have access to WIKI bios and plot summaries that allow us to enjoy a series while keeping the details straight.
Much like we say we need a human in the loop with Ai, I believe we need creatives in the loop with these exciting technologies. People who are passionate about both sides and who can be the conduit and translator between them. I’m hoping to be one of those people. My passions as a writer, a reader, and a TV creator and viewer have given me a vision for both the Future of Text – and the Future of Media.