This is my wishlist for that next iteration of Scrivener that’s not out yet. Whatever it is that comes out, it needs to work with AI seamlessly. That probably means markdown, something similar, some other approach, or it must be built-in.
First, Scrivener is undoubtedly the best story writing software to have ever been created – I’ve tried a few others and they don’t come close. It’s really good at what it does. It’s been the only software I’ve used to create stories after purchasing it, and I’ve loved every minute of it.
Then things changed, and AI became a real thing. The beginning of this year was the turning point, and since then, things have evolved even more quickly. It’s no longer “AI is coming” or “AI is failing”, it already happened. My own experience is that across multiple industries, AI use is now a core part of the job description. In some cases, if it’s not used, it becomes a problem related to efficiency and production. Having used AI myself quite a bit, I can attest to its benefits. Naturally, some people will use it responsibly, others won’t.
I use AI for planning, and keeping track of, well, everything: character bio’s, plot, timeline, chapters, acts, etc… It will never write a single line of the story I’m working on, but it will analyze, and validate, what I have and provide valuable feedback. For $200 a year, I have access to a very capable assistant, one that still makes mistakes. This assistant can keep up with the rate of ideas that I feed it, and help organize and update everything.
Up until this point I haven’t really used AI in this capacity, but now I have. In the past, I tried to be a plotter, and failed every time. The reasons being exactly what AI has done for me above. When I need to make changes to the plan, AI speeds that entire process considerably, and makes things so much easier. Now I focus almost entirely on ideas, and less on book-keeping. I’m no longer trying to use a workflow that works better for someone else, it’s all my workflow.
The consequence of this, and I hate to say it, is that I’ve had to stop using Scrivener. Now mostly working with markdown files, and I dislike markdown; however, integrating with AI far outweighs my preference against it. In two days of non-stop work I’ve completely planned, outlined, created character sheets, created a timeline, etc, and am now ready to start writing. If I attempted to do the exact same thing, all by hand, I wouldn’t even be half-way done.
My workflow now is:
- plan story in markdown files
- write story in Ulysses
Ulysses. Why use the editor that I don’t really like using? Well, it’s starting to grow on me out of necessity. I created an AI skill that will parse through their XML files so the AI can read them as plain text and provide important feedback. That same information is then cross-referenced with my outline, and everything else I’ve done. This is extremely powerful.
What did I lose moving to this workflow? Obviously Scrivener, but I feel like its time to move on to a different approach. While RTF is a good choice, it makes my workflow harder to deal with. I could copy/paste everything, or, I could save myself the step, focus on what’s important, and have AI work natively with what I have. This also saves money when it comes to token usage. I also lost the ability to work wth ProWritingAid, but their Anywhere app is a complete struggle to work with, and I’m not going to trap myself into requiring an online connection to do anything.
EDIT:
just realized that there is a textutil on mac to convert rtf to text files. I don’t think I need Ulysses, and won’t use it, if this works out. Still, it would be much better if it were easier to have AI to modify files directly in the project. Right now, no matter what the workflow, for me, AI never directly touches my story, but if it could, that would mean everything could be in one place.