Subtxt isn’t just AI—it’s trained thinking.
For a long time, the idea of a “moat”—a unique, defensible edge—was the gold standard for businesses and creators alike. If you could build something no one else could replicate, you were set. But in a post-AI world, there is no moat. At least, not in the way we used to think about it.
With the advent of AGI (Artificial General Intelligence) and ASI (Artificial Superintelligence), even the most complex narrative structures—like those found in Dramatica Theory—will be easily justified and executed by machine intelligence. The act of generating a Storyform will become trivial. The real challenge? How you think about and interpret that Storyform.
And that’s where Subtxt is heading.
If you’ve ever attended a Dramatica User Group Meeting over the past 30 years, you know that even among the most seasoned theorists, interpretations of a story's true Storyform can differ. The underlying reasoning behind a Storyform isn’t always the same—some understandings are more accurate or more insightful than others.
That’s exactly where the next frontier of AI-powered storytelling is headed. It won’t just be about generating a Storyform—it’ll be about who’s thinking behind it.
When we talk about AI-assisted storytelling, the future won’t be some monolithic system spitting out the perfect Storyform every time. Instead, we’ll see individualized implementations of narrative models—some better, some worse, some simply different. One person’s RFT-informed (Reinforcement Fine-Tuning) take on Dramatica will differ from another’s. And as these approaches develop, so will a measurable distinction between the best thinkers in narrative structure and the ones who are simply making things “workable.”
Right now, you could feed a structured Storyform into an LLM (Large Language Model) and get something that looks intelligent. But put that same output in front of an expert in Dramatica Theory? They’ll tell you that 90% of it is wrong.
And that’s fine—for many, a workable story structure is all they need. That’s why frameworks like Harmon’s Story Circle and The Hero’s Journey persist. They’re accessible. They’re “good enough.”
But for those looking for precision—for something that isn’t just passable but actually correct—the difference will be in the execution. It will be in how an AI system thinks through the narrative process—not just the fact that it can generate a Storyform.
And that’s where Subtxt stands out.
With Subtxt, every single response, every single model adjustment, is carefully measured, evaluated, and fine-tuned on a daily basis.
When RFT-based reasoning is fully implemented, that process will become even more refined. We’ll be rewarding the AI for thinking like we do here at Narrative First—which means the difference between someone who puts in the effort to shape the system and someone who doesn’t will become massive. The quality of narrative insight won’t come from AI’s ability to generate a Storyform (which will be easy). It will come from who has taken the time to teach it how to think.
Writers and directors don’t operate with moats. They don’t spend time worrying about defensibility—they focus on being creative. They work on expressing their unique perspective of the world and honing their craft in a way that resonates with others.
The same applies here.
The future of AI-driven storytelling isn’t about building barriers. It’s about pushing deeper into the thinking behind a narrative—refining, improving, and evolving the way we construct and interact with stories.
And that’s what will set the best apart.
Moats are for those trying to keep others out.
We’re here to bring more people in.
Don't miss out on the latest in narrative theory and storytelling with artificial intelligence. Subscribe to the Narrative First newsletter below and receive a link to download the 20-page e-book, Never Trust a Hero.