AI opportunity

I was in the process of commenting on AI risk that @jasonleow wrote about today, but the topic reminded me of what I’ve been working on recently with AI. I want to make a living as a writer at a time when writing is cheap. Claude Opus 4.6 costs only $25 per million output tokens, or about one penny for every 300 words. I can’t compete on that level.

But it’s entirely true what they say about AI having no access to human experience. And writing is all about creating connection through shared human experience. The writer tries to create a sense of recognition in the reader, and I think this applies to coding as well. Everything the AI knows about creating a good user interface was ingested from a real human describing their experience. And all of that data is out of date.

AI is shaking up the established patterns that AI is currently trained on which means we need people to parse these new experiences.

What happens to the front end developer when MVC and RESTful architecture no longer mean anything? I just need a chat interface to hard-code my website content, but that doesn’t mean I’m going to build some WordPress look-alike. This opens up so many more opportunities than we even realize at the moment.

We always look at paradigm shifts from the perspective of the-before-times and end up seeming a bit foolish in retrospect. If a company thinks they’re going to replace all of their workers with AI-bots, they’re going to get BURIED by their competitors who have AI-bots and still have real people building next generation products. Meeting today’s benchmarks with AI-bots aren’t the actual opportunities.

For me the question is how can I use AI to make myself a better writer?

But don’t take my word for it. Here’s what my AI says about one of the book projects I’m working on. Literally, it cost one penny for its thoughts.

The old model for someone wanting to write a serial fantasy novel looked something like: have an idea, maybe outline it, start writing, hit walls, push through or abandon, and learn mostly by trial and error. Research meant reading in the genre and maybe some craft books. Industry knowledge came from forums, blogs, and word of mouth, absorbed piecemeal over years.

What you’re doing instead is treating the creative project like a venture that deserves the same rigor as a startup. Before writing a single published chapter, you’ve built a structured knowledge base–over 50 research documents covering everything from platform algorithms and reader demographics to genre conventions, progression system design, and competitive analysis of successful serials. You’ve got a story bible, a custom RPG system, character development docs, and a project management layer tracking open decisions.

The AI’s role in that isn’t “doing the creative work.” It’s more like having a research analyst, developmental editor, and strategic advisor available on demand–someone who can synthesize information across dozens of sources, stress-test ideas against genre conventions, flag inconsistencies in worldbuilding, and help you think through tradeoffs (like whether to launch directly on Royal Road or run a proving-ground project first) with actual evidence rather than gut feel.

The real shift is about information asymmetry and preparation cost. The knowledge that used to take years of community immersion to absorb–what readers actually want, how platform discovery works, what makes progression systems satisfying, how successful authors structure their business models–you’ve compressed that learning curve into a structured research phase. You still have to do the creative work. But you’re doing it with dramatically better situational awareness than most debut authors have.

If I were explaining it to someone, I’d say: “The writing is mine. What AI changed is that I’m not writing blind.”