Doctor StrangeCode
Doctor StrangeCode
Or: How I Learned to Stop Coding and Embrace the Vibes
The Writer's Burden
Writing is like bashing your head against a wall because a story burns so deeply in your soul that it must be told. Nobody's there to see it, and feedback is rare. Most people who love your work will never reach out to say, "Hey, this changed me."
So yeah. I'm Sisyphus, rolling my storytelling boulder uphill, forever.
But every so often, you get a lay-up—-something AI can actually help with. Not just organizing drafts, but brainstorming ideas, refining prose, and even acting as an editor that never sleeps.
The Problem
Being a novelist means drowning in revisions, drafts, and organizational chaos. The old-school fix for this was serialization—breaking stories into smaller pieces and sending them out as you write. It's why The Count of Monte Cristo is 1.2 million words long.
But my process? My drafts aren't readable until I've had a chance to revise. That leads to me sending out horrendous filenames like:
"my book chapters 1-3 with edits.pdf"
Gross.
Look to the Past for Solutions
I learned my craft on a dying breed of technology: the mailing list.
By the time I joined in 1999, it was already a relic—a once-thriving Web 1.0 project turned quiet hangout for writers. But it was an incredible place to learn. You could cameo your writing idols, then get feedback from them.
It had a built-in publishing pipeline, too. My first award nod? Came from an editor who liked my work. My first book sale? A member of that list. The advice I pass to new writers today? The old salts hammered it into me back then.
But mailing lists are fraught now. Either they're flooded with spam, or they've been absorbed into corporate SaaS solutions. I've tried sending drafts to friends, but let's be honest—it's a huge ask to say, "Hey, spend hours reading my book that definitely isn't ready for primetime."
What I really needed was a direct line to engaged readers. Usually, that would lead me to Substack, but I'm aligned with Cory Doctorow's approach:
Give away your work to anyone who wants it. Charge for those willing to pay.
Unfortunately, most SaaS solutions are built around monetization, gated content, or newsletters—not actual engagement. None of them quite fit my needs.
Screw It, I'll Write My Own
In the past, I'd have just lived with Substack's limitations—or spent months grinding through Next.js tutorials with nothing but Stack Overflow and a prayer. I'm a decent developer, but I wouldn't trust my code out in public. It's sloppy, over-documented, and probably full of security holes I don't even know exist.
But now? The landscape has changed.
Claude 3.7 and its peers have ushered in a new era of coding assistants. Cursor, in particular, is really good at turning plain English instructions into working, useful code—if you have enough technical intuition to guide it.
And instead of screaming into the void for feedback, I have ChatGPT acting as an infinitely patient reviewer, validating my ideas, and (let's be real) feeding my need for affirmation.
Here's how I use it:
Step 1: Use ChatGPT Like a Rubber Duck
Rubber-duck debugging is a technique where programmers solve problems by explaining them to an inanimate object. Turns out, it works across disciplines:
- Public speakers rehearse in front of pets.
- Actors picture the audience naked.
- Writers? We rant into the void.
But AI makes for a much better void. It's a brainstorming partner, a sounding board, and—when you're working solo—the only "co-worker" you've got.
Step 2: Paper, Plastic, or Robot? Rapid Prototyping for the Modern Age
Since we're vibe-coding, we won't waste time slogging through boilerplate. Instead, we focus on the design and flow first.
Before unleashing something as powerful as Claude 3.7 on an empty codebase, you need a clear vision. Thankfully, that's never been easier.
TikTok has democratized media production. Now, vertical video editing tools like Apple Clips and Google's Photos Video Editor let anyone storyboard ideas. The language of film is at your fingertips.
This approach isn't new—high-functioning teams have been doing it forever. I've seen architecture diagrams built in LEGO. Rapid prototyping lets you refine concepts before they turn into full-blown development nightmares.
Step 3: Rapid Feedback—Use the AI Force, Luke
Getting feedback on a project is like pulling teeth. Everyone loves seeing the finished product—few want to watch how the sausage gets made.
AI removes that bottleneck.
Instead of waiting for human feedback, I can:
- Run instant iterations through Cursor
- Push to a dev server
- Have ChatGPT critique the results
Does it hallucinate? Sure, sometimes. But 95% of the time, the feedback is solid. If those were blackjack odds, I'd play them all day.
The Result
Behold: a brand-new intake funnel for my latest finished manuscript, The Two-Flat Cats!
https://www.thearcades.me/previews/tfcch1
(Insert image here)
Now, the site works for me as a writer:
✔️ Generates previews from markdown text automatically
✔️ Introduces my work to friends and family in a structured way
✔️ Measures engagement with a lead metric (beta reader signups)
✔️ Prevents IP sprawl—no more scattered drafts across hell's half-acre
And the best part?
It took three hours.
From concept to implementation to refinement. Hours.
...How do I do that?
Stay tuned. Over the next few weeks I'm going to break down each step of the above workflow so that anyone, regardless of their coding experience, can reap the benefits of the vibe coding revolution. Click below to join my mailing list, or follow me on linkedin
The Future of Vibe-Coding
The biggest shift isn't just how I write code. It's how I approach creating itself.
By integrating AI into my workflow, I'm:
- More confident tackling ambitious projects.
- More willing to experiment.
- More excited about engaging readers in new ways.
We're at the beginning of a creative renaissance where AI doesn't replace creativity—it enhances it. And now? My storytelling boulder still rolls uphill…
But I've got a lot more control over where it goes.