Let’s go back deep in time, remembering that I got to see the era of personal computers begin. I ask you to picture an artist-writer-storyteller, a girl who mostly drew unicorns and pretty aliens, and then also put her in front of an Apple IIe facing a command line and hearing ‘You can do anything you want.’

What did I want? I didn’t know. But we had a computer class in school (which was very avant garde at the time!) and that was how I learned about the BASIC programming language, and ended up creating my first very simple programs… the kind of beginner tasks that cause a computer to echo back ‘hello world’. From there, I went straight to “I want to make a choose-your-own-adventure game,” which I did, and based it on Star Trek. You were commanding a starship, you ran into trouble, you had to decide what to do, and then eventually your ship blew up. (All choices led to the ship blowing up. I thought that was funny. No one who played the game liked this storytelling choice; I thought they needed to lighten up and embrace the boom.)

Writing games like this was pretty easy: it was basically a tree of “If this choice, show this text. Prompt for next choice. Then if this choice, then this text…” all the way down the chain to the possible endings.

It seems weird to me to realize that I don’t call myself a coder or a gamer, and yet at 12? Years old? About? I was excited about using a computer to tell an interactive story. All of which is to say: I guess I’ve always sort of coded. But I also always sort of coded as a means, and the end was telling a story. I was never fascinated by computers qua computers; there are coders who code because they like the puzzle of putting together instructions, or because they’re fascinated by the computer itself, or the process of coding massages their brain in a specific way that I can only imagine. To me it’s just a tool, and an often tedious one… and the scope of my game design dreams (or web design dreams) always meant that I would have to spend too much time on the annoying parts to get to the end I wanted (the game, the story, the website).

Time went by, and I got jobs where I had to write websites that had database interactions, and I did this with giant tomes about coding resting on my lap, or with my (primitive back then) web browser open on newsgroups about coding where I could copy code snippets into my terminal window and adapt them to my use case. I often asked for help; I learned from the people helping me; I became fairly adept at reading code and understanding what it was doing. But I reused code constantly, and I never loved writing it. I loved succeeding when something worked the way it was supposed to, but this was not my beautiful house, my beautiful wife, any of it. I was a storyteller and an artist; I do not have coder brain. I have wordcel brain. (Yes, I own it.)

This long story I tell you so that you understand that when AI vibecoding became a thing, I was all over experimenting with it. And my delight was boundless. At last, a tool that removed the tedious parts and left me only with the parts I enjoyed and am good at! For this reason, AI is my foremost tool in the gamedev process. Because of AI, I can concentrate on drawing and writing stuff for games and websites and not have to do the part I don’t like and am not natively interested in.

I understand there’s a lot of controversy about AI use. I have fans who won’t play my games or use my quizzes because I used AI to code them. But I remember the me who was cutting and pasting newsgroup code snippets into her own terminal windows, and the me who was laboriously typing sample programs into those windows from books, and the me who was already leaning on the enormous base of knowledge compiled by people, sharing freely on the internet, and all I think is that AI makes it faster, and faster makes it possible. I made my peace with that, but I recognize other people will have other opinions.

There’s a lot of confusion around how you use AI to code; lots of “it forgets things” and “it hallucinates” and “you can only code to a certain complexity level and then it fails”, etc. Here’s how I do it:

IMAGINE: I think about a feature I want the game to have. I imagine how it works in my head, what delights me about it, how I interact with it.

PLAN: I have a design session with the AI where I say, “This is what I want… what are my options for accomplishing it?” I have an instruction there that says ‘I have some coding background but I want to learn and particularly best practices and security-conscious design. Please educate me!” So we talk, and I learn. We go through a few ways to make what I want to happen work, and then – this is important – we create two documents. One is an overview of the feature and what I want it to do; the other is an implementation plan.

IMPLEMENT: We use the implementation plan to start work. I like to work from the invisible-to-the-user-backend outward to the UI. Each step of the implementation plan involves a checkpoint and a test. We put in the first layer. We test to make sure it works. Then we proceed to the next, etc. Only when I’m sure that all the invisible pieces work do I start the last piece, the UI that makes it visible to the player.

DOCUMENT: After the feature is implemented, we update the design document to reflect what it does, how it works and what it touches, and what my plans are for it for the future now that it’s working.

Rinse and repeat.

In the course of seven months I created over 168 pieces of documentation. That’s only counting the files I kept, not the ones I discarded because they were no longer up-to-date, or the ones I consolidated. I often did “doc review” days, and those days I would often do things like turn 8 small implementation documents into a single overall feature document. So that’s just 168 that I have kept, currently, after months of iterating.

I do this not because I don’t trust the AI, but because my own memory is poor. I know I’m not going to be able to hold the codebase in my head. I take good notes because good notes are the only way you keep yourself aligned to your goal. Breaking a big project into bite-sized pieces and then doing each of those, one by one, is the only way I have ever accomplished anything, as scatterbrained as I am. And as it turns out, being scatterbrained is excellent training for the AI age, because all the coping mechanisms I developed to become a modestly functional adult are exactly what I need to stay focused and productive with tools that accelerate you in whatever direction you’re already facing. If you’re disorganized, unmotivated, or easily distracted, you can become more so instantly and at scale. But if you’re even a little bit ready to compensate for those problems with habits and hacks, you’ll go far with the new tools. It really is all up to you.

At the point at which I began this, I was using primarily Grok and Claude, the chatbot versions. There were no CLI/cowork tools at that time. I’ll talk in more detail about how that worked in forthcoming entries.

AI, then, is what I use to code. I’ll bow briefly to the other tool I’m using, which is Clip Studio Paint on my ancient Microsoft Surface, with a cheap off-brand stylus. I’ll say this immediately: I hate drawing digitally. Given the choice between a pencil and a stylus I’ll take the pencil every day. But for a game that has paper doll mechanics, it’s nonsensical to draw and scan the amount of art assets I need to do. This is the one place where the digital art tool is legitimately the right choice for the job, no matter my personal feelings about it.

The Surface is the only windows machine left in my possession, and is too old now for significant upgrades. I’m going to continue on with the software/OS revisions I have for as long as possible! But I am dreading the day I need to do a cutover. That is a problem for Future Jaguar, though which brings me to an important philosophical belief of mine… that you just have to do the thing. The time is never perfect. The future is full of pitfalls. You will never learn enough to do the thing without doing the thing: the only way out is through. So the faster you start traveling toward the goal, the better.

So those are my tools! Next time, we’ll talk about my initial foray into the Jokka game, the prototype! Unless you have questions about my process, and want me to go deeper into it. I’d be happy to devote future entries to things you’d like to know… just let me know.