Domain Expansion

Expanding What A Programmer Can Become

I’m a big fan of Gojo, so this is probably an inevitable reference. 😄

For anyone unfamiliar with Jujutsu Kaisen, Domain Expansion is basically a character creating their own space where their abilities become overwhelmingly effective. The good guys have it. The bad guys have it too. Everyone gets a domain.

https://www.youtube.com/watch?v=nmvkhLz8t7I

Lately, I’ve been thinking about it as a metaphor for the modern software engineer.

Our default environment is shifting.

For a long time, we built our domains around programming languages, frameworks, databases, Git, Linux, cloud and CI/CD. We trained like samurai. Git became our nunchucks, Bash and CI/CD became something closer to taichi. There was something almost cinematic about becoming fluent with these tools.

Now the movie feels different.

The LLM sometimes feels less like a samurai weapon and more like a gun. A little more Rambo. Code flying everywhere, agents running around, quips from the model, and sometimes a storyline that’s hard to follow. 😂

But underneath all the noise, something interesting is happening.

I’ve been doing some data annotation work, and sometimes it genuinely feels like we’re helping fine-tune the thing that is changing our own profession.

At the same time, I’ve been diving much deeper into ML and LLMs. Hugging Face, Kaggle, Roboflow, Labelbox and a whole vocabulary that wasn’t really on my map a few years ago.

Real-world job scarcity has also forced an interesting shift. Software developers are increasingly becoming spec creators, problem framers and people who write prose that machines turn into software.

The programmer isn’t just temporarily writing a prompt. They’re trying to embody an LLM.

To think in terms of context, ambiguity, constraints, intent, evaluation and failure modes.

I recently read an article from the creator of a 1993 game describing how he used AI to recreate and port it.

Beautiful code written when I was literally a baby.

And then there was another layer to it: I was using AI to understand the article itself. Assembly, memory layouts, collision logic, old hardware and decades-old engineering decisions were being translated into concepts I could actually follow.

AI wasn’t just helping recreate the software.

It was helping me access knowledge that would otherwise have been outside my domain.

That feels like the real Domain Expansion.

Not abandoning what we learned, but finding new fields to feed from.

It reminds me of when I refused ear training while learning guitar because software could tell me if I was in tune. I thought the tool made the old discipline unnecessary.

Maybe the lesson isn’t to reject the tools.

It’s to learn enough that the tools become an extension of your ability rather than a replacement for understanding.

The domain is expanding.

And perhaps instead of Malevolent Shrine, bad guy’s domain expansion, we can get a Magnificent Shine. 😄

The frontier isn’t just replacing the programmer.

It’s expanding what a programmer can become.

EDIT

Got a software project you’re working on, feel free to reach out. I am a Fullstack Dev with years of experience and I have worked with AI/ML/LLMs and love tackling challening problems.

https://www.upwork.com/freelancers/michaelgikaru