I've come to a realization.
I have to use AI as part of my workforce.
Not because I can't code.
Not because I don't understand what I'm building.
And certainly not because I want to surrender my projects to a machine.
I need to make progress.
I have a lot of things I want to build, and unfortunately, I still only have 24 hours in a day.
KiwiEngine.
Blackwater Sound.
Music software.
Artist platforms.
Stores.
Guitar pedals.
Amplifiers.
Electronics.
Games.
Writing.
Teaching.
And that's before we get into actually making music.
I can either spend the next several years trying to do every little thing myself, or I can use the tools available to me to start bringing these ideas to life.
I've made my decision.
I Don't Have The Luxury Of Being A Purist
There's a perspective in software development that I understand and respect.
Learn your tools.
Understand your language.
Write your own code.
Master your specialty.
Become exceptionally good at what you do.
There's nothing wrong with that.
But I think we sometimes make the mistake of treating one approach to professional development as though it's prescribed to everyone.
Not everyone is trying to accomplish the same thing.
Some developers want to specialize deeply in a particular area of engineering.
Some want to work for established companies.
Some want to become experts in a particular language, framework, or discipline.
Those are legitimate goals.
But they're not necessarily mine.
I'm a builder.
I have a vision for things that extend well beyond software development.
Software is one of the tools I use to bring those things into existence.
And if I'm spending all my time proving that I can personally implement every line of code, I'm not necessarily getting closer to those goals.
I'm just getting better at doing everything myself.
Those aren't the same thing.
I Can Put My Code Where My Thoughts Are
One thing I've learned about myself is that I'm a big thinker.
I see systems.
I see how things connect.
I see how one piece of infrastructure can support five different projects.
I can look at a music business, a software platform, a storefront, and an electronics workshop and start imagining how they could operate together.
And for the most part, I can put my code where my thoughts are.
I've been doing that with KiwiEngine.
The problem isn't that I don't know how to build things.
The problem is that I can't personally implement every idea at the speed I can envision it.
There are only so many hours available.
And I don't have an entire engineering department sitting behind me.
So why wouldn't I use technology that helps me extend my capacity?
AI Doesn't Design My Vision
This is where I think an important distinction needs to be made.
I don't hand AI a vague idea and tell it to go build my company.
I design the systems.
I establish the architecture.
I determine the responsibilities of the libraries.
I make decisions about how the pieces interact.
I write and develop the foundations.
Then I use AI to help with implementation, repetitive work, documentation, tests, and other tasks that would otherwise consume significant amounts of time.
That's delegation.
And delegation doesn't eliminate responsibility.
If anything, it makes responsibility more important.
Because now I have to understand not only what I'm building, but what I'm approving.
No Ghost Code
I have a very deliberate boundary when it comes to AI working on KiwiEngine.
AI does not have independent authority to commit code, change architectural direction, or make decisions on my behalf.
It doesn't get to decide that a library should be rewritten.
It doesn't get to introduce dependencies because it thinks they're convenient.
It doesn't get to change the design philosophy because another approach is more conventional.
And it doesn't get to silently commit code I haven't approved.
There is no ghost code.
If AI is going to take an action that changes the project, it needs my explicit authorization.
I review the work.
I approve the direction.
I remain responsible for the result.
AI can be part of my workforce without becoming the owner of my work.
The Grunt Work Is Still Work
Here's something I think gets lost in conversations about AI-assisted development.
Not every programming task requires the same level of creative investment.
Some tasks are genuinely interesting architectural problems.
Others are repetitive implementation work.
Writing another adapter.
Creating another set of tests.
Updating documentation.
Implementing a pattern that's already established elsewhere in the project.
Working through boilerplate.
Those things matter.
They still need to be done correctly.
But they don't all require me to spend hours personally typing every character.
If I've already established how a system should work, why wouldn't I use a tool to help implement that established pattern?
My time is better spent reviewing the result, improving the architecture, and moving on to the next problem.
That's not avoiding engineering.
That's managing engineering work.
AI Doesn't Eliminate The Need To Understand Your Code
There's a difference between using AI to extend your abilities and using it to avoid developing them.
I don't believe AI should become an excuse to stop learning.
If I don't understand a change, I shouldn't approve it.
If I can't explain why something belongs in the architecture, it probably doesn't belong there yet.
If AI generates code that violates the design principles of KiwiEngine, that code doesn't get a pass simply because it works.
Working code isn't automatically good architecture.
And generated code isn't automatically correct.
AI can make mistakes.
It can misunderstand requirements.
It can introduce unnecessary complexity.
It can confidently implement something that doesn't fit the rest of the system.
That's why oversight matters.
I want AI helping me move faster, not helping me accumulate problems faster.
I'm Not Trying To Win A Coding Endurance Competition
This is the part I've been thinking about the most.
What exactly am I proving by doing everything manually?
That I can work longer?
That I can memorize more syntax?
That I can rebuild the same implementation for the fifteenth time?
I already know I can build software.
I want to use that ability to accomplish something.
I want to create tools for musicians.
I want to produce music.
I want to develop physical products.
I want to experiment with electronics.
I want to build businesses around things I genuinely enjoy.
I want to spend more time creating cool things and less time getting bogged down in technical work simply for the sake of purity.
If AI can help me accomplish that while I retain ownership and oversight, I'm going to use it.
KiwiEngine Is Supposed To Create Leverage
This is actually very consistent with why I'm building KiwiEngine in the first place.
The whole point of a reusable framework is to avoid solving the same problems repeatedly.
Juice helps me avoid recreating common styling behavior.
Nectarine helps establish reusable configuration patterns.
Other libraries and engines provide foundations I can build upon.
AI fits into that philosophy as another form of leverage.
KiwiEngine reduces the amount of infrastructure I need to rebuild.
AI reduces the amount of implementation work I need to personally perform.
Together, they allow me to spend more time on the problems that are unique to what I'm creating.
And that's the entire point.
I Want To Build Things, Not Just Talk About Building Them
I have enough ideas to keep myself busy for several lifetimes.
What I don't have is several lifetimes.
I have this one.
And I would rather spend it bringing as many meaningful ideas into reality as I reasonably can.
I'm not interested in pretending that using AI somehow makes those ideas less mine.
The architecture is mine.
The decisions are mine.
The responsibility is mine.
The vision is mine.
The tools help me execute it.
That's how I see AI.
Not as a replacement for my abilities.
Not as an excuse to abandon craftsmanship.
And not as an autonomous authority over my projects.
It's part of my workforce.
I intend to use it that way.
The future isn't something I'm waiting around for.
It's already here.
And I have things to build.
