There are a lot of stacks I could have reached for.
And because I wasn’t a master of any of them, I didn’t have a bias. That sounds like a weakness and mostly it is, but here it helped. No favourite framework to defend, no muscle memory pulling me toward the thing I already knew. I was free to just pick what met the conditions.
The conditions were short:
- It should be easy to build with.
- I don’t want to spend time building an interface unless I have to, and only then for the efficiency of testing.
- It should be free, and I don’t want to install a new engine or another piece of software.
Run those three against the options and Xcode just walks through. It was already on my Mac, so nothing new to install. It’s free. And SwiftUI ships with a pile of built-in components and animations, which means unless I need to build a genuinely new animation or a custom UI, the framework has me covered out of the box. Less interface work, more testing. That’s condition two, handled.
It also helped that I’d been spending time with Swift anyway. ( I restarted Hacking with Swift this year , slowly, the repetition kind of learning.) So the tool wasn’t foreign. It was the one I happened to already be warming up to.
I’m not going to pretend this is the right stack for everyone. It’s the one that fit my three lines. Yours might be different, and the constraints are what matter, not the logo at the end of them.
Next, the actual method, how I break a task down before AI touches it. This is the part I think matters most.