Unbothered
The difference between using AI to skip the work and using AI to do it better
My husband Sean read Boring Is a Feature and had a critique. He is in a job search right now, and he said he thinks the discipline I outlined is what is slowing him down. Every posting wants AI first or AI native. He does all the things I wrote about. He is not AI first. He said: I think AI first is a way to skip all those parts.
The formal definitions are about systems and organizations. AI first means prioritizing AI when building products and services, while AI native means designing systems so AI is embedded into the architecture itself.
Here is what being AI Native means to me.
I use AI every day. It is woven into how I design, how I build, and how I think through problems. But I do not use it to skip steps. I use it to do the steps better than I could alone.
The thinking still happens first. I spend time designing in markdown before I write a line of code. That document becomes the context my AI thinking partner works from. The documentation does not get skipped because I am using AI. It gets better, because the design is already there and the comments write themselves from it. Fewer keystrokes. Not fewer decisions.
I will be honest about testing. I am a typical engineer who loves the happy path. Left to my own devices I will test the thing I expect to happen and quietly not think too hard about everything else. AI is changing that. I am writing more tests than I ever have, covering more scenarios than I would have thought to cover on my own. I still need to run a code coverage tool to see where my numbers actually land, but the trajectory is clear.
I once inherited a codebase with less than ten percent coverage. We were tasked with getting it to sixty. It was one of the most painful stretches of work I can remember. With AI, that same job would have been hard. Not brutal.
Here is what I keep coming back to: initial code was always the cheapest part of a product. Maintenance is where you spent the real time and the real money. That has not changed. What has changed is that some people are now saying: if it breaks, just use AI to rebuild it. Maybe. But why run at computer speed when you could run at slightly less than computer speed and actually understand what you built?
AI native does not mean frictionless. It means using AI to close the gap between the engineer you are and the engineer you want to be. More thorough. More tested. More observable. Still responsible for the code.
Being AI native to me means using every tool available to do it better. Not faster at all costs. Better.
And Sean? He is still on his way to AI native. But he already has the part that matters most. The rest is just tooling.