I know I said I was going to keep writing about concepts I use every day as a frontend engineer. Previously, I wrote about things like SPA routing, client-side vs server-side rendering, CDNs and all that good stuff. Life happened, a month went by, and honestly my workflow changed so much that it felt wrong to just jump back in without talking about it first.

Not that I stopped learning or building, far from it actually, but the way I approach features and fixes now is very different from what it used to be, mostly because of AI tools like Claude Code.

These days, I spend more time supervising code than actually typing every line myself.

At first, that feeling messed with me a little bit to be honest. It almost felt like I was cheating the process somehow. Like, if I’m no longer manually implementing every single thing, what exactly is my role now?

But after a lot of reflection, I’m starting to understand that engineering was never just about repetitive implementation. We were primarily created to think. To make decisions. To identify tradeoffs. To know why something should exist and what approach makes the most sense.

One thing AI does brilliantly though is helping ideas move from zero to one very quickly.

Previously, you could have an idea, think deeply about it, maybe even start building it, then life happens. Work happens. Priorities change. Maybe you just lack discipline (sarrry sarrry 😭) and the project just ends up in the dump of unfinished side projects.

I think a lot of engineers can relate to that.

I got an idea a few months ago that was very personal to me. I started actively taking notes during church services and events, but I noticed I rarely went back to read them. Sometimes the next time I opened a note was the next Sunday.

Discipline issue? Maybe. But I started thinking the real problem was visibility. It’s easy to forget things you never see again.

So I built something. A reminder based mobile app that resurfaces my notes intentionally throughout the week.

And building it was the first time I really felt what AI in a workflow actually means.

I had the idea, I knew the problem, AI helped me move fast on implementation. It gave me a solid template to work with. But decent wasn’t enough. My experience with UX meant I could look at what it gave me and immediately know what felt off. Certain flows weren’t natural. Some interactions didn’t match how users actually behave. So I made the adjustments, not because AI did a bad job, but because I already had a standard in my head.

That’s when I really started understanding taste.

And honestly the code AI produces is very decent my bro 😭

Decent enough that if you’re not careful, you stop thinking deeply about the decisions being made. Sometimes I catch myself just accepting changes quickly without really sitting down to understand if it’s actually the best approach.

But a colleague told me something recently that stuck. He said one thing AI will not take away from engineers is taste.

For me, taste is feeding your mind well enough to make sound judgments. Experiencing enough to know what good actually looks like. Because AI can generate templates but taste is what takes things from generic to actually good.

There has to be a standard in your head already, otherwise AI becomes the standard, and that guy can give some very generic stuff 😭

And honestly staying open minded is not a bad thing. It exposes you to a lot. The good stuff, the bad stuff, all of it. And that’s actually how taste gets built. You have to see what doesn’t work to know what does. So when AI puts something in front of you, you already know.

Thinking gives you something to drive towards. Leave your mind completely open and you will accept anything. That’s what taste protects you from.

So yeah, let’s keep building our taste, keep building things, and keep growing into better builders.

Till I come across something again… see you soon. Byeee 👋