your 3fo all-star: meet rex

The most obscure bug you've had to debug?

They're all obscure for me. We spend a lot of time doing quality assurance so when we do get a bug, they’ve slipped past the normal process because they are special. The truly evil ones are the bugs that refuse to present themselves, either you can't reproduce it at all, or you can reproduce it perfectly and the logs contain nothing but a cheerful silence. It’s the bug the client sees every time and you’ve never seen it once. It only happens on their nephew's iPad, on Tuesdays, in landscape mode.

The dirty secret of complex systems is that one symptom can have a thousand causes, and one wrong line can cause a thousand symptoms. The only real defense is simplification—the less machinery you have, the fewer places a gremlin can hide. That, and logs, those beautiful, verbose, totally judgy logs.

What everyday task would you automate with a custom script?

Coffee. Not "press a button on the Keurig" coffee, I mean the full ritual. Fresh grind, water heated to exactly the right temperature, and a slow, meditative pour-over. Philz (Bay Area chain) style. Then a bit of honey and some frothy heavy cream.

I want a robot that does all of that while I stand there watching with my arms crossed, nodding approvingly. I've done the math, and building this would take roughly 400 hours of engineering to save myself 10 minutes a day, which means it pays for itself in... a number of years I’m not calculating. This is what we in the industry call "a passion project." And that my boss (probably?) won’t pay me to develop.

Overhyped vs. underhyped tech trend?

AI (LLMs) is both. It's overhyped by companies claiming it can do everything short of raising your children (give it a quarter), and it's genuinely feared by people worried about their jobs — which, unfortunately, isn't paranoia for everyone. So there's a lot of noise on both ends.

Meanwhile, underneath all that, there's a quieter group of people just using it. For creative things, for mundane things, for the unglamorous Tuesday-afternoon work that never sees the light. That's the underhyped part. My hope is we blow through the honeymoon phase soon, the job market stabilizes, and we all land somewhere reasonable. Meanwhile, the billionaires will continue getting more billionaire-y, which appears to be the one constant in every cycle with new technology.

Most common web development mistakes companies make?

Thinking it's going to be easy. There's a classic line in engineering that goes something like: "We do these things not because they are easy, but because we thought they would be easy." Everyone has lived this.

The other big one: treating a website launch like a finish line. You don't buy a plant, water it once, and then call it landscaping. Websites left alone quietly rot and you won’t notice. It’s usually something simple like the content goes stale, plugins age, or the contact form has been silently failing since March.

But honestly, the list of possible mistakes is infinite, so I'd rather talk about the short list of things that actually matter: real thought about what users are actually trying to do on the site, navigation that doesn't require a map, solid visuals, and copy that's relevant to the audience. A solid technical foundation with excellent performance and accessibility. This sounds like a lot but fixing this stuff doesn't always mean a full rebuild. Often you can improve the experience one page at a time.

How has web dev changed, and where's custom design heading?

AI has hit the accelerator on everything—development, design, all of it. What used to take weeks can take days. But here's the catch: acceleration is not direction. If you are vibe coding something in a day when you don’t understand the underlying structure and exploits, that will bite you in the rear. So if you actually care about the result, you still need humans doing the strategy and the refinement. A fast car with no driver is just an expensive way to hit a wall.

Where's it all going in 2–5 years? I genuinely have no idea, and I am suspicious of anyone who claims they do. The only safe prediction is that it'll look different than it does now. But I'm optimistic—I think it lands well for us as an agency and for our clients too. Mostly because the alternative is being pessimistic, and I've already got the coffee robot project to worry about.