Convenience at Our Doorstep

On vibe coding, company culture, and the questions nobody's paid to ask

Nothing grows in a machine that never stops

For a year I commuted three hours a day. I gave work everything — every morning, every commute, every hour in the office. And when I got home I’d eat something and fall asleep by nine. Wake up, do it again. There was a room in my basement that was becoming a home office — theoretically. I’d gotten as far as putting a computer down there, but I never fully set it up.

A couple days after I was laid off, I was in that room, ready to jam. I don’t remember deciding to set up the room. It just became possible, the way new things become possible when the thing that was consuming everything just suddenly stops. Space. Opportunity.

I opened Claude with an idea, or, more honestly, half an idea. I knew enough about prompting to know I should have a real plan before I started vibe coding. I didn’t have one. What I had was an instinct and a direction. I wanted Claude to help me draft the plan and build the thing in parallel, which is almost certainly not the sophisticated, correct, or applaudable way to do this.

But I started. And within a few hours I had something. An app. A working thing.

It might not be good. Probably isn’t. The idea might be right and the execution might be wrong. The execution might be passable while the idea isn’t. I don’t know yet. What I do know is that within an hour, I started catching things in this snazzy-looking app: what didn’t matter, what did (or might), where the framing was off, where the concept and the flow and the UI and the copy were all tangled up in ways I couldn’t separate cleanly. Who are the users? What do they actually want versus what I think they need? Does the value proposition actually resonate with anyone besides me?

I’m a UX researcher. My entire practice is built on the premise that you can’t think your way to understanding in isolation. You have to get outside your own head, test your assumptions against other peoples’ reality. The work should be shaped by perspectives you didn’t anticipate, and usually can’t. Instead, I was in my basement, in a closed loop with an AI, iterating on an idea that no other human saw. The evolution of the concept lives entirely in a conversation between me and my buddy Claude. Nobody else is in the room to say “Wait, what about?” — or “Dude, no!” or even the dreaded but imperative occasional “Yikes, that’s ridiculous.”

There’s something seductive about that kind of freedom. And something I will never trust.

The freedom to fail is unevenly distributed

So why couldn’t I do this before?

The easy answer is time: the commute, the exhaustion, the fact that there was literally nothing left in my brain or bones by the end of the day. That’s true. But it’s not interesting and it’s not the full story.

The full story is that even if I’d had the time, even if I’d been remote, well-rested, with a functioning home office and hours to spare, it still wouldn’t have been rational for me to build. Every hour spent experimenting would have been an hour I wasn’t doing the work I was constantly being evaluated on.

The work I was hired to do isn’t the kind you can point to in a demo; it’s the kind that keeps bad decisions from getting made — slow, invisible, and only noticed when it’s missing.

If I’d spent a weekend vibe coding something and it worked, nobody would have rewarded me. If it didn’t, I’d have neglected the thing I was actually accountable for. The math here doesn’t math.

Recklessness that ships is still recklessness

I was a senior UX researcher at Amazon, now a line item on a balance sheet flashing red. There was no time to spare. What there was, was a kind of frantic performance — everyone running around trying to prove they were “using AI,” trying to show they were speeding up their workflows, just about to launch something new, more bets, more things. What those things were didn’t seem to matter much. We were measured on AI usage, and just using it registered as a positive. For what purpose? Pulling us toward what outcome? Nobody stopped long enough to ask. Probably none of us could. Tools were being updated and deprecated faster than anyone but maybe engineers could learn, while also still doing the jobs they were hired to do. The velocity made sense from a distance; it didn’t make sense for the people inside it.

And here we are. This isn’t just where I was. This is where we all are, in a world created faster than anyone can question. Engineers build what they’re told to build, inquiry removed. And then they’re rewarded for it — not subtly. Publicly. By name. With awards attached, breadcrumbs to promotions. Shipping is the metric, and shipping fast on high-risk bets is courage — a “good bet.” And it’s still a good bet even if it fails. Because the act of building is the value. Whether what gets built serves a real need or locks people into a product they have no choice but to use, even if it’s confusing or damaging. This is someone else’s problem. Or no one’s.

I stayed curious from a distance that never felt safe; it felt like a slow leak with the threat of bursting at any moment. I’d been on edge for at least a year.

What frustrated me wasn’t that engineers were building. The new possibilities here are full-stop sparkling. It was that too many aren’t asking whether what we’re building should exist at all. Too many take an idea at face value and build it. Then they demo it to the people the thing was ostensibly designed for and say the UI isn’t quite there yet — as though the UI were the only thing that wasn’t there yet. And those people nod, because that’s how all this works.

You put a working shiny thing in front of someone, they’re gonna oooh ahh and be bought in. They start to imagine living with it. And now everyone is caught gravitating around a future that might not be the right one.

This is what building without rigor does. It commits resources without discernment and it narrows what anyone is able to imagine. This way of working locks people into a deterministic future, and we haven’t even skimmed the surface of possibility.

I knew this. I watched it happen, front-row seat. And that knowledge made it impossible for me to build, to learn new skills, to demonstrate that research could have a place in this new way of working. Because I understand what it costs to vibe-code without disciplined thinking. The system I was in had created a double bind — the people most aware of the stakes of building recklessly were the same people least able to build at all. Then the bind broke.

The 5am text popped into view and my access was revoked. And the first thing I did with my free time was the thing the system — my job —had made irrational to try, while rewarding it in others.

Eliminating the question is the point

In the middle of the largest technological transformation most of us will live through, the people whose job was to ask does this actually work for humans are the ones being let go. The questions still matter — arguably more than anything else — but the system decided the questions were slowing things down. And so the roles that exist to represent the person on the other end of the product are treated as friction to be eliminated.

Yes, your package still arrives in twelve hours. More or less, the product still works. The interface is fast and the price is low, and the whole thing is so one-click frictionless that it’s easy to mistake efficiency for care.

Somewhere along the way, the system decided that the fastest path between a product and a person is a straight line. All miles are the last mile — no detours, no doubt, no one asking should we? That feels like convenience. It’s seductive in every way. But it’s also a total takeover of how we buy, how we work, how we live, packaged neatly, arriving with a thud on our doorsteps.

I’m in my basement now, vibe coding something I’m not sure about, asking the questions nobody’s paying me to ask. Not because I chose this, but because the system made the choice for me — and for all of us — and most people won’t notice until the thing they lost is something convenience was never going to replace.