The Removal

On twenty years of making things worse by making them simpler

Don’t want to read? Listen to the audio in an interactive discussion (all generated through Google Notebook LM)

The invisible becomes the inevitable

My car has a screen the size of a cutting board and it cannot tell me what it wants from me without scaring me half to death.

I’m driving a few days after a snowstorm. The roads are mostly clear at this point, the world still white where the sky meets the road. I’m just cruising. Heat on, windows up, the car doing what it’s supposed to do, carrying me somewhere without asking anything of me.

Then something flashes on the dashboard. An orange icon, sharp against the black screen. A screech repeating two, three times. My hands tighten on the wheel, my heart rate spikes, my eyes drop to the screen — at seventy miles an hour — trying to decode a symbol that doesn’t look like anything. By the time I understand it’s telling me my tires flew over a patch of ice, I’m already past it. The ice is way behind us. The only danger was actually the interface telling me about it.

I think about this a lot. Not just because it might be the worst interface I’ve ever encountered. But because it might be the most revealing.

It pretends to help me. What it actually does is transmit noise and leave me to figure it out while driving at highway speed.

The more I sit with it, the more I think this is where the entire field has been heading for a long time. And I’ve been inside it for a decade, watching, afraid and angry.

Recognition was the first thing we lost

When I was in kindergarten, my dad trained me to raise my hand at the end of every storytime and ask the same question: “Dennis, is this a metaphor for life?” I was five. I didn’t know what a metaphor was. But I learned the rhythm of the question before I learned its meaning, and eventually the meaning caught up.

Everything could stand for something else. Everything had a layer underneath that has always been worth asking about.

I think about this when I think about interfaces.

There was a period when digital interfaces looked like the things they were replacing. The notepad looked like a notepad. The bookshelf looked like a bookshelf. The calculator had buttons that looked like buttons, with shadows underneath them, as if your finger might actually press them down. The design world called this skeuomorphism, and then spent a decade making fun of it. The fake leather. The stitching. The drop shadows. Oh, how quaint.

The industry moved on. And in moving on, it stopped asking why any of it had worked in the first place. Too many designers learned the tools without learning what the tools were for. What they never understood was that the metaphor was doing the work. The interface looked like something you already knew, which meant you already knew how to use it. The knowledge you carry in your body from living in the physical world — how a shelf holds things, how a button presses down, how a page turns between your fingers — the metaphor brought all of that with you. That was a bridge between the world you felt and knew, and the screen you didn’t.

Then flat design came along and rearranged the meaning. It didn’t destroy it. It moved it. Meaning was taken out of the hands of the person using the interface and put into the hands of whoever built it. The shadows disappeared. The textures vanished. What replaced this actually wasn’t “clean” or “minimal,” though that’s what we told ourselves. What actually happened was a hodgepodge of lego slop — icons that meant different things on different platforms or pages, layouts optimized to sell not something but anything (or everything). Every company built a design system and called it maturity. But maturity isn’t a system; it’s understanding. And understanding was exactly what got taken away from all of us on the other end of the screens.

Nobody asked who it was given to. We treated the removal like an arrival.

That was over a decade ago. Closer to two, depending on where you start counting. An entire generation of designers has been trained in a world where interfaces aren’t supposed to mean anything. They’ve never built a metaphor. They’ve only built surfaces.

Legibility is a right, not a feature

It was 2020. The pandemic had shrunk my world to a living room in a condo I shared with a roommate. There was no separate office. There was a couch, a laptop, and a small screen that was also where I did my design work.

I was working at an agency supporting the Small Business Administration during the PPP loan program. The application for the first round had been built fast — technical language nobody had checked against the people actually filling it out, pages with no clear signal of what mattered most, a form that functioned like a locked door for anyone who didn’t already speak the language of government lending.

The program helped people. But it also helped Potbelly’s. The interface couldn’t tell the difference between a restaurant owner with twelve employees and a chain with a thousand, and it wasn’t built to try.

I stayed up through the night on that couch, more than once, working on a redesign. I was trying to make it so a person — someone who might not speak English fluently, who didn’t have a lawyer, who was terrified of losing everything — could fill out this form correctly and simply. So that when I eventually walked outside again when presumably the world would return to normal, the storefronts I passed would still be open to meet me. The neighborhood bakery. The Dominican hair salon. The Thai place you forgot the name of but used to rely on twice a week. The places that make a street feel like it belongs to the people who live on it. Not just the big box stores that would survive anyway.

Then the politics shifted. Uncertainty over whether another round would be approved took over. Start, stop, start again, stop again. I was standing at my kitchen counter on a call when I heard the redesign was cut. Not tabled. Not saved for later. Scrapped. Nobody even looked at the redesign. The team just moved on to the next thing, as reactively as everything that had come before it.

I quit the agency not long after.

That couch, that small screen, that PPP loan form — I think about it every time someone tells me the best interface is the one you don’t notice. I very actively wasn’t trying to make something invisible; I was trying to make something legible. An interface that said: this is where to start, this is what matters, and here is how to get through it. Not frictionless. Clear.

The kind of clear that keeps someone’s life from falling through a crack in a form they couldn’t read.

That’s what design is for when it’s working. And the field has been moving away from it for a very long time.

Disregard is not the same as immaturity

My very first job in design was at an agency. I chose it because I wanted to learn fast — see as many expressions of the work as I could. I was juggling five projects at once, because that’s how agencies work. One of them was for the General Services Administration. The opportunity was to improve the pages government employees use when they travel — how they set up their trips, how they figure out their per diem, and how they get reimbursed so they can get back to the work that actually matters.

The pages stood between government employees and the reimbursement they were owed like a gate with no visible lock — just hours of reading and guessing until something finally gave. Whether that was by design or by indifference, the effect was the same.

So I did what designers actually do. The job description says it’s about learning how to make something better. But the real work is closer to legal: it’s building a case strong enough that someone with the power to change it will actually give up the resources to fix it. I recruited people willing to record themselves trying to get through the pages alone and set up scenarios, then let it loose. Imagine you’re traveling to Denver for a two-day conference and need to figure out your per diem, how would you do it? — and watched. I listened to each recording probably five times, maybe more. So many times that, ten years later, I can still hear the parakeets squawking in the background of one man’s session. The parakeets, and his frustration, and the sound of someone clicking the same wrong link for the third time, slower each time, like the screen was actually wearing him down.

The design industry has a name for this kind of failure. They call it “low design maturity.” My senior mentors used it constantly. I spent months translating what I was seeing into that language, packaging it into decks and presentations, framing disregard as a growth opportunity, because that was the only way to get anyone in the room to nod along.

But it wasn’t a growth opportunity. The woman I worked with on the inside believed in the work. She brought it to the people above her and they didn’t engage. Not because the evidence was weak. Because caring about whether government employees could navigate a webpage wasn’t something that had ever cost anyone anything.

Nobody got fired for a confusing per diem page. Nobody got promoted for fixing one. That’s not immaturity. That’s just what happens when power organizes itself into hierarchies and the people on the other end of the screen trying to get reimbursed are also at the bottom of the priority list.

I didn’t have the language for that then. I was too new. What I had was angst and the instinct that the framing was wrong.

It took a decade — working at agencies, then health care, then the largest tech companies in the world — to see the full shape of it. Different organizations, different products, different scales. The same structure underneath. And every framework, every maturity model, every best practice the industry offered was laid on top of the thing that actually determines what gets built: power and money.

I tried to write a book about it once. Called it “Biased by Design.” I couldn’t finish. I was too deep inside the machine and didn’t have enough altitude to see what I was describing.

I see it now. And I have the language.

The metaphor is gone. The cursor remains.

I was on the couch in our side room this morning, coffee and almond milk, my partner next to me, our dog between us, scrolling LinkedIn, when I saw a post from a designer I respect. “So the UI of the future turned out to be the command line,” he wrote, with a shrug emoji.

I laughed. And then I couldn’t stop thinking about it.

He’s right. And he’s wrong. The command line is a future, it is not the future. A blank text box, a blinking cursor, and the assumption that someone already knows what they need. No metaphor, no bridge, nothing to help someone understand what they’re looking at before they have to do something about it. That’s one reimagining. But companies with too much power and too much money have made it look like the only one. They have not done this by accident, but by funding this way of thinking, hiring for it, rewarding it, and now eliminating the people who imagine differently.

There is a way of building products that has become so dominant it feels like gravity. It says the point of a product is to help someone get something done. Figure out what they need, then get out of the way. We have a term for this, but I won’t bore you with it.

This is the logic underneath how we search on Google and buy on Amazon and scroll through other people’s lives on Meta and listen on Spotify and manage our calendars and finances and health records and relationships. Every one of these products is supposed to help someone get through something basic, and most of them are subtly horrible at it. Design was coopted by this logic before anyone finished imagining what else it could be.

Getting something done isn’t the problem. A light switch should be simple. The problem is when that logic becomes the only logic — when it scales across every system people depend on. We’re talking about basic, important things now left to products that do not keep everyday people in mind: navigating our health, our relationships, our money, our lives. And nobody stops to ask whether any of it actually works for people.

The way we’ve been designing our products has cut the moment that actually matters — the moment when someone is staring at a screen and hasn’t yet figured out what they’re looking at. And certainly haven’t formulated the right question, or maybe any question at all. The government employee with the per diem page. The small business owner with the loan application. The driver with the beeping dashboard. The people who needed the interface to help them think, not just do. And we are the people.

I know how to use the blank box. I open Claude and I type and something comes back and I iterate. But I’m someone who has spent a decade learning how to ask questions for a living. The cursor blinks for me and I know what to type.

That’s one possible future of design. The cursor is still blinking…