I. A Line That Has Been Quoted Too Many Times
In 1998, Steve Jobs gave an interview to BusinessWeek and was asked whether Apple had done any consumer research while developing the iMac. It's very hard, he said, to design products by focus group — and then he left behind the line that countless designers and product managers have been repeating ever since:
People don't know what they want until you show it to them.
Much of the time, people only know what they want once you've put it in front of them.
That was an era when a designer could "lead" taste without apology. Translucent iMacs in fruit colors, products in shapes nobody had asked for — bold, even a little willful, but every one of them announcing to the world: what looks good and what works well, we get to decide.
But two pieces of context around that line tend to get dropped. In the same answer, Jobs added that Apple also watches industry trends very carefully. And in the same interview, he summed up his own creed as "focus and simplicity."
The Jobs that later generations flattened into "users don't know, listen to me" actually did both: he watched the trends, and he believed in subtraction. That's where I want this essay to end up.
II. "Apple Has Gone Mediocre"
On September 1, 2026, John Ternus officially succeeded Tim Cook as Apple's CEO. He had joined Apple's product design team in 2001 and was Senior Vice President of Hardware Engineering before taking over. A week later, at the launch event, Apple unveiled its first folding iPhone, the iPhone Duo, alongside the iPhone 18 Pro lineup.
The verdict online was nearly unanimous: Apple is just piling on specs now — bigger screen, bigger battery, more cores. A folding phone is reheated Android leftovers. The Apple that dared to make a fruit-colored computer is dead.
I don't see it that way.
III. There Is Only One Way to Score Full Marks
In the 1990s, the phone industry was in its wild-growth years. Nobody knew what a phone was "supposed" to look like, so candy bars, flip phones, sliders, and swivel screens all bloomed at once. Thirty years on, the industry is thoroughly mature. Screens, batteries, camera modules, and hinges come from the same handful of suppliers, obey the same physics, and serve the same-shaped hands and pockets. Even Apple's first foldable gets its folding panel through a supply agreement with Samsung.
Under those conditions, taste naturally converges. On any exam, there is usually only one way to score full marks; it's the 60-percent and 30-percent answers that come in endless, bizarre variety.
Cars are the same. It's not that Honda wouldn't like to make the Civic flashier; it's that under the combined constraints of cost, regulation, drag, crash safety, and supply chain, a flashier answer simply scores lower. Biology has a name for this: convergent evolution. Sharks and dolphins are only distantly related, yet they evolved nearly identical streamlined bodies, because water resists everyone the same way. Convergence in a mature industry isn't everyone copying each other. It's everyone closing in on the same full-marks answer.
Arriving late is itself an old Apple tradition. The iPod wasn't the first MP3 player; the iPhone wasn't the first smartphone. Apple's strength was never being first — it was waiting until a category had matured enough to be done well, then stepping in. Entering foldables only after years of refinement in the Android camp isn't a halt in innovation so much as going with the grain. Once the possibilities of form have been compressed to their limit, innovation naturally migrates to where you can't see it: silicon, materials, hinges, software.
IV. Why Everyone Wants Their Physical Buttons Back
Phones got me thinking about cars. Every time I watch a car review, the reviewer and the comment section are pleading, almost in unison, for the same thing: give us back the physical buttons on the center console, and stop cramming every function into three or four screens. A touchscreen is nearly impossible to operate by feel while driving, its menus run several layers deep, and a driver shouldn't be juggling complex tasks unrelated to driving in the first place.
This isn't just nostalgia. Starting in January 2026, Euro NCAP's new rating scheme means that a carmaker that doesn't provide physical controls for the turn signals, hazard lights, horn, wipers, and emergency call can't earn the top safety rating. Volkswagen has publicly admitted that removing buttons was a mistake and, after a wave of customer complaints, has begun restoring physical climate and volume controls on new models. Mercedes' head of software has said the same: the data shows physical buttons are better, so they're going back in.
It's a glimpse of a larger pattern. For the past decade, plenty of carmakers have been eager to "educate" consumers on what counts as premium and what counts as futuristic, rather than listening to what users were actually complaining about. The BMW G70 7 Series, with its split headlights (nicknamed "squinty eyes" in China) and its oversized kidney grille, has been controversial since the day it launched.
The Mercedes-AMG C63 is the more telling case. In 2022, AMG swapped the C63's twin-turbo V8 for a four-cylinder hybrid. The new powertrain was stronger and more advanced. Yet in Germany, dealers reported that the new car was barely selling at all; in the United States, dealers had to put up $10,000 lease incentives to move it. AMG's response at the time was revealing: its chief admitted they had lost some customers who only wanted a V8, yet framed the problem as a failure to explain the technology well enough. That's almost the textbook version of "the users just haven't been educated yet." In the end, AMG conceded it had overestimated how willing customers were to accept a four-cylinder C63, and the car was retired early. Motor1's verdict was apt: the C63's early retirement proves that specs aren't everything.
V. Users Can't Tell You the Answer, but They Can Tell You Where It Hurts
Slowly, I've worked out a distinction: the user's voice actually comes in two kinds. One is "here's the solution I want." The other is "here's where it hurts."
On the first kind, Jobs was right — users usually can't say. Ask consumers in 1998 what kind of computer they wanted, and nobody would have answered "translucent blue." But on the second kind, users are almost never wrong. Not being able to find the wiper stalk by feel while driving is real pain. A C63 without the V8's soundtrack and drama is real pain too — it doesn't hurt the spec sheet, it hurts the reason the car exists.
So "leading taste" and "listening to users" aren't in conflict. Good designers lead on solutions and listen on pain. Arrogant designers do exactly the opposite: the user is clearly saying "this hurts," and the designer answers, "you don't understand."
Back to Apple. A bigger screen, a bigger battery, more cores: these are real needs that mainstream users have expressed, over and over, with their purchases, and here Apple chose to listen. But in design, it is still subtracting. You won't see Apple bolt a physical keyboard onto the iPhone, cram in three screens, or stack on fifteen cameras. Follow the user on needs, stay restrained on form: the two don't conflict.
There's nothing wrong with playing music too refined for the crowd, and popular taste is certainly not the same as good taste. But that's no reason to dismiss the voice of ordinary people wholesale. This kind of arrogance is especially common today in consumer electronics, cars, and luxury goods. Designers can absolutely lead taste, but they can never completely ignore how users feel.
VI. In the Age of AI, Restraint Is What's Scarce
As an AI engineer, I'm seeing a different kind of failure more and more often in other people's projects: not arrogance, but accumulation. Letting the AI exercise its "creativity," adding one feature after another, because the cost of adding a feature has dropped to almost nothing.
It reminds me of Baidu Maps. A map app that has swallowed navigation, hotel booking, restaurant reservations, food delivery, endless coupon campaigns, social features… As an engineer, I get a headache every time I open it. How does a product like this survive? My answer is competition. When there are few rivals, even garbage can live on in the market; the moment a cleaner option like Amap showed up, the people around me voted with their feet almost immediately. A market vote only works when users have a choice, and that's as true for phones and cars as it is for apps.
AI amplifies the accumulation problem. Features used to be hard to build, and engineering cost itself acted as a natural filter. Now that generation is cheap, that filter is gone, and judgment is the only gate left. I feel the same thing in agent systems: the more tools you hand the model and the longer its instructions, the harder it is for it to make the right call at the critical moment, and the more ways it has to go wrong. Capability isn't better when there's more of it. It's better when it's more precise.
Look back at Jobs' "focus and simplicity" through that lens, and you realize his real legacy may not be "users don't understand taste." It may be the willingness to say no.
VII. Closing
If you want to show your taste, one watch that genuinely reflects your eye is enough. Ten Rolexes stacked up both arms only makes people think you're a fool.
In a mature industry, converging taste isn't mediocrity — it's the answer coming into focus. In an era when anyone can pile on endlessly, real taste is knowing what to leave out.
I hope the product-makers of this era carry a little less of the arrogance of "let me teach you what good is," and a little less of the indulgence of "it costs nothing to add it anyway" — and learn to subtract.
