Most are just copying
Many designers believe they’re building taste when they’re really just collecting aesthetics. They save screenshots from Dribbble, organise Pinterest boards, follow people with polished portfolios. It looks like effort. We all do it. I have 100+ boards probably with 1000+ saved pins and bookmarked posts on IG.
Most don’t reverse-engineer what they collect. They don’t rebuild those screens from scratch or ask why something works. They get good at spotting trends. Neumorphism one month, brutalism the next but not at making tradeoffs. I’ve seen people reject working layouts because they “don’t feel fresh,” or chase a flatter hardware enclosure simply because another DTC brand launched something sleeker. And when style changes, their so-called taste disappears with it.
This shows up across product teams too. Founders request “Apple-like” interfaces for products that don’t deserve that purity. Business teams propose logo redesigns without touching brand strategy. It’s all mimicry with no memory of why the original worked. Taste, when built this way, has no spine. It’s pretending to know what good looks like by replicating the last thing that worked for someone else.
2. Why this happens
Part of the problem is how easy tools have made things look finished. You can stack a few blocks in Figma, apply a decent type scale, and the result feels clean. Midjourney and AI renderers let you generate product ideas that look ready for a billboard even if they’re impossible to manufacture or difficult to maintain or service. When something looks good fast, it becomes harder to tell how shallow it really is.
This illusion encourages everyone to participate. Founders start giving aesthetic feedback. PMs curate “premium” reference decks. Someone always knows a “cooler” font. And because taste is treated like opinion rather than judgment, everyone assumes theirs is valid. I’ve seen business stakeholders veto a product decision based on a single AI generated shot, or push for a feature because it “feels more modern,” never considering the cost of implementation or the break in flow.
The root cause is a belief that taste is visible. That you can see it in the polish. But polish is the last 5%. Real taste shows up in what you choose to omit, in the order of decisions, in the friction you prevent before it happens.
3. The wrong methods
The most common mistake people make is studying only the final output. They look at the finished product and assume they’re learning how it was made. But they never see the ten discarded directions, the reversed decisions, or the moments where someone chose to simplify instead of add.
Another flawed method is collecting references without studying them. Designers and PMs save beautiful screenshots, packaging layouts, or industrial form factors, but rarely ask what tradeoffs made them possible. They don’t redraw the layout or replicate the mechanical structure. They don’t even list out the decisions they would’ve made differently.
When AI (generating visuals on command) enters this loop, the problem gets worse. The output might look good, but there’s no judgment behind it. No reason something is placed second instead of first. No idea what was deliberately omitted. Without judgment, you don’t develop taste. You just produce more plausible decoration.
Polishing too early is another common trap. Teams rush to styling before they’ve clarified the structure. AI accelerates this further. You can now get dozens of high-resolution outputs in seconds, all of which feel finished but none of which have actually solved the problem. You get the look of good taste without ever making a real decision.
The most subtle failure is when feedback becomes vague and aesthetic: “Can this feel more premium?” or “Let’s make it pop more.” AI tools can easily oblige but it doesn’t know what you’re building, what tradeoffs you’re willing to accept, or which constraint matters most. It’s not holding the whole system in its head. You are.
4: What is good taste
If you watch people who actually have taste, you’ll notice they’ve solved the same type of problem again and again. They’re not relying on instinct. They’re relying on accumulated judgment. They’ve seen which designs break, which flows confuse users, which finishes don’t survive manufacturing. Their taste is vocational. It’s not a sense of style. It’s a sense of what works, what holds, and what should be rejected even if it looks good.
That kind of judgment can’t be generated. It has to be earned. AI can give you thousands of variations. It can remix the best of everything. But it can’t tell you what matters in your specific case. It doesn’t know which features were hard-fought, or why you might leave something out because of cost.
One of the clearest signals of real taste is subtraction. People with taste remove more than they add. They cut steps, trim explanations, reduce options. They don’t chase cleverness. They chase clarity. A designer with taste will cut an onboarding screen instead of redesigning it. A hardware lead with taste won’t add LEDs for drama if a subtle indicator solves the job better. These aren’t aesthetic preferences which is the common misconception. They’re judgment calls that come from having seen what happens when you get it wrong.
And those calls get sharper under real-world constraints. If you’ve had to design around heat, service, packaging, or delivery time, your relationship with aesthetics changes. You start asking what’s necessary instead of what’s impressive. That’s the shift no AI will make on your behalf because it won’t be the one to take responsibility when the product doesn’t hold up.
5: How to build taste
I had written about this in detail here. Give it a read and come back.
If you want to build taste, stop studying portfolios and start studying the things you already use. Pick up your water bottle and ask why the cap stops where it does. Look at your laptop hinge and notice when it engages. Watch someone use a microwave and pay attention to which buttons they skip. Ask what makes something feel obvious, or invisible, or complete. Write it down. Refine it. Use it. You’re trying to build consistency in your judgment.
Recreate classics. Because they’re still around. Redesign the Braun calculator from scratch. Rebuild a Gmail thread UI using only text. Recreate a Native purifier tap mechanism and simplify it further. Don’t do it for a portfolio. Do it to understand why they work. If your version feels off, that’s the lesson. You’re not trying to match the shape because that's the easy part. You’re trying to reach the same level of decision clarity.
Don’t use AI to replace the hard parts. Use it to stretch your options. It’s fine to generate ideas, explore materials, or simulate layouts. But then you stop and ask: which one fits the real world you’re building for? Which one can be assembled, scaled, serviced, explained, repaired, and still feel right? That’s the judgment AI won’t give you. That’s where your taste starts to form.
6: In conclusion
If you’re in ops and want better packaging, don’t start with finishes or materials. Start by checking how the product actually fits in the box. Can a customer repack it without watching a video? Can a technician open it without damaging anything?
If you’re in any other function and giving design feedback, avoid saying things like “make it more premium” or “add some wow.” Instead, focus on what feels off. Are too many things fighting for attention? Is something over-explained? Does one feature take too many steps to activate? If you feel the need to give feedback, learn to articulate it.
If you’re building a hardware product, start with decisions. If a purifier has too many lights, decide which ones the user actually needs. If your smart lock has a screen, ask what it’s showing when no one is around. Clean interfaces come from subtracting.
If you’re using AI to generate ideas, don’t confuse variety with good judgment. Just because a design looks good in a render doesn’t mean it makes sense in the real world. Can it be manufactured? Will it survive 2 years of daily use in a dusty country? Can it be repaired if it breaks? AI can’t answer those. That’s your job.
Taste is how you make fewer decisions with more impact. The teams that build great products know what not to do. Apple taught us that. That’s what real taste looks like.