Cher's Closet Algorithm: Why Taste Still Beats Tools

You know that scene in Clueless where Cher's computer generates outfit combinations? I wanted that thing my entire life. As a kid watching that movie, I was mesmerized. A computer that understands aesthetics? That generates options based on coordinated taste? That's not just a closet system — that's peak technology. Adult me still wants it. But here's the thing: the computer generates 50 variations instantly. Technically coordinated. But then, a mismatch error. The computer can generate a lot of options, but it doesn't always get it right. That's the thing, though. Cher doesn't just grab the first acceptable outfit. She stands there looking at the screen, thinking. Evaluating. Because there's a difference between "this technically works" and "this is actually right." That's taste. As if she'd pick the first one.

Speed used to be the actual problem.

You'd sit there waiting. Weeks. Feedback cycles that moved like molasses. Engineering estimates that change every Friday. A three-week design revision because someone had an opinion at 4:47 pm. We all got very good at speed optimization because we had no choice — the bottleneck was time, and it showed no mercy. Then AI showed up and dramatically loosened the constraint. And honestly? It's sick. You generate 10 layout variations in 30 seconds. You mock up interactions in 2 minutes. You prototype features without touching code. A timeline that used to be weeks is now hours. Hours are now minutes. The pixel pushing is absurdly fast. The grunt work got lighter. The tools are genuinely incredible. It freed up so much space. Which created a completely different problem.

Here's what's actually happening in product design right now.

The tools can generate 50 technically perfect button variations. 50 color palettes. 50 layout options. All of them will work. Perfectly coordinated. Balanced. Aligned to the grid. But taste is what looks at those 50 and knows. Knows which one fits the problem. Which one feels considered? Which one actually solves for the person on the other end of the interface, not just the grid system? It's exactly what Cher is doing. She's not overwhelmed by the options. She's not picking the first good one. She's applying judgment — context, experience, an understanding of what "right" looks like — to a set of technically perfect choices. You can't rush taste. Can't shortcut it. It lives in thinking about why a choice matters, what you're trading off, and what you're giving up. It's the difference between a working surface and one that actually feels good to use. The designers I know with real taste spent years building it. Not in a romantic way. In a "I've made a thousand small decisions and lived with the consequences" way. That slow work looks like:

Sitting in a research session and watching someone click the wrong button because of something you designed. Feeling that sting. Letting it rewire how you think about hierarchy.
Showing work to your fellow designers and them asking "but why?" until you realize you don't have an answer. Being forced to go back and actually earn the choice.
A mentor pointing out a detail you missed — a micro-alignment that kills the whole thing, a color relationship that's just off — and suddenly your brain learns to see it.
Sleeping on a ship decision at 5 pm. Waking up and realizing it's wrong. Having time to fix it.

That's deliberation. That's friction. That's the weird discipline of refusing to move forward until it actually feels right. It's the compounding effect of a thousand moments where you chose careful over fast. The brutal part: when speed was scarce, everyone was forced to do this. You had to think deeply because you couldn't afford endlessly iterating. Now you absolutely can. And most people do. They're generating sporadically — throwing options at the wall without really thinking about why. 50 options, 5 of them might work, 45 of them are noise. Most designers are just Cher's computer at this point. Generating options at infinite speed, but without the judgment to choose.

But the people with taste use the same tools differently.

They generate 30 options and know which one actually matters. They can prototype 5 variations in 30 minutes because they understand the problem well enough to test the right things. The fast tools don't slow them down — they amplify what they already think. They're being Cher. Standing in front of the screen. Looking. Choosing. Here's where it gets interesting: the market is starting to notice. The products that feel intentional instead of optimized. The interfaces where every detail makes sense, not just technically works. The experiences that feel like someone chose things carefully instead of iterating until something stuck. Those products stand out. They feel different. That's not luck. That's taste. And in a world where literally anything can be made fast, the scarce thing isn't speed anymore. It's judgment. It's the willingness to slow down. It's knowing the difference between "fine" and "right." The luxury now is time to actually think. The constraint is having the taste to use it well.

Cher would be buggin' about this.

Because here's the thing about Cher's closet moment: she's not just picking an outfit. She's looking at the computer's answer and going, "No. Try again." She's rejecting perfection on paper because she knows something the algorithm doesn't. She has context. Experience. Judgment. She'd never look at those 50 options and just grab the first one. Ugh, as if she would ever. The tools can generate 50 options. They can make them all technically correct. But they can't make the call. And the moment you ship something just because it's the first acceptable thing the computer generated? The moment you stop standing there, thinking, choosing? That's when you start looking like you got dressed in the dark. And worse — that's when your users start noticing it too. They can't articulate why, but something feels off. Something feels like it wasn't chosen. Like it was just generated. Like you're using taste sporadically instead of actually applying it. She'd stand there as long as it took. She'd look at all 50 options. She'd say "mismatch error" to the ones that didn't work. She'd reject the technically perfect ones because they weren't right. And she'd keep going until she found the one that was actually right. That's the difference between a designer and an output machine.

Here's the irony.

The closet computer Cher had in 1995 is finally real. We all have it now. Infinitely fast. Infinitely many options. Infinitely generating. And somehow we're all just grabbing the first thing that works and shipping it. Your users don't know why, but they sense it. Something's off. You look like you got dressed in the dark. And that's the design they'll remember you for. The tools are incredible. But they're not the designer. You're standing in front of 50 options right now. At least one of them is wrong. Maybe all of them are. Your job isn't to generate faster. Your job is to know the difference.