Vibe Designing: Are digital designers obsolete?

You've probably noticed AI is everywhere now. It writes code, generates images and video, runs data analysis. So what happens when you point it at UI design? With new software promising to produce whole interfaces from a prompt, the uncomfortable question is worth asking: are UI designers becoming obsolete?

I spent a year trying to find out. I tested a handful of AI UI generators, Figma AI, Stitch, Pencil, MagicPath, and Banani, on conversion-focused landing pages, logic-heavy dashboards, and brand consistency redesigns. I went in hoping AI wasn't about to replace me. The results were reassuring, and they pushed me to rethink how I work. They also gave me a term I now use constantly: "vibe designing."

Generic prompts don't tell you much. I needed real design problems with real constraints, so I gave them three.

1. Creativity. Design a high-converting landing page for a brand I picked at random, The Honey Department. I wanted to see how much visual storytelling, typography sense, and actual creativity the tools brought.

2. Logic and hierarchy. UI design lives or dies by information architecture. I asked the tools to build a dense desktop "Command Center" for a logistics firm tracking 500+ active assets. Could a data-heavy screen still feel breathable?

3. Consistency. Could AI act as a design critic? I gave it an existing e-commerce product page, Sponser's Collagen, and asked it to improve the design and the conversion rate without trampling the brand.

Here's what I took away.

The new wave of design tools is pulling us away from manually clicking and dragging rectangles around a canvas. Design software is catching the "vibe coding" wave, and in our world that becomes "vibe designing."

I'd suggest adopting the term. Vibe designing is when you act as the creative director instead of the pixel pusher. You describe the intention and the "feel" of an interface in natural language (or with visual references), and an AI handles the implementation. Your job is to set the direction, watch what comes back, and make sure it actually solves the problem.

It's tempting to assume AI is supposed to spit out a perfect, production-ready screen on the first try. It can't, and that's not really where it's useful anyway. What it's actually good at is fast exploration.

Tools like Stitch and Banani are useful for generating variations. You can explore an idea without starting from a blank file, and A/B testing gets a lot less painful. They speed up wireframing and shorten the gap between iterations. Some of them, Stitch in particular, even produce predictive heatmaps, which is a surprisingly handy way to sanity-check a layout before building it out.

The blank-page problem mostly disappears, and your iteration loop gets tighter.

The current AI design landscape is messy and the tools are wildly uneven.

Figma AI is fine for structure, but the designs come out flat and personality-free. Right now I wouldn't pick it as a standalone generator. Other toolkits do that part better.

MagicPath is more impressive. The easiest way to describe it: what you'd get if Figma and Lovable had a baby. The toolkit is flexible and the canvas handles multi-screen design well.

Google's Stitch (formerly Galileo) puts out denser, more detailed work and gives you several versions to pick from. They ship updates constantly, so it's worth keeping an eye on. Their Twitter is the easiest way to follow what's new.

There's no "should I use AI" decision anymore. The decision is which specific tool you reach for at which point in the process, and getting that right is starting to matter as much as your design fundamentals.

So is it time to walk away from digital design? No.

What I came away with is that AI is a partner. It can do the technical setup and generate layouts, but it can't be the creative director. Someone has to look at what it produces, throw out what doesn't work, and shape the rest into something that fits the brand.

These are tools. New ones, and good ones, but tools.

Vibe designing doesn't kill the designer role. It moves it. We lean further into creative direction, use AI for fast prototyping and variations, and spend more of our time on the parts of the work that need a human: taste, judgment, and the emotional layer machine output can't fake yet.

This isn't a moment to panic. It's a moment to ask how your strategic thinking and creative direction need to evolve.

These tools have interesting features. They're not going to replace us. The output still feels flat, often visibly machine-made, and it's missing emotion, which is the thing that makes design land. First results can look better than they really are, but if you dig past the surface there's usually something useful in there: a starting point, an angle you hadn't considered. Where these tools shine is speed on complex, content-heavy layouts. They'll get you a usable first pass much faster than starting from scratch. Consistency on existing designs is still where they fall apart.

If you need an average result, sure, go ahead and use them. But somebody is still going to spend real time vibe designing the work. If you want something good, hire a designer. They'll know how to prompt, and more importantly, they'll know which output to throw away.

If you want more real-world tests, tool breakdowns, and practical takes on where design and AI are going, subscribe to the Double Tap newsletter.