Everybody’s talking about vibe coding, not enough about vibe design.

FYI by “design” I’m referring to digital product design, here and henceforth.

First, what is “vibe coding”?

The most recent rise of vibe coding coined by Andrej Karpathy, is about coding with AI. Embracing the variables, accepting all AI suggestions and going with the flows to focus more on problem-solving than the “coding” aspect of it.

Coding is different from Software engineering.

One is an act of translating logic into a computer language, other is an approach to designing, building, and maintaining scalable and reliable software systems.

One is a skill, other is a process.

tweet screenshot with highlights — image source author

It’s like you’re talking to a friend about an app you want to build, the features it’ll have, how it will be the coolest app — and now guess what? As you were talking, it’s already built. Ready to use. You can continue the conversation if you want to change, but it’s all there at your fingertips.

Not to follow the hype, but I actually witnessed this first-hand.

My most recent belief and confirmation of this was my sister, a first-year college student who had just started watching “Intro to Python” on YouTube, with no previous experience, created a fully functional web app in hours. Just by chatting with Claude, because coding is English. 

Demo video. It’s not just a frontend, it stores the responses too.

screenshot of the web app — image source author

So then, what’s the equivalent in design?

Software engineering is a field in which coding is a part, the medium. There are many other parts, all to achieve the mission of scaling softwares — “building stuff” to make it reliable, cheaper and faster. Similarly, product design is a field with the main mission of personalization or scaling while being inclusive. There are four major parts in the process (also the areas where AI can be implemented) — 

> Research,
> UX, 
> UI/Prototyping,
> Taste

image source author

And what’s the part that’s like coding? UI/Prototyping.

The part where AI will generate aesthetics, or for the pun of it all — vibes.

I believe everyone has come to a consensus that “taste” should be left to the humans. Or maybe, it’s a temporary belief to feel relevant? Anyways...

I’ll come back to UX later. Focusing on the UI/Prototyping for now, we have already seen something similar happen in the graphic design field. The coding-like aspect of creating visuals is now done in English rather than a pen tool via Midjourney or Visual Electric.

Now it’s product design’s turn.

What is “vibe design”?

You’ve probably guessed it by now. 

Instead of focusing on *current* figma processes; the tokens, the pixels or playing Tetris, vibe designers will focus on the full presentation/feel.

AI could roughly understand and change the feel of the web app when my sister said, “it feels a little dry”.

It’ll be mixing UX, UI and interactions to create fully testable flows.

Instead of static screens, they focus on the interactions. 

Vibe designers can create mini components that can be directly shared with devs to be plugged in and deployed. 

Mockups existed because coding was difficult, but now that it’s easy, at least to create mockup-level prototypes, that’s going to be the future. As a designer, I’ve always only been able to critique better when I experience the feel. 

There is now a need for a new tool, 

where designers can play around with code as they do with rectangles and instead of adding colour, they add functions and behaviours. It’s almost like designing a game in Unity but for 2D software. (I am literally very close to just start making it myself since I haven’t yet come across one. Isn’t it obvious?)

Now imagine how powerful designers will become.

There’s a little caveat.

As there always is. While UX and UI/Prototyping merge into one, and research and UX also merge into one, there would come a need where UX is needed on both sides but not fully focused on by either.

If vibe design is the future, how can we ensure UX quality?

Psst… that’s where my *new research* comes in.

Foresaw a trend a while back, 6 months ago to be precise, that the rise of software creation, i.e., being able to create fully functional apps in minutes, will lead to a craving to be differentiated and the focus on user experience will grow. Every product would want to be more personalized for their users, more engaging, keeping them put and of course, abiding by the design principles of aesthetic-ness.

So at the end of the day, everyone would need a design mindset.

How do you democratize this mindset? 

What does it even consist of? 

Can this mindset be plugged and played?

Does everyone need to be a designer?

That’s what I’ve been thinking about for the past few months.

I’m calling this mindset “Design Intelligence”.

I believe there’s a way for this mindset, “design intelligence”, to be used by everyone, even those who are not designers. AI enables us to do that.

Will be going more in-depth on this here. But if this direction of thought excites you or you’ve already been thinking towards, would love to chat — Schedule a jam session with me :) 

Coming back to vibe design.

What can designers do today?

Prototype. Simple.

I’m seeing a trend of many designers already using Claude, Bolt, or other new AI tools to just create prototypes. Share their ideas in “flows” rather than static screens.

As Ben South put it well “Designers handing static images to developers will soon feel as dated as drawing rounded corners pixel by pixel (yes we had to do this)”.

My own workflow is changing. I create all the animated components of the landing page and my cofounder just plugs and deploys. I could have used video editors, or Figma static screens with ‘smart animate’, but it was way easier and faster to have Cursor make it and I just tweaked it a little.

Having worked at different scales, I think I can help with how to get started — 

• Freelancers MUST prototype to share their ideas. Use Claude + Cursor. Period.

• Designers in small teams SHOULD prototype. They mostly handle the research, usability testing and execution. So sometimes quick Figjam/sketch mockups are needed, but all later stages should be prototyped for a faster handoff to the dev team. You can just share fully-vibe-fitted components.

• Designers in large organizations COULD prototype. It’s more of a proactive brownie point. Creating prototypes in your spare time, of similar things you did at work, to accelerate the integration of AI at your firm would go a long way! That’s what I used to do. Create prototypes to get the chatter going, and form AI use cases for our design team, leading to a faster AI integration for us to experiment.

For a lil motivation, my sister’s doing it and she’s not even career-focused yet. And if I can get my sister to do it, you sure can start by yourself.

Lastly, the case against.

The case where vibe designing fails — big companies focused on designing for everyone rather than shipping out ideas. Jess puts it perfectly and I somewhat agree. Will the tools stay as are in the future? No. Will the same methods stay? Yes.

tweet on why vibe designing isn’t for everyone — image source author

Vibe design requires a new tool. Figma is for the current design process, which is changing.

Right now though, there’s a place for both sides to exist, and both types of designers to create. Just pick one you like and go wild. Rest is transferrable.

HMU if you want to bounce ideas or just argue