Why I Build AI Prototypes in Code, Not Figma

by Jojo && Aavi · 2025-07-21

After thirty years of pixel pushing, I still love a clean grid and a beautiful type ramp.
But these days, design lives somewhere deeper — in behavior, timing, and intent.
That’s why I build AI prototypes in code, not Figma.

Figma is incredible for what it does: sketching structure, shaping hierarchy, building shared language.
But it freezes intelligence into pictures — like a stop-motion animation of something that’s supposed to breathe.
Every frame is meticulously crafted, yet lifeless between the gaps. You can see motion, but never feel momentum.
And the process itself is tedious — dragging each pixel into place to imitate what real systems already know how to do in motion.

When you prototype in code, you’re not faking the dance — you’re running it through the actual (or simulated) pipeline.
You hear how the model thinks, how latency feels, how feedback loops form.
Instead of animating the illusion of intelligence, you experience its behavior directly.

AI design is temporal. It unfolds in time, not space. You don’t just draw how it looks — you listen to how it feels.

When I code a prototype, I’m not “making a mockup.” I’m creating a living fragment of intent — a small world where timing, emotion, and language can collide.
That’s where real UX truth lives: in the milliseconds between prompt and response, in the weight of a pause, in how uncertainty is handled.

After decades of shipping interfaces, I’ve learned the best ones don’t need explanation. They reveal themselves through use.
That’s what code does: it collapses the gap between design and experience.
When a prototype runs, everyone — designers, engineers, clients, stakeholders — experiences the same reality.
No translation. No “handoff.” Just one shared loop of feedback and curiosity.

Prototypes like these carry intent at the atomic level.
Every line of code asks: what does this intelligence sense, decide, and express?
How does it respond when it doesn’t know?
That’s design now — not in the arrangement of pixels, but in the choreography of understanding.

“After three decades of design, I’ve stopped drawing interfaces. I build behaviors that show me what the interface wants to become.”

When I look back, the work that mattered most wasn’t always the most polished — it was the most alive.
You don’t just sculpt the surface; you collaborate with emergence.
That’s the designer’s new frontier — not just to build for intelligence, but to create with it.