MTV launched on August 1, 1981 with "Video Killed the Radio Star." Fitting title, wrong prediction. Video didn't kill radio, it created something entirely new. But not right away.
For the first few years, most music videos were just artists lip-syncing in front of a camera, radio with pictures. The medium was new, but the thinking was old. It took directors like Michel Gondry and Spike Jonze to show what the format could actually do, a grammar and a set of possibilities that had nothing to do with radio.
Every technology shift follows this pattern. First you use the new tool to do the old thing. Then, slowly, you discover the old thing was shaped by the old tool's limitations.
We are at that exact moment with AI and design.
The process was built around scarcity
The design process as we know it exists to protect one thing: expensive development time. You research first because building without research wastes months. You wireframe before you prototype because prototypes cost real capacity. Every phase is scaffolding around a scarce resource.
AI doesn't speed that process up, it makes the scarce resource abundant. When a working prototype can exist by the end of a single conversation, the scaffolding doesn't protect anything anymore, it just slows you down.
Which means the value shifts fast to what remains scarce. Judgment. Empathy. Taste. The ability to look at a user and understand what they actually need, not what they say they need. Those capabilities don't get automated, they get more important.
Building in the room
So what does the new way of working look like? Fewer phases.
Instead of researching, briefing, wireframing, prototyping, and testing in sequence, you sit with the person who has the problem. You listen. You watch. And then you build something while they're still in the room. Not a mockup, a working thing they can react to.
The feedback loop collapses from weeks to minutes. The prototype stops being the deliverable and starts being the question: is this what you meant?
Three things to get right before you try this.
First, watch before you build. Spend twenty minutes observing someone do the work you're about to redesign. Not interviewing them, watching. People describe their work the way they think it should be. They do their work the way it actually is. The gap between those two is where the opportunity lives.
Second, test the value before you test the technology. If a human expert performed this task flawlessly, would the user care? Would it change their day? If the answer is no, building it faster won't help. Don't automate indifference.
Third, give someone the authority to say no. The speed of AI-assisted building creates momentum that's hard to stop. Someone needs to be able to say "this isn't what we need" without feeling like they're killing progress. Catching a wrong direction in minute fifteen isn't waste, it's the point.
What gets built next
Somewhere out there, someone is about to design something that couldn't have existed before. Not a faster version of what we already make, something with its own logic, its own possibilities, shaped entirely by what this medium can do. The way a music video wasn't a better radio broadcast, it was a different thing entirely.
That's what I'm waiting for. And I think it starts with the people willing to stop protecting the old process and start finding out what this one is actually capable of.