Yesterday I was on an internal video call, where our designer showed me a fully working prototype, that he had created with help of Claude Code. A timeline component with bars you could drag to resize. He grabbed the right edge of a bar on that timeline and pulled it left toward the minimum width. First the total days label changed it's format to strip the day unit and then the handles disappeared. He said: "Not a bug. They were in the way at that scale, so I wanted them to step aside."
That kind of thinking usually only shows up after someone files a ticket. Here it was already solved, before the first line of production code.
He's a designer and this was kind of the first time he was exposing himself to Claude Code. He built it because Figma couldn't show what he was trying to show.
I watched a video from the Cursor team around the same time, about their internal workflows. They were talking about how the team hierarchy is flattening, that everyone is contributing in ways that simply weren't possible before. A feature pitch isn't always a written spec anymore. Sometimes it's a working prototype, built by a PM to show the thing they're describing. That used to be a different person's job, or it didn't happen at all.
Our designer's prototype is the same thing. A new form of expression. A way to show what words and static frames can't.
When I first received these kind of prototypes, it felt threatening. I'm a developer. If the designer is building the prototype and the PM is shipping the demo, what's left for me?
What I eventually landed on is that the prototype isn't the product. What it takes to ship something stable and maintainable, something that fits into a technical architecture and a company's long-term roadmap — that doesn't come from building fully functioning prototype with perfect drag interaction in an afternoon. It comes from years of building things that break, and learning why.
The designer showed me exactly what the interaction should feel like. That's real, useful, skilled work. It's just not the same as the work of shipping and running in production.
I have no idea, where it all goes from here. Some people I talk to are scared and unsettled about whether their role survives this. I don't have a reassuring answer. Things are moving fast enough that no one's map stays accurate for very long.
I genuinely enjoy working like this. Building a feature, playing with it, feeling how it actually feels, and then shaping it from what you discover rather than from what you planned.
I didn't expect that.