Something has changed in how things get made. A sentence turns into a screen. A brand can start on a phone, on the way home. For most of the history of design, making was the hard part: the hours in the file, the craft of getting every edge right. Now making is cheap, and the hard part is choosing.
That shift is why Kite exists. These are the things we believe about where design is going.
Making things is easy now. Knowing what's worth making is the rest of it.
When a layout takes seconds, you can have twenty of them by lunch. That sounds like freedom, and it is, but it moves the work. The question is no longer “can we make this?” It's “which of these is right, and why?”
That question was always the real job. It used to hide behind the effort of producing things. Now there's nowhere for it to hide.
Design is splitting in two.
The drawing is going to agents, and it's already on its way: the layouts, the resizes, the twentieth version of the same banner. What stays with people is the part that was always theirs: knowing what's worth making, and knowing when it's right.
For now, that means working side by side. The agent is fast and has no taste. The person has taste and not enough hours. Each needs the other, and most tools haven't caught up. Designers paste screenshots into chats; agents describe designs they can't actually draw.
Taste doesn't scale on its own. Systems do.
One person's judgment can't sit beside every screen a team ships, let alone every screen an agent draws at three in the morning. A design system can. It is how one person's taste reaches a thousand screens they never touched.
That makes the system more important than it has ever been. It stops being a style guide people forget to open, and becomes the first thing every agent reads before it draws.
When drawing is cheap, critique is the craft.
If an agent can produce ten directions in a minute, the valuable skill is looking at the ten and saying which one is right, and why, in words precise enough to act on. Good critique was always part of design. It is becoming most of it.
That changes what a design review is. It isn't the end of the work anymore. It's how the work gets directed.
One canvas, no handoff.
The future needs a different kind of place to work. Not a chat with a canvas bolted on, or a canvas with a chatbot in the corner, but one surface where people and agents work on the same thing. Whatever a person can do, an agent can do too, in the same file, where everyone can see it, change it and talk about it.
Bring your own agent.
The best agent will keep changing. The one that's best this month won't be the best next spring. Your tools shouldn't change every time it does, and your work shouldn't be locked to whichever model a design tool happened to pick.
The designer isn't going away. The job is moving up a level.
The balance will shift. Agents will make most of the pixels, and people will make most of the decisions: what the system is, what the brand stands for, what ships and what doesn't. Being a designer will become less technical and more creative.
We don't think the fully automated version is close, and we're not sure it should be. For a long while, the best work will come from people with a point of view, working with fast, tireless collaborators who don't have one.
For the years in between
Kite is our attempt at a canvas for those years. First you design next to your agents. Then, mostly, you direct them. The short version of all this is our manifesto.