Design is making decisions
Design is not a static asset. Design is making decisions.
Design has always been around. If you're planning something, if you're making decisions about how a thing should work, you're designing. What changed over the years was never that. It was the tools.
And new tools usually mean new processes, sometimes changing completely the way we approach our work. With AI at hand we're at one of those moments.
It enables us to do what we do best: think, orchestrate, take the time to process all the information we acquired. This is where product design stops being a professional box drawer. The floor is elevated. Producing a competent screen is free now, so the gap moved: it's challenging the status quo, asking whether the thing is worth building, whether the flow everyone accepted three years ago still makes sense, whether the constraint you were handed is real.
That's uncomfortable if turning a spec into screens was your edge.
And it changed so much that now we might even start with the prototype, to help us think about the edge cases and see what the initial look actually feels like, so we can move to testing and validating. We were always learning. The difference is when. It used to happen after we committed, when changing course meant throwing away work someone already paid for. Now it happens before anyone has spent anything.
The communication isn't at the end anymore. That's the whole advantage of prototyping first. The thing you build on day one is what gets everyone aligned, and everything after that is a conversation about something real.
Polish comes last, and it's the hardest part now. When everyone can generate something that works, the difference between a well designed product and everything else is the final ten percent. The transition that feels right instead of merely being there. The empty state someone actually thought about. The copy that isn't a placeholder. That part didn't get cheaper. It got more visible, because the baseline everyone else is shipping is competent and forgettable.
Plato had one ideal world we were all reaching for. The problem in product work is that everyone has their own. Six people leave the same meeting agreeing on the same words, each holding a different picture of what they just agreed to, and nobody finds out until weeks later. A prototype drags those private versions into the open. Now there's one thing on the table to point at and argue about.
One thing is still missing: collaborating. Commenting, sharing, versioning.
Developer tools aren't meant for that kind of fast iteration, they're meant for production code, and besides wasting tokens it makes everything slower. Every version lives somewhere different. Feedback arrives as a screenshot in Slack with no context around it. You end up re-explaining the same thing to your agent because the conversation is scattered across four places.
That's the part we're fixing with Artor. If design is making decisions, the decisions should live next to the thing they were made about. Not in a thread that scrolled past, not in a screenshot nobody can find. Every version in one place, every comment attached to what it's actually about.
The process already changed. Most of us are still running it on tools built for the old one.
