From the garden · Budding
What AI actually changes about design tools
Not everything changes. But the things that do change are significant.
There’s a lot of noise about AI and design. Most of it is wrong in one of two directions: either AI is going to replace designers entirely, or it’s just a fancy autocomplete that doesn’t change anything fundamental.
The truth is more interesting and more specific.
What changes
The cost of exploration drops dramatically. Generating ten variations of a layout used to take hours. Now it takes minutes. This doesn’t replace design judgment — it accelerates the front end of the design process, making it cheaper to explore before committing.
The artifact shifts. For a long time, the Figma file was the primary deliverable. AI tools are starting to make the prompt (or the specification) equally important. Knowing how to describe what you want precisely enough to generate something useful is a design skill.
Feedback loops get shorter. When you can go from idea to prototype in minutes, you can test more hypotheses in less time. The constraint moves from “how fast can I build this?” to “how fast can I learn from it?”
What doesn’t change
Taste. AI can generate variations, but it can’t tell you which one is right. That still requires a human with context, judgment, and an understanding of the user.
The hard design problems. The hardest design problems aren’t about pixels. They’re about understanding what users actually need, navigating organizational constraints, and making bets about the future. AI doesn’t help much with any of that.
The relationship with engineers. Great product design still requires close collaboration with engineering. If anything, the ability to prototype more quickly raises the importance of that relationship — because more ideas will make it to “should we actually build this?” faster.
Where I’m landing
I use AI tools as thought partners and velocity multipliers, not replacements for design judgment. The designers who will thrive are the ones who learn to use these tools precisely — knowing when to use them, when to put them down, and how to evaluate what they produce.