We are living in a time when artificial intelligence can generate interfaces, components, and even entire products based on a prompt. But amid this rapid pace of change, a key question arises:
If anything can be built, how do we decide what’s worth building?
With the new integration between Claude Code and Figma (via Figma MCP), it’s now possible to send the browser’s rendered state directly to Figma and convert it into fully editable layers. Just type “Send this to Figma,” and the code is transformed into a design you can edit.
Figma reinforces a powerful idea: in a world where AI can create almost anything, the real value lies in judgment, craftsmanship, and perspective. The problem is no longer a shortage of ideas. The problem is an excess of possibilities.
An IDE or command prompt is great for production. But the canvas is better for exploring and iterating on your design.
In Figma, you can
: – Compare approaches side by side.
– Think divergently.
– See the big picture.
– Tweak details with direct manipulation. –
Refine before committing.
And then, once the design is clear, you can feed the changes back into the code. This workflow breaks away from the traditional linear model (brainstorm → design → development). Now you can start at any point: terminal, sketch, prompt, or visual UI.
In my opinion, AI has a silent side effect: it generates momentum. It’s very easy to keep iterating without stopping to ask ourselves if we’re really building the right solution. The first working version often ends up being the final version—not because it’s the best, but because it already exists.
AI is giving us superpowers when it comes to production.
But the true “senior” level in this new era won’t be the one who produces the fastest, but the one who selects the best. In a space of infinite possibilities, design is no longer just about execution—it’s about direction once again. And that, more than ever, is a strategic responsibility of the designer or programmer.
