Blending new and old product design

It's been a year since I called the death of Figma. I'm still not sure if I was right, but the stock market seems to agree with me at the moment.

Since then I've spent most of my time designing, and developing, with Claude Code.

I've done a lot in that time. Built projects, shipped work, tested, collaborated, and done more than ever before. I also think that it's been better work, more interesting work, work that's been more thoroughly used and tried and pushed—because so much of it has been done directly in the browser, in the native app, in an environment matching what the end user would see—and not locked to a static Figma canvas.

Now, with the latest Figma update (still in beta and I don't have access yet), it looks like they're pushing back, hard. I'm not sure what that means yet. I sort of think the ship has sailed on that front, but I'll give it a solid try once it's available.

Something I've noticed though—as I've fully embraced a world of developing design locally, as opposed to laying out components in frames and hoping developers can interpret the meaning when it comes time to move into production—is that I'm yearning to stretch my design muscles in ways that can't be fully utilized with this new way of designing.

I spent the last decade refining my ability to sketch, to think with a pen, to imagine interfaces and bring them to reality with tools like Freeform. With the last year of Claude Code I've mostly paused that work, preferring instead to converse back and forth with an agent until I reach a point where I'm happy. That's been fine, I've managed to design novel interfaces, follow best practices, and create some really good work.

But, sometimes it's like pushing against a wet noodle. Sometimes it's not enough. So, in the last few weeks I've started reintegrating Freeform on iPad back into my Claude Code flow. When I need to work on an interface I'll start a conversation with Claude, but then turn to pen and paper to flush out how the pieces fit together.

Because LLMs are just a bag of words, an average of all the information instead them, they'll manage to create average interfaces. Sometimes that's fine, but it's not the type of work I want to be throwing myself into year after year. I don't want to just make average. That's where adding my skillset of laying out pages the old way matters.

And that's exciting. I'm trying to figure out how to pull these two together—an old method of product design into a new tool with Claude Code. And that's got me excited. I think there's new ground to tread here, and in the coming year an opportunity to do some of my best work.