Recently I started doing design sprints for myself. The idea is very simple: take an idea, block 1 hour and work on it, then stop once that time is up and share it somewhere.

The sharing part is really just part of a broader attempt to overcome a fear of visibility and sharing my work. The broader point of these sprints is to keep in touch with the design canvas in a world that increasingly tells designers that LLMs will make the canvas obsolete (I don't believe that it will, yet). I've spent the last few weeks working on the production side of things, collaborating with engineers to translate design to code via LLMs, and I can already begin to feel the knowledge of auto layouts in Figma seeping away from me.

The picture below is the first of these sprints, it's a landing page for a fictional fintech named "Motion." For this one, I used Paper instead of Figma as my canvas. I'm a big fan of Paper, but I always use an LLM to prompt it, and I've never tried to make anything with my own hands from scratch. I'm glad I did because not only did I learn so much about the Paper canvas layout, I also discovered a lot of interesting tools (the "create SVG" tool for example).

With all that being said, I think the most interesting takeaway from this exercise for me is something that has been repeated to the point of cliché: static assets don't quite capture the full feel of digital experiences. I think this is a good cliché, and the senior designer who recommends you add a little motion to your portfolio has a point.

I look at this landing page, and I think it's a nice starting point for something that could become an actual product, but the image does not even begin to capture the full range of expression the final site should capture. What happens when the user scrolls? Hover states for the buttons? Should the hero image be an image at all? A hero video instead, perhaps? What should happen when the user hovers over this hero image/video?

The one thing these things all have in common is motion. The digital experience has motion at its core, the basic thing we do all day is drag a cursor around and react to what is in front of us based on changes that happen when that cursor moves. (Or flick a screen with our thumb and react based on the changes that occur when the screen is flicked. Pohtato. Potahto.)

And I think this is why static assets just don't cut it (of course, they do have their uses) for showcasing the full range of the digital experience. Nothing is moving in a screenshot.

I suspect this very reason is why a lot of designers were fascinated by LLMs to begin with. In Figma, you draw a frame, and you then have to draw a couple of lines and do some manipulation to simulate what happens when the user hovers over a button. With Claude, you just say, "switch background colour to accent on hover" — it's an incredibly heady feeling.