The Noise Around AI Design

It’s been around 5 months since I fully transitioned from architecture to product design, and it has been the most fruitful 5 months of my career. I got to experiment with many new tools and build for some really interesting problem spaces. Something that was common across all the companies I’ve worked with is the increasing demand for AI literacy.

Just because we are all given access to AI does not mean we are all going to be building deployable, high quality designs. I remember that when I first started exploring AI designs (in tandem with an entirely fresh lens to tech and product design as well due to my career pivot), I felt frustrated at the lack of direction I felt when vibe coding. A feeling that what I create will be generic and not actually tangible or consequential at the end. I was generating artifacts after artifacts, but I had no clear vision at the beginning to what I should be expecting, or if there was a “proper” way to prompt.

Lack of experience was a part of it. But something about me is that I will dig deep to learn how to overcome these information gaps. There’s so much stuff online to teach yourself everything (and the real good stuff is often not hidden behind paywalls.)

Seek to build something lasting, something replicable. A structure or framework. A top down approach. Of course, I’m thinking in terms of applying AI design to full blown products, infrastructure, and company design systems. Some would find it fun and satisfying just to explore without a clear end goal in mind.

There is value in exploring with more ambiguity and spontaneity. Personally, I find myself to be pretty goal oriented. If I were to make something, I’m going to research it, spend time talking to people about it, and build it. I leave the exploration to spontaneous side quests such as hackathons and designathons. This year, I did around 5 of those, and everything you’ve read so far was largely informed by my experiences partaking in these.

We won a few prizes and a couple thousand dollars along the way. What’s more is that I am extremely lucky to have shifted to UX at such an eventful time. There are tools and information that even senior designers are seeing for the first time.

I consulted for two YC startup this year; both had vibe-coded interfaces and demos. For one of them I spent an entire semester interviewing users, plucking out friction points with the current UX, and offering design recommendations. In the end, they’re spending a big sum of money on contracting a branding designer.

Taste matters. UX matters. AI is a tool to automate, prototype, and systemize designs. A serious business will only go so far with unadulterated AI artifacts. In the words of Figma’s founders; in an age where everyone can build and ship fast, taste and design becomes the biggest differentiator.