Welcome to Unknown Arts — I’m Patrick, your field guide to tech’s creative frontier. Join thousands of builders around the world navigating what’s next.
I started writing online in January 2022, about ten months before ChatGPT came out.
At the time, I felt like building my own software wasn’t possible. I had started my career as a front-end developer, but I had been out of the coding game for more than half a decade. If I had once been fluent in the language of code, I wasn’t anymore. I was stumbling to put sentences together, and getting that fluency back felt too daunting. I didn’t have the time or energy to re-educate myself to the extent necessary to build the way I wanted.
Writing felt more approachable. It wasn’t easy, but I knew I could start doing it immediately. I could get my ideas about making software down on the page and share them without running into a technical blocker before I even started.
That was the real reason I started writing: it was the most direct way I could create something of my own and put it into the world. I couldn’t build software in the way I wanted to, but I could write about it.
And that same logic explains why I’ve been writing less recently.
Because building became accessible to me again.
Ironically, it’s often easier to build the thing now than it is to write about it.
ChatGPT started this transition, Cursor got me back into an IDE, and agentic coding tools like Claude Code, Codex, and Cursor Agents fully flipped the script. It still takes plenty of trial and error, but I can build and ship code again. It’s honestly gone way beyond simply recovering an old skill. I can now build things that would have been impossible for me to make even at the height of my coding days.
It’s not hyperbole to say that the way I do my day-to-day job as a software designer has changed more this year than it has at any other point in my 14-year career. It’s been an exhilarating and, at times, exhausting learning journey, but I feel like I’ve gone far enough down the rabbit hole to come up for air and report back.
Since January, I’ve been putting almost all of my extra time and headspace into building an AI-assisted design prototyping tool for my team at Sublime.
The initial problem was simple: my team was making useful but one-off prototypes in Claude Artifacts and we didn’t have a good way to share them and build on them over time. Our collective effort wasn’t compounding.
So I built a very simple shared environment for prototyping designs in React code. It didn’t do much or look like much to begin with, but it added enough value to motivate me to keep working on it.
Over time, it’s grown from that simple shared environment for me and one other product designer into a tool used by designers across product and marketing, as well as by product managers. While there are still gaps to fill, it has become the first tool I reach for when I’m doing lo-fi, wireframe-style exploration, making high-fidelity prototypes, synthesizing customer research, or even writing a design brief.
This work has changed how I think about design, software, and my own career. Building with code is no longer something I’m trying to get back to someday. It’s central to my daily work once again. As I keep pulling on the thread of my curiosity, I’m learning what this new era makes possible for designers, and what it demands of them.
The tool is now reaching a level of maturity where I can step back from the trenches a little and start documenting what I’ve learned. I’m still writing. I’m just writing from a different place now.
In the weeks ahead, I’ll share more about what I’ve been building, how it works, and what it might mean for how design gets done going forward.
Until next time,
Patrick
Find this useful? Share it with someone who might also get value.
Not subscribed yet? Get my writing directly — 7k+ already do.
Great to see you're back!