The Method Before the Philosophy
I spent three years working on a design system where every component had to feel both familiar and new. The team kept going in circles because we were trying to invent everything from scratch. It was slow and the work suffered. Someone finally suggested we reverse-engineer existing systems instead, and it changed the entire process. What we ended up building was heavily influenced by Material Design, the Apple Human Interface Guidelines, and a few open-source UI libraries. But we didn't copy anything directly. We broke down why certain patterns worked, then rebuilt them to fit our specific constraints. The concept itself isn't complicated. Austin Kleon popularized it with the idea that nothing is entirely original, so you should actively study the work you admire and make it your own through transformation. The real practice is more granular than that. You don't just absorb work casually. You deconstruct it intentionally. I keep a simple reference system where I catalog pieces of work I connect with, note what specifically draws me in, and then document how I might apply those mechanics to my own projects. The deconstruction phase is the part most people skip because it feels like homework. It's also the only part that matters. When I was working on a custom CMS for a client who ran a publishing platform, I encountered a specific bottleneck. The client wanted the article layout to feel modern but also carry a sense of tradition. Every template I designed felt generic. I ended up pulling apart the typography and grid systems from five different editorial designs spanning from The New Yorker to Medium to Substack. I mapped out their type scale ratios, line lengths, and spacing conventions in a spreadsheet. The breakthrough came when I noticed all of them followed a consistent mathematical relationship between heading size and body copy that I'd never paid attention to before. I used that exact ratio in my own design but with a completely different typeface pairing. The result looked intentional instead of templated.
The actual workflow I use has five steps. First, collect. Save or screenshot work that makes you pause. Second, annotate. Write down exactly what element caught your attention — is it the color combination, the spacing, the hierarchy, the tone. Third, isolate. Strip that element away from its original context in your notes. Fourth, translate. Rebuild that element using your own tools, constraints, and subject matter. Fifth, iterate. Repeat the process with different source material until the result feels like yours. I've seen people misuse this approach in ways that get them in real trouble. The most common mistake is stopping at step three and calling it inspiration. Copying a layout structure without understanding why it works is just plagiarism dressed up as homage. Another mistake is collecting too broadly. If your reference library covers everything from fashion photography to industrial design to software interfaces, you won't have enough depth to actually translate any of it. I learned that the hard way when I tried to merge three completely different visual languages into one project and ended up with something that felt confused. Narrow your sources to two or three related disciplines until you have a solid grasp of the patterns. There are also edge cases where this method doesn't work. If you're working in a field where originality is the primary value — experimental music, conceptual art, certain types of research — the Steal Like An Artist approach can actually hold you back. The framework assumes you're working in applied creative spaces where the goal is to produce something functional and polished. It falls apart when the objective is novelty for its own sake. In those situations, direct experimentation tends to be more useful than reference analysis.
For anyone who wants a structured way to actually do this instead of just reading about it, there are a few resources worth looking at. A lot of people reference the Steal Like An Artist website as a starting point, which has some free worksheets and exercises. I also use a modified version of the Creative Commons license search to find source material that's actually safe to reference when working commercially. The key is understanding fair use and attribution requirements in your jurisdiction before you build anything public-facing. One thing I wish I'd understood earlier is that stealing like an artist is not a shortcut. It's a discipline that takes longer upfront because you're doing the analytical work that most people skip. The time you save by reverse-engineering instead of reinventing usually shows up after a few months of consistent practice. In my experience, it cuts the ideation phase down from about two weeks of brainstorming to roughly four days of focused deconstruction and reconstruction. The tradeoff is that you need to be comfortable being very explicit about what you're studying and why. If you want to start without any special tools, just use a physical notebook and a pen. Write down what you see before you forget. The best reference material is the stuff you actually engage with, not the stuff you save because it looks impressive. Your future self will thank you for the specificity.
Get the Full Details
