I have never understood the instinct to specialize at the expense of everything else. The people whose work I find most worth studying — da Vinci, Buckminster Fuller, Walter Russell — were not dilettantes flitting between disciplines. They were people for whom every domain they touched was illuminated by every other domain they knew. The range was not a liability. The range was the mechanism.

My work operates the same way. Installing a commercial security camera system is a systems problem: how do you design something that survives its builder, that a stranger can maintain in five years without a phone call to ask what you were thinking? Building a website is the same problem with different materials. Writing about either one is an attempt to make the thinking visible — because the thinking is the part that transfers, not the technical specifics.

I believe that most tools are rented and most methods are borrowed, but the way you think about problems is yours. That thinking is what I try to bring to every engagement, which is why Zonebrite installations don't look like anybody else's and Safaii Studio sites don't look like anybody else's. They solve for the long run because I've never been interested in the short one.

On craft

Craft, as I understand it, is the practice of making decisions at every level of a thing — not just the visible ones. A camera installation where the cable runs are clean and labeled is not just aesthetically better than one where they aren't. It is functionally better: it fails more gracefully, diagnoses more quickly, costs less to maintain. The craft is not decoration. It is structure.

The same is true in web development. CUBE CSS, vanilla JavaScript, static site generation on Cloudflare Pages — these are not preferences for their own sake. They are positions about what makes a website durable, maintainable, fast, and owned by the person who commissioned it rather than the platform that hosts it. Every technical decision is an argument about what matters.

I build things I want to still be proud of in ten years. That constraint eliminates most of the available shortcuts.