The technology disappeared. Mostly.

The idea was simple. One page. No framework. No monthly hosting bill. A site that loads in under a second and says something real about the person behind it.

I built it at night, after the day job. The tool was Claude Code, an AI that writes code when you describe what you want. I had never shipped a personal site before. Not because I lacked opinions about how they should work, but because the gap between having the idea and finishing it was always wider than the time I had.

That gap closed. Not because I learned to code faster, but because the nature of the work changed. Instead of writing every line, I described what I wanted and refined what came back. The role shifted from typist to editor. From implementer to architect.

The constraint was the point

HTML, CSS, and fifty lines of JavaScript. No React, no Tailwind, no build step. Every framework I considered would have added complexity to solve problems a single page does not have. The site needed to be elegant, precise, and minimal. The stack had to embody the same qualities.

This turned out to be the interesting decision. Constraints force you to understand what each piece actually does. Native CSS now handles dark mode, fluid typography, and nesting without a preprocessor. A static file on Cloudflare Pages loads faster than anything a framework could produce. The technology disappeared, which is what good technology does.

What I actually learned

Typography carries more weight than layout. The difference between a site that feels considered and one that feels generated lives in font pairings, line heights, and the space between paragraphs. I spent more time on typographic details than on any structural decision.

Copy is architecture. Every sentence on the page went through more revisions than the CSS. Writing in your own voice, without hiding behind corporate language or borrowed frameworks, is harder than any technical problem on this project. The words had to be precise without being sterile, personal without being indulgent.

The meta observation surprised me. This site claims I see patterns across domains. The process of building it required exactly that: understanding how typography, engineering, copywriting, and design interconnect. Each decision in one domain constrained the others. Getting the color palette right meant understanding accessibility contrast ratios. Getting the copy right meant understanding what the layout could support. The site became a demonstration of its own thesis.

Imagination as bottleneck

The hero section says my imagination is now the bottleneck. That came from building this site. I had a working homepage within hours of starting. Not a placeholder. A real, styled, accessible page. The friction that used to kill side projects evaporated. What remained was the hard question: what do you actually want to say?

That question has no technical solution. It requires sitting with it, iterating, throwing away sentences that sound good but say nothing. The tools removed every excuse. Now the only limitation is whether you have something worth building. I find that more interesting than debugging CSS at midnight.