I've built my own portfolio three different ways over the years: hand-coded from scratch, assembled in Wix, and this current site, directed through iteration with Claude. Comparing them says less about the tools than about what stayed constant.
My first portfolio was hand-coded HTML and CSS, built on a free template I customized myself: real markup, real jQuery plugins, real trial and error. Years later, I moved to Wix for speed: a visual, drag-and-drop builder that let me focus on content instead of code. This current site is different again, built through direct collaboration with Claude, where my role was less "typing code" and more "directing, critiquing, and deciding."
Three tools, three very different amounts of hands-on building. The interesting question isn't which one is "best"; it's what each one actually asked of me.
| Criteria | Hand-coded | Wix | AI-directed |
|---|---|---|---|
| Speed to build | Slow: weeks of manual work | Fast: a few days | Fast: hours to a couple days |
| Design control | Full, but limited by my coding skill at the time | Constrained by templates and layout blocks | Full, bespoke: nothing pre-built |
| Coding skill required | High | None | None, but strong critique & direction skill |
| Iteration speed | Slow: every change was manual | Medium: visual editor, but template limits | Very fast: describe the change, review, refine |
| Ownership | Full: plain files, host anywhere | Locked into the Wix platform | Full: plain files, host anywhere |
| My actual role | Executor and designer | Assembler, working within constraints | Creative director and critic |
As the tool changed, how much of the manual execution I personally did dropped: from writing every line, to filling in template blocks, to describing what I wanted and reacting to drafts. But one thing didn't change across any of the three: deciding what "good" looks like. Every version still required me to catch what wasn't working, push back on choices that didn't fit, and know when a design was actually done.
With AI in particular, that judgment became the whole job; execution is nearly free, so the only thing left for me to contribute is knowing what's right. That's the same skill I bring to research and design work generally, just applied to my own site instead of someone else's product.