It was past midnight when I decided to build a website. The baby was finally asleep. The dogs were finally asleep. My husband was asleep. The house was quiet in that specific way it only gets after 11pm, and I was sitting at the kitchen table with a cold cup of tea and a thought I couldn't shake: I want somewhere to put all of this.
I had no idea how to build a website. I mean — none. I know how spreadsheets work. I understand a pivot table. I can write a board memo that makes senior partners actually read it. But HTML? CSS? I didn't even know what those stood for.
What I did have was Claude Code, a vague vision, and enough stubborn energy to stay up until 2am on a Wednesday.
Here's how it actually went — the prompts, the pivots, the moments of complete confusion, and the part where it actually worked.
01 · The first prompt
I started with a screenshot and a vibe
I didn't start with a wireframe or a detailed spec. I pulled up a website I liked — clean, warm, editorial — took a screenshot, and dropped it into Claude with a message that was essentially: "I want something like this but more me. Here's my color palette idea: cream, rust, a little bit of sage." That was it. That was my brief. And Claude just... started building. The first version wasn't perfect, but it was a real website. On a screen. That I had made. At 12:30am.
02 · The learning curve
I had to learn just enough to ask better questions
The first few hours were a lot of "why does it look like that" and "can you make the spacing less weird." But somewhere around the third session, something shifted. I started understanding what I was looking at — not how to write it, but how to describe what I wanted. I learned that "padding" means breathing room. That "serif" means the fonts with the little feet. That when something looks off, you can just describe it plainly and the right fix usually follows. The skill isn't coding. The skill is seeing clearly and saying so.
03 · Where it broke
It broke. Several times. That was fine.
There was a stretch where the nav disappeared on mobile. A period when the footer refused to stay at the bottom of the page. A genuinely baffling afternoon where adding a new section somehow deleted a different section. Each time, I told Claude what had broken, described what the page looked like vs. what I expected, and we worked it out. The process is iterative — it's not magic, it's conversation. What surprised me was how much patience that required, and how much patience I actually had for something I was genuinely excited about.
You don't need a CS degree. You don't need to know what a div is before you start (you'll figure it out). What you do need is a clear sense of what you want and the willingness to describe it out loud, over and over, until it exists.
Start with something real — a page you love, a color you keep coming back to, a feeling you want someone to have when they land on your site. Bring that to the conversation and iterate from there. The bar for "good enough to ship" is lower than you think, and shipping is the whole point.
This website is still a work in progress. The Tinkerings page needs more posts. The mobile nav has opinions. But it exists, it's mine, and I built it — mostly after midnight, one prompt at a time.
The skill is seeing clearly and saying so."