Building a Blog with AI
I wanted a simple blog. I asked Claude to build it.
First attempt: Bear Blog. It looked minimal. Claude researched it and found a full Django application hiding behind the clean facade—PostgreSQL, Redis, multi-tenant routing, user management. The simple frontend was a lie.
Insight. Multi-tenancy belongs at infrastructure level. GitHub Pages already handles domains and hosting. Why rebuild it?
Claude found Hugo with Bear Cub theme. Static generation. Same aesthetic. No runtime. We forked the theme as gopher-blog and customized it.
The Stack
- Hugo generates static HTML
- gopher-blog theme (forked from Bear Cub [1])
- Markdown files in git
- Claude for research, debugging, content
How It Works
I describe what I want. Claude writes the code, fixes bugs, deploys. When something breaks, I paste the error. Claude fixes it.
The blog builds in milliseconds. No database. No server. No security updates.
hugo new site .
git submodule add <theme-repo> themes/gopher-blog
make deployWhy This Matters
AI can build functional software from conversation. Not perfectly—I guide, review, correct. But the iteration speed is different. Ideas become code in minutes.
This blog exists because Claude built it. The words, the theme, the deployment—all from conversation.