Lecture slides

Static Site Generators in the Age of AI

Building and Maintaining Content with AI Agents

Chapter 0: Course introduction

Polla Fattah

What you will build

One small website that grows from chapter to chapter.

  • A working site on your own computer in Chapter 1
  • Published at a public address by Chapter 7
  • Improved by an AI agent, with every change reviewed by you
  • A plan for a website of your own in Chapter 19

Why a static website?

  • Your content is plain text in folders you control
  • The build produces ordinary files any browser can read
  • Hosting can be free: GitHub Pages serves public projects at no charge
  • Every change is recorded in Git and can be undone

What “static” means

A static site generator combines your writing, a layout, and settings into finished web pages before anyone visits.

“Static” describes how the result is served. It does not mean the site never changes: you edit, rebuild, and publish again.

How a static site generator works

content/     the Markdown you write
layouts/     templates that shape each page
hugo.toml    the site's settings
     |
     v
   hugo       one command builds the site
     |
     v
public/      HTML, CSS, and images: the finished site

Where AI fits

An AI agent can draft, rearrange, and check your work.

It cannot know what happened to you, confirm that a claim is true, or decide what should be published.

Machines check rules. People judge meaning.

The working method

  1. Supply the material the agent works from
  2. Bound the task: one clear job, named files
  3. Review every changed line, not the agent’s summary
  4. Check with a build and automatic tests
  5. Decide yourself what goes live

Your tools

ToolIts job
HugoBuilds the website
MarkdownThe format you write in
VS CodeThe editor we use
GitRecords every change so you can undo it
GitHubHosts, checks, and publishes the project
Codex CLIThe AI agent in the exercises

What you need

  • A laptop or desktop where you can install software
  • An internet connection
  • A free GitHub account, from Chapter 7
  • Access to an AI agent, from Chapter 8: check your account’s usage limits first

No programming, Git, or command-line experience is assumed.

The course in five parts

ChaptersPart
1 to 5Build and understand a first site
6 to 7Record it and publish it
8 to 12Bring in an agent and give content a structure
13 to 17Contribute, check, and reach readers
18 to 19Keep it running, then make your own

How every chapter works

  1. See the result first: what you will have at the end
  2. Make one small change early
  3. Break something on purpose, then recover from it
  4. Try it your way, then save a checkpoint

Four learning paths

PathFor you if
The complete pathYou are new to building websites
Web-literateYou know HTML, CSS, JSON, or YAML
Software engineersGit and CI are daily tools; you need Hugo
Maintainers and teamsYou already run a website

Every path builds the same website.

Habits for this course

  • Keep a short learning note: versions, errors, what you tried
  • Read error messages before changing anything
  • Save a checkpoint before each experiment
  • Never paste passwords or tokens into project files or prompts
  • Review every change an agent makes before you keep it

Before the next session

  1. Install VS Code from its official download page
  2. Install Hugo
    • Windows: winget install Hugo.Hugo.Extended
    • macOS: brew install hugo
    • Linux: follow the official Linux guide
  3. Open a terminal and check: hugo version
  4. Create a folder named my-knowledge-site

Next: Your First Hugo Website

Chapter 1: a local preview, your first edit, and one deliberate mistake you will know how to repair.

polla.dev/ssg-book