AI Class

By next week you can take a change from idea to live on the internet without looking anything up.

Below is the list. Nothing here explains how — finding out is the assignment. Use AI, videos, docs, each other. There’s no cheating on the how. You’re judged on whether it works, whether you can explain it, and how you got yourself unstuck.

Each row has two halves. The left is the thing to make work. The right is the thing to understand — if you can’t answer it, you haven’t finished the row, no matter what’s on your screen.

The list

Accounts

The Git loop

Run that loop start to finish, twice, without notes. That’s the bar.

Live on the internet

Secrets

Everything you commit is permanent, and on a public repo it’s instantly public. Bots scrape new commits within seconds of a push. API keys, tokens, passwords, .env files: never.

Hand in

done when All four handed in, and you can walk through the Git loop out loud with the screen off.

What you’re actually graded on

Three things, and only three:

Everything above is how you get ready for those. Nothing on the list is worth marks on its own — it’s worth marks because you can’t do 2 or 3 without it.

Limits worth knowing before you assume you’re stuck

• GitHub Pages needs a public repo on the free plan. Other hosts don’t.
• Your profile needs to look like a real person for the Student Pack. Nothing beyond that belongs in a repo — not your address, not your schedule, not your phone number.

When you’re stuck

That’s not the assignment going wrong. That’s the assignment.

Before you ask a person: search the exact error text, read the real documentation, ask an AI to explain it three different ways until one lands, find a video. Note where the answer came from — hand-in 4 is that note.

I need to learn X. Don’t explain it yet.
Find me the three best resources — the official docs page, one video, one worked example.
Tell me what order to use them in and what I should be able to do after each. Then quiz me
on it and tell me what I still don’t have.

The portfolio — due Friday 21 August

The list comes first — there’s no point styling a page you can’t deploy yet. But the portfolio is due at the end of this week. What it has to do and how it’s marked →

That live page becomes your portfolio, and you’ll add your work to it every unit. Nobody’s handing you a template — working out what makes a portfolio good, and what makes yours worth looking at, is the assignment.

Which makes this the place to plan before you build. Every project this year starts the same way — Goal, Scaffold, Evaluate — and you write all three down before you make anything. Write it in a file in the repo, so it’s committed and you can hold yourself to it later.

This isn’t paperwork. It’s the handoff. An AI will build whatever you ask, instantly, and has no idea what you actually want or when it’s finished — so these three, once you’ve got them aligned, are exactly what you hand it so it can go and work without you standing over it. Goal says what to build. Scaffold says in what order. Evaluate is the test it has to pass before either of you calls it done.

Get all three right and you’re driving. Leave one vague and you’re just accepting whatever came out.

Do this before you write a line of it, or you’ll end up with whatever the AI defaulted to. That’s the point of the prompt below — same trick as the one further up, make it find you the map instead of walking it for you.

I’m building a portfolio site that I’ll add school and personal projects to all year.
Don’t build it, and don’t give me code or a template yet.
First: find me four or five real portfolios worth stealing ideas from — a mix of students,
designers, and developers — and for each one tell me what it does well and what it does badly.
Then ask me the questions you’d need answered to plan mine: who’s meant to read it, what it’s
for, what I actually want out of it. One at a time, and don’t move on until my answer is
specific.
Then help me write three things down, and push back where I’m being vague:
GOAL — who it’s for and what they should think after ten seconds
SCAFFOLD — the pieces in build order, what depends on what, what I’m not building yet
EVALUATE — how we’ll tell whether the finished thing is any good
Tell me which parts you’re least sure about and why.

Then, and only then, hand it over and let it work:

Here is my plan. Goal, Scaffold, Evaluate:
[paste your three]
Before you build anything: tell me where this plan is ambiguous, and where two people could
read it differently. Ask me about anything you’d otherwise have to guess. I’d rather answer
five questions now than get the wrong thing.
Once it’s clear, work through the scaffold in order and build it. Then — separately, and
strictly against my Evaluate section, not against your own taste — go back over what you
built and tell me where it falls short of the standard I wrote. Be hard on it.

If it built the wrong thing, that’s information: your plan let it. Fix the plan, not just the page.

done when It’s live, something of yours is on it, and you can say why it looks the way it does — and “because the AI made it that way” isn’t a reason.

Extra — get an agent in your terminal

Only if you already have a paid Claude or ChatGPT plan — both agents need one, and neither free tier includes them.

Get it running inside your Codespace, not on your laptop. Things to look up, in order:

• The official install command for Claude Code, or for the Codex CLI
• How to authenticate it from inside a container. This is the awkward part. The first link you’re offered won’t work, and working out why is the exercise
• How to get it to read your whole project before you ask it for anything
• How to stop your login vanishing every time you create a new Codespace

done when It has explained your own repo back to you in plain English, and you’ve had it change something that you then reviewed line by line.

Codex is OpenAI’s coding agent that runs in your terminal, in your actual project folder, on your actual files. Not a chat window you copy out of — it reads the repo, writes to it, runs commands, and you approve or reject what it does.

This page is unusual for this class: it tells you the how. Not because the how matters, but because a broken install is a bad place to practice figuring things out. Getting it running is twenty minutes. What you do with it afterwards is the rest of the year.

Before you start

#Know which machine you’re setting up, and stop if the answer is “the school one.”

Codex itself has almost no prerequisites. It’s a single compiled program — no Node, no Python, no npm install, no admin password. The installer needs curl and tar, which both systems already have.

What you do need is Git and a copy of your repo on the machine, and that’s the part people get stuck on. Assignment 01 lived in a Codespace, so your code is on GitHub, not on your laptop. Codex reads the folder you launch it in — point it at an empty one and it has nothing to work with. Both paths below clone the repo first, deliberately.

Mac

#Codex installed, signed in, and answering questions about your own repo.

curl -fsSL https://chatgpt.com/codex/install.sh | less

Windows

#WSL2 installed, then Codex inside it — signed in and reading your repo.

OpenAI’s supported setup on Windows is through WSL2, and that’s what you should use. WSL is a real Linux system running inside Windows: same terminal, same commands, same everything as the Mac instructions above, on the machine you already own.

You need Windows 11, or Windows 10 build 19041 or newer. Winver in the Start menu tells you. The good news is the Ubuntu that WSL installs already includes Git and curl, so there’s no equivalent of the Mac’s command-line-tools step.

If that didn’t work — the other routes

There is no single correct installer. There are four, they all end with the same program on your machine, and if one is blocked or broken you switch. Try them in this order.

School network

#Know whether the network is the problem before you blame the instructions.

This is a real possibility and worth checking first, because a filtered network produces errors that look like broken instructions. Codex needs to reach:

Test it in ten seconds, before installing anything:

curl -I https://chatgpt.com
curl -I https://auth.openai.com

Any response at all — even 403 or 404 — means you got through; that’s the server talking. No response, a timeout, or a certificate error means the network stopped you. After installing, codex doctor checks the same ground and reports it in English.

The workaround is a Codespace, and it’s a genuinely interesting one. In a Codespace the agent isn’t running on the school network at all — it runs in Microsoft’s data centre, and that machine talks to OpenAI. The only thing crossing the school firewall is your browser talking to github.com. If GitHub is allowed and ChatGPT isn’t, that gap is the whole reason to work there. Which is also the answer to “why do we use Codespaces in this class” — see the bottom of this page.

Failing that: install it at home, or on a phone hotspot. Once it’s installed and signed in, it still needs network to run — the model isn’t on your laptop.

Both: is it actually working?

Not “did it install.” Working.

Codex ships with a self-check. Run it from inside your repo:

codex doctor

It reports on your install, whether it can find Git, whether you’re authenticated, and whether it can reach the network — with a ✓, ⚠ or ✗ on each. Read the failures; they’re written in plain English and they tell you which of the steps above didn’t take.

Then the part doctor can’t check for you:

Later on, codex update pulls down new versions — worth knowing, because this tool changes fast enough that a guide written today will be slightly wrong by the time you read it. Including this one.

When it breaks

Run codex doctor first — it names most of these for you.

I’m setting up the OpenAI Codex CLI on [Mac / Windows 11 with WSL2]. Here is exactly what I
ran and exactly what came back:
[paste the command and the full error text, not a summary of it]
Don’t give me a list of five things to try. Tell me what this specific error means, what the
most likely single cause is, how I can confirm that’s the cause before changing anything, and
then the fix. If you’re guessing, say you’re guessing.

Understand it when you can answer

Installing something is not knowledge. These are.

done when Codex is installed on your own machine, signed in with your plan, and it has explained your repo back to you correctly — and you rejected at least one thing it wanted to do, on purpose, because you read it.

On a Codespace instead

No paid plan, or a locked-down laptop? A Codespace is Linux, so the Mac commands work as-is — run the curl installer in the Codespace terminal.

The sign-in is where it gets interesting, and it’s the same trap as the Assignment 01 extra: the normal login hands credentials back to a browser on localhost, and localhost inside a container is not your laptop. So the link you’re given goes nowhere.

There are two ways out. One is to forward the port so the container’s localhost reaches your browser. The other is a flag on codex login built for exactly this situation — machines with no browser of their own. codex login --help lists it. Which one you use is up to you; working out why the default fails is the part worth having.

The other thing to know: a new Codespace is a new machine, so your login is gone again. Figuring out where credentials live and how to stop re-doing this is the follow-up.

What this is for

The tool isn’t the point. Being able to hand work to something that will do it without you watching is the point, and that only works if you can say what you want and tell whether you got it.

Which is the same three boxes as always — Goal, Scaffold, Evaluate. A terminal agent makes the cost of skipping them obvious and immediate: vague goal, and it builds the wrong thing in four seconds instead of four days. No Evaluate, and it announces it’s finished and you have nothing to argue with.

Read this entire repository before answering anything.
Then tell me, in plain English: what this project is, how it’s structured, and the three
things about it that would most confuse someone seeing it for the first time.
Then — don’t fix them — tell me what you’d change and why, and which of your suggestions you’re
least confident about.

1. The portfolio — due Friday 21 August

You hand in a URL, submitted via AXIS. Have fun with the design.

You have agents installed — use them

This assumes Claude Code or Codex is working on your machine. If it isn’t, that’s Assignment 02, and it comes first.

Here’s my plan for my portfolio site.
GOAL: [one sentence — who it’s for and what they should think]
SCAFFOLD: [the pieces, in build order]
EVALUATE: [how we’ll both know it’s finished]

2. The book — buy it by Monday 24 August

AI Engineering — Chip Huyen.

done when You have your own paper copy with you on Monday.

3. The quiz — Friday 4 September

Chapter 1 only. Closed book.

What gets handed in

Due Friday 21 August, submitted via AXIS. You hand in a URL — not a repo, not a screenshot, not a zip.

What it has to do

done when You paste the link into a browser on a computer that is not yours, in a window that is not logged into anything, and your site loads.

The decisions

Nothing below has one right answer. Research your options, pick one, and be able to say why you picked it.

On the writing

The words are the assignment as much as the layout. “Hi I’m a student who likes coding” tells a reader nothing. Say what you’re actually into, what you’ve built, what you got wrong and fixed.

Build the smallest version of my portfolio site that satisfies the plan below and gets
deployed to a public URL.
GOAL: [paste yours]
SCAFFOLD: [paste yours]
EVALUATE: [paste yours]
Ask me anything you need before you start. Leave every sentence about me as TODO — I’m
writing those myself.