Chapter 0: Set Up Your Machine

Time: 15–30 minutes, mostly download time. Cost: $0 if you stick with Ollama, which is enough for all of Foundations.

Before you go on a road trip, you pack the car first: snacks, a phone charger, a map. You don't figure that out while you're already driving. This chapter is the packing part. It's not exciting, but skipping it means every later chapter stops to fix a setup problem instead of teaching you something new. Do this once, carefully, and you won't think about it again.

You don't need to know how to code yet. You just need to follow each step in order.

What you're installing, and why

Tool	Why you need it
uv	A free tool that manages Python for you, including Python itself if you don't already have it, plus a private, isolated folder for each lab's packages so they can't clash with anything else on your computer. Every lab in this course uses Python, the language most AI tooling is built around, and uv is how you'll run it.
VS Code	A free code editor. You'll use it to open and run the lab files.
Ollama	Lets you run an AI model directly on your own computer. Free, no account, no internet needed once it's downloaded. This is how you'll do almost every lab in Foundations and Intermediate at zero cost.
(Optional) An OpenAI or Anthropic API key	If you'd rather use a hosted model instead of (or alongside) Ollama, this is how. Optional because Ollama alone gets you through the free path.

Here's the shape of the decision you're setting up for:

You can install both and switch per lab. Nothing below locks you into one path.

Step 1: Install uv

Open your terminal (Mac/Linux: the Terminal app; Windows: PowerShell) and check if you already have it:

uv --version

If that printed a version number, skip to Step 2. If not:

macOS / Linux

curl -LsSf https://astral.sh/uv/install.sh | sh

Windows

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Close and reopen your terminal, then re-run uv --version to confirm it worked.

Why uv instead of installing Python by hand? Normally you'd need to separately install Python, then learn a second tool (venv) to keep each project's packages from clashing with each other, then remember to "activate" that environment every time you open a new terminal. uv folds all three of those into one tool.

If you don't have Python, uv downloads the right version for you the first time a lab needs it. For isolation, it automatically creates a private .venv folder per lab the first time you run something in it, no separate command, nothing to activate by hand.

Step 2: Create a project folder

Pick a spot for this course's code (your Desktop or home folder is fine) and create a folder:

mkdir academy-labs
cd academy-labs

That's it, no environment to create yet. Each lab folder in this course ships with its own pyproject.toml file that lists exactly which packages that lab needs.

The first time you run uv run <script>.py inside a lab folder, uv reads that file, quietly creates an isolated .venv just for that lab, installs the packages into it, and runs your script, all in one step. Run it again later and it skips straight to running, since everything's already there.

Step 3: Install VS Code

Download and install VS Code, it's free. You'll use it to open the lab folders and run the Python scripts. Any code editor works, but the labs' instructions assume VS Code, so it's the easiest path if this is new to you.

Step 4: Install Ollama and download a model

This is the step that gets you a free, private AI model running on your own computer.

1. Go to ollama.com, download the installer for your OS, and run it.
2. Open a terminal and pull a small model:

ollama pull llama3.2

This downloads about 2GB, so it takes a few minutes depending on your internet speed. This is a one-time download. After this, the model lives on your machine and needs no internet connection to run.

3. Check it works:

ollama run llama3.2

Type something like hi, are you working? and press Enter. You should get a reply printed back in your terminal. Type /bye to exit.

If you got a reply, Ollama is fully working and you can already do most of this course's labs for free, forever, with no account and no credit card.

Step 5 (optional): Get a hosted API key

Ollama is enough to complete Foundations and most of Intermediate. Some learners prefer a hosted model instead, because it's often faster or more capable. This costs a small amount of money per request (usually fractions of a cent for these labs), billed by the provider directly to you. This project never sees or handles your key or your money.

• OpenAI: create a key at platform.openai.com/api-keys
• Anthropic: create a key at console.anthropic.com

Either way, you'll be asked to add a payment method on their site before the key works. Save the key somewhere safe. Each lab will ask you to paste it into a local .env file that never leaves your computer and is never uploaded anywhere.

Checkpoint

Why does each lab have its own isolated environment instead of just installing packages normally?

So this course's packages stay isolated in their own folder and can't conflict with other Python projects on your computer (or be broken by them later). uv creates and manages that isolation for you automatically, no separate command to remember.

What's the difference between Ollama and an OpenAI/Anthropic API key?

Ollama runs a model directly on your computer: free, private, no account needed, but limited to your computer's hardware. An API key sends your request to a company's servers over the internet: usually faster or more capable, but costs a small amount of money per request.

Do you need to "activate" anything before running a lab, like you might have heard about with Python virtual environments?

No, that's the whole point of uv. Just run uv run <script>.py inside the lab's folder, and it handles creating/activating the isolated environment for you behind the scenes.

Check Your Knowledge

Click to start the quiz

1. When you set up a hosted OpenAI or Anthropic API key in Chapter 0's optional Step 5, who actually handles billing for your API requests?

This course, since it generates the key for youThe provider (OpenAI or Anthropic) bills you directly on their own site; this course never sees or handles your key or your moneyOllama, since it manages every provider behind the scenesThere is no billing, hosted keys are always free to use

2. Step 4 has you run `ollama pull llama3.2`, a roughly 2GB download. What does that download actually buy you, according to the chapter?

A day pass, the model needs re-downloading every 24 hoursA one-time download, after which the model lives on your machine and runs with no internet connection neededA decompression tool for running OpenAI models locallyNothing changes until you also install VS Code

3. The first time you run `uv run <script>.py` inside a lab folder, what does uv do that a second run later skips straight past?

Nothing, it behaves identically on every runReads that lab's pyproject.toml, creates its isolated .venv, and installs its packages, before running the scriptDownloads a fresh copy of OllamaDeletes the previous lab's environment to free up space

What's next

With your toolbox packed, Chapter 1 starts the actual subject: what people mean when they say "AI," "machine learning," and "generative AI," and how those words relate to each other.