SHORT ANSWER
Open Terminal, use cd to get to the folder you want to work in, then type claude and press Enter. That's the Mac version of everything Joe does to start a session. From that point on, you're inside Claude Code and you give instructions in plain English. You don't need to know what mkdir, touch, git clone, or any other command in the video actually does — Claude runs them for you.
Where Terminal lives on your Mac
Terminal is an app that ships with every Mac. Three ways to open it:
- ›Spotlight. Press ⌘ Space, type terminal, press Enter. Fastest way in.
- ›Finder. Open the Applications folder, then the Utilities subfolder. Terminal is in there.
- ›Launchpad. Open Launchpad, find the Other folder, open it, and Terminal is inside.
When Terminal opens, you see a window with a blinking cursor and a short line of text that ends in %. That's the prompt, waiting for you to type something. The Windows equivalent — the one in Joe's videos — looks different visually but works the same way.
The two commands you actually type
In this whole course, these are the only commands you ever type into Terminal yourself:
1. cd — go to a folder
cd stands for "change directory." You use it to tell Terminal which folder you want to work in. Example:
cd ~/my-course-folderThe ~ is shorthand for your home folder. This says "take me to the my-course-folder folder inside my home folder."
2. claude — launch the agent
Once you're in the folder you want to work in, launch Claude Code:
claudeThat drops you into an interactive session with Claude Code in whatever folder you just cd'd into. Everything from this point on is English sentences, not shell commands. If you're using Gemini CLI instead, the command is gemini. For Codex CLI it's codex. Same pattern: cd somewhere, then launch your agent.
That's it. No other commands. If you ever find yourself tempted to type something else into Terminal, stop and ask Claude to do it for you instead. That's the whole shift this course is teaching.
But Joe types things in the video. What are those?
Two situations, and neither one means you have to type the same thing on your Mac:
- ›He's inside a Claude Code session, typing a prompt. Those are plain English sentences he's giving to Claude. Everything he says to Claude works identically on Mac, Windows, and Linux. The language model doesn't care what operating system you're on.
- ›He's setting up the demo at the very beginning. Even here, he really only does
cdandclaude. Anything that looks like he's creating folders, touching files, running git, or installing something is usually Claude doing that inside the session in response to his prompt — not Joe typing it himself.
Watch the video as a movie, not a recipe. You're watching what a session can look like, not memorizing keystrokes. Your own session will look different because your prompt will be different.
What if something goes wrong?
The course's official answer to any tool error, no matter how small: take a screenshot or copy the error text, open the web version of your preferred LLM (claude.ai, gemini.google.com, or chatgpt.com), paste the error in, and ask the model to walk you through fixing it. Paste new screenshots as the conversation continues. That's the standard debugging approach this course uses, and it's how Joe handles his own errors too.
If your claude command itself isn't launching, see the "command not found" FAQ. That's a first-install issue rather than a Mac-specific one.
IF YOU WANT THE QUICK MENTAL MODEL
Terminal is a room. cd is you walking into a specific folder in that room. claude is you calling a colleague into that same folder and handing them the work. Everything after that is you and the colleague having a conversation about the work. You never pick up the tools yourself.