Getting Started
Your first steps with AI-assisted development
No experience required. No computer science degree. Just curiosity, $20/month, and 20 minutes of setup. This guide gets you from zero to productive.
If you only do one thing
Stop reading this page. Go install Claude Code right now. That is the single most important thing you can do today.
You do not need to be an expert. You do not need to understand everything on this page first. You do not need Python, Node.js, or any other tool installed. Just Claude Code. It is safe to run out of the box. Anthropic built it that way.
Once it is running, it becomes your guide for everything else. It will teach you the command line. It will help you install Python. It will explain what Markdown is when you need it. Every single step below becomes easier because you have an AI sitting right there, ready to help you through it.
If Claude Code is not your thing, install Cursor, Gemini Code Assist, or Codex. I do not care which one. Pick one. Install it. Start using it. The gap between people who use these tools and people who do not is growing every single day, and the only thing standing between you and the other side is 20 minutes of setup.
Come back and read the rest of this page after. It will all make more sense once you have tried it.
Start Here
You want to use AI for coding, business, or both. You have heard the buzz. You have seen the demos. You are not sure where to begin.
Good news: you can be up and running tonight. Not next week, not after a course, not after reading a stack of articles. Tonight.
This page is your starting point. It will grow over time as I add more resources, but everything here is enough to get moving right now.
Step 1: Install Claude Code
Claude Code is available for both Windows and Mac. The installation is slightly different depending on your platform, but either way you will be up and running in minutes.
On Windows
Windows now has a direct installer for Claude Code. No extra steps needed. Download it from Anthropic's Claude Code page and run the installer.
On Mac
On Mac, you install Claude Code through npm (the Node.js package manager). You will need Node.js installed first, then run npm install -g @anthropic-ai/claude-code in your terminal. Full instructions are on Anthropic's Claude Code page.
Python and Node.js (optional but recommended)
You do not strictly need Python or Node.js to get started. Claude Code works without them. But having them installed makes your experience much better. Claude Code uses Python behind the scenes when it needs to calculate something or process data, and Node.js for browser automation and web tools.
Here is the thing: you do not need to figure out how to install them yourself. Get Claude Code up and running first. Then just ask it to install Python and Node.js for you. It will walk you through the whole process. That is the pattern from here on out. Get the AI running, then let it handle the rest.
For the full toolbox
If you want to go beyond Claude Code and set up the complete suite of tools I use daily, check out the CC Director installation guide. CC Director is free and open source. It installs 20+ productivity tools on top of Claude Code. But that is optional. Claude Code alone is plenty to get started.
Pay for the $20/month Anthropic subscription. That is your only cost.
I teach Claude Code because it is the tool I use every single day. There are other great options out there. Google has Gemini, there is Cursor, Copilot, and more. I have no interest in arguing about which is best. I picked one, I know it deeply, and that is what I teach. If you prefer a different tool, go for it. The principles are the same. But if you want to follow along with everything on this site, Claude Code is where we start.
Drop any other chatbot subscriptions you are paying for. Not because they are worse, but because they cannot interact with your computer the way Claude Code can. The difference between a chatbot in a browser and an AI agent on your machine is the difference between reading about swimming and getting in the water.
Step 2: Learn Three Things
Before you do anything fancy, you need to be comfortable with three concepts. None of them are hard. None of them require a computer science degree. You can learn all three in an afternoon.
The command line
That black screen with the blinking cursor. It looks intimidating. It is not.
The command line is just talking to your computer with text instead of clicking icons. You need maybe five commands to get around: change directory, list files, move things, copy things, and run programs. That is it.
Grab any beginner YouTube tutorial on "command line basics" or "terminal for beginners." Spend 30 minutes. You will have everything you need.
Once you have Claude Code installed, you can even ask it to teach you the command line. It is remarkably patient.
Markdown
Markdown is a fancy text file. A hash sign makes a heading. Two asterisks make something bold. A dash makes a bullet point.
When ChatGPT or Claude gives you a nicely formatted response with headings and bold text -- that is Markdown being rendered. It is the lingua franca of AI tools.
You do not need to memorize it. You will pick it up naturally as you use Claude Code. But knowing it exists and that it is just a simple text format will save you a lot of confusion.
File paths
You need to understand how your computer organizes files. Folders inside folders, with a path that tells you exactly where something lives.
On Windows it looks like C:\Users\you\Documents\project. On Mac it looks like /Users/you/Documents/project. Same concept, slightly different notation.
This matters because when you point Claude Code at a folder, you need to know which folder you are pointing it at. Think of it like giving someone directions to your house -- you need to know your own address.
Step 3: Start Exploring
Point Claude Code at a directory you want to understand. Your documents folder, a project you are working on, a spreadsheet you have been meaning to clean up. Ask it questions. "What is in this folder?" "Summarize this spreadsheet." "What is the average in column three?"
If you want to learn about code, go grab a well-known open-source library from GitHub, clone it to your machine, and point Claude Code at it. Ask it to explain how the code works. Ask it to walk you through the architecture. You will learn more in an hour than in a week of reading documentation.
If you cannot figure out how to clone a repo or navigate GitHub -- that is fine. Once you have Claude Code running, just ask it to help you. That is the whole point. It is your companion, not a tool you need to master before you can use it.
Here is a trick: if you get stuck on a website like GitHub and you do not understand what you are looking at, take a screenshot and give the file path to Claude Code. It will look at the image and tell you exactly what to do next.
Who to Watch
Look up Boris Cherny on YouTube. He is one of the creators of Claude Code. I cannot recommend him enough. So many things he says line up exactly with what I have learned the hard way over the past year. If you want to understand how to think about AI-assisted development, he is the place to start.
Start here: "Mastering Claude Code in 30 Minutes"
Watch on YouTube. This is the single best beginner resource I have found. Boris walks you through everything step by step: installation, configuration, asking questions about a codebase, editing files, managing context, keybindings, and scripting.
He says it himself in the video: "codebase Q&A is the easiest way for new users to start." I agree completely. Point Claude Code at a folder and start asking questions. That is how you build trust and learn what it can do.
It is 30 minutes, structured, and practical. No fluff.
Then watch: "Claude Code Live" (Anthropic Webinar)
Watch on YouTube. This one shows three live demos: building something from scratch, fixing a real GitHub issue, and exploring an unfamiliar codebase. About 35 minutes.
It covers CLAUDE.md files (instructions you give the AI about your project), plan mode, permissions, and running parallel sessions. It assumes a little bit of terminal and Git familiarity, so watch the first video above before this one.
There are a lot of great resources on YouTube, but they take some filtering. Not everything out there is good, and some of it is actively misleading. I will keep adding recommended links to this page as I find content worth sharing.
Should You Still Learn to Code?
This is the big question, and I will give you an honest answer: it depends on what you want to do.
The traditional software developer role -- someone who writes code line by line as their primary job -- is changing fast. I have 30+ years of development experience, and I have not written a single line of code by hand in months. I still build software every day. I just direct AI agents to do the writing.
What matters now is understanding what good software looks like, how to troubleshoot problems, and how to think about products. Go watch Steve Jobs talk about building the iPhone -- how every detail was considered, how the whole experience was orchestrated. That kind of product thinking is more valuable than ever.
If you are a business student, you are actually in a great position. You understand problems that need solving. You understand customers. You understand value. Those skills combined with an AI agent that can build things for you is an incredibly powerful combination.
My honest take: learn enough about code to have a conversation with the AI. Understand the basics so you can evaluate what it produces. But do not spend years learning to write code by hand if your goal is to build things. The tools have changed. Use the new tools.
Trust and Safety
When you first start, you will not trust the AI. That is normal and healthy.
Claude Code has a permission system. You can run it in a mode where it asks before every action -- before writing a file, before running a command. Start there. Watch what it does. Read every change it proposes.
As you build confidence, you will start giving it more freedom. Eventually you might run it the way I do -- with most permissions turned off, letting it work autonomously on tasks while you review the output.
If you want extra peace of mind, sandbox it. Set up a virtual machine, a Hyper-V instance on Windows, or Parallels on Mac. Give the AI access only to that small world. Nothing it does can touch the rest of your computer. As your trust grows, expand its access.
The important thing is to start. You can start cautiously. You can start in a sandbox. You can start with permissions locked down tight. Just start.
Keep Going
This page will grow. I will add specific YouTube links, recommended tutorials, and walkthroughs as I curate them.
If you have questions, reach out to me on LinkedIn. I am happy to point you in the right direction.
The single best thing you can do right now is close this page, install Claude Code, and start a conversation with it. Ask it anything. Ask it to help you set up your first project. Ask it what it can do.
Do not wait until tomorrow. Do not tell yourself you will get to it this weekend. The gap between people who use these tools and people who do not is growing every single day.
Twenty dollars a month and 20 minutes of setup. That is the barrier. Everything after that is just curiosity and practice.
Ready for More?
This page covers the basics. If you want to go deeper -- learn to build tools, manage multiple agents, and automate your entire workflow -- check out the full course or read the blog.
Questions? Connect with me on LinkedIn