Claude Code — Overview

There are many AI tools that claim to help you build software. Most of them will write code at you. Claude Code will work with you — inside your actual project, understanding your actual files, running real commands on your actual machine.

That difference is everything.

---

The Three Categories of AI Coding Tools

Before you pick up any tool, you need to understand what category it belongs to. Using the wrong category for the job is like using a pivot table when you needed a formula — technically possible, but the wrong fit.

Category 1: Conversational AI (ChatGPT, Gemini)

You describe a problem in a chat window. The AI writes some code. You copy it. You open your editor. You paste it in. You read the error. You go back to the chat window and describe the error. The AI writes more code. You copy it again.

This is like faxing instructions to a contractor and waiting for a fax back. Every round trip is manual. The AI has no idea what your actual project looks like. It doesn't know what files you have, what you named your components, what version of React you're using, or what you tried 10 minutes ago. Every message starts from scratch.

For a quick question about a concept — fine. For actually building a real project — painful.

Category 2: Full App Builders (Lovable, Bolt, v0)

You describe an entire app. The AI generates the whole thing. A working prototype appears in minutes.

This is genuinely impressive — and genuinely useful for prototyping. If a client wants to see what something could look like before you commit to building it, these tools are excellent. They'll give you something visual in 20 minutes.

The limitation is real: you lose control of the codebase. These tools work best when they generate everything. The moment you need to deviate — add a custom business rule, connect to a specific API, follow a particular pattern your team uses — you're fighting the tool rather than using it. The code they generate is often structured for the tool's convenience, not for a production team's maintainability.

You'll use Lovable in this workshop for exactly what it's good at: fast prototypes and concept validation. But your production work will not live there.

Category 3: Context-Aware AI Dev Assistants (Claude Code)

Claude Code runs in your terminal, inside your project. It can:

• Read any file in your codebase
• Write and edit files directly
• Run terminal commands (install packages, run tests, run your dev server)
• See the output of those commands and react to errors
• Remember everything that happened in the current session
• Read your CLAUDE.md before starting any work, so it already knows your project's rules

This is not a chat window where you paste code back and forth. This is an AI that is inside the project with you.

---

The Analogy That Makes This Click

You have worked in Excel for years. You know what it feels like when a junior associate — someone smart but new — joins your team.

On their first day, you could hand them a task by emailing them a description and waiting for their response by email. (ChatGPT)

Or a specialized consulting firm could come in, generate an entire new Excel model from scratch in a day, and hand it to you. (Lovable)

Or you could sit that junior associate down at your desk, with your files open in front of them, and say: "Here's the client file. Here's what we need to fix. Here's how we do things here. Go ahead." They can see everything. They can make changes directly. You review their work and guide them. (Claude Code)

The third approach is faster, more accurate, and produces work that actually fits your project — because the person doing it can see the project.

---

Why Context Is the Key Difference

The word that matters in that table is context. ChatGPT has none of your project's context. Lovable only has context for what it built. Claude Code has full context — your files, your rules, your patterns, your history — because it is running inside your project.

---

The Competitive Differentiator

Thousands of people use ChatGPT to write code. Thousands more use Lovable to generate apps. Very few people have learned to use Claude Code properly — with a well-written CLAUDE.md, with skill files that encode their hard-won knowledge, with hooks that automate repetitive actions.

The people who learn this system don't just produce code faster. They produce consistent, professional, production-grade code that follows the same patterns every time — because those patterns are encoded in files that Claude reads on every session.

That is what this module teaches. Not just how to install Claude Code and ask it questions. How to build a system around it so that every session starts with Claude already knowing your standards, your project structure, your decisions, and your preferences.

This is what separates someone who uses AI from someone who has built an AI-augmented workflow.

---

What You'll Have After This Section