The AI Landscape

You've heard people are building things with AI. Maybe you've had a conversation with ChatGPT. Maybe you haven't. Either way, you're here because you want to understand what's actually going on — and whether it's for you.

Good news: it is. And it's not as complicated as it sounds.

---

What AI coding tools actually are

AI coding tools are not search engines. They're not copy-paste generators. They're conversation partners that can write, edit, and deploy code based on what you tell them.

Think of it like this: instead of learning to write code yourself, you describe what you want in plain language, and the AI builds it. Then you look at what it made, say "no, more like this," and it adjusts. Back and forth, like working with a fast, tireless collaborator.

This is what people mean by "vibe coding" — you guide the direction, the AI handles the implementation. You don't need to know JavaScript. You need to know what you want.

---

The major players (March 2026)

The AI models

These are the "brains" — the language models that understand your requests and generate code:

• Claude (by Anthropic) — the model behind this curriculum. Excellent at code, careful with nuance, good at following complex instructions. Currently the strongest for vibe coding.
• ChatGPT (by OpenAI) — the one that started it all. Very capable, widely known. GPT-4o and o3 are the current models.
• Gemini (by Google) — deeply integrated with Google services. Strong reasoning, massive context windows.
• Open-source models — DeepSeek, Qwen, GLM, and others. Free to run locally, improving fast. Less polished but no subscription needed.

Chat interfaces vs. coding agents

This is an important distinction:

Chat interfaces — you talk to AI in a browser, it responds with text:

• Claude.ai, ChatGPT.com, Gemini — general-purpose conversations
• Claude.ai Artifacts — can generate and preview code right in the browser
• Good for: exploration, learning, quick prototypes

Coding agents — AI that can actually create files, run commands, and deploy:

• Claude Code — Anthropic's command-line coding agent. The primary tool in this curriculum.
• Cursor — VS Code fork with AI built in. Popular with developers.
• v0 (by Vercel) — generates React components from descriptions. Zero install.
• Replit — browser-based IDE with AI. No install needed.
• Good for: building real things that run on the internet

The key difference: chat interfaces give you text you have to copy somewhere. Coding agents build the thing directly.

---

What things cost

Let's be honest about this, because it matters:

Free tier ($0)

• Claude.ai, ChatGPT, Gemini all have free tiers
• Limited messages per day, older models
• Good enough to start learning and exploring
• v0 and Replit have free tiers for small projects

Chat subscription ($20/month)

• Claude Pro, ChatGPT Plus, Gemini Advanced
• More messages, better models, faster responses
• Claude Pro includes Artifacts and basic coding in the browser
• This is where most people should start

Coding agent tier ($100–200/month)

• Claude Max ($100/mo) or Claude Code with API usage
• Cursor Pro ($20/mo) + API costs
• This is where you can build entire projects in a day
• Only worth it once you're building regularly

---

How fast this is moving

A quick timeline to orient you:

• 2022: ChatGPT launches. Most people use it for text.
• 2023: GPT-4 can write real code. Early adopters start building.
• 2024: Coding agents emerge (Claude Code, Cursor). "Vibe coding" coined. People build full apps through conversation.
• 2025: Agents get reliable enough for production work. Multi-file projects, deployment, debugging — all through conversation.
• 2026 (now): The tools are mature. The question is no longer "can AI write code?" but "how do I work with it effectively?"

That's what this curriculum is about. Not the technology — the collaboration.

---

What to expect from this course

After completing the core curriculum, you will have:

• Built and deployed at least one real project
• Developed your own prompting style (not copied templates)
• Understood when AI helps and when it gets in the way
• Published something to the web that other people can see and use

You won't become a programmer. You'll become something different: someone who can build things through conversation with AI. That's a new skill, and it's genuinely powerful.

---

Exercise: Your first real conversation

Before moving on, try this. Go to claude.ai (free) and have a conversation about something you actually care about. Not a test prompt. Not "write me a poem." Something real.

Ideas:

• "I'm thinking about starting a side project that does X. Help me think through what that would look like as a website."
• "I want to understand how Y works. Explain it to me, then ask me questions to check my understanding."
• "I have this idea for Z. What would it take to actually build it?"

Notice how the conversation feels. That back-and-forth — that's the foundation of everything we'll build on.