✦
Welcome to Actually AI

Learn AI by actually using it

This isn't a theory course. Every lesson gives you something real to try — so you build intuition, not just knowledge. Takes about 5 minutes to set up.

Step 1 of 2

Which AI are you using?

Pick the one you have access to. We'll tailor every exercise to match — linking you straight to the right place.

Step 2 of 2

A word about AI usage

AI runs on tokens — roughly ¾ of a word each. Every message you send and receive uses them. Here's what that actually means for you.

What a token actually is
A short message
≈ 50–100 tokens
One back-and-forth
≈ 200–500 tokens
A full short story
≈ 2,000 tokens
This entire lesson
≈ 800 tokens
📋 Check your platform's current limits. Free tier allowances change — always check the source. Check your AI's pricing page ↗
✂️ Start new chats often. Each conversation carries its full history, which uses tokens. A fresh chat is leaner — and often gets better results.
🎯 Be specific, not long. A precise 20-word prompt usually beats a rambling 100-word one — and uses a fraction of your allowance.
Level 1 · Lesson 1

What is AI, really?

⏱ 8 min read ✦ 30 XP 🧠 3 questions
Reading
Exercise
Quiz
XP

It's not what the movies told you

When most people hear "AI," they picture a robot with a glowing red eye, or a supercomputer that's secretly plotting something. Hollywood has a lot to answer for.

The AI you're going to use — Claude, ChatGPT, Gemini — is something much more specific and, honestly, more interesting. It's a language model. And understanding what that means will change how you use it.

The key insight: AI doesn't think. It predicts. And that distinction matters more than almost anything else you'll learn in this course.

What does "language model" actually mean?

An LLM — Large Language Model — was trained on an enormous amount of text. Books, websites, code, conversations, articles. Billions of them.

Through that training, it learned patterns. Not facts stored in a filing cabinet, but statistical relationships between words, ideas, and concepts. When you type something, it predicts what should come next — over and over, word by word — until a response is complete.

Analogy

Think of autocomplete on your phone — but trained on the entire internet, capable of completing entire essays, writing code, or explaining quantum physics. Same core idea, wildly different scale.

This is why AI can seem almost magical. It's not retrieving a pre-written answer. It's constructing something new, right now, based on everything it's learned.

Prediction vs thinking

Here's where it gets important. Because the AI is predicting — not reasoning from first principles — it can confidently say things that are wrong. It doesn't always know when it doesn't know something.

This isn't a bug or a flaw to fix. It's the nature of how it works. And once you understand that, you stop being surprised by mistakes, and you start prompting in a way that reduces them.

Remember this: The AI is always trying to give you a plausible-sounding response. Your job is to give it enough context to make that response actually useful.

What this means for you

Most people treat AI like a search engine — type a quick question, expect a definitive answer. That's not quite right.

A better mental model: it's a very well-read collaborator. It has broad knowledge, can follow instructions, adapts to your style, and improves when you give it better direction.

The difference between someone who gets mediocre results and someone who gets extraordinary results? Almost always: the quality of their prompts. That's what this entire course is about.

Try it yourself

Ask your AI a simple question

Open your AI in a new tab. Ask it to explain what it is in simple terms — as if talking to someone who has never heard of AI before.

Suggested prompt: "Explain what you are and how you work, in simple terms. Imagine I've never heard of AI before."
↗ Open your AI

Read the response. Notice how it describes itself. Then come back here.

Knowledge check

Three quick questions

1. What does LLM stand for?
2. How does an AI generate a response?
3. Why can AI confidently say wrong things?
+30
XP Earned

Lesson complete! 🎉

You now understand what an LLM actually is — and why that matters. Next up: what can your AI actually do?