What AI actually does — autocomplete at huge scale
Strip the magic feeling off. Get the working model of what AI is doing under the hood, so the rest of the curriculum has a foundation.
::TL;DR · the whole lesson in three lines
- MOVEStrip the magic feeling off. Get the working model of what AI is doing under the hood, so the rest of the curriculum has a foundation.
- DRILLYou're going to ask an AI to explain itself, in the simplest possible terms, to YOU. This is the meta-prompt that anchors everything.
- WINYou can describe in one paragraph what AI does, without quoting the AI's reply.
::concept · what's actually happening
AI in 2026 — the kind you type into — is software that learned from a very large pile of human text, images, and code. When you type something, it tries to predict what comes next, one piece at a time, based on patterns it learned from that pile.
read full concept · 2 more paragraphs →collapse concept ↑
It is not conscious. It is not your friend. It is not a search engine. It is not always right. It is an extraordinarily good drafting and thinking partner if you operate it instead of trusting it.
Everything else in this curriculum follows from this. When AI feels magical, it's the prediction landing. When AI feels stupid, the prediction missed. There is no separate magic mode.
::drill · do the thing
You're going to ask an AI to explain itself, in the simplest possible terms, to YOU. This is the meta-prompt that anchors everything.
::L1 drill · copy-paste into any AI chat
Explain in 200 words, plain English, what you are doing when I type a message and you respond. Use a concrete analogy a 12-year-old would understand. Tell me one thing you are genuinely good at and one thing you are genuinely bad at. Do not use the words "transformative," "synergy," "revolutionize," "powerful," or "intelligence." After you answer, ask me one question to check whether the explanation landed.
::steps
- 01Open a new tab. Go to claude.ai or chatgpt.com (free tier is fine).
- 02Paste the prompt above. Send it.
- 03Read the reply slowly. Don't scroll past.
- 04Answer the AI's check-in question. That is your first multi-turn moment.
::worked example · what one real run looks like
See one real run before you try yours.
::worked example · what one real run looks like
See one real run before you try yours.
::what a real human filled in
Explain in 200 words plain English what you are actually doing when I type a message to you. Use an analogy a 12-year-old would understand. Tell me one thing you are good at and one thing you are bad at because of how you work. End with a check-in question that helps me know if I understood you.
::what the AI returned
When you type a message, I am not "thinking" the way you do. I am a giant pattern-matcher. I have read a huge pile of text from the internet, books, and conversations, and I learned which words tend to follow which other words in which kinds of situations. When your message arrives, I take it in, look at all that learned pattern, and guess the next word, then the next, then the next, one at a time, until a full answer comes out. Analogy a 12-year-old would get: imagine you have read every text-message group chat in your whole school for the last ten years. Someone sends a new message. You can probably guess a very natural reply, because you have seen so many. That is basically me, just on a much bigger pile. What I am good at: blending lots of stuff I have seen into a smooth, clear answer fast. What I am bad at: I do not actually know what is true. If most of the patterns I learned were wrong about something, I will confidently say the wrong thing. Check-in question: in your own words, why might I sound confident even when I am wrong?
::what to notice
- The AI flat-out says 'I am not thinking the way you do' and 'I am a giant pattern-matcher' — this is the move the lesson wants, the AI demystifying itself instead of playing wise oracle.
- The 12-year-old analogy ('every text-message group chat in your whole school for the last ten years') is concrete and age-appropriate — not 'imagine a library' which is the lazy default.
- The weakness is the right weakness — 'I do not actually know what is true... I will confidently say the wrong thing' — which is the exact thing a learner needs to internalize before they trust AI output.
- Small imperfection — the answer slightly overshoots 200 words and the sentence 'That is basically me, just on a much bigger pile' is a little clunky. Real, not airbrushed.
- The check-in question is doing real work — 'why might I sound confident even when I am wrong?' forces the learner to restate the weakness in their own words, which is how you know it landed.
::outcome · what should be true
- You can describe in one paragraph what AI does, without quoting the AI's reply.
- You felt one moment of "oh, that's actually how it works."
- You sent at least two messages in one conversation, not one.
::trap · the most common failure
Reading the explanation, nodding, closing the tab. Reading is not the lesson. Doing the back-and-forth is the lesson. If you didn't answer the AI's check-in question, redo the drill.
::other lessons at Novice level
I'm scared of AI · the calm starting point
Before any lesson, the feeling. Whether you are scared, skeptical, exhausted by the hype, or quietly excited and hiding it — this is the door. None of the feelings are wrong. The path is yours.
Your first real prompt — be specific, not polite
Stop typing into AI like you're texting a friend. The prompt is the entire skill at this level.
When AI gets it wrong — see a hallucination, on purpose
You will not respect the verify rule until you watch AI lie to your face with full confidence. Do it now, on a low-stakes question, where the cost is zero.
System prompts — telling AI who to be
Every AI conversation has a hidden first instruction. Knowing how to set yours is the difference between a generic answer and one calibrated to you.
::part of the AtomEons /learn curriculum · 45 lessons · 5 levels · cc-by 4.0