All pages in this guide
Start Here
A short ladder of real tasks. Each rung is something you do in whatever tool you already have open, and each one leaves you a little better at using AI on your own work.
This isn't a course you finish and file away. It's a handful of small wins, done one at a time on the things you already care about: a paper, a job search, a lesson plan, a side project, the admin that eats your evenings. By the time you've climbed a few rungs, using AI well will just be something you do.
Two promises before we start. You can begin completely free, no subscription, no card, not yet. And you're not going to fake your way through this: the first thing we build is real understanding, not button-clicking.
The reframe: read this before anything else
You are not "studying to become an AI expert." That framing makes anyone feel like a fraud. Here's the true version instead: you are using AI on your own real work. That's the whole job right now.
Once it visibly works for you, once the first draft writes itself and the thing you were dreading takes ten minutes instead of an hour, you'll understand these tools better than almost everyone around you. Not because you memorized anything, but because you used them on something that mattered to you.
Life is unpredictable, so nothing here runs on weeks or deadlines. Every step below ends the same way: you're done when a specific thing exists. Not "by Friday." Not "in three weeks." When the thing exists, you've cleared the rung.
That means a busy month where you touch none of this costs you nothing. The ladder waits. There is no behind to fall.
First: understand how it actually works (free)
Before any setup, spend a little time just understanding what AI is. Not to become technical, but to become the person who knows what these tools are doing instead of poking at them and hoping. That understanding is what turns "I use ChatGPT sometimes" into real, transferable skill.
The good news: the best explanations in the world are free videos, and you can play them in the background while you do something else. No subscription, no notes, no homework. Just let them wash over you a few times. Here's the short list, easiest first.
Start visual and short (3Blue1Brown). Calm, animated explanations:
- But what is a neural network?, what an AI is, drawn out in pictures.
- Transformers, the tech behind LLMs, how ChatGPT-style models work, still visual and gentle.
Then go deeper (Andrej Karpathy). Longer and plainer-spoken, good background listening:
- Deep Dive into LLMs, the big one. If you only half-absorb it the first time, that's fine. Play it again another day.
- How I Use LLMs, practical: how someone fluent works with these tools day to day.
Want a fuller visual series to keep going? This playlist walks through the ideas step by step.
You don't have to finish all of these before starting the hands-on steps below. Do both at once: understanding grows in the background while your hands learn the tools. But don't skip it. The understanding is what makes the skill real.
What's free and what costs money
You can go a long way without paying a cent, so start there and only pay once it's clearly earning its keep.
Free, start today: all the understanding videos above; ChatGPT, Claude, and Gemini all have a free tier; and Google's NotebookLM is free with a Google account. That's enough for Milestones 0, 1, and 2.
Paid, for later: the paid tier of any of the three (roughly $20/month as of writing) gets you the better models, higher limits, and the coding-agent step in Milestone 3. Don't pay yet. You'll know it's time when the free tools are already saving you real hours. At that point the subscription pays for itself in a week.
The first few rungs are deliberately easy, free, and pressure-free: just "look what you can already do." Start there.
-
0
TODAY · FREE · NO SETUP
The 15-minute win
Before you set anything up, before you finish a single video, let's just feel it work once. Today. In whatever tool you already have open.
Do Open the AI you already use: ChatGPT, Claude, or Gemini, all free, no card. Paste in something real you'd send or hand in: a few emails you keep rewriting, a rough paragraph of an essay, a message you're nervous about, the bones of a cover letter. Then type: "Rewrite this clearer and more like me, plain, warm, not stiff." You can even talk it in with voice-to-text. Read what comes back.
You're done when there's one AI-drafted thing sitting in front of you that you'd genuinely use as-is.
Pages for this step: Why This Matters, The Basics
-
1
FREE · ONE AFTERNOON, SPREAD OUT
Load your work in
The 15-minute win was the AI knowing nothing about you. Now you teach it, once, about the thing you're working on, so it never asks again. This is the afternoon that changes how the whole thing feels, and you can do it in ten-minute pieces.
Do First, pick a tool to make your home base. If you already lean on one, just keep it. Then set up a saved workspace that knows the thing you're working on. That's the notes for a course you're taking, the details of a job search, the material for a unit you're teaching, or the specifics of a project or a small business. Follow the setup and projects pages, and use Your Work to decide what's worth loading in.
Keep it moving on a no-time weekNotebookLM can turn your own notes into an audio overview you play back on a walk. Keep the understanding videos going in the background too; the two threads feed each other.
You're done when you can open a fresh chat and start working without re-explaining who you are or what you're doing.
Pages for this step: Setting Up ChatGPT, Making Projects, Your Work, Setting Up Gemini
-
2
FREE · A COUPLE OF WEEKS OF NORMAL WORK
Live in it
Now you just use it. This isn't a lesson, it's your actual work, done inside a tool that already knows you. The goal here is for it to stop feeling like a tool and start feeling like a habit.
Do Run everyday work through it: the email you're avoiding, the reading you need summarized, the outline you're stuck on, the "here's my situation, what am I missing?" question. Read the prompting and checking pages along the way, so your answers get sharper and you can spot when the AI is confidently wrong.
Keep it moving on a no-time weekBoth of those pages work as audio if you'd rather absorb them while your hands are busy. And this is a good stretch to finish the deeper understanding video if you haven't.
You're done when reaching for the AI on a real task is automatic, you do it without thinking about it.
Pages for this step: Writing Prompts, Checking the Output
-
3
THE PAID STEP · CODING AGENTS
Make something real
Only once Milestone 2 feels completely normal. This is where AI stops talking about things and starts building them, and it's the first step where a paid plan earns its keep. Do it when the free tools are already saving you real hours, so the subscription pays for itself almost immediately.
Do Build one real thing with a coding agent: an AI that can create and edit files for you, no prior coding needed. Pick something small and yours: a simple one-page website, a tidy set of study notes turned into a quiz, a folder of old writing rewritten in a fresher voice. Read the codebase and next-level pages first so you know what you're walking into.
If you like a deep dive firstFor the curious, there are long, free walkthrough courses that make good background listening, but none are required. You learn a coding agent by doing one small thing, not by watching ten hours first.
You're done when a real thing exists that you made with a coding agent: a page you can open, or a folder of finished files.
Pages for this step: Coding Agents, Step by Step, Codebase, IDE, CLI, The Next Level
Where this can go next (optional)
Those four rungs are the whole path. Climb them and you're good at this, better than almost everyone who only ever chats. Everything past here is an optional branch, taken only if it pulls you in:
- Go deeper on your own world. Your Work is about pointing the AI squarely at whatever you personally do, your studies, your job, your project, and loading it in properly so the AI knows your world cold.
- Do this for money. If it interests you, Getting Paid covers how people turn these exact skills into income: what others pay for, and how to take on a first small job without drowning. Not the destination. Just a door, if you want it.
One last thing
Notice how you got here. Not by memorizing jargon, but by understanding the tools and using them, one real task at a time, each one leaving you a little more fluent. That's why it can't be faked: you understand what you're doing because you did it on something that mattered to you.
And if life gets loud and you don't touch this for a month, that's allowed. The ladder is built out of "you're done when," not out of dates. It'll be right here when the next quiet hour shows up.
You're not learning AI in the abstract. You're using it on your own real work right now, for free, to start, and every rung you clear leaves a real thing behind. Do the next rung. Nothing more.