The Course
Linguist
Learn to use AI well — when to think alongside it, when to hand it a task and how to stay in control of both. No experience required.
Find your starting point →Two axes
Use the tool well. Or build with it. Or both.
Use is the original ladder — Beginner, Enabled and Advanced. It teaches you to talk to AI well, as a human. Build is the new axis — API, Tools, Agents, Production. It teaches you to build on top of Claude when the work has to run when you're not there. The axes are independent. Sit at any rung on either.
Your starting point
You are not behind. You just haven't started yet.
Most people starting new jobs right now are in the same place — they've heard the word "AI" more times than they've actually used it deliberately. The first three modules will get you to a point where you can use AI for real work tasks before you've been in the job a week.
Start at B-01: Your First Message → Retake quizYour starting point
You've already started. Now find out why it's inconsistent.
Skip B-01 and B-02 — you know how to send a message and make a starter prompt your own. Start at B-04, which explains the single most common reason experienced users get unpredictable results: the way you phrase your request is directly shaping what comes back.
Start at B-04: Your Input Shapes the Output → Retake quizYour starting point
Skip ahead. Your foundation is already there.
The A-track is not a more advanced version of the chat interface tutorials you've already read. It's a different mode entirely — Claude Code runs in your terminal, reads your actual files and can delegate discrete tasks to separate sessions. After this track you'll be able to control what tools it uses, write session briefs that produce evaluable outputs and decide when delegation saves time versus when it just costs more.
Start at A-00: Getting Started with Claude Code → Retake quizYour starting point
Different axis. Same building.
The Build axis is the one where you write the loop the model sits inside. You decide which model. You decide when it stops. You pay for every byte. API-01 is the cleanest test of whether building on the API is the right call for the thing you're trying to do — sometimes the answer is "no, use Claude Code instead." The honest decision saves you weeks. Sits on top of Anthropic's own free courses, not in competition with them.
Start at API-01: First call, the loop → Retake quizPrefer to set something up first?
The hands-on exercises.
Two short guided setups that give you a pair of working projects — one for exploring with the AI, one for dispatching tasks to it — on whichever platform you use. They pair well with the modules below and take about ten minutes total.
Open the exercises →
Stage 1
Beginner 10 modules
Ten modules from your first message to a strong opening seed. No prior experience required.
↓
Stage 1
Beginner 10 modules
Ten modules from your first message to a strong opening seed. No prior experience required.
Stage 2
Enabled User 7 modules
Seven modules that sharpen your verb choice, context loading and recovery skills — building on the Beginner foundation.
↓
Stage 2
Enabled User 7 modules
Seven modules that sharpen your verb choice, context loading and recovery skills — building on the Beginner foundation.
Stage 3
Advanced Practitioner 10 modules
Ten modules for working with Claude Code: cost awareness, session hygiene, context management, tool selection and agent delegation.
↓
Stage 3
Advanced Practitioner 10 modules
Ten modules for working with Claude Code: cost awareness, session hygiene, context management, tool selection and agent delegation.
The other axis
Build 4 modules · phase 1
Where you write the loop and pay the bill. API, Tools, Agents, Production. Phase 1 ships four modules — two in API, one each in Tools and Production — to prove the axis. More to follow.
↓
The other axis
Build 4 modules · phase 1
Where you write the loop and pay the bill. API, Tools, Agents, Production. Phase 1 ships four modules — two in API, one each in Tools and Production — to prove the axis. More to follow.