How this shelf is organized
The courses are arranged from universal foundations to deep building. You can read a whole part, or jump to the one chapter your work needs.
| Part | What it covers | Who it's for |
|---|---|---|
| 1 · Foundations | How generative AI works, how to collaborate with it (the 4D framework), and everyday Claude. | Everyone |
| 2 · AI Fluency for your field | The AI Fluency framework applied to a specific world — students, educators, small business, nonprofits, creative work, builders. | Pick the one that's you |
| 3 · Working with Claude, hands-on | The products: Cowork, Claude Code, agent skills, and subagents — mostly no-/low-code. | Practitioners |
| 4 · Building on the Claude Platform | The developer path: Platform 101, the Claude API, and Claude on Bedrock and Google Cloud. | Developers |
| 5 · Extending Claude | The Model Context Protocol and the agent ecosystem, intro through advanced. | Advanced builders |
Every chapter is an independent study guide — a concise summary of one course's key concepts, followed by a quiz to test recall. The full, hands-on courses (with videos and exercises) live free on Anthropic Academy.
Choose your track
Find the row that sounds most like you and follow its route. Tracks overlap — Part 1 is the shared starting point for almost everyone.
| If you are… | Start with | Then |
|---|---|---|
| A curious newcomer — "what even is this?" | AI Capabilities and Limitations | Claude 101 → try Cowork |
| A knowledge worker who wants to use Claude at work | AI Fluency: Framework & Foundations | Claude 101 → your Part 2 course → Cowork |
| A student | AI Fluency: Framework & Foundations | AI Fluency for students |
| An educator or teacher | AI Fluency: Framework & Foundations | AI Fluency for educators / Teaching AI Fluency / pK-12 |
| A business, nonprofit, or creative lead | AI Fluency: Framework & Foundations | Your Part 2 course (Small Business / Nonprofits / Creative Work) |
| A developer new to Claude | Part 1 Foundations | Claude Code 101 → Platform 101 → the Claude API |
| An AI engineer / builder | Building with the Claude API | MCP intro → MCP advanced → agent skills → subagents → cloud |
Key concept · Foundations first, then diverge
Almost every track begins in Part 1. Understanding how the model behaves (Capabilities & Limitations) and how to collaborate with it (AI Fluency) makes every later course land better — whether you go on to write prompts, build agents, or teach a classroom. Read the two foundation chapters, then branch to the part your work lives in.
One idea runs through everything: the 4Ds
Across the catalog, one framework recurs — the 4D framework for AI Fluency (Delegation, Description, Discernment, Diligence), developed by Prof. Rick Dakan and Prof. Joseph Feller. It's the spine of the AI Fluency courses, it underlies Claude 101's prompting advice, and each of the four AI "properties" in the Capabilities course maps to one of the D's.
Key concept · Calibrated trust, not blind trust
The foundational stance this book teaches is calibrated trust: AI isn't uniformly reliable or unreliable — it's strong and weak along predictable axes, and your job is to locate a task on that continuum and apply the right amount of scrutiny. Delegation decides what to hand off, Description shapes how you ask, Discernment evaluates what comes back, and Diligence keeps the whole thing responsible. Keep those four in mind and the rest of the shelf is just detail.
Orientation check
Eight questions on how the shelf is organized and where to start. Answer first, then expand Show answer.
Part A · Multiple choice
Which part is the shared starting point for almost every track?
- A Part 1 · Foundations
- B Part 4 · Building on the Platform
- C Part 5 · Extending Claude
- D None — start anywhere
Show answer
A. Foundations (how AI works + AI Fluency) makes every later course land better.
A high-school teacher should, after the AI Fluency framework, go to:
- A The Claude API
- B AI Fluency for educators / Teaching AI Fluency / pK-12
- C MCP Advanced Topics
- D Amazon Bedrock
Show answer
B. The educator-facing Part 2 courses.
An AI engineer's recommended route starts with:
- A Building with the Claude API
- B Claude 101
- C AI Fluency for students
- D Enterprise Search
Show answer
A. Then MCP → agent skills → subagents → cloud.
What are the four competencies of the 4D framework?
- A Delegation, Description, Discernment, Diligence
- B Design, Develop, Deploy, Debug
- C Data, Diagrams, Docs, Delivery
- D Detect, Decide, Do, Done
Show answer
A.
Which part covers the products you'd use hands-on (Cowork, Claude Code, skills, subagents)?
- A Part 3 · Working with Claude, hands-on
- B Part 1 · Foundations
- C Part 5 · Extending Claude
- D Part 2 · AI Fluency for your field
Show answer
A.
Part B · True or false
You must read every chapter in order to get value from this shelf.
Show answer
False. Read a whole part, or jump to the one chapter your work needs.
"Calibrated trust" means deciding wholesale whether to trust AI or not.
Show answer
False. It means locating a task on the capability–limitation continuum and applying the right amount of scrutiny.
Part C · Short answer
In one line each, what does each of the 4Ds govern?
Show answer
Delegation — what to hand off; Description — how you ask; Discernment — evaluating what comes back; Diligence — keeping it responsible.