▲ Claude Academy
Sign in

Vault note

Prerequisite Track 0A — For People Who Aren't Technical

Welcome. You don't need to code. You don't need to have used AI before. You don't need a computer-science background, and you will not touch a terminal, an API, or a single line of code on this track. If you can type a message and read a reply, you can do everything here.

This is the gentlest on-ramp in the whole course. It exists so that nobody is left behind. Take it slowly — there is no prize for rushing.

Total time: about 4–6 hours, comfortably spread over 1–2 weeks at one short session a day. Go slower if you like.

Before anything: what Claude actually is, in plain language

  • Claude is a writing-and-thinking assistant you talk to in plain English. You type what you want; it types back. No special commands, no syntax.
  • It is made by a company called Anthropic, and it is trained to be helpful, honest, and harmless — meaning it tries to be useful, tries not to make things up, and tries not to do harm.
  • Think of it as the world's most patient assistant who has read an enormous amount, never gets tired of your questions, and will happily redo something "but shorter" or "but friendlier" as many times as you ask.
  • It is not a search engine, a calculator, or a source of guaranteed facts. It can sound confident and still be wrong. For anything that matters — a date, a number, a medical or legal point — you double-check, or you give Claude the document to work from.

That last paragraph is the single most important idea on this whole track. Everything else is practice.

The mindset that makes this easy

  1. It's a conversation, not a search box. Tell Claude the goal, who it's for, and what you want the answer to look like. "Vague in, vague out."
  2. You can't break it. There is no wrong button. If an answer isn't right, just say "no, more like this" and try again.
  3. Iterate, don't restart. Instead of rewriting your whole request, reply "make it shorter" or "sound warmer." Claude remembers the current chat.
  4. You are the editor. Claude drafts; you decide what's good. You're always in charge.

One-time setup (5 minutes, no developer tools)

Do just the first section of the setup-checklist:

  • Create a free claude.ai account. That's it. You do not need Node, Python, an API key, Claude Code, or anything a developer would install. Skip all of that — it is not for this track.

Your path (in order)

Do these four courses and three projects in sequence. Each course is a set of short, readable "hyperframe" lessons (see how-to-use-hyperframes) — watch Anthropic's video, then read the matching frames here, then do the project. Reading without doing the project does not stick.

Step 1 — Meet Claude · ⏱ ~60 min + project

  • Course: claude-101 — what Claude is, how to talk to it, models, Projects, Artifacts, and when to verify.
  • Project: p01-first-conversations — run your first 10 real conversations (summarize a PDF, draft an email, plan something) and keep a before/after of one prompt you improved.

Step 2 — Learn to work with AI, not just use it · ⏱ ~60 min + project

  • Course: ai-fluency-framework-foundations — the 4 D's (Delegation, Description, Discernment, Diligence): a simple framework for deciding what to hand to AI and how to judge what comes back.
  • Project: p03-ai-fluency-delegation-audit — look at your own real tasks and decide, on paper, which to delegate to Claude, which to keep, and how you'll check the results.

Step 3 — Know what to trust · ⏱ ~45 min

  • Course: ai-capabilities-and-limitations — where Claude shines, where it slips (hallucination, knowledge cutoff, math, bias), and how to get reliable results anyway.

Step 4 — Let Claude work alongside you · ⏱ ~60 min + project

  • Course: introduction-to-claude-cowork — Cowork is a way for Claude to work with you on a task in a shared space, not just answer one message at a time.
  • Project: p04-cowork-workspace — set up a small Cowork workspace and complete one real task in it end to end.

You're ready to move on when…

  • You can hold a multi-turn conversation and improve the answer with follow-ups ("shorter," "warmer," "add an example").
  • You can explain, in your own words, when to use Opus vs. Sonnet vs. Haiku.
  • You can name at least three things Claude should not be trusted to do unaided — and what you'd do instead.
  • You can apply the 4 D's to decide whether to hand a task to Claude.
  • You've finished p01, p03, and p04 and journaled one thing that clicked from each.

Where to go next (no pressure)

You have two perfectly good options, and both are fine:

  • Stop here. Seriously. With these four courses you can already get real, everyday work done with Claude — drafting, summarizing, planning, analyzing your own documents. Many people never need more than this, and that's a complete, successful outcome.
  • Keep going. If you're curious, the natural next step is track-1-claude-foundations, which firms up everything you just learned. From there the tracks gradually build toward prompting, and only much later toward anything technical — and you can stop at any track boundary.

Either way, you've done something real. Tick these off in progress and be proud of it.

See also