Course catalog
Use Claude as a thinking workspace, not a search box.
Claude.ai is a product with surfaces, settings and limits, and most people use about a fifth of it. This course covers that product concretely: when a Project beats a one-off chat, what project knowledge and custom instructions actually change, how an artifact lets you revise a document instead of regenerating it, and what a file loses on its way into a message. It then turns to the harder half — writing a brief that produces a usable first answer, asking for disagreement instead of encouragement, and closing a long thread into a written decision. It ends on the practical week: picking a model by stakes, spending a usage budget on purpose, keeping secrets and other people's data out, and recognizing an answer that is fluent and false. No programming knowledge is assumed.
Section 1
The surfaces Claude runs on — chat, Projects, Cowork and Chrome — and the places you can park standing context so you stop repeating yourself: project knowledge, custom instructions and memory.
Section 2
Artifacts as revisable objects beside the chat, what a PDF or a photo loses on its way into a message, and how to draw more reasoning out of a model that now decides for itself how long to think.
Eligible first-time members can get 7 free days · Cancel anytime
Get startedSection 3
Using a conversation to think: a brief in three parts, sources reconciled rather than blended, critique you asked for on purpose, and a thread closed into a written decision.
Section 4
The week-to-week practicalities: matching a model to the stakes, treating whichever allowance or spend cap your plan exposes as a budget, keeping the wrong material out of the box, and catching an answer that is fluent and false.