The first session in an internal AI workshop series I run. A 30-minute crash course designed so that, by the end, attendees know what Claude is, where they can use it, and what the buzzwords mean — without the jargon hangover.
Audience: anyone new to Claude. No prerequisites.
Topics covered
- LLM fundamentals — what an LLM is, who makes them, what a “model” is, thinking modes, the context window, tokens.
- Claude products — Claude.ai (chat), Cowork (desktop assistant), and Claude Code (developer terminal).
- The vocabulary — the jargon, decoded.
- How context actually works — what’s loaded at session start and why it matters.
- Memory across conversations — project instructions/knowledge, account-wide instructions,
CLAUDE.md, and automatic memory. - Where each thing lives — the simple rule for picking the right tool.
- Reference examples — global epistemic rules and project-specific instructions.
Each session ships with a self-contained HTML handout and a recording.