AI Fundamentals

How LLMs actually work — enough to be a calibrated user, not a magical thinker.

Vocabulary primer — the words you'll hear every day. No fluff, no jargon-for-jargon's-sake.

What goes in the window, what stays out, and how prompt structure changes the answer.

Reusable shapes that work: role, examples, format, scaffolding, anti-patterns.

When to reach for Sonnet vs Opus vs Haiku. Cost intuition without a spreadsheet.

Breaking a fuzzy ask into pieces an agent can actually finish.

Ask vs Edit vs Agent vs Plan — what each does, when each wins.

Pragmatic comparison for teams choosing or running both side by side.

.cursor/rules and AGENTS.md — repo-level settings that compound across the team.

The Cursor shortcuts that earn back the most minutes per day.

Use the agent loop as a TDD partner — write the test, let it iterate to green.

Stack trace → reproduce → bisect → fix. Where the agent helps, where it hurts.