Install Claude Code (or Codex), meet your agent, and read the nine foundations notes: what a model is actually doing, tokens, the context window, why answers differ, what makes an agent, what an API is. Then build the warm-up: a job-watch agent on our live job board, with a two-line prediction written before you start.
Many learners arrive having already done this as free users. If that is you, week 1 starts at full speed.
It ends with
Working tools, the vocabulary, and your first agent in your own repo: a watch that reports what changed on the board since it last looked, run twice without duplicating anything, plus one sentence about the gap between your prediction and reality.