Labs — Work on Your Own
Labs — Work on Your Own
The chapters teach the ideas. The labs make you sweat.
Each lab is a self-contained exercise drawn from the original course assignments. They are deliberately stripped down — a brief, pointers to the data, a prompt or two, a deliverable — so a reader without an instructor in the room can still do them. Think of a lab as a half-day to a day of focused work; the harder ones (Labs 6 and 8) take longer.
How a lab is structured
Every lab has the same five parts:
- The brief. What you are being asked to do, in business or research terms.
- What you bring. Data, prior chapters, prior labs.
- What to do. A short list of steps. Not a recipe — choices to make.
- What to hand in. A concrete deliverable: a notebook, a one-page report, a graph, a rendered PDF.
- AI and me reflection. Three questions, written down. Required.
Mapping labs to chapters
| Lab | Topic | Read first |
|---|---|---|
| Lab 1 | First contact with an LLM | Ch. 2 (LLMs) |
| Lab 2 | Documenting code and data | Ch. 5–6 (Documentation) |
| Lab 3 | Drafting a report with AI | Ch. 8–9 (Reports) |
| Lab 4 | First agentic-CLI session | Ch. 11–14 (Terminal → CLI) |
| Lab 5 | Reproducible CLI workflow | Ch. 15–18 (Advanced CLI) |
| Lab 6 | Data-to-PDF on real CPS data | Ch. 16 (End-to-end) |
| Lab 7 | Reading text with an LLM | Ch. 19–20 (NLP, text) |
| Lab 8 | LLM sentiment vs. human sentiment | Ch. 21 (Sentiment) |
| Lab 10 | IV thinking with AI | Ch. 24 (IV) |
The numbering follows the original course assignments and is not perfectly contiguous (no Lab 9 in this draft).
A note on rules
Two non-negotiable rules across all labs:
- Use AI freely. That is the whole point of the course.
- Submit nothing AI created end-to-end. You must be able to defend every line and every claim. If you cannot, you over-delegated.
Then, three questions. Always:
- How did AI support me?
- How did AI fail me?
- How did AI extend me?