← 101 BuildWeek 1 · Nov 9 – Nov 13
Cursor and the IDE
working locally with AI
Student works in Cursor on their own laptop: knows the layout, has caught the AI fluent and wrong, has an AGENTS.md the agent obeys, one custom command and one skill, and a one-page product running on localhost.
Monday · 90 minutes live
- 5′Open: six weeks, and this week nothing leaves your laptop.
- 22′Install Cursor together, privacy mode on (on camera), open the course folder. Tour: files, editor, chat (Ask vs Agent), @filename mentions, accept/reject a change, undo with a checkpoint, the browser preview.
- 15′First conversation with AI, Ask mode, about something you know cold. Catch it fluent and wrong; one line in the shared doc.
- 8′What you just saw: predicts the next word, wants to finish, model vs product. The chat is a guess; the page is the truth.
- 20′AGENTS.md: rules the agent reads every time, because rules typed in chat fade. Write six lines (including 'ask me 3 clarifying questions before you build'). Then ask the agent for a one-page product site using @PRODUCT.md and open it on localhost. Test one rule: ask for a key; it should refuse.
- 15′Commands and skills: a command is a saved prompt you run by name (/done-check); a skill is a folder of know-how the agent pulls in when it is relevant. Build /done-check together; demo one skill.
- 5′Habits for a local week: end every session with the page open in the browser; keys never in chat; transcripts are kept. Buddy pairs posted.
Tuesday to Thursday · you build
- PRODUCT.md: who it is for, what they do today, what would change. No tech words.
- Three changes to the page, each in a new chat, each checked in the browser and with /done-check.
- Two-run table in week-1/: same ask with no definition of done, then with a written one. Screenshot both.
- One skill of your own: my-product, holding PRODUCT.md and how you want the page to sound.
Office hours · two 30-minute blocks
- 1Tuesday: install problems, locked-down laptops, the IDE tour again for anyone lost.
- 2Thursday: show me the page on localhost, your AGENTS.md, and /done-check running.
Exam whisper · five minutes of vocabulary
Create the OpenAI Academy and Anthropic Academy accounts. Exam Whisper: AI vs. ML vs. GenAI, foundation models, training vs. inference, context window.