Conversational
Programming
A hand-drawn skyline: the Illinois campus inked on the left gives way to an imagined future city, then to pencil sketches, then to a bare horizon line running off the page.

Walkthroughs

An activity is hard to picture from a description, because it happens on several screens at once. So each one is recorded end to end: every screen in the room side by side, a step at a time, with a caption saying what just happened and why.

These are rehearsals, not recordings of a real class. The students are invented and the conversation with the agent is generated by the same model the course uses, but everything else is the software students actually use, driven the way they would drive it. We make these to check a flow before we run it; they're here because they're the clearest answer we have to "what do you actually do in there?"

  1. 4 min · 21 steps · 4 screens

    Fresh Eyes

    A partner looks at your website the way a stranger would and says what they see; you change it while they watch. Then you swap, then you find someone new.

    The activity itself

  2. 3 min · 19 steps · 4 screens

    Pair Prompting

    One student talks to the agent while a partner follows the conversation, asks, and keeps notes. Then they switch.

    The activity itself

  3. 1 min · 10 steps · 2 screens

    Your Website

    Connect Codex to the course with one line, then ask the agent the question almost nobody asks it.

    The activity itself