Session 3 · Video
Choose one small route from what actually happened in Session 3: inspect a sequence, translate an activity, or critique a claim. Building is an optional extension. No live model, polished project, or direct AI use is required.
Choose one frozen clip, or create a new pair of drawing runs in Coherence Animator. Follow one feature across time and mark the first exact change. Separate what you observed from why you think it happened. If you compare Run A and Run B, name the practice-and-order caveat instead of calling the result controlled.
~10–20 min · one observation · no account required
Make a one-page activity for your learners using drawing, a frozen clip, or no AI at all. Include what the activity makes visible, what it does not model, one piece of evidence learners will record, and one question about defaults, labor, consent, or source verification.
~25–45 min · one reusable activity card
Choose a clip or a claim such as “smooth video is evidence” or “nobody knows how AI works.” Label what you are reading: observed activity, participant hypothesis, guest argument, documented mechanism, or unknown. Then name the source, corroboration, provenance, consent, or technical documentation still needed.
~15–30 min · one claim audit · no generation required
Make, adapt, question, or deliberately refuse something from any of the three sessions. Session 3 left several concrete starting points: add onion-skin support so static details need less redrawing; explore timing and spacing in one subtle motion; sketch a “change one part of the frame” workflow; build a provenance activity; or bring a set of questions instead of software.
Prepare a three-minute share: the thing itself, one piece of evidence behind it, the limit of your claim, and the question you want the group to help with. Rough work is welcome; slides are not required.
~20–60 min · any form · one thing you can point to