The same material, sequenced three different ways. Each route says what already exists and what is still being written.
Editable decks built from the guide text, so the slides and the reading a student is set never contradict each other.
None of these decks is downloadable yet. The outlines below are the commissioned scope — slide counts and section breakdowns are what each deck will contain, and each one waits on clinical review of the guide text it is drawn from.
Unlike the decks, all of this exists now.
Questions with no settled answer, for a room that is meant to argue. Each one has a page on this site taking a position you can push against.
If a model performs worse for one subgroup but better on average, should it ship?
AURORA's answer is no — subgroup performance is a release gate, not a limitation paragraph. Ask the room what that costs, and who pays it.
Who is liable when a clinician disregards a model that was right?
And when they follow one that was wrong. The override is permanent by design; liability is not settled by design choices alone.
Does federation actually protect privacy, or just relocate the risk?
Aggregates can still leak. Have students attack the boundary rather than accept it.
Should patient advocates vote on technical decisions they cannot evaluate?
Observer status was proposed and rejected. The rejection reasoning is worth reading before the seminar.
What is lost when patient education is written at age-11 reading level?
Precision, sometimes. Have students find a sentence in a guide where simplification changed the meaning — then rewrite it.
Tell us what you taught and what fell flat. Course feedback changes the guides faster than anything else we receive.
Guides are versioned and dated. Cite the version you taught from, not the live page.
Citation formats →