Why we chose this

A hundred and thirteen million photographs of New York, put through a model and rendered as motion. The city handled as a dataset. Early enough (2021) to have set the visual grammar everyone else in this room is still working against.