Review for NeurIPS paper: ShapeFlow: Learnable Deformation Flows Among 3D Shapes

I very much enjoyed reading the technical part. On style: The only downside was the made-up praise of non-parametric aspects. It is not that the method "preserves" any details. I think authors know better and poor students and industry readers will suffer not benefit from such marketing. What happens is that the pick up truck has a few little details and fenders and screws which make it look quite cool.