I’ve seen this trick used in a few Epic examples.
This first was presented as a way of “collapsing [the mesh vertices] to an infinitely tiny point”, which then helps to achieve the dynamic mesh texture effect the presenter is going for. Source: [GDC 2017, Content-Driven Multipass Rendering, 5:00:40][1]
This second example I found in the Math Hall of the Content Examples project.
Both example use the same “subtract Absolute World Position, add Actor Position” trick but I don’t understand how it works, or what it does. Can anyone shed some light on its functionality?

