compared to old-school bone animation where you send the info to the vertex shader every frame?
If I understand correctly this really isn't a thing anymore. Instead most modern engines copy the original vertices to a temp or skinned vertex buffer with a compute shader and then use normal non-skinned rendering techniques to display them. This means you don't need 2 copies of every shader, one with skinning and one without so less combinations.
That doesn't change the other answers to your question.
The paper describes that you still need to skin the cage (if you want to approximate skinned animation). The Youtube video describes that the rays get deformed instead of the geometry. Rays get always transformed from world space to the local space of the currently traversed instance in the TLAS.
Comments
If I understand correctly this really isn't a thing anymore. Instead most modern engines copy the original vertices to a temp or skinned vertex buffer with a compute shader and then use normal non-skinned rendering techniques to display them. This means you don't need 2 copies of every shader, one with skinning and one without so less combinations.
That doesn't change the other answers to your question.
The paper describes that you still need to skin the cage (if you want to approximate skinned animation). The Youtube video describes that the rays get deformed instead of the geometry. Rays get always transformed from world space to the local space of the currently traversed instance in the TLAS.