Yeah I think its not going to work with just a conversion layer due to the types of effects that you can have in audio vs video. Since from reading I gathered that they are using the shader programming stuff... it won't be easy to translate whatever random code is written in a vst to be able to yield the same results on those shaders... but who knows.. if they abstract the wrapper enough it could work. I'd have to check out some APIs and such before coming to a good conclusion... but hey that's their proprietary code now :/

Maybe there will be an open standard for this sort of thing in the near future.