Presumably there are plenty here happy to eat their words. 😀
Quadros are poor for rendering (they don't have as many CUDA cores), eg. A GTX 580 leaves
a Quadro 4000 in the dust for RT3D rendering in AE.
Lots of RAM is very useful for pro tasks that do need it such as GIS, medical, defense sat analysis,
etc., but also media-related apps such as AE. Titans are aimed at budget professionals who can't
afford Teslas, or for whom the additional features of a Tesla are not relevant, such as the full speed
return path, additional caching mechanisms and unlocked 64bit fp CUDA support.
Note that if the Titan-Z announced today employs the standard method of expoiting two GPUs,
then its 12GB RAM is almost certainly just 6GB per GPU, not a single shared 12GB, in which case
IMO it's just a 6GB card since each GPU can still only make use of 6GB. Calling it a 12GB card is
the same marketing fluff as used by most vendors who sell dual-GPU cards.
Ian.