The way I understood it (and I could be wrong
![:lol: :lol:]()
) was that the pixel and vertex shaders are fully programmable, but the rest of the card is not. A fully programmable GPU would be like a CPU that is heavily optimised for 3D rendering and with a huge amount of bandwidth.
The idea behind having less lag time for developers is that they could start building features into their games before there is any hardware to support it, since they will be able to specify which functions the chip carries out using a firmware image.
"Ignorance is bliss, but I tend to get screwed over."