I did read the analysis and my point is still valid. While the GF104 isn't just a GF100 with even more cores fused off, but a GF100 derivative, many (most?) of the accomplishments of GF104 are simply due to it being a cut-down version of the original GF100. In any case, "revamping" the GF100 is pretty much a no-go. There's too much time and work involved to pull this off; time and work better spent on doing the actual follow-up to the GF100. I'd say nVidia's best bet for a high-end high-performing card right now would be to do a "GTX460X2", or "460GX2" if you will.