Basically what he is implying is that the 6800 will vastly improve as new drivers are released which, of course, is pure speculation. He's also trying to defend them by implying that the drivers are immature, which leads us to ask "why would they release something that isn't performing anywhere near its potential?" And as far as NV30 goes, nVidia's CEO acknowledged that the NV30 had some serious engineering bungles, and in this era of prolific shading and DX9 the R300 cards continue to demonstrate top-notch IQ without sacrificing performance.
The battle between R300 and NV30 is far from over as the next-gen titles that these cards were designed for are only beginning to hit store shelves; it will only get more interesting as the year goes by. Sure the 6800 supports some new stuff that R4x0 does not, like PS3.0, among other things- things that won't be implemented in games for quite a while and that have questionable real-world differences in image fidelity. That's nice if you're working on Unreal 3, but let's see if the 6800 gets trumped in Half-Life 2 by its enemy's little brother the X800 Pro. Next-generation feature support is for developers- by the time something like Unreal 3 comes out, we will be pre-ordering Quake 4, and the 6800 will be a mid-range card and there will be a new gen. of cards for enthusiasts to choose from.
Wolfram Tismer criticizes the new ATi cards for their "today performance, an overglorified R300 with double the pipes..." People don't upgrade for games that are coming out in two years, they upgrade to enjoy improved performance in current and upcoming titles, and ideally, hopefully, the card they choose will age gracefully *ahem*. And anyway, in the end, price-to-performance is the ultimate benchmark, and in many cases the X800 Pro beats the pricey 6800U. ATi is no longer the underdog, they are a serious competitor and a threat to nVidia.