I think you are giving them too little credit. Even Intel and AMD have had missteps and arsed up designs. (Itanium and K6 anyone? Rambus system RAM?). Does Itanium prove that Intel engineers SUCK? No, sometimes people try something different, and market conditions, timing, and competitor's products control whether or not it is seen as a glorius victory or suckage. AMD went conservative, and not Intel is forced to follow suit. NV30 wasn't a bad design, it simply wasn't good enough. If ATI hadn't shipped the R300, people would be quite satisifed with the boost the NV30 delivered over the NV25. It is only because ATI did such an incredible job that we view the NV30 as a suckage. ATI went for full FP, Nvidia thought long instruction limits, stencil, and fast integer was the right strategy. ATI guessed the market right.
If you've been on this board long enough, you know how quickly tables can turn. People once thought 3dfx was unstoppable. Then they thought NVidia was unstoppable. Now a new group of kids is once again proclaiming that a large faceless corporation has an unbeatable lead, and won't make any mistakes to give it up, nor will any new competitors arrive. How ironic it will be if PowerVR comes up with some magic VS/PS3.0 chipset that blows the doors off both R420 and NV40. Then we'll all be debating how ATI and nVidia both "arsed up" their 3.0 pipelines.
But see, you've already come to the conclusion that the NV40 is arsed up, without any real knowledge. And here you are in another thread instructing me to "wait and see how PCI Express performs", yet you've already passed judgement on anything NVidia can possibly do.