I basically agree. 2004 was setting the K8 foundation which is now complete, 2005 will be when AMD and the stock starts to capitalize. The only good thing about still having half of K8 starts on 130nm in Q4 is that they can still use the existing mature 130nm wafer volume exclusively for high bins until 90nm catches up whilst scaling down the K7 starts and replacing them with 90nm. Each quarter could be substantially different from the next. People are so used to Intel taking the lion's share of x86 spoils each quarter it maybe quite a shock if and when the status quo changes.
I wonder with all these continuous smoke and mirrors whether we will actually see anything new in 2005 from Intel, i.e. Prescott and Dothan may be it for 2005 too, a probability of 50-50 I would say.