I don't think that article was as conclusive (well, at least as surprising) as some are making it out to be.
My understsanding of the Prescott design was not that it would be able to be faster on a per-clock basis than Northwood, but that it would actually be able to be clocked significantly higher.
Now, I guess the jury is still out on whether Intel can reach their clock speed goals, but THG was able to get more overclocking out of the Prescott test chip than the Northwood one, so perhaps the design is not terribly bad (yea, yea, this is only one test sample, but still).
So, it seems to me this article isn't telling us anything that we shouldn't have already known if we had been paying attention for the last several months.
-D'oh!