Regardless of your philosophy on purchasing at the high, middle, or low end (and I advise my clients not to pay the premium for top models), Chris is wrong in the analogies he used:
A 2.16Ghz Macbook Pro has an 8% clock advantage over the 2Ghz model, which is likely to produce perhaps a 4% speedup, all else being equal. But that SE/30 (which cost me less than $3400 in 1989, so it could not have been a $3400 difference!) offered both twice the clock speed AND twice the datapath of the 68000-based SE or Classic. In practice it was at least 3x as fast. That opened up a whole category of applications that weren't feasible before, basically anything that ran on the Mac II, which was equivalent to the research workstations of the day.
The same situation existed when the 8500/9500 replaced the first PPC machines. A next-generation processor gets you longevity as a small bonus--the only good reason to buy it is for the new types of work it will make feasible or profitable
After all, if your work isn't changing, why get a new computer at all?
Specification Obsession: Is a Souped-up Computer Worth It?