Statistical Technicians are those who are well versed in the science of statistics. They're the ones who can fire up a command line and text editor, implement a classifier in a couple of hours (knowing what sort of classification to do), do a couple of regression tests, fire up a couple of test runs, mathematically comment on the accuracy of their data, and then go about gloating about the merits of their school of thought (Ha! Bayesian's rule at spam filtering!!!).
Statistical Philosophers on the other hand aren't necessarily that good at the technical aspect of Statistics. Sure, they can knock off some simple hypothesis testing with R or make some API calls to a library, but that's not their forte.
No, their true strength lies in a peculiar perspective: While their technical cousins can model data to fit their world view, these philosophers have the ability to model their world view to fit data.
ie: They can project themselves unto the statistical analysis, and accurately answer the question, "Given this randomness, what do I do now?"
Needless to say, the most dangerous people are those who are both.