I think you were gone when I started 16 threads, one for each type, to look at the cognitive processes results of people allegedly of each type. The results were really interesting, but, as an Ni-dom, I still stand back, and don't choose whether I believe strongly in the camp that people do normally develop their first two shadow functions, or whether they really only use their four normal functions for the most part, and that people just learn to do similar things that their first two shadow functions are usually known for, but they're actually just utilizing their first two normal functions to do so (@uumlau
used to be a believer in the former, but recently said that he's switched over to the latter -- I remain a little more neutral on the matter).
Wherever the truth may lie: I think the second shadow function is undoubtedly much more difficult to develop than the first
(which is not to say it cannot and does not happen), so, in the same line of reasoning as that which I said at the beginning of this post -- that I don't think you'll normally see heavy Ti usage in an INTJ -- I don't think you'll normally see heavy Ni usage in an INTP.