I do, however, believe that g can be a somewhat useful indicator.
In the auxiliary, however, it is, iirc, the second most highly correlated with intelligence (i.e., ITJs).
The most highly correlated function with intelligence, when in the dominant position, is Ni.
Interestingly, when it's in the auxiliary, though, Ni is not all that highly correlated.
And, for honesty's sake: IN is more correlated than Ni dominance (i.e., P>J).
(And, ftr, these conclusions are mostly drawn from one study.)