In short, that a small correlation with performance in beurocratic tasks is not enough to make IQ representative of overall intelligence.

I assume you got confused since the article references his other body of work, I don’t think it’s meant to be read on its own.