If I am understanding your argument correctly in context, based on what you quoted, you are saying that the impact of genetics on g was found to be insignificant in the Hart & Risley study. That objectively seems not to be the case. First, the measurement outcome of the study was language use. They didn't attempt to measure IQ or g except when the children were 3. As far as I can tell from references, without access to the primary source, they did not perform any IQ tests on the children at later dates. I verified this with google searches within the book which only returned mentions of a test given at age three, but this is not iron-clad given google books' search mechanism. As far as I can tell, any claims that IQ was shown to be increased by the intervention are based on the very high correlation found between IQ and vocabulary at that age three test. Second, they only evaluated children (for language use) until the age of 9-10, which would be broadly consistent with the consensus understanding of psychometricians, which is that the heritability of IQ increases throughout childhood (implying in simplified language that non-heritable environmental factors have lesser correlation with IQ at adulthood, even if they have a greater impact during childhood).

The study wasn't designed to measure the impact of genetics on g, so we shouldn't be surprised that it fails to measure it. It was designed to measure the impact of in-home language use on vocabulary acquisition, which it clearly does well. That effect has been further validated in meta-analysis, e.g. [1].

1: http://www-personal.umich.edu/~sbneuman/pdf/marulisNeuman.pd...