It's reasonably consistently measurable and has demonstrated correlations with both causative factors and outcomes. You can legitimately question the nature of what is being measured, but it's obviously something real (even if it's a composite).

?

> ... whether general "intelligence", or g, even exists.

I'm talking about the one that is for example measured in standardized IQ tests.

That's a little bit circular, though: "Intelligence is what is measured by an intelligence test."

And he is saying the IQ test measures very little, though the test appears at least to be self-consistent.