> my first thought is "wow, that's a lot. And some of those aren't even in the same family of language?"

The meaning of the historical linguistics term "family" doesn't map directly to our common usage.

A language family is a collection of languages for a which we can identify a significant number of shared core historical grammatical and lexical features.

Just because languages belong to different families doesn't mean they don't have a shared ancestry, just that we can't see it in the modern languages.

If anything, in PNG population genetics identifies the highland populations as closely genetically related, but their languages have diverged to the point that signs their shared ancestry have been lost.

There is an theory that the linguistic diversity PNG represents the state of human societies in Southwestern Asia just before the emergence of agriculture - possibly explaining why many ancient attested languages like Sumerian, Elamite, and Assyrian which were in close proximity appear to be entirely unrelated to each other.