Merritt Ruhlen was my guy. He not only claimed that all languages sprang eventually from a single initial source, he provided evidence that he said made that the mostly likely reality. He was almost universally reviled a generation ago, and is still disdained, though younger historical linguists are giving him some grudging acknowledgement. A PhD linguist who was a student of Joseph Greenberg of Stanford, likewise dismissed as terrible. Ruhlen recently died, and I think missed the new tidbit of DNA evidence that speaks in his favor. It is not the first. Research from other field, increasingly including archaeology and genetics, have vindicated them. Plus other linguists had always sulkily agreed that both were excellent at classification.
I have written about the controversy many times before, if you are interested. About half of these don't pertain, though. You will have to skim.
For the moment, the interesting piece is his claim that the Kusunda language was related to Juwoi in the Andaman Islands. They are a thousand miles apart, with lots of mountains and oceans creating barriers. Worse, there are no languages in between them that would suggest a connection, and at the time of the claim, it was believed that these peoples had not been in any contact for 80,000 years. Linguists believe that language relationships cannot be detected beyond 10,000 years, so this was hooted at. (When I was in school anything more than 5000 was considered suspect. So things change.) Things got a bit better over the years and the time distance dropped to 60,000 years, but still excessive. But new DNA evidence suggests that because one population pushed out another on the way to the islands, the real number is now only 25,000 years. Even that would be a lot, but because one of the main similarities is a pronoun ("He, she, or it" gita versus kiteh), it is looking like less of a stretch, as pronouns are some of the most stable words in families of languages. Ich in German and ego in Greek are separated by a few thousand years, for example.
Genetics and language relationship are not proofs of each other. One only needs look at America to see that a lot of genetically diverse people are all speaking English. But it happens often enough that it is the first line of inquiry when there is any question.