Everett's claims have provoked a firestorm of debate. I was lucky enough to take a seminar with David Pesetsky from MIT shortly after Everett's story began appearing in the popular media, and he discussed what he perceived as serious flaws in Everett's findings and methodology.
Nevins, Pesetsky, and Rodrigues have since published a lengthy and detailed rebuttal (
http://ling.auf.net/lingBuzz/000411) that shows that Piraha does in fact have recursion. In fact, the surprising thing about Piraha is how unremarkable it is. It shares traits with German, Hindi, Hebrew, and many more languages. Pesetsky, et. al. also believe that the argument that quantifiers do not exist is not supported by the facts. More details are in the paper.
As far as their culture is concerned, there is evidence from other anthropologists that the Piraha do indeed have creation myths (the rebuttal paper actually includes one on page 44).
People jump on stories like these because it makes for a good news story ("Rebel Linguist Takes Down Noam Chomsky"). But the evidence I've seen doesn't sound so convincing.