This isn't really news; there have been many projects with similar goals. As a closest example I can point you to Adobe's Alchemy. It was never really 'hot', but it is similar in concept. It was a step in the 'near-native speed' direction, though it was still far.
An important difference is Alchemy compiled C++ to Flash bytecode, and Flash support, while decreasing is still way higher than the Google Chrome market share.
A personal (and subjective) observation: developers don't trust Google that much. The hyped language Dart that Google released a few months back, which promised to do wonders for the development of large web projects,
was pretty much a debacle. It gained little adoption, and drew (sometimes undeserved) criticism to Google.
EDIT: As pointed by Javier, Dart is still in very early stage (alpha release), though it did manage to gather a lot of criticism.
So the success of NaCl will pretty much depend on the quality of the products that depend on it, and the effort Google throws to make sure it stays around for sure (support, marketing, 'strategic partnerships', etc). And at least the second one is questionable: Google have thrown out many underdeveloped projects that had promise.
Moreover, virtually every other browser has similar tools: in IE, you can write an ActiveX; in Firefox, you can create an XPCOM object; Safari has support for this, too, though I don't know the technologies involved.
Bottom line: it's not revolutionary and it's probably good only for:
- Legacy software that has to make a fast jump to the web
- Computations inside a web-browser (e.g. BOINC).
So I don't expect it to revolutionize anything. JS already can do (and does) everything but these pretty well.