The barrier to adoption of Flex has always been the requirement of Flash Player and the feeling by many that developing with it isn't "standards-based" and instead opt for an DHTML / AJAX solution. Therefore I don't think that Blaze DS is going to push a large number of people towards Flex-based RIAs. Even before Blaze you could write some pretty large-scale applications using of XML-based services (RPC, SOAP, RESTful). It only became problematic from a performance issue when you were dealing with some very large data sets (thousands of large records at a time). Blaze is also a Java-only solution and there are lots of other types of backend technologies being used with Flex: PHP, .NET, Python, Ruby, ColdFusion, etc.
Speaking to the market, I can tell you that I joined one of the largest Flex shops in the world about three years ago and have seen our company grow at least 5x during that period, due almost exclusively to increased demand for Flex. We are also seeing significant growth with Silverlight and WPF as Microsoft rolls out these solutions to their huge developer community. Either way, I think that "player-based" RIA is a lot more than a flash in the pan (no pun intended) and is going to continue to experience incredible growth. The web browsers won't be able to keep pace with the innovation of the player-based runtimes and HTML 5 is nowhere in sight.