I think we've had enough of those in the past decade. Each one has something new to offer, but at the same time it disperses attention of the programmer crowd and so less effort is put into each language's development.
I think we need to experiment more with single-purpose languages. I'm not saying DSLs because those are mostly embedded syntactic constructs in existing languages. By single-purpose I mean an independent language (not just new syntax, but semantics as well, and possibly new runtime) that does one thing and does it well.
I believe such a language will have much more luck in becoming the single good enough solution for a particular problem (any problem it chooses to address, but only that one). There is really no hope for any general-purpose language to be a good fit for the whole wide spectrum of problems modern computing needs to solve.