Keith has some discourse on what he'd like to see done with VB.NET.
As I mentioned back in July, and as the article he links to confirms, the language divergence seems to be less of a real technical divergence than a divergence in perception and marketing.
But I don't think that's such a bad thing.
MSIL is the lowest-common denominator. At some point, no matter how much the languages diverge, you eventually have to generate MSIL. This necessarily implies that you can't do anything in a .NET language that you can't do with MSIL alone. And for the same reason, any higher-level constructs provided by way of an object model can be consumed by any language (by design).
I still don't see a compelling argument to pigeon-hole functionality such as generics and edit & continue to a single language. As Krishna points out in the comments, doesn't that kind of defeat the purpose in having a CLR and common BCL in the first place?
It seems to me like a form of affirmative action for .NET languages. Oh, VB.NET doesn't have a great reputation and not as many people use it? Not to worry, we'll add some exclusive features to give people a reason to use it.
Differentiation for the sake of differentiation is a bad thing, for both VB.NET and C# developers.
Another point to consider is that many enterprises, in an effort to commodify their developers, will standardize on a language for maintainability and support reasons. Should developers be penalized and deprived of functionality because of these decisions over which they probably have no control?