I've moderated and participated on a bunch of panels where the topic was something like "HTML5 vs. Native apps, which will win?"  And I've always said that native apps aren't going away, and my co-founder Sean has often pointed out that HTML5 won't be replacing complied apps so long as mobile hardware is changing drastically every 6 months (HTML is a trailing standard that can't keep up with innovation on the hardware side).  

But here's the real smoking gun:

The Facebook engineering team, in a blog post, writes, "we realized that when it comes to platforms like iOS, people expect a fast, reliable experience and our iOS app [that was heaving leveraging HTML5] was falling short."

This is really significant.  Facebook has a lot of incredibly good reasons not to rebuild its app natively.  Facebook doesn't want to be beholden to Apple for its distribution channel and access to its users.  Facebook doesn't want to have to create and maintain completely different and incompatible codebases for various distribution channels.  Facebook arguably has the very best web engineers on the planet.  And yet they've moved away from a heavily dependent HTML5 strategy in mobile.

So for anyone saying they're going to completely move to HTML5 and stop (or not start) native app development, I say that approach will only work if:

1) You're only going to be a "fast follower" and not an innovator on mobile, and/or

2) The mobile channel isn't that important to you, meaning that things like in-app purchase (and other ways to package, monetize and distribute your content) aren't that important to you

If you think you can succeed where Facebook couldn't, I'd love to hear more about it.