But I work making games. And so this promised land is still way over the horizon. We're only just getting to the point where a HTML5 game can be notable as anything other than a tech demo on the desktop. It'll be another few years before we reach that point for mobile.
But a lot of mobile apps are not like games. They are adaptations of web sites for a particular mobile device. This is where what he is saying surely makes a lot of sense. Really, apps like twitter should be able to be done with HTML5.
But more involved applications like games make sense to be native applications.
It seems that standardized frameworks aren't the problem, performance is. So what's your take on hardware acceleration, canvas, and WebGL?
There are always going to be classes of applications that need every bit of performance they can squeeze out of their host platform, and will be written native to that platform.
Web apps are very useful things, but this "native code is dead" line has been touted before and it was wrong then too.