The author hints at good point - is the browser going to be around in 10 years? The trend definitely seems to be cloud + rich client. In that case, is the browser needed?
If the browser platform is not around in 10 years, I hope that means it has gotten superseded by a more powerful open platform that is not owned by anyone.
But I do not see any potential for a revolutionary change in the way we consume the web. What I expect (or maybe hope for) is just browsers that keep getting more powerful, I expect evolutionary progress in the browser space. And I expect the browser model to remain the dominant web platform, because it is open and it is good enough--there are no other paradigms that blow it away; Adobe Air, Java, and iOS are just more of the same (whether it runs in a browser, on Air, JRE or iOS, it is still an app).
Therefore since all of these platforms are based on the same principles and they all do the same thing, what we need is to improve the web platform to be the best one.
If we find out that there are certain bottlenecks that will make it impossible for the web to be as fast and shiny as the others, we'll need to re-engineer the parts that are causing the bottleneck. What is it that makes iOS apps faster and shinier than web apps? Couldn't we add new technologies or improve the current ones to allow web apps to be just as good if not better?
The browser is 'needed', is valuable, because it is an independent, unowned 'medium' for other products and activity. It allows something more like real competition. That independence is part of the open architecture of the internet that is probably what makes it all so great.
. . . and what would an independent rich-client look like? Would that not be a browser pretty much? . . .
When you isolate WebKit and the other rendering engines as OS-level services, it's even fuzzier.