The roles will be flipped. The browser will run inside the App, and not the other way round.

Jotted down some thoughts inspired by this post here:
http://openmobster.blogspot.com/2010/04/killer-mobile-app-will-not-be-browser.html

First with respect to barriers. It’s common sense that if you want people to use or purchase your product or service, you present a minimum number of obstacles to the user. What happened to desktop applications when they started to compete with web applications that required no installation, no potential bugs to disrupt the OS and are easy to trial. Desktop applications died.

The same principle applies from the developer side as well, the easier it is to develop and deploy applications the more applications will be created. Device vendors, especially in Apples case, require developers to learn their proprietary systems and develop within their strict guidelines, require approval and sometimes surrender their work to the vendor. Compare that to web development where developers can leverage their existing skills and are not subject to the whims of draconian vendors.

Currently you can create a better experience by developing natively for the device, but that will inevitably change. Javascript performance improvements, local storage, SVG and hooks to device information such as orientation will dramatically change the usability and utility of web applications for devices. When it does, you can expect mobile application to die off in the same way as desktop applications.

There are also the other economic considerations. When you develop a web application, you can be developing for a range of form factors, desktop browsers, mobile browsers and so on. This way you are maximizing your market and minimizing your costs. There is simply no way you can ignore the utility of web applications outside niche areas. As it is a practical necessity to create a web application why duplicate the effort with a mobile device?