Google did one thing, and they did it well. Sell advertising. What, you thought I was going to say search? Search is one vehicle through which they can market their advertising. That vehicle, however, is quietly becoming obsolete. Google search, in its time, was one of the best ways of finding static information on the web. Wikipedia has been quietly overtaken Google in that area.
This is interesting considering that Yahoo and other web indexes previously had tried human powered collecting of information. We have to wonder, why did they fail when Wikipedia succeeded? I think that it was their dependence on a centralized approach whereas Wikipedia utilizes a more distributed approach. Yet indexes were distributed with regard to the content, but they were centralized with regard to the controls. The next thing that destroys Wikipedia will be both decentralized information and decentralized control---but yet, still indexed, still useful.
I've heard that there are three webs: 1. The static web, 2. The realtime web, and 3. The hyper-local web.
Google has controlled the static web, but this seems to be slowly shifting to Wikipedia. Twitter, Facebook, Digg, Reddit, and others have been the leaders in the realtime web. As for the hyper-local, I think I would rather call it the social-web, and we know Facebook leads there.
What is going to be the next big breakthrough? We are going to again distribute. I think that if companies such as Reddit, Digg, and yes, even Google, want to survive, they need to once again give up control. They need to be able to easily integrate with other platforms. The content is going to be distributed. The comments, feedback, etc will be distributed. The next big thing is going to seamlessly broadcast to other web media as well as seamlessly receive feedback from a variety of sources (and of course, in realtime ;).