Google posted a blog a couple days ago taking users for a peek 'under the hood' of the mechanics of its search engine: talking about the origins, improvements and the future of search, all from what started as a research project in 1996.
"Our goal is to get you to the answer you’re looking for faster and faster, creating a nearly seamless connection between your questions and the information you seek." The post reads, describing how the experience needs to be ubiquotous regardless of what updates they make to it (users don't need to know what's changed under the hood, so long as it follows the same path of connecting them more reliably to what they're searching for).
Interviews with Google staff in the video show how the search engine was started around the 'Pagerank calculation' algorithm to form a better way of cataloguing this explosion of content on the web in the mid-late 90s. From this developed all the key points in their timeline, from Adwords and image search. They even delve into the time when they didn't have any of the stories from the September 11th World Trade Centre attacks catalogued in their search on the day, as pages were crawled a month earlier. This prompted their makeshift solution of putting live story feeds on the homepage.
To better present the points being made, they pointed out Universal results and how all the different items they had in place separately just made sense to combine, and the Quick answers functionality to instead answer a question to the best of it's ability above the links that would also help.
The third point being made, and probably the most interesting, was about the future of search: talking about pushes in the technological field ahead with the likes of Google Instant and voice search. The future seems bright; but as Google put much better than we ever could: "If the past is any indication, we don’t know what search will look like in 2020, but we wouldn’t be surprised if it looks nothing like it does today." See the video below.
Source: Official Google Blog