He's talking about giving people the tools for the social graph. Some tools help you stay connected; others help you make new connections. Most people think of FB as the former--helping you stay connected with people you already know
Today, he says, going to discuss a service that helps users make new connections.
Like any database, you should be able to query it. There are a few major use cases ....the first -- they want to know in the world, w/ the people around them. That's news feed. Another common query : who is this person? Tell me about them. That's timeline.
Those are first two pillars. Today, we're talking about the third.
What's more interesting is giving people the power to take any cut or query of the graph they want -- we're callnig it "graph search"
What is graph search? It's not web search.
There rare more than 1 billion people, more than 240 people photos, 1 trillion connections.
He's describing the "graph search" -- and indexing all this content to it can be retrieved.
The search we wanted to build is privacy aware.
It's really powerful. On FB, most the things people share w/ you aren't public. Which makes search challenging.
- Every piece of content has its own audience; most content is not public; you can only search for content that has been shared with you.
He's talking about the difference between web search and graph search.
I.e., a search for "hip hop" -- web search will take any open ended query and return links. Graph search is intended to give you the answer -- not links that might give you the answer.
One big design problem, how can we answer questions that are intuitive.
Answer we came to is: filters. (And he gets laughs.)
We came up with something we thought was much more natural. After a few months, the team came up with ....
He's showing a video of search. I.e., photos of my friends taken in 2009. Or photos of me and my friends. Or: restaurants in Chicago; Or: Music my friends like.... And so on.
It's early, but graph search is a way of answering questions.
Focused on people, photos, places and interests. Those are the four different use cases.
One example: He asked for "My friends who live in Palo Alto, Calif., and like "Game of Thrones."
And the results list the top people in Zuckerberg's network.
Another example: Photos of me and Priscilla Chan. The results pulled up the best photos.
Another search Zuckeberg did: "Mexican restaurants in Palo Alto, Calif, my friends have been to." And the results pull up friends and their opinions of restaurants.
He's introducing Tom Stocky and Lars Ramussen
They're here to show more examples.
They're showing how Graph Search works. This is not keyword search, he points out. Graph search is structured. I.e., "friends who like star ways and harry potter."
The results come back -- 16 of his friends who like Star Wars and Harry Potter. It shows a list of his friends that like those movies, plus he gets to see other movies they like.
Tom: These results are entirely unique to me.