Gadgetory

i'm dan farber editor at cnet and i am joined by tom stocky the product manager for graph search at Facebook and Lars Rasmussen who runs the engineering group at the team welcome yeah thanks for having us so grab search launched yesterday and it's an early beta version of the product it's going to be in beta for months what are the next steps to really make it robust and roll it out to more people well so look the first thing we need to do is get a lot of feedback from you guys who know can use it and the people we're rolling out we're rolling out pretty slowly we have only a few thousand people in there now as the feedback comes in will be spending a lot of time responding to that and just making it better and better but there are a few major pieces that we know are missing you can search for people for photos for places for interests but we can't yet search over poster the stories the status updates people have made we don't have a mobile version yet and it's only available in English and so those are the three main things that we're going to be working on going forward and when when Mark Zuckerberg was talking about the product he was saying would take years and years to make this product whole what does it look like at the end of those years hmm that's a really hard thing to predict I mean it'sit's we have some idea of where we want to go and Mike Mars mention there's a few things we know we want to work on but to be honest like the thing that's going to be most interesting to us is to see how people use it to see what they complain about it to see what it they want it to do that it must have a road map that says you know beyond doing posts and beyond doing languages and and also as well as the open graph so that you can bring all the data from those applications into play it seems to me that that one of the features that missing is yes I can ask my friends about what restaurants they like in New York City but can they tell me more about that restaurant such as show me the menu or can I make a reservation through facebook services yeah I think I mean that's definitely part of it I think at a high level where we want this to end up is everything you can see on facebook we want you to be able to search for that so to make everything searchable and like you were saying I mean there's a lot of stuff that isn't there about restaurants and places and other pages and maybe now those those guys have a reason to fill that stuff out Lars on a technical side what are the big challenge you're facing they say that there's 240 billion photos at trillion connections a billion users how do you take all that massive data and turn it into something that a natural language query can deliver a straight answer on it's a very good question so there's there's two main pieces to the technology there is a pset understands the the natural language that people type in we actually we think of what people type in as the title of the content they're looking for it's as if you were building your own page on facebook you type in the title we try to understand the title and we fill in the content and that's one challenging pieces as we cover more content areas we have to grow this piece to understand more kinds of these titles here and so that's as one thing the other thing is that then once we want to stood what the title is and we have to then answer the question and we have to do it fast and we have to do it in a way that respects all of the privacy settings and all of that data and that's that's just a big systems engineering challenge because of the sheer volume of data you mentioned a bunch of numbers we have a billion active users about a million come new ones every day 240 billion photos about 300 million new every day one of the trillion connections that I mean the numbers are just huge and also we want it to be so that if you go and upload a photo or you can tag someone in a photo you should be able to search for that right away and so we get I think it's more than 5 billion new pieces of content whether it's a like or photo or comment or post every day and we want all of those to be indexed live let me ask you one final question people always saying well this is a you know going to be going up against Google what's your take on it well look the way we think about it is that we're focus on here's the data that people have shared with each other on Facebook can we make that more useful by making it searchable and I think you will find and this is actually what I think makes the product the most interesting that most of the content you'll find through graph search is not public content its content that people have shared on Facebook with a limited audience that you happen to be in and so as a result I think if you compare any web search engine to Facebook's graph search there's very little overlap in the data that we're actually searching for they're really very different product okay oh thanks very much both of ya thang you like for Cena I'm Dan Farber thanks for watching