It’s getting easier to find the information we need in the vast ocean of Web data. Google recently launched its new Knowledge Graph search tool, which aims to connect users with a “real world map of things,” including information that might not turn up in a traditional text search. Meanwhile, programs such as iPhone’s Siri, Wolfram Alpha and Quora seek to respond to specific questions and information requests. We explore the fast-changing world of search technology and how we’ll find information in the future.

Guests

  • Barry Schwartz CEO of RustyBrick, Inc.; Executive Editor of Search Engine Roundtable
  • Amit Singhal Senior VP of Google Search
  • Alex Howard Government 2.0 Correspondent, 
O’Reilly Media
  • Peter Corbett CEO, iStrategyLabs

Introducing Google’s Knowledge Graph

Transcript

  • 12:06:45

    MR. KOJO NNAMDIFrom WAMU 88.5 at American University in Washington, welcome to "The Kojo Nnamdi Show," connecting your neighborhood with the world. It's Tech Tuesday. Online search and query tools have come a long way since the early '90s. They've gotten faster and smarter. Most are able to recognize what you're looking for and give you related topics. Many say that tools like iPhone's Siri that can answer questions spoken in a natural voice are the future of search.

  • 12:07:30

    MR. KOJO NNAMDIAnd programs like Quora and Wolfram Alpha can answer questions or do complex computations. But, even so, there are more than 50,000 searches done every second, and about a third of those are abandoned in frustration. The rest generally have to tweak their query to get what they're looking for. Microsoft, Yahoo, Google and a whole raft of companies are working toward the ultimate goal: to connect us with any information available on the Web when and how we want to discuss it.

  • 12:08:00

    MR. KOJO NNAMDIJoining us to have this conversation first is Amit Singhal. He is a senior vice president at Google in charge of Google Search. He joins us by phone from Mountain View, Calif. Amit Singhal, thank you for joining us.

  • 12:08:14

    MR. AMIT SINGHALGood to be here, Kojo.

  • 12:08:16

    NNAMDIGoogle is the dominant search engine in the U.S., but the numbers are higher in Europe where four out of five people use Google as their search tool. European regulators said yesterday -- he jumps straight to the news -- that Google is promoting its own businesses at the expense of competitors. They're threatening an antitrust lawsuit if Google doesn't address the issue. What is your thinking about this?

  • 12:08:39

    SINGHALSo it's still early in the process, Kojo. We are looking at the commission's report, and we're actively working with the commission to address their concerns. So it's just early in the process.

  • 12:08:48

    NNAMDISo we don't know how this is going to play out as yet?

  • 12:08:52

    SINGHALIndeed. It's too early in the process.

  • 12:08:54

    NNAMDIWell, Amit, Google has been refining its search engine all along. What is the latest change to Google Search?

  • 12:09:01

    SINGHALMost recently, we just introduced our Knowledge Graph, and that would be the foundation for tomorrow's search. Today, computers don't understand language and things like you and I do. To a computer, Taj Mahal is just two words next to each other, and that's it, whereas, to you and I, it's either a beautiful monument or even a casino in Atlantic City. With Knowledge Graph, we are laying the foundation for computers to understand things in the real world, like you and I do.

  • 12:09:35

    NNAMDIGive me an example. Say I'm searching for Pluto. What comes up with the current Google Search tool, and what would the new Knowledge Graph search tool add?

  • 12:09:48

    SINGHALYou know, with Pluto, what comes up on the regular search is also indeed a list of very relevant articles which talk about Pluto which used to be a planet and so on. But with the new Knowledge Graph, you get to see some of the most common things that the world wants to know about Pluto. And this is based upon what we have observed our users looking for when they look for Pluto, and we bring up a beautiful panel on the right which tells the important set of facts that people want to know about Pluto.

  • 12:10:30

    NNAMDIWell, a lot of links and relevant information like maps come up right now when we search for something. What would be different?

  • 12:10:37

    SINGHALWell, you would still get all the relevant information that you're used to getting from Google alongside for things in the real world or the universe that we understand as human beings as either planets or celestial bodies or, you know, beautiful buildings. We would bring alongside our search results to the right of the result page, a nice panel that has what that thing really is in the world, what are some of the most looked for facts about those -- that thing and what are some of the other things people, places and so on that you should know about if you're interested in that thing.

  • 12:11:23

    NNAMDIThis is still in its early stages, but what are some of the possibilities for the future of this kind of search feature?

  • 12:11:31

    SINGHALWell, Kojo, when computers start understanding things in the real world like you and I do, you can just imagine the next generation of search will emerge which would understand and reason with things and facts like you and I can. Today, in our Knowledge Graph, the first implementations thereof, we have over 500 million things and over 3.5 billion facts and relationships about those things that we know.

  • 12:12:04

    NNAMDIWell...

  • 12:12:04

    SINGHALOnce you know all this, you can start reasoning with this data. For example, you can start answering queries like: What are the 10 deepest lakes in Africa? Because you know what a lake is, you know what depths of lakes are and where they are in the world. So that kind of reasoning search engines cannot do today. And we are just laying the foundation and taking baby steps towards that reasoning that you and I expect our search engines to do.

  • 12:12:33

    NNAMDICouple of other things: Bing just rolled out its social search, and there's also a social search component on Google Plus. How does social search work?

  • 12:12:43

    SINGHALSocial search is a way to enhance search where people that you care about and you trust, content that they have either written or endorsed is brought to you, even relevant, right in the search results page. And Google was a pioneer in social search when we launched earlier versions of our social search three years back. And with more recent advances that we have made, we are able to find very relevant information from people you know or trust right in your search result page from the same wonderful interface.

  • 12:13:20

    NNAMDIGoogle has had voice commands for a while, but it's my understanding you're also working on a voice assistant that some are calling an answer to the iPhone Siri. What will this voice assistant be able to do?

  • 12:13:32

    SINGHALWell, I can't talk about what products we are building right now. But Google was a pioneer in voice search. And voice search is just the natural way to search when you are on a phone and it's hard to type. Voice is a very natural interface to human beings. We pioneered some of the key methods that are used in voice search, and we'll continue to invest in this technology, of course.

  • 12:13:59

    NNAMDIAmit, before I let you go, we're hearing a lot about Google Goggles and augmented reality, a kind of search that recognizes words, objects and images. What can it do?

  • 12:14:10

    SINGHALWell, this is our new direction that we are exploring as to what happens when the boundary between (word?) computing and human interfaces -- computer-human interfaces is (word?). So, today, you -- to know something, you have to take a cellphone out of your pocket and then type a bunch of words or say a bunch of words and then get back information.

  • 12:14:35

    SINGHALNow, imagine tomorrow, if that interaction was made completely seamless. Now, this is a very experimental project that we are actively working on right now, and I'm very excited about the future possibilities that such (word?) computing can bring to the world.

  • 12:14:54

    NNAMDIWe also have in studio with us Alex Howard. He is the government 2.0 correspondent with O'Reilly Media. Alex, good to see again.

  • 12:15:00

    MR. ALEX HOWARDIt's great to be here, Kojo. Thank you.

  • 12:15:02

    NNAMDIAnd joining us from the Argo Studios in New York is Peter Corbett. He is the founder and CEO of iStrategyLabs. Peter Corbett, thank you for joining us.

  • 12:15:12

    MR. PETER CORBETTMy pleasure, Kojo.

  • 12:15:13

    NNAMDII don't know if either of you have any questions for Amit, but I do have a caller on the line, Paul in Arlington, Va., who does. So I'll start with Paul. You're on the air. Go ahead, please.

  • 12:15:24

    PAULHi. I don't know if there's anybody on the air right now who would be the person to answer this question, but I'll just ask it anyway for what it's worth. A lot of people, they remember that 411 calls to the phone company used to be free. And then people overused it, and so the phone company started charging. And one thing I've noticed about Google Search is that a lot of people do them unnecessarily.

  • 12:15:44

    PAULThey know perfectly well that they want to go to the Wikipedia article about something. But, instead of going through Wikipedia directly, they search the entire Internet and then pick the Wikipedia link off of the, you know, first couple of links that come up. And I'm wondering, is -- I know that the cost of a Google search is infinitesimally small, but, still, is there any possibility that the overuse of searching could lead to problems in the future?

  • 12:16:08

    NNAMDIAmit?

  • 12:16:09

    SINGHALWell, from our perspective, Google Search will remain free. And searching is a process by which people gather a lot more information. Wikipedia is a wonderful resource. And people get a lot of information from Wikipedia and other sources on the Internet. But from other perspective, searching is fundamental activity on the Web that people use to gather information. And information is empowering in the end. So the more people can get information, the better prepared they are for whatever they need to do.

  • 12:16:45

    NNAMDIPaul, thank you very much for your call. Also joining us by phone from Suffern, N.Y., is Barry Schwartz, CEO of RustyBrick. That's a custom Web software company. He's also the executive editor of Search Engine Roundtable, an online forum. Barry Schwartz, thank you for joining us.

  • 12:17:01

    MR. BARRY SCHWARTZThank you for having me.

  • 12:17:02

    NNAMDIHere's Alex Howard, Amit.

  • 12:17:04

    HOWARDYeah, just a quick question for you. When I'm interested in what's happening right now, I'll often turn to Twitter to look for real-time search. There was a while that Google had those results in its search results pages, and then, for a number of reasons, they went away. Will users be able to see any kind of real-time stream from Twitter or other similar services anytime soon?

  • 12:17:24

    SINGHALSo, Alex, we are always looking for more partnerships to bring the best products to our users. We had great real-time products. Unfortunately, the negotiations between Twitter and us broke down. And we couldn't renew the data deal with Twitter. But we're always looking for sources to bring the most real-time, most fresh information to our users, and we're actively looking at partnerships in that space.

  • 12:17:50

    NNAMDIPeter Corbett, any questions for Amit?

  • 12:17:52

    CORBETTYeah. I'm wondering if the Knowledge Graph is a specific response to Facebook's Open Graph rollout, which, for those who are not aware, is a specific protocol for indexing objects in the world, so really understanding what the world is talking about on the Web similar to what Google is doing?

  • 12:18:07

    NNAMDIAmit?

  • 12:18:08

    SINGHALSo from our perspective, we are not motivated by what others do. I have worked in the field of search for over 22 years. And, fundamentally, as an academic researcher in the field of search, we're always troubled with understanding real-world things -- people, places, monuments and so on -- because it has been really hard to represent real-world things in computer memory.

  • 12:18:33

    SINGHALComputers are very good with words, not so good with things. For the first time with Knowledge Graph, we have built some breakthrough technology where we can represent things in a computer memory and do some logic with it. So, from our perspective, this is the foundation of the future of search, and we're very excited about what we have built.

  • 12:18:57

    NNAMDIBarry Schwartz, any questions for Amit Singhal?

  • 12:19:00

    SCHWARTZYeah. Hi. How are you doing? My question is -- it's interesting to see that the way search is going. One is you have the (unintelligible) you know, the (word?) links, the Web page crawl and understanding, you know, unstructured data. And with social search, we're more opening up to the public and asking our friends for, you know, relevant -- we have queries. We're asking our friends for information.

  • 12:19:21

    SCHWARTZAnd now, with this, you know, Knowledge Graph, it's more about giving you factual information. How does that all, in your mind, blend in the search results, you know, in two years from now? How do you see that happening where you ask a query and all that information is in one interface and the most relevant results for that user?

  • 12:19:37

    NNAMDIAmit?

  • 12:19:38

    SINGHALSo, Barry, you know well. You are an expert in the area as well. We, here at Google, have had a single focus on relevance for our users. Our primary responsibility is to answer what the user's asking us, be it a map of a location, be it simple facts about objects, things, or be it the most relevant Web page. And relevance is Google's close friend. We keep experimenting with our relevance formulas to bring even more relevant search to our users.

  • 12:20:15

    SINGHALJust last year, we did over 50,000 experiments with our search engine, and we launched over 500 improvements to our research. So it is in our DNA to make search more and more relevant. And with all these new things coming online, these new forms of information available to us, be it social connections or important facts about things from our Knowledge Graph, we are just going to see a much more relevant, much more useful search engine in Google going forward.

  • 12:20:48

    NNAMDIAmit Singhal is a senior vice president at Google in charge of Google Search. Amit Singhal, thank you for joining us.

  • 12:20:55

    SINGHALThank you. It was my pleasure, Kojo.

  • 12:20:57

    NNAMDIWe're going to take a short break. When we come back, we will continue our conversation on the future of searching the Web on this Tech Tuesday with out other panelists. But you can call us, 800-433-8850. What kind of information or data would you like to be able to search on the Web? 800-433-8850. You can send us a tweet at #TechTuesday, email to kojo@wamu.org, or simply go to our website, kojoshow.org. Join the conversation there. I'm Kojo Nnamdi.

  • 12:23:17

    NNAMDIIt's Tech Tuesday. We're discussing the future of search on the Web, and taking your calls at 800-433-8850. Do you ever have trouble finding what you're looking for when you search online? 800-433-8850. We're talking with Barry Schwartz, CEO of RustyBrick. That's a custom Web software company. He's also the executive editor of Search Engine Roundtable, an online forum. He joins us by phone from Suffern, N.Y.

  • 12:23:42

    NNAMDIAlex Howard is the Government 2.0 correspondent with O'Reilly Media. He joins us in studio. And joining us from studios in New York is Peter Corbett, the founder and CEO of iStrategyLabs. Peter, what do you think of the direction that Google Search is heading in?

  • 12:23:59

    CORBETTI think it's terrific. I saw some early feedback on the Web last night on a site called Reddit, and someone said, oh, this new actor feature is amazing. And Reddit, if you know, there's a lot of commentary because it's not an actor feature. It's a Knowledge Graph. It's actually showing this object, this person we know all of this stuff about, and now the data is linked across all the information Google has about that specific person. I find it to be very useful, and I think that, over the course of the next few years, people will become very accustomed to it.

  • 12:24:26

    NNAMDIExcept that the term Knowledge Graph is so, well, blah. Anyway, Barry, there's a lot going on right now. Are we on the verge of a revolution in search? What new tools are you excited about?

  • 12:24:39

    SCHWARTZYeah. I mean, the Knowledge Graph is nice and everything. It doesn't mean Google was the first to go this route or think this route. I mean, they've been doing this for a long time with quick answers. Ask.com has been doing this early on. But they're taking it to the next level, and a lot of search engines are thinking about taking this to the next level. The real exciting part in search is really not just typing into a search box.

  • 12:25:00

    SCHWARTZLike Amit was saying from Google, he -- it's really, you know, just talking your query, thinking your query. We have these Google Glasses that are coming out soon. We have self-driving cars. We have, obviously, Google Voice and Apple's Siri. This type of stuff, we're out and about. You're on the street. You're doing your day-to-day stuff, and you want to pretty much ask a question while you're doing your day-to-day business.

  • 12:25:20

    SCHWARTZAnd you're not behind a computer. And you don't want to pull out your phone. It's just there in front of you. It's integrated into your day-to-day lifestyle. And that is what I find most exciting. (unintelligible) obviously scary as well, but it's definitely an exciting time.

  • 12:25:33

    NNAMDIAlex, there's an enormous amount of government information out there that people want to find. How helpful are search tools now in getting to that information?

  • 12:25:42

    HOWARDDepends upon the data. They are certainly increasingly helpful for some kinds of information if it has been released in a way that it can be baked into the search into the search tools that people use in their daily lives. One of the best examples is transit data. If I go to Google maps and I look for directions somewhere, I can click upon a tab that gives me search results for transit, for the local train, for the local bus. And that works when government has worked to make that data open, accessible in a machine-readable format that can be used in the map.

  • 12:26:14

    HOWARDThere certainly all kinds of other data that's out there, things like SEC filings. You can go to a website that the SEC has set up and browse that kind of information. But, overall, this data is only going to be as useful for most citizens as when it's baked into this -- the search engines they already use, whether it's a -- mobile device is increasingly -- something like Siri may have a hard time returning good results, and...

  • 12:26:43

    NNAMDIBut I'm wondering whether new search tools like Google's that include encyclopedia-like information and data that -- can they help highlight government information that might, right now, get buried?

  • 12:26:53

    HOWARDAbsolutely. There's -- if the data can be crawled, then it can be indexed. And then Google or other engines that are evolving will be able to add more context and clarity to it. So the question is simply whether the data itself has been exposed or not -- and exposed, not in the sense of a data breach, right, and a leak, but actually voluntarily put up onto a platform online where there's -- whether it's a bulk data download or an API, something that lets the search engine see what's there, and then convey it to the user.

  • 12:27:26

    HOWARDAnd in, I think, the future, we'll see a lot more of data become more social as there are vast repositories that are full of people, whether they're journalists or advocates or big people in business, who are helping to clean it up and then categorize it. You know, add metadata to it so that when citizens look to do something, whether it's travel somewhere, file something, get more information about their local legislator, it's been made more sensible. And you can see already these infomediaries acting in many cases.

  • 12:27:59

    HOWARDSomething like govtrack.us for thomas.gov, which we know about, something like OpenCongress and an increased number of civic startups that are trying to make this data more useful than it might perhaps be buried in a government website where you have to know exactly what page to go to and exactly what query to put into.

  • 12:28:18

    NNAMDIOn to the telephones. Here is Elizabeth in Herndon, Va. Elizabeth, your turn.

  • 12:28:24

    ELIZABETHHi, Kojo. Thanks for taking my call. I'm calling because my recent -- or seems recent struggle with search engines is that, especially on things that are obscure, I feel like I get just one after another of -- I'm not really sure what they are, some kind of data or query aggregators or something. And I'll just feel like, you know, 10, 12 -- it's obviously just some company that's trying to get my attention. And so it's really hard to tell whether there's just not anything after about the subject or whether it's, you know, going to be, you know, entry number 50 or 60 of what I'm looking at.

  • 12:29:02

    NNAMDI(word?)...

  • 12:29:02

    ELIZABETHSo, you know, is this new search technology going to, like, get these people out of the way (unintelligible) ?

  • 12:29:07

    NNAMDIWhat do you say to that, Peter Corbett?

  • 12:29:11

    CORBETTI think you're right. Most search results that I have experienced, and especially in Google, can be pretty awful. That might be surprising to say. Google's been obviously a very successful company. But the people that are engaged in daily minute-by-minute SEO activities to get there, their result up to top of the list just to get you there to then click on their ads, that's not what you're looking for.

  • 12:29:31

    CORBETTThat's why I think the Knowledge Graph is very promising. That's a feature that Google will take you directly to knowledge about the specific subject. Now, it's within their interest to drive advertising around that content, those results. The question is: Will you still have an iffy experience? And that's just going to be in the details as we see it roll out.

  • 12:29:51

    NNAMDIElizabeth, thank you very much for your call. I'm glad you mentioned advertising because, Barry, if people are getting to what they're specifically looking for faster, they'll spend less time searching. How will that affect online advertising?

  • 12:30:07

    SCHWARTZRight. So Google's goal obviously with the search results is to get the most relevant results at the top. And so we hear from the caller that's not always happening. Ads come in obviously in that place, and at the same time, Google's ad team, which is totally separate from the organic search team, wants the most relevant ad as well. So when they rank ads, it's not just who's paying the most to show an ad.

  • 12:30:26

    SCHWARTZThe ads are based off of various data, such as quality score, which is based off (unintelligible) and some other math metrics. So Google -- so while you might have one ad over, you know, the -- to first position, the second ad might be paying more than the first ad just because that the first ad is more relevant. It's ranking higher. Now, Google's thought is that if you provide the most -- best user experience on the ad side as well as on the free listing side, people will come back more and more and use their search engine.

  • 12:30:54

    SCHWARTZSo they're looking for repeat searches as opposed to -- you know, as opposed to just somebody getting their answer right away. They want people to get the answer right away, so they come back more often with more questions. And what Google's doing is they're expanding what people are thinking about in terms of when they should come to Google.

  • 12:31:11

    SCHWARTZAnd that might be over using Voice Search or using Google Goggles or whatever types of things they might be using, such as AdSense, in order to expand their query base in order to increase the amount of revenue they could generate from a single user. That might be less search, you know, unique -- you know, repeat searches, but, at the same time, it's more unique searches, if that makes sense.

  • 12:31:30

    NNAMDIHere's Eric in Silver Spring, Md. Eric, you're on the air. Go ahead, please.

  • 12:31:34

    ERICThank you so much for taking the call. I have two questions that were prompted by the comments from the vice president at Google, these things I've been wondering about. The first is, some of the directions at which they're going, it sounds like it's been pioneered by other somewhat experimental search engines like Wolfram Alpha.

  • 12:31:48

    ERICI can go to Wolfram Alpha, take the example that he used and have quoted the 10 deepest lakes in Africa, and up it comes right away. That's really not an innovation. I don't understand why Google continues to treat these things as if they're pioneering them. And the second question I have is: Why is it that Wikipedia entries show up so prominently in Google search results?

  • 12:32:07

    NNAMDIFirst you, Peter Corbett.

  • 12:32:09

    CORBETTSure. I think the reason why Wikipedia entries show up so high in Google search results is probably twofold. One, they're linked to by so many millions of people because people have said over time that Wikipedia articles are incredibly relevant and that this is a data source that you can trust. And then, secondly, purely from a technology point of view, the way that articles are structured in Wikipedia are much more friendly to the way that Google actually indexes content.

  • 12:32:36

    CORBETTSo they're essentially optimized for search very well, so those will show up at the top. Now, whether or not the Knowledge Graph is a true innovation or not, I guess the real question is, has any innovation ever really been purely from scratch, right? You know, the best artists steal. Things like that are commonplace in American and human innovation. We base our innovations off of things that have come in the past.

  • 12:32:59

    CORBETTNow, it may be purely derivative based on what we've seen from Wolfram Alpha now coming into the Knowledge Graph. But I would say Google has put a lot of time and energy into building a better search engine. And I would say it is very innovative. I think it's going to be very useful. Is this the end of the road? Absolutely not. We haven't really talked about social search, and I think that's probably one of the areas I'm most excited about.

  • 12:33:20

    NNAMDIEric, thank you very much for you call. And Eric and Peter both mentioned Wolfram Alpha. Alex, there are a number of search and query tools that are based on specific questions, Ask Jeeves, being an early one. There are some others looking at new ways to find specific information. You find Wolfram Alpha to be very interesting. Why?

  • 12:33:38

    HOWARDWell, Wolfram Alpha is tapping in to something you've talked about here before. It's the world of big data, the idea that there is vast amounts of unstructured information out there. It's not just machine readable in one sense, but it's just expanding volumes of it. And that the thing that increasingly people in science or research want to do is find the needles in those haystacks and to look at data from many different angles and to see how it's linked together.

  • 12:34:05

    HOWARDThe founder of the World Wide Web, Tim Berners-Lee, has talked a lot about the next generation of the Web -- some people call that the Semantic Web -- but where there'll be ways to link the data together. When we talk about the Knowledge Graph or the social graph, we're really talking about ways that we link data together so that it can be seen in the context of other objects. Instead of a static Web, it's a more interactive Web.

  • 12:34:27

    HOWARDAnd things like Wolfram Alpha can allow people to do more complex functions when they search, not just operations like doing math -- although that -- these things can be quite useful to the extent that some classrooms have actually banned access to Wolfram Alpha for exams, things like this, 'cause they can do quite a lot -- but that they can be used to get a great deal of insight using the computational power of these engines.

  • 12:34:51

    HOWARDAnd one other thing people may not realize is that Siri uses Wolfram Alpha, right, that this black box, that whenever you're talking to your phone to ask it to do something, is tapping into the computational power. And, in many ways, the devices that we carry around are going back to another generation of computer history where there are thin clients on top of vast networked computers, right? We dialed into the mainframe. Well, we're not dialing into anything now. We're maybe connecting through 3G or Wi-Fi.

  • 12:35:19

    HOWARDBut we're putting in queries that algorithms are looking at and trying to determine not just what we might be looking for but what we might be interested in and interpreting that to give us better results. Now, we're still in the early days of this moment, but people have already talked about how social signals give us more ideas, location gives us more information and past search gives us a lot of information, too.

  • 12:35:47

    HOWARDIf we're going to talk about Google or Bing, we'd have to know that they're keeping a Web history, and they have some idea of the things you've searched for before, who you're connected to through your social graph and then to give you results that are personalized. Now, that's going to be quite scary to people. It often is when you show them how much information Google has.

  • 12:36:05

    HOWARDBut the tradeoff is that you can have a significantly increased utility. And there's going to be a lot of push and pull over the next decade, I think, as we look to those tradeoffs and people think about exactly what they want to give up in return for being able to get to information where and when they need it.

  • 12:36:25

    NNAMDII know all the panelists want to weigh in on this, but I'd like to invite our callers -- our listeners to join the conversation, 800-433-8850. Have you used Ask Jeeves, Wolfram Alpha or Quora, which we'll be discussing, or other tools to answer a question? Call us at 800-433-8850, or go to our website, kojoshow.org to join the conversation there. Barry Schwartz, what is new about how Wolfram Alpha works?

  • 12:36:52

    SCHWARTZWell, yeah, so, you know, basically Wolfram Alpha is a huge database of data. And what they're able to do is understand -- you know, you could type in inquiry into Wolfram Alpha, and what it will do is take all their data sources they have, all that data, and try to give you back information that is scientific in nature, that is data-driven and that is more -- it's factual. It's not just, you know, just Web pages that are useful or Wikipedia pages that are useful. It really provides, like, a quality answer to your scientific or data-driven query.

  • 12:37:25

    SCHWARTZSo if you type in something like, you know, how much time elapsed between this date and that date, it will give you -- it'll break it down to you in all these different types of metrics so you can understand it in a better way. If you want a graph, it has 3-D graphing calculators, right, built into the actual engine. And Google has these features as well. The cool part is that one of -- people don't know to go -- most people don't know to go to specialty search engines. That's what these are. You know, Wolfram Alpha, Wikipedia, these are all specialty search engines.

  • 12:37:55

    SCHWARTZAnd people don't know to go to these search engines. The one that's going to win -- Google wants to win in this place -- is that this whole Knowledge Graph is obviously using these specialty search engines, like the caller said before, and it's nothing -- you know, there's nothing great about that. The question is, can a search engine go ahead and pull from all these sources and understand which specialty search engine, either the Web or data sources or mapping or local restaurants, like Yelp or whatever, could that search engine figure out what the proper specialty search engine is for that query?

  • 12:38:28

    SCHWARTZAnd that's the real tricky question. And like we said before, it depends on the user, his friends, his previous search history, where he's based. All this information is taken into account -- personalized search is taken into account when a search engine is trying to give you the right answer.

  • 12:38:44

    NNAMDIPeter Corbett, Microsoft's Bing rolled out a social search tool last week. What's new in the new Bing?

  • 12:38:52

    CORBETTI'll be honest. I'm not sure exactly what's new in the new Bing from a social search point of view. But I did use Bing two days ago when I was here in New York City to do a little mapping. And I saw a store-by-store analysis of Soho, which I thought was so interesting. You never really know what's on a street corner in Washington, D.C. or in New York or what have you. And to see that level of detail, I thought it was incredible. So I might have to punt the social piece of this to someone else.

  • 12:39:17

    NNAMDICare to pick it up, Alex?

  • 12:39:19

    HOWARDSure, I could pick it up a little bit. The thing to remember in one case here is that Microsoft is an early investor in Facebook, not as early as people like Peter Thiel perhaps, but they're -- there's a partnership there. And so when people have long -- you know, have done searches in Bing for a long time, there's been some amount of social signals baked in there. But the future of this space, I think, is that we should expect those social signals to become more important.

  • 12:39:48

    HOWARDOne of the challenges right now is that if you do a search in Facebook, it might be useful for finding people, friends and other objects like that but not so much products or activities unless you want to, you know, like a brand, like a company, like a service. So the question will be how search engines in the future use our social connections and our personal details to deliver results that seem more relevant to us.

  • 12:40:12

    HOWARDAnd, you know, for now, Facebook is just kind of relentlessly focused on staying ahead in search. And if they're able to grow that comfortable lead and people stick around there, then Bing will be able to use a lot more of that -- those social signals, those we call emphatics, right, that little signals that show that someone likes this or they've consumed that or otherwise to give us more information. So if you -- or interested in searching for a local restaurant or a local plumbing service or searching for the best cab...

  • 12:40:43

    NNAMDIWell, I got an example that Peter might be able to help us out on. Peter, there's another aspect to these social search tools that Alex was talking about here. They connect you to particular thing, but they also connect you to your personal networks. You use the example of buying a Specialized bike. How does that work?

  • 12:41:02

    CORBETTSure. So if I were to search for Specialized bikes, Specialized being a brand, and I find that search result, I could see that maybe 20, 30 of my friends actually liked that product, right. And if I saw that, then maybe if Alex is my friend and is one of the people that have bought that, I could say, hey, Alex, what did you think of the bike? Do you like it? Should I buy it? And that's really the promise of social search.

  • 12:41:21

    CORBETTIt gives you a little bit of context overlapped with a product or service or otherwise that you might be interested in. And, therefore, you can connect with your network and say, hey, what do you guys think of this? None of that stuff is really "federated" in a great way today. As Alex was talking about Bing is integrating the social signals now from Facebook. Google has done that for some time both obviously with Google Plus and then with Twitter and to a lesser extent with Facebook.

  • 12:41:47

    CORBETTWhat I'm interested in seeing is what folks like knod.es, which is K-N-O-D, dot, E-S and ark, A-R-K, .com, will be doing with hardcore social search. 'Cause, right now, if you're in Facebook, you search for Specialized bike, you might, as Alex said, find that brand page. You're not going to find that much more.

  • 12:42:06

    NNAMDIGot to take a short break. When we come back, we will continue this conversation on Tech Tuesday, the future of searching the Web. If you've called, stay on the line or you can send us tweet at #TechTuesday or email to kojo@wamu.org. I'm Kojo Nnamdi.

  • 12:44:12

    NNAMDIIt's Tech Tuesday. We're looking at the future of searching the Web with Alex Howard. He's the Government 2.0 correspondent with O'Reilly Media. Barry Schwartz is CEO of RustyBrick, a custom Web software company. He's also the executive editor of Search Engine Roundtable, an online forum. And Peter Corbett is the founder and CEO of iStrategyLabs. We've been inviting your calls at 800-433-8850. And many of you have responded, so I will start with Francis in Warrenton, Va. Francis, you're on the air. Go ahead, please.

  • 12:44:43

    FRANCISYes. How are you doing, Kojo?

  • 12:44:45

    NNAMDIDoing well.

  • 12:44:46

    FRANCISHow are you doing, sir? My question is regard -- I just want to get your opinion. The internet technology has advanced much to AI technology well. The sites can basically search for things that are, you know, you're interested in with, you know, in your life. And anyhow, do you think that's created a tool for the government to basically have a psychological profile on basically every American, every Internet user? Do you think they could use that as a database for, per se, to, you know, I guess, look for certain type of activity and then monitor these people?

  • 12:45:22

    NNAMDIThe privacy concern comes up again. Francis, thank you for your call. Alex, isn't that a concern that, here it is, your and your friends' activities are coming up in searches, private companies have all these information, the government has the ability to coerce private companies? What say you?

  • 12:45:38

    HOWARDWell, there's two different ways to look about this. And, first of all, listeners should know that this exact question is being debated before Congress right now in terms of information sharing between private companies and government. There was a bill that passed the House called CISPA that will now be up for debate in terms of the Senate itself. If you're interested in these matters, you should definitely go and learn more about them and talk to your congressman about them. Have a voice in it. There are many different tools to learn.

  • 12:46:03

    HOWARDThere's a speech by Sen. Ron Wyden, taking his approach to it, et cetera. So, look, there's been what they call open-source intelligence for a long time. There are intelligence analysts who read the newspapers and watch cable news, and they get information. Now, what's different now is there's vast amounts of data about citizens online, from -- with respect to social media, these are voluntary updates, in many cases. With respect to search, it's a little bit different because then you get a sense of what people are looking for. Now, that can be incredibly useful.

  • 12:46:34

    HOWARDIf you look at flu searches, for instance, you can get ahead of epidemics. But it can also be quite private if you look at someone, say, looking for privileged information about their health. It is very difficult for law enforcement not to want to tap into that information. In many cases, it can be quite relevant to staving off a potential terrorist attack or criminal activity. At the same time, it also poses significant threats to civil liberties.

  • 12:47:00

    HOWARDThat's why there's been a lot of contention over whether we should have access to someone's book reading habits, for instance, in a library, whether we should be able to have access otherwise. Now, unfortunately, the laws have not been updated since 1986. The Electronic Communications Privacy Act has not been updated. So, for the moment, more of your information is accessible to law enforcement out in the cloud, out on the Internet, than it would be on your home server. People may not realize that. That's the truth of it.

  • 12:47:29

    HOWARDThis information from search, certainly information about your location from mobile, is being asked for, you know, by law enforcement agencies, of these different entities, some of them more forthcoming than others. Google gives raw data about government requests for information. Twitter has tried to give some information. Facebook, to this point, does not tend to give very much information about how much law enforcement is asking for. But we do know that there's great interest.

  • 12:47:56

    HOWARDAnd we do know that there's great potential for good in terms of government being able to protect us, being able to help with health issues, being able to help at better transit and also, you know, unfortunately, for significant overreaches. And these debates are very important to have in public, and I certainly encourage anyone listening to make sure that they are engaged with them.

  • 12:48:15

    NNAMDIOn to Bill in Herndon, Va. Bill, you're on the air. Go ahead, please.

  • 12:48:20

    BILLHi, Kojo. Good talking to you this afternoon.

  • 12:48:23

    NNAMDIThank you.

  • 12:48:24

    BILLI was wondering -- I'm a little late to the party tuning into your show, but I was wondering if you or any of your guests had mentioned DuckDuckGo as a Bing and Google search competitor and kind of a good example of how there's still a lot of room for innovation and growth in that space.

  • 12:48:39

    NNAMDINo, we haven't. Barry Schwartz, care to comment on that?

  • 12:48:43

    SCHWARTZYeah. DuckDuckGo is a great up-and-coming search engine. I don't know if it could really compete in terms of market share with a Google or a Microsoft Bing, but it's definitely one of those, you know, new -- not so new, but it's a fairly -- I think it's run by one or two people, very low-budget search engine. It does a great job in terms of relevancy. But there's no way a search engine like that could compete with budgets like Google or Microsoft have. So, yeah, there's a lot of these types of search engines. There's also blekko, which is a very popular up-and-coming search engine as well.

  • 12:49:16

    SCHWARTZUsers should definitely take a look at these search engines. DuckDuckGo is very into their privacy, so in terms of the -- so they don't store any history in terms of what people are searching for. I think blekko does. But all the search engines -- Google, Microsoft -- they give users a way to delete their history. Just so you keep that in mind, you can delete your history, and they all have their policies in terms of how quickly they purge it. But DuckDuckGo has been increasing in their market share, but it's still just a blip on the radar.

  • 12:49:44

    NNAMDIAnd we're going to put a link to DuckDuckGo on our website at kojoshow.org. Bill, thank you for mentioning it. Peter, you find the tool known as Quora particularly interesting. Tell us what it is first.

  • 12:49:56

    CORBETTSure. So Quora is a site where you can ask a question, and real humans, subject matter experts of all kinds, will begin to answer that question, vote up those answers and then you, as the question asker, curate the final answer. And I love Quora. I'll give you a specific example. I sat on Mayor Vincent Gray's transition team, basically building an innovation agenda for the city. And it's a typical kind of room, you know, 20, 25 people in the room. I said, you know what? We may be experts, but there are so many other people that should inform into this. Let me ask a question on Quora about it.

  • 12:50:30

    CORBETTSo I asked Quora, and I said, how should we incentivize technology innovation in the District of Columbia? We had 98 answers in about five hours, and it was seen by over 4,000 people. Some of the answers, former CTO of the city, other former government employees, technology entrepreneurs in the region, venture capitalists. And over the course of about four or five hours, I curated what was essentially the summary eight-bullet point answer, right. So that's why Quora is so interesting for searching for knowledge and gaining it.

  • 12:51:02

    NNAMDIWhat was your experience when you asked Quora who brought hackathons to Facebook in its early days?

  • 12:51:08

    CORBETTSo I didn't ask that question. Someone else did. And it turned out that the collective memory couldn't figure it out for, I guess, a few weeks, and then someone answered. One of the founders, one of the original founders -- I'm not sure if it was Dustin Moskovitz or otherwise -- said, you know what? We had gone to this one hackathon. It was just an open thing in the Bay Area.

  • 12:51:26

    CORBETTAnd someone approached Mark and said, hey, should we do this at Facebook? And he said, yeah, great. Let's do it. And so that's a way of tracking down institutional memory inside of what's now a very big company that just went public last week. Or otherwise, how would you find that information?

  • 12:51:42

    NNAMDIAnd, Barry Schwartz, being able to cross-reference different kinds of information and sources, put it together in a coherent way -- that's what we're talking about here -- is this the Holy Grail in search?

  • 12:51:55

    SCHWARTZYeah, definitely. It really is. When you're looking and you're doing a search and query and you have a question, being able to understand what's out there on the Web, being able to tip-tap into your friends on Facebook or on Twitter and being able to use open structure datasets, as well as opening -- asking these questions on Quora or other places like answers.com or LinkedIn Q-and-A, and getting basically one snapshot of all the advice across all your Web profiles from, like, all of -- like I discussed, that, I believe, is the Holy Grail in search.

  • 12:52:25

    SCHWARTZAnd being able to, for search engines, to actually provide that in a way that people could consume and understand quickly, that is what we're looking at in the future.

  • 12:52:33

    NNAMDIAlex?

  • 12:52:34

    HOWARDAnd there's really interesting new applications that are coming online now that are made possible by mobile phones and geolocation. You know, if I'm interested in going out on a particular night, you might just go to Google and search. But now you can actually use things like Localmind, which is a new application I've seen, where you can search where people have checked in around town using location-based social networks like Foursquare and then ask them a question about what's happening there. Right.

  • 12:52:59

    HOWARDSo in real time you can say, is there a cool scene? Are there a lot of people there? Are there pretty girls? How are the drinks? I mean -- and I've gotten questions back. What's the best, you know, sushi restaurant to go, you know, nearby here? And you can get local experts on this. Now, this is just the first step. There will be a lot of people who will be vying to do something like this.

  • 12:53:19

    HOWARDBut the idea is you can tap into the distributed network intelligence of people in a given area by location. They've opted in to share it -- though sometimes people share these things without realizing they do -- and then get back better answers from the people combining location, mobile and social.

  • 12:53:36

    NNAMDIHere's Katrina in Cheverly, Md. Katrina, your turn.

  • 12:53:41

    KATRINAYes. Hi. I'm calling because I'd like to know, a while ago, I was doing a search. I was trying to find the top 100 movies of all time. And when I did the Google search, I came up with a whole bunch of porn movies. And so I wanted to know -- I realized that -- or I found out that I should have been using the word film instead of movie. So I would like to know, how is that filtered out? And has that -- I know not to do, like, things like white house. I know not to do that, but just in that recent situation, you know, what can you do? Has that been filtered out?

  • 12:54:20

    NNAMDIAre you suggesting a porn movie couldn't be among the top 100 movies of all time? I don't know. Here's...

  • 12:54:25

    HOWARDProbably is.

  • 12:54:28

    NNAMDIHere is Peter Corbett.

  • 12:54:31

    CORBETTSure, turn it to me on that one. You know, the thing about the white house was that whitehouse.com, for number of years, was owned by a pornographer, and it was a porn site. That's why you wouldn't really want to search for that necessarily. So, as I said earlier, I don't think that Google actually provides great results. If I'm going to look for top 100 movies of all time, I'm probably going to go to Quora. I'm assuming that Quora has got a really good curated answer to that.

  • 12:54:55

    CORBETTAnd if someone in the audience is listening, check it out. And then I'd love to see the link myself. I do think, again, the Knowledge Graph is going to help with these things. If you were to search for top movies, I would assume that, you know, the porn industry is smaller than Hollywood in general, so it should have a different ranking. It sounds like, to me, that search engine optimization kind of struck again, and the pornographers won. So -- unless that really was what you were looking for.

  • 12:55:22

    SCHWARTZIf I can interject, I don't want to blame search engine optimization for those problems. Searching for top 100 movies in Google does provide a nice list of results. I don't think any of it is really pornography right now. I'm looking at it right now. What happens a lot is a lot of these computers get injected with malware and spyware.

  • 12:55:42

    CORBETTSure.

  • 12:55:43

    SCHWARTZAnd it changes the search results when you surf for certain types of keywords just to generate money. So what you have on your computer is you probably -- you might have a virus. You might have to do a virus check to make sure that you don't have any spyware or malware. So a lot of these -- I wouldn't call them search-engine-optimization people because what you have is -- your computer was probably hacked, and the search results, once you go to Google, were probably hacked because you have a virus in your computer. And that is probably why you're seeing these pretty bad results.

  • 12:56:09

    HOWARDI can make two points on this.

  • 12:56:09

    NNAMDIPlease, Alex.

  • 12:56:10

    HOWARDOne, if you're looking for the best movies of all time, you can actually go to a very Web 1.0 site, which is imdb.com, right, and that's actually used the collective intelligence of all its users to rate these together. Now, search, ideally, should spit you back that top 250 site. Now, whether if it does or not, it's because it's probably not using all of the signals it might be.

  • 12:56:33

    HOWARDWith respect to this issue of pornography and finding search results full of it, one of the unfortunate features of search is that there are lots of black hat SEO people, which is to say people who are trying to use these techniques to get you to land in malware. And one of the most evil things online is when they actually will use a very popular search, say for disaster relief for Haiti or something similar, and then put fake links up -- or they look like real links, they look like real sites -- and then fish people to give them money instead of the relief organization.

  • 12:57:06

    HOWARDAnd in many, many cases, I think we're going to see more of that where people will look for what they know. Someone will be looking for it and have an intention, right? Someone's intention graph is an idea here and then provide information. So, for right now, it's being used politically, not necessarily an evil way, but it's just happening. People are searching for Corey Booker after his comments on "Meet the Press."

  • 12:57:27

    HOWARDAnd now you have political organizations who are putting ads against those searches because they know there's interest. That's something we'll see much more. When people are interested in something, you'll see searches come up for it with people selling services, political ideas, et cetera.

  • 12:57:42

    NNAMDII'm afraid that's all the time we have. Alex Howard is a Government 2.0 correspondent with O'Reilly Media. Alex, thank you so much for joining us. Barry Schwartz is CEO of RustyBrick, a custom Web software company. He's also the executive editor of Search Engine Roundtable, an online forum. Barry, thank you for joining us. And Peter Corbett is the founder and CEO of iStrategyLabs. Peter, thank you for joining us. And thank you all for listening. I'm Kojo Nnamdi.

Related Links

Topics + Tags

Comments

comments powered by Disqus
Most Recent Shows

Winter Reading 2014

Thursday, Dec 18 2014Whether you like horror stories or cookbooks, poetry or works in translation, we consider a range of titles that will keep you turning pages. And we want to know what's on your reading list, so join the conversation on air or on our website to share the best book you've read this year.