Tip:
Highlight text to annotate it
X
>>Tim O'Reilly: So one of the hot areas right now, obviously, is mobile location-based services.
And I think the guy who's really the poster child for location-based services, because
he's been at it so long and with so much intensity and infectious enthusiasm and because he's
got the hot company of the moment is Dennis Crowley, cofounder of Foursquare.
[Applause] So, Dennis, please join us.
>>Tim O'Reilly: And the final guest for this panel is the president and CEO of a company
that I think has absolutely revolutionary potential is at the intersection of more trends
than you can count and is setting out to change the world. I'd like you to welcome Steve Cousins,
who is the president and CEO of Willow Garage. And you'll learn from Steve what he does.
[Applause] So let's start with you, Dennis How many of
you here use Foursquare? Oh, wow. Not enough of you. Not enough check-ins here.
>>Dennis Crowley: Working on it. >>Tim O'Reilly: So maybe then I should ask
you just to tell us the story of Foursquare, how you happened to come to it, what you're
trying to accomplish.
>>> Dennis Crowley: Do people know what Foursquare is? That's good. I used to give these talks,
and no one would know. And it would take me a half an hour to describe the whole thing.
But the gist of it is we built this platform from mobile phones that enables people to
check in to places and share the places that you go to and share the places that you like
with your social crowd. So we've been doing this for a while. I used to have another startup
called Dodgeball. And I've been experimenting in the space since, like, 2001 or so. Initially,
it was all about knowing where your friends were. I live in New York, and it's all about
trying to make New York more efficient and make, like, cities easier to use. And, if
you have this general awareness of where your friends are, this imagined map of where they
happen to be, it makes it easier to meet up with them. That was the general idea behind
a lot of this stuff. But then as we started getting people checking
in, you know, one day at a time or couple times a day or a couple times a week, you
start amassing all this data. And the data is really interesting. It's like who is hanging
out with what people over time? And you can start using --
>>Tim O'Reilly: And where are they hanging out?
>>Dennis Crowley: People go to this coffee shop, people shop at these places, people
like to go to these places with their friends or got to these places alone. You can cut
that data up to do really fascinating things with it.
>>Tim O'Reilly: Yeah. So -- actually, I'll come back to a question for you in a minute.
Let's get Steve introduced. Steve, tell us a little about what do you at Willow Garage
and why. >>Steve Cousins: So we're a robotics company,
and we're about four years old. And we've, basically, set out to try to bring automation
out of factories where it's been very effective over the last 20 years and out into the real
world. So you can see one of our robots, the PR2, out here on display. Eric and Keenan,
who designed that robot, are here and doing a great job of showing off their vision of
getting from a prototype at Stanford to a -- we've got now 11 robots that we've given
away into the world, and we're starting to sell them. You say what are these robots good
for? And we're looking at a time frame that's longer
than a typical Internet startup. In fact, we're not a typical startup at all. If I go
back a little bit, we're set out to do impact. And we believe that having impact -- and our
funders believe -- that having impact is more important than return on capital. Return on
capital becomes a second goal. An impact -- the best way we can think of
to have impact, being software guys, Scott Hassan, who founded the company, and myself,
is either software or automation, in general, is a huge lever. So, if you pour money into
that and you're doing it in a good thing, you can cause more impact to happen.
So that was kind of the premise. And we started off playing in the robotics space a little
bit in autonomous cars. There's a DARPA Challenge and a DARPA Grand Challenge. And we played
around with his autonomous boat that was supposed to sail around the world.
And then we discovered Eric and Keenan at Stanford who were building this prototype
of a robot that's like Rosie from the Jetsons but without the attitude. It's got two arms;
it can move around; prepare your breakfast, push a vacuum cleaner, help somebody out in
a factory situation. It's really a way to, again, automate.
And it's a little bit early days for that kind of robots for a few reasons.
One of the things we liked about their project was that they had identified a way to make
the arms safer. If you look at industrial robot arm, if you work for Toyota, there's
a red line in the factory or, basically, a cage. If you go inside the cage without all
the safeties disabled, you're fired, that's the rule. End of story. It's just too dangerous.
Toyota doesn't want somebody killed by a robot inside their plant.
If you're going to have robots going around the home, you need a completely different
situation. So the robots become designed differently. But we're not quite ready to put those robots
out in the world. It's a 10- or 15-year process before they're ready. So the question is how
do you get the technology -- how do you work on a problem that far out in a startup? So,
if you have the funding, that's a good start. And then if you have -- if you want to really
take off -- and we talked about being a catalyst in the last session. If you really want to
be a catalyst for an industry to get started -- and that's really what we're going for
-- the way you do that and the way we found to do that is just go open source. So we,
basically, built a community of all the best robotics people around the world who are all
now working on a common software platform that came out of this robot development project.
We've given away 11 of those robots to the best universities that we can find. They all
have them now. And we're getting to the point where there can be kind of a robot app store.
Now, it's on a much, much smaller scale than the iPhone, right? There's 11 of them. We
have 10. So there's 20 in the world. not like iPhones.
[ Laughter ] But, nevertheless, you know, understand that
in robotics the way things were before we got started was we'd go to a conference and
you'd show somebody a video of what you made your robot do. You know, if you go to the
conferences, they only did it once. Right? All the other times are not on the video.
You don't have outtakes, typically, at the conference. I show you my video; you show
me your video. And we go, "Oh, my video is cooler" and you can argue about it, right?
That's not really science. Science is about repeatability of experiments. So, if you get
to a point where you have a common hardware platform and software platform and you can
say, I packaged up my application and somebody at University of Pennsylvania was able to
run it, then they can actually do the experiment over again. And that is a huge change. So
we've done this sort of open source operating system that runs on lots of robots plus this
robot that has a lot of capability and put all those things together to really try to
transform the field of robotics. >>Tim O'Reilly: All fright. You might ask
we've got the original web entrepreneur. We've got the guy who's the cutting edge of location-based
services on mobile and a robotics guy, what do these guys have in common? What do you
see as a common thread? I see you've got a big vision. You started before there was anybody
else really trying to make a -- you're not "me too" companies. Everybody else is "me,
too'ing" after you. And you are, in fact, changing the world.
>>Dennis Crowley: I love the story you're telling, and I'm like that's crazy. That's
the position that we've all been in. We've told the Foursquare story 100 times over the
last two years. Started off with "This is never going to work. You guys are crazy" to
hey, we get to sit up on these panels with all these amazing people. I have a feeling
you're probably in the same spot. You tell the stories over and over again and people
think you're crazy until you manage to prove them wrong.
>>Tim O'Reilly: Dennis, one of the things you've done a lot with Foursquare is you've
done some pretty clever social engineering of the friend network that has been built
up. Can you talk a little bit about how you think about encouraging user behavior?
>>Dennis Crowley: Yeah. So it's funny, like, for those of you who have used Foursquare
and heard about it, there's a game component behind it. It's not a game. But what we're
trying to do is make the tools that we use all the time, you know, like the social graph
and maps and city guides, those things should be playful. Going out with your friends and
finding restaurants, those are fun behaviors. So a lot of our thinking was not just how
do you make the city easier to use from a utility perspective, but how do you make the
tools we use all the time, how do you make them more fun? And how do you make your Saturday
nights more fun? That's where a lot of the ideas came from. Let's turn life into a game.
Let's look at a Saturday night as a leaderboard. How do I compete with my friends and see who
is going out and doing the most adventurous things. We found the game mechanics on Foursquare
work really great at motivating people to do more interesting things. We make these
little badges and we reward people with these little badges--
>>Tim O'Reilly: Gym rat, for example. >>Dennis Crowley: Yeah. Go to the gym 10 times
and get the gym rat badge. Go hike 10 mountains and you get the great outdoors badge. Some
people get so nutty over the badges and wanting to collect them all, like Pokemon, that they
go out and do these things. They'll go to 10 different pizza places or 10 different
art galleries. It's like this idea of using software to change the way that people experience
the real world and encourage people to do stuff they wouldn't normally do, there's something
very powerful about that. >>Tim O'Reilly: That's right. That idea is
spreading. Michelle Obama has been behind the idea of can we come up with games to get
kids to do more healthy behavior? A great example, I think, inspired in part by what
you've done. >> Dennis Crowley: Yeah, yeah.
>> Tim O'Reilly: So let me talk a little bit about crowdsourcing again. Jerry, we started
talking about it. But I think each of you has an interesting crowdsourcing story. In
any order. >>Steve Cousins: I'll give you an example.
Not exactly crowdsources because, again, it's on a much, much smaller scale. But we asked
for people to do just a cool video. So we tried to do things really fast because robotics
typically takes a long time. It's really hard. You have to build the hardware. You have to
build the software. You have to integrate it all together. It's electronics. So we asked
people just put together a video of the robot doing something cool. This is to the 11 places.
You have six weeks. That's it. Six weeks. Some people didn't realize there was a contest
until two days before, and they put it together anyway. You can do things really fast. And,
when the robot is actually manipulating the world, it's cool. So the number one application,
the one we judged the winner, is sock pairing. So it takes a sock and turns it inside out.
>>Tim O'Reilly: That assumes you can actually find the other sock.
>>Steve Cousins: There's two socks on the table.
>>Tim O'Reilly: That's cheating, if you give the robot the socks.
>>Steve Cousins: One is inside out. It's, like, fascinating to watch something work
autonomously. If you see the video on towel folding that Berkeley did. You know, it's
fascinating. It picks them up. It did 50 towels in a row. There's a stack of 50 towels. Picks
them up, shakes it out, folds it up using a table. Smoothing out. That smoothing action
is really great on the video. Looks so great, right? And come up with a stack of towels.
That's amazing. How can you deal with that when industrial robots are like I can get
to this millimeter position and go to that position and everything's very precise. And
this is just the opposite. And the thing -- it's just so much fun to watch. One of
the teams did a playing a band. So they had a keyboard and a drum over here, and they
had their robot playing songs. They feed it Beatle music. And they did this over a weekend
at the University of Pennsylvania. And they're playing Beatles songs and things. Just feeding
it in, and they're playing the drums. It's really cool.
>>Tim O'Reilly: This goes back to the robot app store. what you have in the app store
is behaviors, effectively. >>Steve Cousins: The theme is fun. That's
the common theme. It's fun. >>Tim O'Reilly: But in some sense they're
being shared. The robot becomes more capable because somebody is saying I built a behavior
for the robot. Here it is. You can have this behavior, too. So the robots are going to
get smarter over time as more and more people start building behaviors for them.
>>Steve Cousins: And before anyone says it, and not take over the world.
>>Tim O'Reilly: That's right. Sarah Connor is gunning for you at some point.
>>Jerry Yang: In our example, at least Yahoo! Flicker was one of the first versions of crowdsourcing.
We didn't know it then, But I think it was that. But right now whether it's -- you know,
we bought a company called Associated Content. And they have people who are contributing
content of all forms on a part-time basis, on the demand basis.
You know, we would put out, you know, a request to have something written about the 5th anniversary
of Katrina. And we would get back firsthand accounts. And what I think is happening with
crowdsourcing in general, not only around content but around, you know, sales forces
or anything else, policing or sort of protecting the communities, there's this notion that,
hey, if you can figure out how to drive some quality, drive some relevance, whether it's
algorithmic or human produced, and try to create a marketplace where people who are
participating in these crowdsourcing exercises are compensated through the right licensing
or however you do it, it's really powerful. And, actually, it goes to the notion of innovation
to the end to the individual where we need to give them the tools. We need to open up
the code. We need to open up the infrastructure to them to be able to do it. But I think people
inherently want to be productive, want to have fun, to Dennis's point. If you can make
it into more of a game rather than a chore, you'll get more productivity out of them.
So these things are all related. And I love the robotics example. If I can get a towel
folding robot, I will be a hero at home. [Laughter]
>>Tim O'Reilly: So innovation, actually, can come from crowdsourcing. I think I'd love
to have you, Dennis, talk a little bit about, again, I don't know how far along you are
with developing the Foursquare business model. But it seems to me like some of your participants
in your ecosystem invented your business model, as far as I can tell.
>>Dennis Crowley: We've been crowdsourcing pretty much everything. When we would launch
city by city, we'd turn on Vancouver and there would be no bars and restaurants in the system.
Nothing to check into. And 24 hours later there's 2500 places to check into. Users create
the content for us. That's amazing. It happened all over the world. There's 8 million --
between 6 and 8 million Foursquare venues now created by people all over the world using
the service. >>Tim O'Reilly: People adding tips also.
>>Dennis Crowley: People adding tips. People adding venues. Random things. There's a whole
bunch of 9/11 memorials popped up as venues. But, you know, we really see a lot of innovation
coming from the people who use the products. It's not just the users, the people checking
in. It's the venues, the people that see the people checking in to their places.
And so, you know, what's starting to become our business model, like offering specials
to merchants and to local users, depending on who is the mayor of a place, who's been
there 10 times -- we didn't invent that stuff. That stuff came from the venues. Last summer,
summer before, we, you know, as Foursquare was just starting to pick up, we saw coffee
shops in San Francisco hanging flyers around the neighborhood like: "Come here. Show us
you checked in on Foursquare, and we'll give you a discount."
We're like, wow, that's kind of a great usage. We reached out. We talked to some of the merchants.
We realized a whole bunch of people were doing this, doing the ad hoc flyer stuff. And we
talked to those folks. And we ended up building a product around it. Now we're seeing our
mission is to try to bring that product to millions of businesses across the U.S. And
it's the users that are doing that. We get these smart *** users that are awesome. They
go into coffee shops and say, "I'm the mayor here. I've been here 10 times. What do I get
for free?" And they're like, "I don't know what Foursquare
is. I don't know what it means to be the mayor." >>Tim O'Reilly: I bet a bunch of people in
the audience don't know what it means to be the mayor either.
>>Dennis Crowley: If you're the mayor of a place on Foursquare, it means you've been
there more than anyone else. So people fight over being mayor of the park or their office
or the coffee sop. There are these big things, these big mayor battles. They get written
up in papers, people make flyers, and people go nuts over this stuff.
But this happens all the time. We hear the story, "I'm the mayor of my pizza place. What
do I get?" And the pizza place is like, "I have no idea
what you're talking about." And it happens four or five times, people
are saying, "What is this foursquare thing?" And it's the users teaching the venues that
the venues should be on Foursquare. So we're crowdsourcing a sales force, in a sense. We're
in the very early days of figuring out how that works. But it's worked for everything
else we're doing, so why wouldn't it work for that?
>>Tim O'Reilly: I think it's really interesting. Because it really does suggest that, when
you have a technology transition, you end up with new business models that come out
of left field. You know? When we started with the web, it was my first idea of web advertising
was influenced by those bingo cards you used to have in industry publications. Some of
you probably remember those. It was like, oh, wow we can do Web sites that will be the
destination for what used to be a bingo card. And you'll be able to have a catalog. And
people will go look at stuff online. You know? And then along came all the other forms of
advertising. But it was key word advertising on a search engine that really sort of ignited
this next big phase. But nobody thought of that in the beginning.
But we're seeing that now as we are starting to embrace location-based services on mobile.
And there's a lot of sort of thinking about what will the advertising model look like?
What will the business model look like? I think we're going to be surprised.
And I think what's happening organically with Foursquare is a great example of the market
teaching us where that business model might be. But I'd love any thoughts that you have,
Jerry, on that, if you've been thinking about --
>>Jerry Yang: I think you're exactly right in that probably the best business models
haven't been figured out yet. And consumers -- I know A.G. Lafley talked about that this
morning. But consumers, ultimately, are going to tell us where to go. Even at Yahoo!, when
we force a business model that doesn't quite work with consumers, it never really works.
And so you have to ultimately trade that balance between do you monetize early or do you go
where the consumer is? Especially this day and age, being able to really leverage where
the consumer wants to go and make the monetization with that seamless is really critical. But
in a different way I think people who are getting what they want are willing to pay
a lot more. So you have this almost bimodal where, if you're happy with something, you're
going to pay a lot more to that place. And, if you're unhappy, you want to do everything
you can to get away from it. So I think consumers are becoming a lot more intelligent as well.
>>Tim O'Reilly: So another aspect of crowdsourcing is open source. Gary, you've been a big open
source -- not Gary. Gary works with you. Gary Bradski wrote "Open CV," which is this fabulous
machine learning package for computer vision and works on this. One of the really fascinating
things we were talking -- Jerry and I were talking earlier about machine learning is
that computer vision and a lot of stuff that's happening on the web actually use the same
algorithms. And that's really worth thinking about. But sorry, a little aside there.
I just sort of want to ask, you know, how much do you think open source drives innovation?
>>Steve Cousins: It's pretty interesting. Open source -- to me IP is friction. Right?
So friction. Friction isn't all bad. Right? I was coming down the mountain this morning,
and without friction it was going to be really bad.
But it's not all good either. Right? And IP, when you open things up, things change dramatically.
So, for example, we give away the software and completely open BSD license so anyone
can use it for commercial or noncommercial. And suddenly, because it's not ours, because
it's a community thing. And we actually put it on our robotics.org site instead of willowgarage.com.
Because it's on a dot org site, because it's everybody's, all these different robotics
people are willing to contribute in a way that they wouldn't be willing to if it was
our code. Right? So, even though it's open source, it's not just that it's open source.
It's open source, and you can take it and form a company on it. We're not the ones that
are going to commercialize it and you're doing slave labor. That makes a huge difference
to people. And there's definitely a place for companies to take off.
But, you know, one of the companies that has a PR2 robot now is Bosch. And Bosch, the reason
that we gave them one in our beta program was because they committed to open source.
And think about a company the size of Bosch, which makes auto parts and drills and things
and all kinds of different products. They're a worldwide conglomerate. A company of that
size deciding that they're going to do open source is really surprising. It doesn't happen
that often. Intel has done it. To some extent, they're the ones who originally supported
Open CV. IBM has talked about it and done it in some ways. But it's very unusual.
The reason that they do it is because we have a very clear business model. Robots could
take off in 10 years, and we're going to be the robots parts makers. So we can give away
software all we want because we're never going to be a software company. So let's contribute
to this open source software and become part of the community. And it really makes a difference.
>>Tim O'Reilly: Okay. It goes back to that create more value than you capture?
>>Steve Cousins: Yeah, exactly. >>Tim O'Reilly: Be part of building something.
So if -- I just wanted to open this up to questions from the audience. If anybody wants
to ask any of us here a question, feel free to go to the mics.
Lacking that, I'm going to ask you guys, you guys think about the future. You've got to.
So when you think about the future, what's going to surprise us the most over the next
few years? >>Steve Cousins: For me that's easy. It's
what application will come out of robotics. People ask me all the time, What are you going
to do with these robots, right? It is not just folding towels. You can do lots and lots
of stuff. But what's the really important thing? And the answer is you can't even think
about that until you have robots. Once you have robots, then you can go out and build
it. The spreadsheet didn't come out until there
were PCs readily available. And we don't actually in robotics say "killer application" because
it has a different meaning there. [ Laughter ]
But what is the equivalent of a spreadsheet for robotics?
>>Tim O'Reilly: Yeah, yeah. >>Jerry Yang: I think with all the opportunity
we're talking about here and obviously all the stuff that's going on in our world, I
do think there is a -- how do I say this without being negative about it? I do think that there
is going to be more challenges around how we govern and how do we make sure that --
assuming this trend continues where everybody, every individual can be part of a creative
or innovative process, I think in some way that's really good.
I also think that that may mean that other people with the same tools can do bad. And
I think how do we create a system where this stuff is sustainable is a really critical
question that we should think upfront. It's not that robots or location services
or Yahoo! by itself could be part of that, but I do think the more you empower individuals,
the more you do things, the one person, the one entity could do some damage in a system.
And I'm really struck by this notion of, you know, how do we -- how do we create impact
but for our next generation, you know -- I was saying to Tim earlier. I'm sort of living
through my kids' eyes, seeing the world through my kids' eyes and there is nothing in my mind
more special but also more worrisome about seeing the world through a 6-year-old's eyes.
The question she would ask me really worries me about the future, whether it is on the
energy, environment side, whether it is on the biology side, whether it is in the robotic
side. And I think to me doing good is not just giving
money away. It is about solving some of these problems that we have the technology, and
I think we have the brains to do it now. And it is something I hope we all take away
and think about. >>Tim O'Reilly: They are giving us the red
light. So I think that means we've got to wrap? Do you have something quick to add?
>>Dennis Crowley: Well, yours is really deep. We always talk about the stuff we are building
is, like, technology that facilitates serendipity or makes these little chance encounters happen.
And so whether it is a room of people that don't know each other that have information
to share, whether you are walking down the street in a strange city, like, those are,
like, situations that present problems that could be solved with some of the stuff that
we're doing. I desperately want the application on my phone
that scans for everyone here and tells me the ten people I need to talk to before I
leave today. And when I have two hours to kill down in Phoenix before my flight, I want
to go to my social graph and tell me what are the five things I got to do if I have
two hours to kill. So that's the type of stuff that we are trying to solve. We are getting
to the point where the phones are smart enough. The graph is out there. The geotools and the
social tools are there. It is putting all the pieces together, just making that stuff
happen. >>Tim O'Reilly: All right. Sounds awesome.
Thank you very much. [ Applause ]