usvsessions1_fulltranscript_ref.doc

WikiHome ยป usvsessions1_fulltranscript_ref.doc

UNION SQUARE VENTURES SESSIONS

October 20, 2005

540 West 21st Street New York, New York 10: 20 A.M.

Reported by Edward Leto

ATTENDEES:

CHRIS ANDERSON

YOCHAI BENKLER

MATT BLUMBERG

BRAD BURNHAM

JOHN BUTTRICK

TOMMY COHEN

DICK COSTOLO

TOM EVSLIN

SHANA FISHER

CHRIS FRALIC

MICHAEL FRUMIN

SETH GODIN

SETH GOLDSTEIN

HEATHER GREEN

UMAIR HAQUE

SCOTT HEIFERMAN

PETER HERSHBERG

MARY HODDER

CYRIL HOURI

JEFF JARVIS

RONY KAHAN

JOSH KOPELMAN

SCOTT KURNIT

HOWARD MORGAN

DAVE MORGAN

MARTIN NISENHOLTZ

CHARLIE O'DONNELL

TIM O'REILLY

MICHAEL PAREKH

NANCY PERETSMAN

JONAH PERETTI

MARC PINCUS

JOSHUA SCHACTER

PETER SEMMELACK

CLAY SHIRKY

BRUCE SPECTOR

JOEL SPOLSKY

JOSHUA STYLMAN

RIKKI TAHTA

MATTIAS TURCK

ALBERT WENGER

FRED WILSON

BOB YOUNG
MR. BURNHAM: Folks, thank you for coming. This is what I guess is the first of what we call Sessions, which is just an opportunity to get some of the smartest people we know together to talk about what's happening in the market. We are very, very excited about this group and the set of topics -- I hope everyone's had a chance to see some of the things on the Wiki that we put up. By the way, since we have this group here, we can help decide, is it Wiki or Weeki?

THE GROUP: Wiki.
MR. BURNHAM: It's Wiki. Thank you. There's a whole lot of things that are going to be made clear to me today that have never been clear before, like that. So there is a Wiki that most of you have probably seen that has an agenda on it. There's an agenda in front of you. We have a lot of things that we want to cover today, and so we're going to try to move through it pretty quickly. Fred and I are going to be -- since you know we paid for the green cloth here and that's about it, but since we did, we're going to be pretty ruthless in trying to keep to one conversation, which is going to be very hard, but it's going to be impossible to hear people if we, you know, have two or three conversations going on. Great opportunity during lunch and during the breaks to have breakout conversations, but when we're sitting down here, let's try to keep it to one conversation. There's a lot of people here who have a lot to offer. With that, I'd just like you to just dive into the first topic.
MR. WILSON: Brad, let's hang on. We have some rules, we should go over them, but also, I'm not sure that the acoustics will be great in here. In fact, I'm sure that they won't be, so try to speak up so that the people on the other side of the room can hear you, and if you have something you want to get out and you can't get it out, feel free to raise your hand and kind of do this and Brad and I will try to play, you know, business school professor or whatever the right word is and call on you. The rules are that we would prefer people not blog or take transcripts or notes of the event, because we have a stenographer who is transcribing the whole thing. That's going to be up on our Wiki, and hopefully that will be a good source of notes and fodder for you to go grab and blog tonight or tomorrow and over the weekend. We're also going to post all the pictures that Mark is taking on the same place, so if you want to grab some pictures, and you know say this was Jeff Jarvis getting apoplectic, that would be fine. Also, cell phones, we prefer people not taking cell phone calls during the session. If you're expecting one, Kerri's here and we'd be happy to hold your phone and take the call for you and pull you out of the meeting. The other thing is that we think it's going to be pretty easy to get cabs on 21st and 10th, but if you're particularly nervous about getting somewhere, I know a few of you have to get somewhere, feel free to ask Kerri to schedule a car service to come take you. If it's something that's really important, we're happy to do that for you.
MR. BURNHAM: So the other thing we should probably mention is what our objective is here, and our objective really is only to try to push the conversations forward. There are a lot of very interesting things happening in the market. There's a lot of transformative things that are happening, and we don't have any proprietary interest in the output of this. We want everybody to feel free to blog anything. We're going to put the entire transcript up on the Wiki. That's not promoted, so it's not going to be broadly publicized or anything, but it's going to be there, so if there's something that you don't want to be known, you should not introduce it here, but hopefully, what we're going to do is help clarify some thinking around the trends that we're seeing, the events that we're seeing, and help us all think a little bit more clearly about the kinds of things that'll help move that what I think is generally a very positive set of developments forward. So that's our goal. The first topic is really what is peer production. All of us have some ideas on peer production. We're very fortunate today to have Yochai -- was I close?
MR. BENKLER: Yes.
MR. BURNHAM: Yochai Benkler here with us today who coined the term in his paper, Coase's Penguin, common space peer production. He's thought a lot about it. Thought a lot about it early. I was wondering if I could ask Yochai, just by way of introduction, to talk for a minute about what he thinks it is, because he probably has a better idea than the rest of us.
MR. BENKLER: Except that for me to talk for a minute is really hard. I think the critical characteristics, as far as I can see, is that it is a set of social behaviors that have economic outputs, but are organized along motivational and organizational forums that are developed in parts of life that traditionally were on the periphery of the economy, that we understood ourselves as being engaged in producing social relations and producing psychological well-being, and so to the extent we thought about production of things of value and organizing them in terms of markets and firms, when we thought of motivations, we thought of material motivations and we optimize our systems of thinking about how we get people to do the things we want them to do -- whoever those other people are -- what it is that we expect will emerge out of these centralized behaviors, all of that was optimized on the motivation dimension on material well-being, and on the organizational level, either in terms of the price system or in terms of organizational lines of authority. I think what peer production is intended to capture is the fact that a whole set of other behaviors that have grown up in the household, in friendships, in communities, the motivations that they capture, the signals that get people to explain what it is that they desire, how they desire, what they want to do, what they're trying to do, all of these things are suddenly becoming integrated into the core economic activities of the most advanced economists, and all of the players inside of these economies need to begin to think. It's a new set of social competition. It's a new set of opportunities. It's a new solution space for ways to solve production problems. And we need to start learning how to live with, use, provide platforms for, use the outputs of without undermining this new set of social cultural practices.
MR. BURNHAM: So, you know, I think in your paper, Yochai, you talk about Wikipedia, you talk about the NASA project mapping, the craters on the moon. Are there a couple of other representative examples that -- I mean, do you think of, as an example, file sharing, MP3 file sharing as a form of peer production or do you think of that as something else?
MR. BENKLER: Well, it depends on what aspect of it. Peer to peer file sharing networks are peer production of two different things, neither of which is music. The first thing, more recently in 2004, I wrote a piece called Sharing Nicely. It's about sharing of material resources, computation bandwidth and processing, and the economics of that, those which is the basic economics of those are different, because these are not public goods. These are not long rival goods. And so there was some working out to do with that. But the first thing that P to P networks are is social provisioning of a distributed data storage and retrieval system. So in 1999, if you were a professor interested in a robust storage system, or a CEO interested in a robust storage system and you came to your engineers and said -- in the middle of 1999 -- build me a data storage and retrieval system that would be available for something like 100 million users anywhere in the world, all the time, that would be robust to attacks even to the level of the central indexing server would be shut down, that armed guards would know that malicious files would be inserted, and nonetheless, all the data would be there and would be easily adopted, again, after each of these attacks. You would say this will cost me X 10s of millions of dollars and, of course, Shawn Fanning did it easily because even when NASA was brought down, the data was still there, all you needed was a new search algorithm. When Major Super Notes were brought down, there were new Super Notes. So it's peer production of a distributive data storage and retrieval systems. If you ignore the copyright issues and you ignore the moral issues and you just ask "What have we learned about peer production?" We've learned about peer production that part of the solutions make the problem. If the problem of production is a robust storage system, P to P is an answer. And to some extent, the other thing is accreditation. So the kinds of things that come out of the P to P networks that are distributedly produced are the aspects of what the recording industry does in regard to saying "this is good, this isn't good," in terms of people basically saying, "Here, if you like this, you'll like this as well."
MR. BURNHAM: Go, Marc.
MR. PINCUS: Well, one other spin on peer to peer production that hopefully is relevant to peer to peer is I keep wondering, every time at the beginning of a new market segment, industry segment, and I see everyone scurrying around building a piece of it and building a lot of the same pieces, it's always frustrating to me that there's various levels of participants in every new segment, so I started with the social networking segment and there might be a friend who's relatively far along, there might be Yahoo who isn't in it but could have a lot of power, and then there's a lot of companies doing individual pieces of it, like Flickr was and like Tribe was trying to classify, and what was frustrating me then and still frustrates me now is that what was clearly in the interest of an end-user was for us, as service providers, to work together in a peer to peer fashion so that -- and Tribe tried to create something called a friend of a friend open standards so there can be one social network. That was an interest of the end-user, and in terms of getting the innovation to market, it would have created better services for each company to provide a layer or interesting service that interconnected with the whole, and instead, everyone takes this approach that persists today of their own wall garden for a while. And now I've been spending time looking at the advertising, online advertising industry, and it's the same thing again, that so many of these companies are trying to create their own ad networks, and I've talked to most of them -- hopefully talk to Dave some more today -- but I talked to most of them, and it seems like the exact same thing that I went through in social networking where the interests of both the advertiser and the publisher would be best served if there were one open network, meaning in my humble opinion, and then there were lots of companies that would try to innovate on top of that open network. But, instead, you have, again, everyone at various stages of ownership of the industry not wanting to let go --
MR. BURNHAM: You're talking about a subject that I would like to get into some depth actually a little later today, which is the relationship between standards and the role that standards play in a peer produced system, but I want to get Tim in this.
MR. O'REILLY: It seems to me that there's several different classes of peer production, you know, and the formative essay for me on the subject was one written by Dan Bricklin called the Cornucopia of the Commons, which he published in 2001, and he basically pointed out the ability to get volunteers to do it or you can architect the system in such a way that it's produced as a byproduct of people's individual selfish activity. And I thought that was a really profound insight, because a lot of peer production examples that we see, like open source software or Wikipedia are in that second category, people coming together, network enable to actually build something consciously contributing to it. Whereas Napster was a great example of a system that was built in such a way that people were just pursuing their own activities, but because of the way the system was designed, you created, as a side-effect, this, you know, this peer good.
MR. BURNHAM: Del.icio.us would probably be in that second category.
MR. O'REILLY: That's right. The architecture of the web, or, for that matter, the fundamental architect of the internet is such that people can pursue individual activity. People put up their own web site for their own purposes, and yet, there is a collective good that was created. And it's kind of interesting, because if you compare Napster with say Google, they're both leveraging that same insight, but in the case of Napster, it was actually built in from the get go as the driver for how the network got built. In the case of Google, it was somebody coming along later and saying, "Okay, we can actually leverage this system that's actually been built without us," and they were able to design some of the potential in that system and then leverage it. So I think that's maybe kind of a fourth level where there is an underlying architecture. Then there's a set of second order phenomena that are driven by that architecture that were implicit in it that someone can then capture to create something --
MR. BURNHAM: Yes. Jeff.
MR. JARVIS: If you give people control, they'll use it, and if you don't, you'll lose it. It's possible. That's exactly what Marc's talking about here, and Tim as well, is that the natural order of things is if I'm doing this as part of my consumption, and I create my iTunes list, I'm consuming music, but I'm creating a radio station. When I'm consuming peer to peer, I'm helping create the network. When I'm tagging something, I'm now creating a value, but if somebody comes in and tries to control that in a way that doesn't serve me, then they may win for a short while, but in the long run, we're all going to lose. But that's so against our culture. That's so against our culture. It's hard to do the things that Marc's talking about, but if you do, when those things happen, if you're smart enough to let them be and then build on top of them, there's tremendous value to be had. That's the new secret success here, is to build on what people are doing in the normal course of their lives.
MR. PINCUS: And the value that would be achieved there accrues way more to the end user and to all the small participants in the industry than to the biggest ones and builds the whole rather than splitting up.
MR. BURNHAM: We're going to come back later in the afternoon and really talk about what the impact on the structure of these industries are and how you could potentially invest or not invest in these industries, where the economic value is. But let's stay kind of in the definition of what we're talking about, peer production.
MR. EVSLIN: I think an interesting blog you wrote about amateurs meeting with professionals, I think the thing we found out is that anthropologically we have a much greater urge to cooperate and to do cooperative things than we knew that we had as a species. Other examples you asked before the help forums that grow up around every possible service where there's a bunch of volunteers basically providing export because they want to. I think what happened is is that we assumed that people only did things that they got paid for doing, in general, or at least the kind of work activities, and that was true, but it wasn't only true because they got a salary, it was true because the employers provided the expense of facilities which used to be necessary to do those things. So you couldn't be a writer before unless you had a publisher, because you didn't have any way to print and distribute what it was that you wrote. You couldn't be a tech support person off in cyberspace because you liked helping people, because you just simply had no access if nobody had any access to you. So all of a sudden, when everybody becomes accessible to everyone else at very, very small cost where you no longer need an employer to provide you a platform for doing those things, we start to do a lot of social things, because we --
MR. BURNHAM: Tom, you're talking about two things. You're talking about, first of all, the availability of the infrastructure, and secondly, the motivation, and one of the interesting things that Yochai said was that the motivation appears to be happening in western societies where, you know, the basic hierarchy of needs is taking care of people. They're not hungry, they're not worried about feeding their kids, and they have this infrastructure, and they can begin to think about what else they can do.
MR. EVSLIN: Right. But I don't think the urge to be cooperative -- once you're well fed, given -- is a strange one. I think we just didn't recognize it before because of the unavailability of the resource, particularly to cooperate with strangers. We did things in our communities. We did things in our homes.
MR. BURNHAM: Let's get Michael and then Howard in here.
MR. PAREKH: Another thing, the term peer to peer, one of the things that's implicit, unfortunately, is that everyone is participating in an equal way, and the reality is in terms of all these others people talk about, from Napster to Wikipedia, the number of people that are actually doing -- contributing content, it's typically a much lower percentage.
MR. BURNHAM: Well, Wikipedia is 2,500 to 10 million or so users.
MR. PAREKH: Exactly. If you look at the number of people who read blogs versus write blogs, the number of people who put comments in and who actually have their own blog, I mean, typically you're talking about percentages that are a fraction of the whole system. And one of the things that I'm trying to figure out in a lot of these examples is which of the systems that have lower friction points. Like Del.icio.us, where you would assume that a lot more people, because it's easier to put tags in, are participating and contributing to the system as opposed to just taking things out, and one of the things we need to be mindful about in all the peer to peer conversation is how do you make these systems more efficient? How do you get more people to participate, whether it's trust factors, shyness factors, learning the tools. Whatever the element is of that system.
MR. BURNHAM: The way you participate is probably different based on which kind of system it is, and how you facilitate participation is probably different.
MR. O'REILLY: Everybody participates in Google, because they figured out how to leverage what people do implicitly, as well as what they do --
MR. BURNHAM: Howard.
MR. H. MORGAN: One of the things that comes out of that is utility function for people now is reputational, so one of the things you're going to get from the peer networks is reputation, and we see everyone contributing, as you said with Google, where they're explicitly adjusting search results based on other user behavior. You're contributing, but you don't know that you're contributing. And inside pages, you're contributing, but you're doing it for reputation purposes, because a review from a friend or a review from someone you know enhances their reputational value. Ebay is obviously a great example of that. And I think looking at the utility function in the western world, we can afford to do that, have a reputation utility function, whereas in other parts of the world they still --
MR. BURNHAM: Chris and Fred.
MR. ANDERSON: It's interesting, to follow up on that point, the reputation, of course, is -- rather than money, and it's interesting that Google got the idea of patroning a citation analysis which was the way they used to measure the hierarchy of individuals in --
MR. EVSLIN: I think the point is that makes even lurkers participants, because they're always voting by the pages that they reference. So even on any forum, there's probably 50 times as many lurkers as there are active posters, but everybody is a participant because everybody is directing attention and where that attention --
MR. BURNHAM: And the more meta data that's available, the more that becomes possible. Fred.
MR. PAREKH: And my point is that there are terms for that. We need to introduce new terms into the language so we know how everyone should comprehend it, and it's not obvious until you think about it. When you think about it, it's like "duh."
MR. BURNHAM: Well, we're going to end up talking about terms and also rights and ownership and things like that at the end of the day, but Fred.
MR. WILSON: I just was fascinated by the point that either Michael or Tim made, which is that there are some peer produced systems where the content is produced automatically, and I've been fascinated over the past couple of weeks with this service called last.fm which simply takes your iTunes listening and uploads it to a social network, so every time I listen to my iPod or my iTunes, I'm contributing data to social network. I do nothing, but I know that I'm doing that, and it actually has changed the way I behave. I actually listen to thing consciously, because I know by listening, I'm contributing. It's a very odd behavior, but I think it's something we should focus on.
MR. BURNHAM: Scott.
MR. HEIFERMAN: As I was in the subway yesterday -- and I'm going to post a picture of this on my photo log -- I saw this 20 year old guy wearing a T-shirt that said "Fuck You, I Don't Need Any More Friends." And so, of course, I had to go up to him and say, "Can I take a picture?" And he was so friendly. But the point of this is, you know, I wonder how this whole idea of implicit peer production, you know, transparent peer production that you don't have to -- there's a company being incubated at Kleiner Perkins called Project Edgar, which is about asynchronous voice, and the way it was pitched in a deck is "Now you can talk without talking." And so it sounds like this whole just idea of peer production where you don't even know you're involved, where you don't even know you're participating in some sort of holy grail, and I would just throw out there that part of the beauty of what we're talking about here is that -- is that the world's a better place when we all -- when we realize that we're in something together. We're all contributing to Google. We're all making something that this transcends the technology and all that, that you're a part of something bigger, and so how does what we're talking about -- just to conclude -- how is what we're talking about going to make it unfashionable to wear a T-shirt that says "Fuck you, I Don't Need Any More Friends."
MR. BURNHAM: So is there a social benefit that's a natural byproduct of this evolution that we're seeing.
MR. HEIFERMAN: In the Napster heyday, I remember -- I mean, there was this certain subtle euphoria that you felt like, you know, you walked down the street -- and maybe it's just me -- but you walk down the street and you said, "You know, these massive strangers, some of them gave me music last night," and like I want to thank the world for giving me all this great music. It sort of -- I think that's something interesting about this notion of peer production, that you're not just this isolated part of the world, but I'm interested in how it's less transparent, how it's more -- I'm sorry, I'm interested in how it's less transparent.
MR. BURNHAM: Tim.
MR. O'REILLY: What I was going to say, one of the things it seems to me as a characteristic of system -- this is returning to Marc's question -- a defining characteristic seems to be how much is the value of the system you try to capture. You know, if you look at something like Ebay or Google, even though they're very, very successful companies, they're actually pretty generous in creating value for people outside the company, and you know, versus say if you look at sort of the traditional wall garden kind of company, there's really not that opportunity for creating value. You know, Pierre likes to point out how many people make a living on Ebay and how the social goal of getting people to trust each other is intrinsic to the value of the system at creating an economic opportunity for people. Ebay gets a small --
MR. BURNHAM: You know, it's funny, there is a relationship between the economics of these systems, but there is a relationship to Marc's point about standards in that it's also the narrower standards that seem to be the ones that are most successful that don't try and define the entire system, but our assessment is a fairly narrow standard. It creates an opportunity for Dick's company to solve a set of problems around that. It creates an ecosystem around that. So it's facilitating these ecosystems that --
MR. PINCUS: But isn't, Tim, I actually in today's -- in last year's world, Ebay was a great example of an open platform. People could have businesses on it. In today's world, they're one of the best examples of a wall garden that doesn't get anymore. Why can't I export my Ebay rating and use it somewhere else? That should be like Paypal, a service they provide me that has an open API that let's you import your Ebay rating and use that in the transactions that has nothing to do with Ebay.
MR. O'REILLY: Well, I think companies are learning that. I think it was interesting, at 2.0 Conference, Yahoo and Google both introduced APIs, and clearly Google got something Yahoo didn't, because Yahoo's was all about putting stuff onto Yahoo, Google's was all about syndicating stuff out, and sure enough, Google is the one that's sort of taking the world by storm. So I think -- I guess to me the faith I have is that there's sort of a science to this. I used to say this about open source, and it either works or it doesn't, and our goal is to find out what works, and in some of these areas, we're trying to figure out. Do you move the dial this way and it stops working? Move the dial back that way, it works better. And our goal should be as sort of the scientists of the social value and economics to figure out where the boundaries are, because you move it too far one way and you've created no economic value for yourself, and therefore, you don't have the incentive to either. Perhaps on the other hand, you try to capture too much and you don't create the social beneficial network effects.
MR. BURNHAM: Let's get Mary in here, and then I want to get back to Yochai.
MS. HODDER: I was just going to comment on what Marc just said. Actually, you can pull data for reputation for Ebay. Opinities pulls reputation data from Ebay, but the thing about the difference between what Tim was talking about and maps and Ebay's reputation information is that the mapping data makes sense when you pull it out of the system, whereas the reputation data, because Ebay is so skewed, it's such a bizarre social environment, everybody is under tremendous pressure to make this sort of, you know, A+++ best sale I've ever had, which, I mean, would only exist if the guy who was selling you the thing drove me the item from Kansas or something, otherwise it's just probably B+. But everybody wants to either make something -- the reputation information perfect or it's terrible, and so when you pull it out of that system and you put it onto Opinity, it doesn't -- it doesn't match up. It doesn't translate with other wall garden reputations.
MR. BURNHAM: It sounds like a failure in the definition of the standard. It's too specific to that environment.

MS. HODDER: Also, the peer production of that reputation system, it's sort of broken down and it's gotten really weird. Whereas the map data from Google makes sense in lots of different contexts.

MS. GREEN: Does that have anything to do with reputation?

MS. HODDER: Well, but the interesting part for me is the Google data isn't peer produced. The map stuff is stuff they make and send out, and the Ebay stuff is something we're all making. But because they've made these really weird social parameters, we're making it a very strange way.
MR. BURNHAM: Let's go to Yochai and then Scott.
MR. BENKLER: Couple of strengths I think that are worth underscoring from the first sort of comments. One, I've actually started to use the term social production to cover the two distinct phenomenon, Tim, that you're describing, with individual action that's based on social psychological motivations, that then gets coordinated by some platform as simply common space production, in the sense that I neither get the inputs out of the property, nor do I create the outputs (inaudible), and peer production to describe the more self-conscious cooperative platforms. And the reason or one reason why that's important for people who are thinking about building platforms or using outputs of this social system as inputs into their own production process is it, again, characterizes, with a little bit more detail, the set of strategies that are open to interact with the social system. Sometimes you need affirmative cooperation. This goes not to the social sense of participation, but to the practical sense of what kind of platform we need to build. Is this kind of behavior, that individual socially motivated behavior can cohere into a useful information product, and if so, what's the intervention, as a separate design question from no, this is something that actually requires affirmative contribution and I need to build a cooperation platform. So I think that's a very important distinction to separate out, even though it's all part of the broader phenomenon of social production where social motivation and organization, of course, are pushing things that are economically valuable. The other thing, though, just with regard to, Howard, when you mentioned issue of reputation, I think you need (inaudible), which is reputation is one social mechanism. It's true that it's very strong in academia, but I think people also participate for a sense of -- as Scott was suggesting -- of being part of a community. And sometimes that actually requires understating the reputation, and sometimes reputation isn't actually enough, and we don't have actually very good research on the extent to which -- we do have very good research on how adding money undermines social motivations depending on contents. We don't have very good research, because there's never been a research question that was important to people in designing economic systems of whether or not you build a very strong reputation system, you're actually undermining some of the solid -- because it becomes "I'm better than you" or whether they work with each other. But the design question is how you build a set of motivations that's outside of the material in a way that's mutually reenforceable. Might be reputation. Might be less reputation. It's not homogeneous.
MR. BURNHAM: Scott was going to jump in.
MR. KURNIT: I just want to go back to Google. One of the reasons that Google is happy to -- they're not just saying "use our APIs and have that because we're good guys." They built a better outbound monetization in their sense than Yahoo has -- so what in fact what Google has done is expanded the wall garden to be the size of the internet, whereas Yahoo's wall garden because they (inaudible).
MR. O'REILLY: I just want to pick up on something on that Yochai said that triggered a thought for me about reputation in the context of open source and I wonder if it's direction for some further research. It seems to me that in open source communities, you actually see two kinds of reputation. One is somebody like Linus Torvalds or Larry Wall creates a system and effectively an architecture by which people can add to their work, and people get secondary reputation by extending what they've done. There are other cases, though, where it is actually rivalrous reputation. So Perl versus Python versus Ruby. And I'm just sort of wondering, it would be kind of interesting to understand the economics and the reputation -- not just of the monetary economics, but the reputation economics of rivalrous systems that are using the same, you know, all the same principles, but, you know, versus ones that are building purely on what was done before. And again, I don't have any answers there.
MR. BURNHAM: Let's get Martin in here.
MR. NISENHOLTZ: I guess I have a question which has to do with what I sort of call the lottery syndrome. There was a Powerball lottery yesterday. Tons of people entered it. We know that someone won in Oregon. I don't know who quite it was yet, but we also know that the chances of winning were one in 164 million, and we know that because we can certainly measure that. I guess what I'm struggling with is how we measure the number of peer production efforts that get started versus Wikipedia, which has become the poster child, the lottery, the one in 164 million actually works. Now it may not be one in 164 million. It may be one in 10. It may be one in 50, but I think that groups of people like this -- and I count myself among you, so guilty -- tend to create the lottery winner and hold the lottery winner up as the norm.
MR. O'REILLY: There's some data for that.
MR. NISENHOLTZ: That's what I'm asking for.
MR. O'REILLY: Look at Source Forge, there's something like 104,000 projects on Source Forge. You can actually do a long tail distribution and figure out how many of them -- but, just to my knowledge, I would guess that one in like, you know, whatever 154 million there are probably out of those 100,000 projects, there are probably, you know, at least 5,000 who have made significant reputation gains as a result of their work. Maybe more. But, again, somebody should go out and measure that.
MR. NISENHOLTZ: I'm not suggesting -- what is success?
MR. JARVIS: That's what's changing now. The definition of success, big enough, is different now. I don't need to -- I can publish a book however I want. I don't need to be big anymore to be successful. And so that's what has essentially changed the economics the way Peter recognized. We're no longer a blockbuster economy. We can be. But if you want to go in and earn a living now, by whatever value you get, plus reputation, plus good heartedness, plus return on my (inaudible), you can now do that in a far smaller scale. However, Marc, when you add all those little things up, I would wager that they're bigger than the old big blockbuster and more sustainable in the long run.
MR. PINCUS: And, Martin, you are in the position you don't have to play the Lotto, because this is really a gift economy, and the more that you give, the more you're going to get back, and that might sound very California new age, but.
MR. NISENHOLTZ: Fuck you.
MR. PINCUS: I did have an experience with a friend I brought and I tried to explain that to her and she just didn't get that, gift economy. But you don't have to wonder whether you'll throw a party and they'll show up. Somebody starting with nothing says, "Look, folks, here's Wikipedia," but you could, the New York Times or AOL, someone who has something of value, could put it out there and give it -- you could give something whether it's like Google with Maps, whether it's some database that you've been closed and hoarding because it's been a value to limit access, you could give it out to the world. You can give people the ability to publish something in your newspaper. There's so many pieces of value you can give out.
MR. H. MORGAN: How about AOL IM. IM is the classic case where we had so many different wall gardens, and, Brad, I have the scars that can prove that. That could have been much better for everyone if it was really open up.
MR. PINCUS: Including AOL.
MR. BURNHAM: We're going to have to break in a couple minutes. What I'd like to do, Scott's got something important to say, and then I would like to hear from anybody who hasn't yet been heard from, and if you haven't spoken yet in this hour long or 50 minute session, we'll give you a chance to speak and then we'll go break. Scott.
MR. HEIFERMAN: Isn't it interesting that in the Powerball lottery, the more people that play, the better the prize, but the more people that play, the less likely for you to win. Whereas network effects, you know, network effects of Napster and Del.icio.us or Ebay or Meet Up, it's the more people that play, the better it is for everyone.
MR. BURNHAM: It's a difference between a rival good and non-rival good in a way.
MR. WILSON: So, Matt, you had your hand up. You got something to add?
MR. BLUMBERG: Just a couple points about reputation. One is I think reputation is absolutely critical to this environment, and it's critical on two different levels. One is about the reputation of whoever is generating the content, which may or may not matter for things like Del.icio.us or for clicking on something in Google, but it clearly matters as you go up the food chain as the investment of consuming the content gets greater. So I'm more likely to read Tom's book because I know Tom has a good reputation because I'm actually going to spend five minutes a day (inaudible). So reputation is important for that, but it's also important to start developing some standards around reputation of greatness, because that meta data is critical, at least going to be so much reputation data, there ought to be a way of sifting through that, and I think to Mary's point, what's interesting to think about is hey, if my reputation on Ebay is kind of portable, but who gives. I think what you'll start to see emerge is this aggregation of reputation data. So to look at an offline analog, look at your credit score, and you may hate how your credit score is compiled because you have nothing to do with it, and it's not transparent, but there will be things like that that emerge in this economy as well where my Ebay reputation is one thing, but my reputation across 15 different services says a lot about me.
MR. BURNHAM: But there are a lot of systems that don't depend on reputation, like the music system that Fred was talking about, and there are probably a number of ways that you can navigate through this without knowing anything about --
MR. JARVIS: Problem is the reputation.
MR. BURNHAM: Right. There's a number of other solutions. Dick.
MR. COSTOLO: I was just going to say that it's interesting that we like to assume that there must be some selfish motivation for participation, but something that Tom said is interesting, right, that there does seem to be anthropological propensity to cooperate or participate without anything in return. We get comments on our feedback forums where people post very detailed replies to technical questions as anonymous. No one's linking back to them. They don't -- why wouldn't you at least log in and say who you are so someone could post a follow-up question. Nothing accrues to you in return, other than you cooperating. So I think there's an interesting distinction to be drawn between creating companies that try to take advantage of reputation or build on reputation and just building companies or things or peer produced goods that just assume that people will cooperate for the purpose of cooperating.
MR. BURNHAM: Dave.
MR. D. MORGAN: I think one of the disconnects that we're talking a lot about the wall garden (inaudible), but I think that there's a big difference between the whole philosophy of peer production, and, you know, low risk potentially higher reward, but low risk low reward is fine. You can put out lots of things and if they work, that's great, and if they don't, I don't know the answer to the 164 million to one, but you can do that. But for people to have to invest cash, particularly if it's corporate cash, the $1 that's spent badly can sometimes overwhelm the 99 that's spent well. Particularly in the media world, it's the one ad that shows up on the Yahoo of the pedophile site that potentially jeopardizes a $40,000,000 Pepsi relationship with Yahoo. And so that leads to a lot of protection, because there's a difference between throwing somebody else's money you're the agent for as opposed to throwing your ideas, whether they be anonymous or not.
MR. BURNHAM: Peter.
MR. SEMMELACK: One of the things that fascinates me is how many people here have a Blackberry or ever used one? There's a lot. And if you think about that, that's a completely closed system. End to end, they make hardware. They make the software. They make everything, but the value of that is pretty good. People pay a lot of money for that sort of thing. And so where the two worlds meet is actually where -- I think we're going to talk about this a little later -- is where pure reputation based systems that are altruistic versus the profit motive where we got to please our investors. We got to make money at this. Where these two things come together I think is going to be a very interesting discussion, because Apple, again, iPod, totally closed system, but rules the world.
MR. BURNHAM: We are going to talk about that in the next session or earlier this afternoon. We're going to talk about it both from the perspective of what does it mean for young companies getting started and what does it mean for the companies that Dave is talking about that are kind of playing defense in this world trying to figure out how to handle that. There's some other people that haven't had a chance to say anything. Josh.
MR. SCHACTER: Just two comments. We're talking about reputation systems, but things like Ebay are not just a reputation, it's also an authority system. So that there's a transaction log and you can look at it. So I sort of wonder, what does it mean to take the reputation and move it somewhere else? Ebay continues to provide the authority or is it something free floating and what does that mean? How does that even work? Here's a file that says I'm really great. You should believe it. I'm not really sure.
MR. BURNHAM: Anybody else want to jump in before we break? Let's break. We're going to talk about the idea of an open data architecture, which is really picking up on what Marc was saying about standards and the gift economy and what it means to be open or closed with your data architecture. 15 minutes we'll be right back. (A break was taken.)
MR. WILSON: Since we're talking about peer production here, I don't know if many of you use the peer to peer files sharing systems, but there's a concept in a lot of those systems of contributors and leechers, and leechers are people who take content out of those systems, but never contribute. So we don't want anyone leaving here today being a leecher.
MR. BURNHAM: So the risk is that you will actually be called on if you're not raising your hand.
MR. WILSON: There's a few people, this guy over here in particular has not said anything. Anyway, Brad is going to introduce the next topic.
MR. BURNHAM: So what we want to talk about now is the idea of an open data architecture and what that means. And so over the years, we've become familiar with what an open hardware architecture is, that's the PC, and what an open software architecture is, Linux, an open network architecture, the internet, and we are beginning to try to understand what an open data architecture is, and there have obviously been examples of Wikipedia. You can go in and pull the stuff out. Del.icio.us, it's fairly easy to pull the stuff out. Interestingly, Craigslist, in the last couple days, has made some -- they've been very open to date, maybe they're not going to be open, but what I'd like to do is ask Tim to just introduce the idea of whether or not that concept even makes sense. Is there such a thing as an open data architecture, and if so, what does it look like?
MR. O'REILLY: Well, Brad just hit me up to offer some thoughts about this three minutes ago, so I don't know that I have anything terribly profound to offer, other than on a very superficial level, we've had a variety of open data architectures and they've been intrinsic to the success of the other open architectures we've been talking about. For example, in the original design of Linux, the idea of pipes and filters was based on the idea (inaudible). And programs cooperating through an open data format. It was simple. I think in the era of the web, obviously HTML open data format, and it drove a lot of the success of the web, but as we've gotten toward more complex systems, it's been pretty hard to figure out. We typically have had attempts to standardize way too much. What we do see is that the data architectures that are simpler tend to win out, and so we've seen this played out -- starting to see this played out in web services, but clearly in a lot of the systems that we think about are heavily data driven. And data in the aggregate is the source of value such that I've tended to use the phrase data is the next Intel Inside. And clearly there are companies whose lock on natural databases, it gives them enormous power, and so I've posited that we will -- there'll be a Richard Stallman among us who will start the Free Data Foundation somewhere about now, because they're going to recognize that we have a problem that somebody owns data, they're not -- again, we heard Bob talking about it already today -- somebody owns some critical piece of data, and it's not portable, it's not modifiable, it's not exportable, and we're going to need to solve that problem. I don't actually know when or how we'll come to a crisis. I do believe that there will be some sort of signal events where somebody basically takes a database, private -- we saw it in small scale with CDDB -- where they'll become peer produced data source that some company will say, "You know, we created this wonderful environment. You built this for us. Thank you very much."
MR. BURNHAM: Michael.
MR. PAREKH: The one point I would ask from a definitional point of view on open data architecture is if we were doing this 10 years ago, it would be fine if everything for the most part was in English. One of the questions I have is look at the next 10 years where over 60 percent of the content out there is going to be not in English, is what open data mechanisms are being created from a translation point of view so that you can take stuff from China or European countries, etcetera and make it malleable, fungible instantly. I know there are some translation efforts going on with Google or Yahoo, but is there an opportunity here to introduce language, translation elements from an open data piece of it that allows more people across countries and cultures to get in on the conversation.
MR. BURNHAM: So there's a couple layers here. We're talking about sort of the gist of philosophical practice of whether your data is accessible or not, and that's irrespective of the standard of the data. So, for instance, do I just allow my site to be crawled. Do I allow the site to pull data out. Then there's the question of are there efficient ways to define interchange of data so you can avoid that crawl and make it more accessible, and then finally there's the question of within that data, there's the problem of languages and formats and things like that. I'd like to try to stay for a minute with that first problem and just say who owns this? You know, who do we think should own it, and particularly of interest in a peer produced system, where the peers at least feel as if they're contributing this to something, like CDDB.
MR. O'REILLY: Or Del.icio.us, to bring it close to home.
MR. BURNHAM: Right. And then how do you -- so what should your philosophy be, and maybe we should just let Joshua tell us what the philosophy should be, and then I want to get to Mary.
MR. SCHACTER: Well, in Del.icio.us, in my belief, is that the users own the data, and this actually brings a couple interesting and technical challenges to the field. Basically the user can say, you know, they put their stuff in. They can take their data and go home. They can remove themselves. They can delete themselves from the system. How many systems actually let you delete your account typically? That's very rare I think. You know, when the user says "delete this item," we really delete the item. We don't mark it as deleted. We actually purge it. So I believe that that data's theirs, and that we have the right to use some of it in some way, but not, you know, personally, identifiably that kind of stuff. This is a tough line to walk in general.
MR. BURNHAM: So if the data is theirs, then do you believe that they have to choose whether or not that data is accessible to a crawler as an example?
MR. SCHACTER: It's interesting. Right now we don't let any crawlers in. This is mostly for, you know, load reasons, because there is a lot of pages in Del.icio.us because of all the tags and tag combinations.
MR. JARVIS: Should you?
MR. SCHACTER: Should I? Well, it's interesting, the question is -- the question is, for example, in this particular case, are the crawlers going to get value they're not going to pass back to us, right? If they certainly keep asking --
MR. O'REILLY: So you've made a decision, though, that therefore you own this aggregate data in some way with respect to other aggregate data.
MR. SCHACTER: Yes. The user owns their data and we own the aggregate.
MR. JARVIS: But the value of the whole community together is greater, the greater the community, right.
MR. SCHACTER: Indeed.
MR. JARVIS: So that if the wisdom of the (inaudible).
MR. SCHACTER: Maybe. I'm talking about the crawlers, like, for example, just the search units.
MR. JARVIS: Right.
MR. SCHACTER: I'm not quite sure what that looks like.
MR. JARVIS: It's not that the individual user owns data, the community owns the data combined. So in that morale thing, what happens?
MR. SCHACTER: Actually, I sort of refute the notion of the community. There are many communities. So that's an interesting question. I mean, a lot of it is dominated by technical. I don't have much of a choice right now, and then there are business concerns. For example, Google keeps asking for the tech data, and should I just hand that over to help them or --
MR. O'REILLY: No. You should make a deal with them.
MR. WILSON: Let's go to Mary, because I think that Mary's involved with something called attention trust, and you know, I think the model behind attention trust is that the user owns their data.
MR. COSTOLO: The user owns a copy of their data.

MS. HODDER: There's more to it that than. Dick is actually also involved and Seth Goldstein, who's not here. Seth and I think basically -- and Dick as well -- I don't know how early you were involved -- but so basically the idea is that users own their data, right, and the trust puts that forward and says that companies own the aggregate data, and a copy of the user data, and then the user owns a copy of their data as well. And just to sort of reference what Jeff and Josh were talking about in terms of what skew, what's of value. Is it the entire aggregate of the data or is it my data at Del.icio.us, plus my data at Amazon, plus my data somewhere else, and that can I make a trade for that somewhere, and if we set up an economic system where that information has value on the other side, does it then sort of change the marketplace so that users have some power to negotiate that information away from different companies because those companies won't remain competitive, and if their competitors are giving users access to their data -- I think that, I mean, one of the things that I see potentially coming out of a system like that, and I think the attention trust, which is the trust side of this, there's sort of the nonprofit that advocates for the user using the data, has to be really careful about is companies could then use that, you know, knowledge that they can get that information from users against them, right. So when my Mastercard company sends me the annual change in the terms, they say "if you don't like them, you can discontinue your service with us in 30 days," but I like all of the privileges that I get from having a credit card, so, of course, I do not refuse the service. So if Mastercard and Visa decided to say part of our news of service this year is that you have to give us all of your online click stream data, we want everything you collect out of our attention recorder, am I sort of helping the bureau in having to give that information away? So I think that there's, you know, there's the positive of that they get, and then there's the negative of can this be used against me, and can the attention trust kind of be advocate for users, but also advocate for users being to keep that data private and having users truly stay in control of that information.
MR. BURNHAM: Mike.
MR. FRUMIN: Just along those lines, the concept of, you know, different people owning whatever your data is and people abusing it, like the Mastercard example, I think that if you think about the fact that Mastercard or American Express or whoever, a lot of people, their business is having your data, and so if you found ways to make your information available to you and selectively to other people, a lot of that opportunity goes away. I mean, you wouldn't get spam about refinancing your house if there was a way for you to tell everyone who was interested in offering you a mortgage that you wanted a mortgage, but there's not really. So in terms of owning your data and abuse, the more you are in control, the more you share, frankly, the less room for abuse there is.
MR. BURNHAM: Peter.
MR. SEMMELACK: I think data ownership is an entirely different question than this.
MR. BURNHAM: Data architecture.
MR. SEMMELACK: That's full of all kinds of privacy issues, but open data architectures, for me, I feel as if like a public library is a great example, and because who owns the public library? The public library is owned by the public, and if you look at the internet, who owns that? Nobody. So isn't -- would it be true to say that the best owner of a standard is a body that is not driven by profits? ASCII is a good example. You have ASCII versus EBCDIC. ASCII obviously won that battle. You have standards bodies who are not the holder, you can argue that, but there's no profit motive per se that is driving -- you know, there's two kinds of standards, there's standards by bodies like that and there's de facto standards like Microsoft. I would argue that the standards bodies are the best owner of something such as this, because you avoid a lot of the strange behaviors that come out of the chase for dollars.
MR. BURNHAM: Tom.
MR. EVSLIN: That's the first thing I've heard here today that I totally disagree with. I think standards bodies is wrong. I've been a part of them. I've been on the outside of them. I think that they do a very good job of slowing down progress, for good and for bad reasons. I think that people who spend their time on standards bodies are being paid by somebody to do that, and that somebody has interest, and I'm not putting down those interests, but they're real. I think that if we're going to have a data standard that's as fast and flexible as we need it to be, and if we give up the idea of 100 percent solution, because if we wait for 100 percent solution, then we'll never have the 90 percent solution, that it almost has to be Wiki in the sense of both the mechanism and of it being fast. If there's a Wiki with perhaps some kind of voting, perhaps not mechanism for defining tags for various kinds of data and formats in that, from a technical point of view, we can afford to tag each piece of data with what it is because storage is so cheap. You used to be able to do that. So I haven't thought about this much, but I think we need something that as flexible as the people who participate in it and as fast as participants are willing to make it and open to all.
MR. BURNHAM: Let's get Chris and then go to Marc.
MR. ANDERSON: Talking about de facto standards versus de juris standards, de facto standards are somebody does something, grass roots. It bubbles up. The marketplace votes on it and it wins. De juris standards is what standards bodies do, and you think of a few successes, but a zillion failures. So my sense is there is no process. There is no body that can establish standards. What you have is a zillion people who are trying things and good ones quickly percolate to the top.
MR. BURNHAM: I'd like to actually see if we can move back up away from sort of the technical process of establishing a standard for the interchange of -- the technical interchange of data and get back to the question which isn't an ownership question, a data ownership question, but what are the issues around allowing yourself to be crawled? I mean, you've got Craigslist, you've got Google Maps, you've got other people who are just pushing the data out there. Does Google Maps have an open data standard or an open set of principles? Is Craigslist changing? Can we talk about that for a minute? Marc was next and then we'll try to get everybody else.
MR. PINCUS: Which is my point, and back to Tim's point I think was spot on for this conversation, I think that all our -- we don't matter in this. Whatever it is we figured out won't matter, and it's really just going to be a continuation of trial and error, and so Craigslist tried being open and crawled, and now they're saying Oodle and whoever, "you can't crawl us," and all of these job databases and other classified databases have so far chosen to not work together and share data because it's competitive. And I think what we're going to find is that people are just going to keep trying different models and out of greed and business interests, the open -- I think the open database standard will eventually win, we just don't know how. And so Hotjobs has said they're going to be kind of open or I think what they mean is they'll be one way open. They'll crawl other people and present it and not allow others to crawl. But I think that we'll see people and see some of the more enlightened or -- that's a biased way of putting it -- we'll see some of the big companies and big databases experiment earlier and some later. AOL, I think it's been painfully obvious for five, six, seven years that they're just slowly dying by being closed, and now they're making open AOL or something. So I think it would be interesting to watch, and I think that it'll be what's in the best interest of the end-user to then how is the end-user best served, and Craigslist right now is a good experimental case, because right now it doesn't look like it's really important to end-users for Craigslist to stop Oodle or others from crawling them, but maybe it will, because maybe other people in the market will open up.
MR. JARVIS: My contention is that Craigslist and Monster and Hotjobs are just way stations. They are cheaper marketplaces than newspapers were. So one version of control is a cheaper version of control. The cheaper version of control has destroyed millions of dollars of revenue money. That's not the end of the game. Somebody's got to come --
MR. BURNHAM: I got to interrupt you just because I don't like that phrasing, because I've heard that applied to Craig a lot by the newspaper industry, that Craig has destroyed value.
MR. JARVIS: Fair. I love Craig.
MR. KURTNIT: What's the term you used?

MS. FISHER: I call it creative destruction.
MR. BURNHAM: Well, if you think about value as being the most efficient delivery of a service, delivering the service more efficiently doesn't destroy value, it destroys pricing.
MR. WILSON: It's wealth transfer, not wealth destruction.
MR. JARVIS: The real point is, we're still thinking centralized. We're still thinking one dot world. So in a distributed world, what can happen? Somebody comes along and can do a micro format to say, "Here's how you put your resume up and here's how you put it on the Wiki. And you know what, you can use Google to search that." And we'll put together a buyer and seller of jobs with no friction and no value at all, except for the people who kept their money in their pockets because they didn't buy the ads anymore on Craig or on newspaper. So if you start looking at it that way, is that Craigs definition -- I'll still say destroy, because Craig is the coal mine here who says "Watch out." Someone can come along. Take the newspaper industry and say, "Bing." Guess what? Well, somebody can now do that to Craig too. So Craig is now -- there's a path that says "I'm open. I'm closed." Somebody else can come along and be more open, and if the user, the job seekers and sellers benefit more, that will win out. That will win, if it can. Some day real estate brokers will get destroyed. Someday that will happen.
MR. BURNHAM: Let me get to Joshua here because he hasn't been too vocal, and he usually is.
MR. STYLMAN: Thank you for that. I was afraid of being called on. It seems like this is a conversation about control and leverage. So initially I guess you didn't like the phrasing of that, and a prime example is look at Blockbuster and now look at Netflix, and obviously Netflix has all the control and all the leverage until the next thing comes along. And what I think will be really interesting is when you got someone like Indeed, and Indeed gets really, really big. Well, at that stage, do they get the control and leverage to say, "You know what, Monster, we're not even letting you into our audience," and ultimately, is that the next power shift?
MR. BURNHAM: There's a question about -- to come back to something Yochai said very early -- is that there may be a sustainable -- if you think about all of these, they're, in a sense, a form of peer production. Craig is able to deliver a service in 190 cities in 35 countries, you know, with about 18 employees, not because -- because he's figured out a way to use the technology really efficiently and get everybody to contribute their own content. So it is a form of peer production. So the question is, is there some sustainable evolution of this that maybe -- I would argue that Jeff isn't correct entirely, because you still have the problem of finding things. So Google has a relevancy ranking algorithm that will bubble up the appropriate jobs. Somebody is going to create some probably peer produced indexing relevancy ranking and there needs to be some centralization of that in order to be able to do that.
MR. JARVIS: Google Jobs, at some point, that won't work, you can't get all the jobs. So you figure my job is stuck in Craig now and I don't want it in Craig. I won't put it there.
MR. BURNHAM: I think there is a pressure.
MR. H. MORGAN: Whether or not there's physical good or whether it's just information. At Heartsearch, we're happy to have our database crawled deeply. We have -- Amazon has it available, because it can eventually lead to a transaction of a physical good, which is money. But if the transaction is purely informational, that's when it gets driven down to how do we get that information out there, and it's cheapest if the person who's providing and wants it out there can get it out there for free, whether it's individually peer produced or aggregated in other ways.
MR. WILSON: Scott hasn't spoken on this topic yet.
MR. HEIFERMAN: I would just add that there's, you know, I'm a huge fan of disruptive, distributed bottom up stuff, but, you know, there really is something about the potential combustion that can happen when these things are sort of cultivated by -- I'm sorry, I'm not speaking well. If Web 2.0 was in the year 2000 and not 2005 or 2006, I would argue that there wouldn't be such a great, wonderful, efficient market called Craigslist today, because you need that place where all the aggregation was happening. That if Del.icio.us was just sort of a completely distributed notion without this sort of a level of protective centralization there, then that little explosion of wonderful stuff wouldn't happen. And just, you know, lastly, I think that there is still the reality, which is 90x percent of the world has no idea of the benefits of Craigslist or Del.icio.us or any of this stuff, and it needs these really well cultivated cultivators to make them available. That, you know, my ability to find something on Craigslist would be dissipated if it was so distributed for some period.
MR. BURNHAM: Tim.
MR. O'REILLY: Couple points. First of all, your question originally was about an open data architecture, and so one of the -- which is not the same as an open data format. First of all, it seems to me there are many, many formats for many, many specialized classes of data, and the question that we have to ask ourselves is what are characteristics of data formats that make them susceptible to being used in an open data architecture. And I don't think any of us really know the answer. I mean, this really is the frontier of the industry right now. But I do think that there are a couple of lessons, and one is an open data architecture is one that is potentially syndicatable, ie. can appear in many places at once. And you know, if you think about the fundamental nature of the hyperlink, you link to many, many -- from many sites. You don't -- nobody's keeping count. I think one of the things that Katarina talks about in Flickr is that sort of she hates sort of two-way links. And that was originally hypertext theory was we need two-way links, and they thought that was appropriate. And Katarina's comment about social network, social networking is broken because it requires two-way links. She said basically just say you set a watch list. So this idea is it one way or is it two way? Actually, one way may be one of the characteristics of an open data architecture, ie. you don't need acknowledgment. There may be some other things like that. So I guess what we need to figure out is looking at the characteristics of systems that seem to have network effects around their data. They're syndicated. They're one way. They're outbound. And, you know, so, again, Google Maps versus Yahoo Maps?
MR. BURNHAM: I wanted to jump on that one-way aspect, because, for instance, one of the things that if you syndicate a piece of data, you know, do you want to have a string tied to that, at least be able to capture the meta data about how it's being used, and does that make it then two way, because if you -- is it the same to syndicate and just let it go and never have no feedback at all about where it's being used, how it's being used, or is the ultimate architecture going to have something to do with --
MR. O'REILLY: I'm making a guess here, but I'm making a guess that if you try, actually quote Blake, catch the joy as it flies, the wing or something or other destroys. Anyway, basically, you have to let -- I think you have to let whatever you do go. If you try to kind of say, "Oh, I have the instrument so I captured it," you're not going to get it.
MR. BURNHAM: That's a pretty significant departure from just saying, "Okay, I'm not going to let it go at all. You can't crawl me. I'm not going to syndicate my data. It's in a box," to "I'm going to syndicate it, but I'd like to be able to at least capture the value that is associated with how people use it," to saying "I'm just going to let it go and I have no idea what's going to happen."
MR. O'REILLY: I'm going to make a leap here, and I don't know if it's actually true, if I look at the web and what happened, it wasn't until the web got to a certain scale that it was possible to discover how to recapture the aggregation. So I think that something happens when you syndicate outwards. When you let things go, you create network effects. If you try to optimize prematurely so that you capture the value and you build the aggregate system too early, you limit the network growth. And that's why often the actual value is captured by a second party, right, because you cannot build it into the system at the outset. You have to figure out how to, you know, how it's actually implicit somewhere in the system. Somebody discovers later where that implicit way --
MR. ANDERSON: Google was formed as an experiment to try to -- it was linking. It was called back rub, and it was the very nature -- in fact, the one-way links created an opportunity for a third party to out of band answer the questions "Who's linking back?" This was the experiment. In doing that, it then created a search engine which was a completely unexpected outcome, which added value in a whole new way, simply because they hadn't predetermined that back turn.
MR. O'REILLY: So the value actually needs to be out of band for the open data system I think is maybe the principle that we were trying to get at here.
MR. WILSON: Let's hear from Bruce, because we haven't heard from Bruce yet. Then we'll go to Josh.
MR. SPECTOR: Thanks. I think that a lot of what we've been talking about, we're really kind of talking around here, and part of my question here goes to I think the afternoon's session, but we're talking really around it now, and that is, what are the, you know, what are the models, what are the economic monetization models that are driving all of the conversation that we're doing right now. And we really haven't been talking much about it. It seems to me that the notion of open data and commoditized data is a done deal. I mean, the web has essentially established the ability to commoditize lots and lots of data in a variety of ways, and many of us sitting at this table have been, you know, struggling over how to earn value, whether it's social value or whether it's economic value, out of the proposition that exists. We're in a commoditized data universe right now, and Michael Bloomberg made millions of dollars by selling proprietary data.
MR. O'REILLY: By selling commoditized data.
MR. SPECTOR: Right. But he found his way, but it was his proprietary version of that commoditized data. So we're now talking about a world in which all the data may be free and open, and I believe that, in fact, it will continue to develop along those lines, then what happens then? Literally, what happens then? If the data is commoditized, whether it's through Del.icio.us, whether it's through web crawling, whether it's through an open source database, then what? I think that's really the question I'd like to --
MR. BURNHAM: Well, let's --
MR. SPECTOR: -- put on the table.
MR. BURNHAM: Let's save the question about the economic models that underlay that, but I think that one of the interesting points that Tim is making is that, you know, it may be, you know, that somebody creates -- sets the table for others. You know, they're going to create an environment in which the data is going to get out there and then others are going to figure out how to do the re-aggregation after the fact. After the models begin to emerge. So maybe we almost can't answer that, but at least we can talk about this afternoon, how one might approach tentatively a couple of those opportunities. Joshua, you want to say something and then we'll come back. Ah, Seth. Let's do Joshua and then Seth.
MR. SCHACTER: It's interesting in that we were previously talking about peer production, which is largely a system where we're able to diffuse the value of the system such that more people benefit, right. Aggregates, aggregators, which we're doing, which is the opposite of a syndication, right? If someone is syndicating someone, someone else is going to be aggregating those things. So I wanted to call bullshit on Jeff Jarvis before who said that when everyone goes off and its micro formats and, you know, everyone transacts for free, but at some point those things have to be found, right. We're stumbling towards markets of information, right, and you say market, you think, you know, is that the (inaudible) or whatever, but that's really the place where everyone gets together, the market square or whatever, right. And the people have to come together to transact. There's no -- the fact that it's distributed and makes a peer produced system or whatever adds a lot of value on the whole, but also, the fact that everything comes together and the liquidity is increased, because we're not talking content in general, right. We were talking about open data where a piece of data is just sort of given away and someone else gets value from that, but these are job listings, right, which is not actually the entire piece of data, right, and that's not the value of the thing. If the pointer's the value of the thing, which is the job, at some point, people have to go somewhere to transact, right. So it's not an "either or" thing. There is tension in the system such that people want to be open, but they also want to gather value from it. So there's the spreading and then the collapsing. It has to find some happy medium, right, and that these -- both of these properties are going to be places that people are going to be able to compete on. Being more open is going to be a place to take a point of competition. Different sorts of ways of aggregating and listing data are all going to be ways that people figure out -- so I think it's going to be not just one person generating and one person aggregating, and the question is do we fall into some system. I think it's going to be continually going through cycles and moving around.
MR. JARVIS: I agree, but I think that's where you get down to you or Jimbo Wales or somebody who is -- what Scott was saying earlier -- the benefactor of that or Craig, as long as they are the benefactor. So it's about trust.
MR. SCHACTER: I agree.
MR. JARVIS: It's like you trust that.
MR. SCHACTER: Right. At some point there's a lot of value. I think that an incredibly distributed system with no central point of authority are very elegant from a bunch of conceptual reasons, but centralized systems also add different values, and that at some point, what we're assembling toward is that the ownership and publication of the data is distributed, but the central point of aggregation of finding this attention router, you type in what you want and it goes to where you want. It gives you something that was at the top of -- some way of finding these things, because we're now transacting in millions of different commodities instead of hundreds or dozens. So the value shifts. So the question is --
MR. O'REILLY: What is the access of aggregation?
MS. GREEN: But you could crawl, you know, Del.icio.us, and we don't want that to happen, because that's why Del.icio.us was created, because Google's flat.
MR. PINCUS: We're all kind of real estate developers. We're living in a world where all the land will be free, and there will be some new kind of land created and maybe whole new building levels that you can build on top of the land, and Del.icio.us might become another kind of land created that buildings can be built on top of, and then the buildings that are going to have value -- the land will have more value because better buildings are built on it, and one way or other Google came and took free land and then stuck a building on it that everybody liked, and now, all of a sudden, through people -- so many people going through their building, there can be new buildings built on top of Google.
MR. BURNHAM: Let's get Seth in here, if we can.

MS. GREEN: Actually, where we're going is very interesting, because we would have a lot of stuff being created by people and (inaudible).

MS. HODDER: So, actually, to tie in what Tim said at the very beginning of this session about open source and making free data, making a free data foundation, I think it's a really interesting idea, because then the question is what's the GPL for free data, right, because the issue is it's not to me. It's not -- I don't see it as --
MR. BURNHAM: What's GPL? Give me a public license?

MS. HODDER: Right. So the idea is, you know, when I use the GPL for stuff that I'm working on, I create --
MR. HEIFERMAN: Define GPL.
MR. O'REILLY: The specifics of that license is that it says if I give this to you, you can only -- basically you can only incorporate into your products if you give it away under the same terms under which you've received it.
MR. PINCUS: Give away under the GPL for data.
MR. O'REILLY: The Berkeley style license, which is I give it to you. You can do whatever you want with it, as long as you acknowledge that I gave it to you.
MR. PINCUS: Why would I let Oodle crawl Craigslist when Oodle won't let anyone crawl Oodle?

MS. HODDER: But there's a couple of properties and data and it's covered under copyright and not under software licensing terms in terms of the GPL. But the point that I was going to make, I mean, there is a way to do that. Craigslist could make a deal where they would say "Well, you know, our data is available for free noncommercial use. But when you go to use it commercially, you have to make a deal with us."
MR. PINCUS: Or you have to make it also public to other people.

MS. HODDER: But that's part of the deal that you make. So the thing that I'm suggesting is that that actually could work here in terms of "here" being the US, because we have a particular structure of laws and what not, under copyright, or some kind of thing where you say for noncommercial use, that people can take things with APIs and they can play around with them. But at the point where you're ready to do a commercial service, you have to come and make a deal. So to me, the difference occurs not when Oodle wants to take Craigslist data right not here in the US, but it's will other places observe, you know, those kinds of I don't know.
MR. BURNHAM: What happens in China is the question.

MS. HODDER: Yes. Exactly.
MR. BURNHAM: Another interesting question is could you actually, in some way, enforce it in an API? Unfortunately, it heads you sort of towards some kind of weird DRM system, that is the GPL DRM system, which says here is the data. As long as it's noncommercial, but maybe you can't. Or maybe there is some way to do it within the API. Seth, you've been trying to get in here.
MR. GOLDSTEIN: I want to try to connect some of the data architecture conversation with what we've been trying to do with attention trust, and I think what I learned from Josh about a year ago, what is Del.icio.us? Now it's going to be or it is certain mass market Del.icio.us is social bookmarks or whatever it might be, but I remember asking Josh, and Del.icio.us is, according to Josh, and correct me if I'm wrong, crystalized attention, and that was what he said, and it sort of stuck with me throughout this past year as I tried to think through all sort of Web 1, Web 2 services, and if you think about that, consumers create content with their attention. So where they pay attention to passively in terms of their click stream or various proxies or actively in terms of tagging or searching for things, they're creating content. Some of that content might be explicit content that you see and some of it may be hidden in databases, but it's still there. One of the things that has been helpful for me related to this is China solved for influence, right. Because influence is really kind of Seth's working on it with lenses. So we defined influence as influence equals the amount of attention you get, divided by the amount of attention you give. So I would argue that -- so you can think about all the different ways that, you know, a service gets attention from people and all the different ways it might provide it out, Del.icio.us is extremely influential, because people can add tags in so many different ways and it controls the attention that goes out of it in terms of it doesn't have access. Access is a huge I guess a deflationary force in the attention, maybe because people that put add tags on their pages just basically -- as much attention as they're getting, they're just sending it out the other way, and Josh and Del.icio.us have been really good at maximizing the influence of the Del.icio.us. So as you start to think about -- I think a lot about what Dick is doing with the feed burner around awareness and the whole web analytic community and the behavioral targeting community is about trying to help publishers understand their awareness, right. And help publishers keep track of subscribers and try to help web sites figure out who's going to web sites. And that's a pretty big industry. And it's getting bigger. And I think to this point, there is no industry for consumers to understand the value of their attention.
MR. BURNHAM: Let's bring Dick in here. What you're basically saying is that when data becomes flat, commoditized, freely available, then the value moves to attention, and then ultimately to influence, which I think is a really interesting model.
MR. GOLDSTEIN: It needs to get priced.
MR. BURNHAM: And that's different than -- and that may make the whole discussion of an open data architecture irrelevant, because it may mean that, you know, the real conversation should be at the next level.
MR. COSTOLO: Well, what I was going to say is so I don't think it makes it irrelevant. I think the thing that it implies for me is that an open data architecture will be one that is purely an API. I think you'll see architectures emerge that have no destination site, they're just APIs into the attention data or content out there, and enormous value of some sort will accrue to the developers and creators of that architecture that have made this API generally available as content flows through that API, and it will be interesting to see what sorts of economics and organizations emerge around that. But I think that what the implication, the bottom line implication is that an open data architecture will be one that is purely API based and not destination based.
MR. H. MORGAN: Semantic Web.
MR. EVSLIN: We're using open in two senses, obviously. One that the architecture's open and one the data's open and aggregated. On the second point, because you were talking about the first, I think we've seen this game before, and we've seen the movie, and it's a cycle, and it'll probably go on forever. If you're the first mover in anything that has a network effect and you create a large enough network, which you partly do by being open for a while, or more open than anyone else, you accrue a great value in that network, and there's no value to you in letting somebody be your N+1 competitor. If you say, "Well, you can crawl me," somebody who has only 10 percent of content you have, but it's disjoint, now has 110 percent of the content that you have, and you've been able to N+1. We're seeing that with Skype. Skype has no reason to be open. Skype has no reason to be open. They have too many subscribers to want to do that until they're trying to add value to Ebay, but then what you have, once somebody gets into this commanding position, and it happens over and over again, is there are two ways to get out of that position, and it's almost never voluntary, or it is never voluntary. One is that they may be the dominant player, but they may only have minority of the market share, and everybody from the second to last has a reason to want to be open and to share, because it's the only way to break the position of the dominant player. And so, if all of us little guys together have 60 percent and the other guy has 40 percent, we have every incentive to be open. The only one who doesn't is the hoarder at the top. So that's one they get broken. Sometimes they don't get broken. Take AOL early on. We had screen names. We had e-mail addresses. Things that were very sticky and had network effect that eventually it's like trees don't grow to the sky, and the wall garden, even if it's dominant in its area, cuts itself off from innovation, because it is a wall garden, and innovation happens on the other platforms, and then soon there's a service of greater value, which N+1 is value rather than content.
MR. PINCUS: It could just die, Tom, because of economics.
MR. EVSLIN: And it dies. But I think we'll see that cycle forever and ever.
MR. WILSON: Let's get Seth Godin in here. We haven't heard from him today.
MR. GODIN: I want to first thank you. I thought that was very insightful. I wanted to just jump in with three Fs, leaving out your F from earlier in the day. The first F is about Friction. If we think about Malcolm Gladwell's book The Tipping Point, how many people in America at this table paid money to read the book? A big number. How many read it in The New Yorker? A smaller number. How many read it for free on gladwell.com? Almost none of us. And if you look at the value accrued to Malcolm, he made no money if you read it on gladwell.com. He made no incremental money if you read it in the New Yorker, and he made a million dollars if you read it in book form. So friction is interesting, because consumers like friction and creators of content like friction, and I don't see any time soon that a desire for friction is going to disappear from most people's -- not the Berkeley community --
MR. BURNHAM: Why do consumers like friction?
MR. GODIN: People are voting with their dollar. They'd rather read the best selling book that everyone else is reading than seek out, figure out a way to look at all the different choices, which leads to the second one, which is Focus. The TV Guide, when it sold at its peak, was worth more than the TV Guides that were selling at the same time, because the directory feature, the directory function of pointing -- I mean, one of the things that open data really is is a billion TV networks. A billion book publishing houses. A billion magazines. And anyone who wants to write now can write. So the TV Guide function becomes really important. And then the third is really a U, is Unfair, not Fair, which is if people create stuff, want more than their fair share. And the feedback that I'm getting more than any other feedback with regard to data is people want more than their fair share of traffic. More than their fair share of people looking at what they're creating. So I think when we put those three things together, just to piggyback on Tom, it's nothing that new. What's happening here is that human nature isn't going to go away, and that this desire to have friction, have focus, and get more than their competition provides a platform for people who want to make a living in the middle, because it's just -- I don't see how if it's completely open and dynamic with no friction and no focus, that people are going to be satisfied. The average person.
MR. D. MORGAN: I would say, following that point, you can look at Walmart as a great example. When it was the number 10 retailer in the country, it had an open database. It participated in IRI. At the point it started to accelerate a lot and became number two or number three like six or seven years ago, they closed down the database, because there's more predictability, and you can grow faster if you control the data, because you can do that with engine if it's growing and its pricing, and that's what's happening here. You play by the open rules to get moving, but as the markets mature, competitive issues become more important than exploiting just the general open opportunity, and you use the data to control that growth better.
MR. BURNHAM: Let me just follow up on that, do you think that cycles is -- and to Tom Evslin -- do you think that cycle's going to happen faster today in this next --
MR. EVSLIN: Yes.
MR. BURNHAM: Why would it happen faster?
MR. EVSLIN: Because everything happens faster. There's less friction and more force. If you asked me that about anything, I'd say yes.
MR. BURNHAM: Let's go to Jeff and then Tim and come back to Peter.
MR. JARVIS: Rephrase what you rephrased before. If you create content, content is data, data is now free, so you can never own enough content. Marc will shoot me across the table, but that's fine, the value of content increases. What increases, the value of trust. So that's what you're saying. When you say it's from data to attention to authority, what you're talking about is trust. What you're talking about in terms of Del.icio.us and what Heather wants is that I choose to swim in the pond of Del.icio.us and I trust that pond and I trust Josh to man that pond, but I also want the value of being in the ocean once in a awhile. And so the content isn't what's valuable. It's the trust and relationship that's valuable, and that to me, in a post scarcity -- what the internet does is it takes away scarcity in terms of both content and distribution, and it changes the value essentially to trust. So the friction is still there. The friction isn't "I own the content and you don't." The friction is, "You're going to keep good content? Okay, then let's talk." It's a different friction, a different value, but that's essentially what we're going towards. So we forget that -- if you're a magazine, your value is that we picked good stuff and you trust us to do that, and the readers will say probably like you, maybe they have value (inaudible). So relationships and trust becomes a new structure.
MR. BURNHAM: Yes. Tim.
MR. JARVIS: I also want to hear Umair on that, too.
MR. BURNHAM: Go ahead, Umair. We haven't heard from you.
MR. HAQUE: Yes. I think fundamentally what -- the way that I approach this is through the lens of simple economics, and what's happening here is we think of a --
MR. WILSON: Speak up.
MR. HAQUE: If we think of a simple value chain with data in the middle and stuff on either side, what's happening is that the middle of this value chain is exploding. Data is commoditizing, as Bruce said. So what happens? Value shifts to the edges essentially. This is not a new concept. Herbert Simon said it in 1971, which is that "What does an abundance of information create?" A scarcity of attention basically, right? So when we talk about data, if we're thinking in e-com terms, what we need to do is understand how data becomes information, and information to an economist means information about people's expectations and their preferences and their utility. That I think is really the heart of this argument, and when Jeff says that relationships and trust are important, certainly they are. Why are they so important? I think the reason is because they give us a very efficient way to reveal other people's expectations and preferences very cheaply, and that allows us to allocate attention fairly efficiently at a cheap price.
MR. BURNHAM: Tim.
MR. O'REILLY: I wanted to pick up on what Tom said. First of all, I really agree that there's a sign wave through the economics of all of these things where we have this aggregation and the like, and I want to just suggest a specific mechanism by which some of this will play out, and I reason a lot by analogy and I've looked a lot at what's happening in software, in particular with Microsoft, and if you think about how developers made their deal with the devil so to speak, it was around the solution to the device driver problem. So what happened was that everybody had to write their own device drivers. Every software company had to deal with many, many problems. Every software developer had to deal with all the possible devices, right, and write the drivers. And basically Microsoft offered a deal. They said, "Wait a minute. We have this great idea. We will deal with all the developers and we'll give you a uniform API." And I think we're in very much of the same situation right now where we have a whole lot of data format, some of them closed, some of them open. And but that doesn't really matter, because what's going to happen is somebody's going to basically figure out that the way to do the interchange between all these data formats, and the aggregation of all these data formats is for there to be some kind of aggregation point where somebody says, "I'll give you a uniform interface." That's what people like Grand Central have tried to do. I don't know who will do it or how it will come about, but I think it's going to be through the transformation of data types. And, again, we see it maybe with Google Maps in that here they closed data providers, Navteq, who meets up with Google, who's just a licensee, but then because of the way Google implements, people say, "Wow, I can mix this with other data types," and it's at the transformation point that the place where multiple data types connect that there is the opportunity for a new sort of economic leverage. So Google, you know, I was told a story of how Navteq was the Intel Inside of all the mapping services. That was until Google Maps came along, and it was by upping the anti to become a conversion point and an aggregation point of multiple data types of data that they became the new power broker.
MR. BURNHAM: It's a powerful thought, because to this point, I've been thinking that data really behaves differently than code in an API, and that data flows much more easily between systems than code does, and what you're pointing out is that -- and for that reason, I was thinking that the Microsoft opportunity doesn't exist in an open data world, but what you're pointing out is there are still some wrappers around that data and some ways of presenting the data and making it available that are unique.
MR. O'REILLY: I still believe that there is a choice ahead of us, and the choice is because of in solving the many, many problems, there really are two solutions that have been proven. And one is I refer to the Microsoft solution as the one ring to rule them all model, and the UNIX and internet solution is the small pieces loosely joined model, and I don't know which of those people will buy into. I think there really is a race. And I think if you value open data, you want to figure out the small pieces loosely joining model before someone figures out the one ring to rule them all model. And on the other hand, if you're the next Bill Gates, you want to figure out the one ring to rule them all --
MR. BURNHAM: You want to find that ring. Let's get Tom in here. We haven't heard from him, and then Peter.
MR. COHEN: I think in listening to the conversation and going back to what Marc said, I think the real estate analogy is a good analogy, because the developer chooses the architecture based on what he perceives to be the need of the audience that's going to go into the development of that construct. The open or closed has to do with how he markets the development. So there are many cases cyclically which we've seen, the AOL example, etcetera, where as a marketing tool, if the end goal of the architect -- of the developer is to make money, and that is what skews the equation, then he might choose to have an open architecture as part of his initial marketing goal, and as for the trust factor, he might choose to give away five houses to influencers so you trust the neighborhood, but then once he has you in his grasp, he closes the architecture. And so I think it really is dictated by what the needs and the goals of the developer are, if end goal is to make money. If it's not to make money, it's a different equation.
MR. HERSHBERG: I was going to make a completely different point, but I guess I'll just follow up on that instead. I think that's a really fair point and I think the real estate analogy is an interesting one, and in thinking about the Craigslist model for instance, I may choose to participate in Craigslist because I like what it stands for and I like everything it represents, and so for that reason I give them my -- and I contribute to the community. Then the question becomes -- and again, going back to this real estate question, and then you're a specific development, do I not want my development put on an 18 wheeler, driven down the LIE and dropped off some place else where I may have not decided to participate in the first place. And the question becomes do I in fact want to end up in a place without my consent, which goes back to this larger question of who owns the data. But I think that really ties into the whole economic question, and Jeff made the point earlier that Craigslist did something to the newspapers that now something like (inaudible). And I think there's some distinction to be made there, because Craigslist didn't take anything from the newspapers. They created an entirely new data. So, at some point, I think you have to ask the question when do the -- I mean, how much intermediation is there among the intermediators, and at what point does it stop, and when can you say all right, Craigslist created this, somebody else is now taking this, and Craigslist should recognize some benefits from it.
MR. BURNHAM: What you're picking up on there, Peter, is I think related to the question of how small pieces loosely joined or one ring that rules them all plays into this sign wave of the transition into this open data architecture. And what I'd like to do here is it is 12:30 now and we're going to break for lunch, but before we break for lunch, I'd love to just bring Yochai in for a second, and because the question I'd like to put to you is, you know, do you think that either the small pieces loosely joined or the one ring to rule them all, is either one of those inevitable, and/or do you think it's possible to create something that defies this sign wave, a platform of peer production in an open data architecture that's sustainable and that has sort of a consensual value proposition in that trust or navigation space that keeps it going?
MR. BENKLER: I'll partly answer that, but partly also characterize a little bit of what I think I'm hearing. You characterized the problem as one of open data architecture, but I think most of the conversation has been around that question as one instance of the broader question of what is the interface between commercial enterprises that are collecting -- that are getting financing and building with the expectation of extracting commercial value, but from a phenomenon that is very much based on peer production. So in the context of data, it strikes me from this conversation that the fact about the world is that lots and lots and lots of value is being created by lots and lots of people, and captured locally by them in the form of content creation and content use without a transaction. And this creates a problem of management, filtering, relevance, accreditation, which is something that Del.icio.us allows people to do, and that function itself gets done not through authority, but through distributed peer production of claims of relevance and claims of things that are like tagging, this is like this, and there are several dozen people that think this is like this, so that becomes value. And then just take the example of crawling Del.icio.us or not crawling Del.icio.us. There is -- I'll take one step back. So the question for the commercial developer of a platform is how do I get the most people producing the most value that I'm providing -- in this case, relevance and accreditation -- while at the same time retaining enough control to be able to extract value to justify my investment, my investor's investment, and this goes to what Tim said in the last session about dialing up and dialing down the level of control of how much do you open things up in order to bring people in and how much do you close it down in order to collect value. What you're describing in the question of do you allow crawling and the constraints on that is if I don't allow crawling, the thing that I have is my group of users now, which somebody else can't then use to get their N+1, as Tom described, and I'm growing it at a certain rate, and then the question was raised well, are you growing it quickly enough or will you grow it faster if you open, but if you grow it faster if you open, are you going to end up loosing that advantage and somebody else is going to come in as a second generation and actually find your wonderful platform. "Thanks for creating this. I now found this extra level." And then when Jeff pushes back and says, "No, it's not the control over the relevance judgement of the people who are inside that are giving you your value, it's the fact that people will associate that value with Del.icio.us." So even though someone else -- lots of other places -- and now it's not money, it's reputation -- lots of other places supposedly create the N+1, and yet that doesn't happen. And instead, what you get is more people seeing Wikipedia is valuable and coming back. So the question in a sense on the table is whether given that you accept the source of insight and rapid information growth is radically distributed interfaces that give a lot of value, all of the value essentially to the individual participants, and then that creates a problem with creating control, is the control then having better technology to sift through it? Is it being the first mover and being understood as the place where you go? Does that mean, for example, that you allow crawling, but you require reference back, so that essentially each one of these users then becomes a replicator of your end, whatever else they have, but then everybody else knows that your end is the place to actually go and that's the place -- but that seems to me to be the interface. The question of the interface is what is the way in which you build your platform most quickly, and I think the answer there is opening it, and do you have a mechanism of capturing the reputation in the marketplace you contribute rather than closing up and treating your currently, relatively slow growing base of users and essentially co-creators of the value of the relevance that you're creating as the thing that's preventing competition. And if it's all true -- sorry for going on so long -- if it's all true that you've got hundreds of millions of producers of content and the way to solve the problem with what's relevant is to get millions of producers, then it's the many distributed loosely company that will grow faster than the one ring, and so that is the way and that forces you to actually solve that problem.
MR. BURNHAM: Boy, why don't we just break for lunch. Josh, we'll pick you up --
MR. SCHACTER: That was sort of directed toward me, so I would like to respond. The interesting question here is an observation that we're providing markets. Del.icio.us is a kind of information market and that market itself has some level of trust to the users which we're talking about and I agree, I think that's a good analysis. I think the issue on the side, however, is that markets compete and that the people who -- if the crawlers were just crawling and providing that stuff, I mean, I'm well aware that I would bring much more attention to the system, because these are big directors of attention, but at the same time, they also are going to compete, and that their end goal is to reduce the size of this market that I have. So it's not just a simple as that analysis in that they are actually not going to just do that. That is the complication. If they were just doing that, then I would be okay with it. That's the issue. Does that make sense?
MR. BURNHAM: Well, they are going to -- markets do compete. If somebody creates an N+1 market, it will compete. What Yochai I think is suggesting is the possibility that there is a frame work within which you can allow them to crawl that identifies you as the source of the value and that what you're building is reputational value that, you know, that may be something that's leverageable broadly.
MR. WILSON: Well, we're going to get to business models after lunch, so I think we're going to try to figure out where the rubber meets the road in all of this after we eat. (A break was taken.)
MR. WILSON: So we have lost a few people. Clay Shirky couldn't make it, so we have a couple of spots around the table that I would encourage the people sitting in the back to take, and I'll identify which ones those are.
MR. BURNHAM: Let's be more ruthless here just so people don't have to feel uncomfortable about it. Peter go sit in that place that says Clay Shirky.
MR. WILSON: Okay. I think the tension that I felt in the room this morning was everybody trying to stay on topic on these sort of more academic I'd say things. We were talking about peer production and open data which are the foundation for the conversation this afternoon, which is, can we make money in this world, and I think a lot of the entrepreneurs and people who run businesses sitting around the table I think found the conversation interesting, but were kind of eager to talk about how do I put money in my pocket or maybe my investor's pockets. So, anyway, that's what we're going to get to this afternoon. And I'm going to turn it over to Brad to get more specific.
MR. BURNHAM: To see if I can start that fire? Well, you know, we have touched on a number of ideas about how one would begin to think about making money, and, actually, maybe I'll just pick on Rikki, because we had an interesting conversation at lunch, and Rikki Tahta has been working in the financial markets for a long time, and he makes the point that that's an open data architecture. That the pricing is actually very broadly distributed, and there have been a number of efforts to create value and sometimes proprietary value around that, and some successful ones. So maybe you can talk for a minute about that.
MR. TAHTA: Just a quick analogy from the financial services industry of data within that. The last 20 years they've been struggling or fighting for open databases, and now, the vast majority of financial markets data is pretty much peer produced or commoditized data, produced centrally and distributed, and yet there are still I can think of five multi-billion dollar companies that make money out of data, run businesses out of data adding value to the data. Now those are traditional aggregation models, which can be cleaning the data, normalizing data, but there's also been a history of financial markets some quite interesting fuse as to where the value of adding to data comes from. In 1987, Thomson's first call of business demanded from investment banks, research organizations the exclusive right to commingle their information. So they took data that was being distributed freely, but they captured value around effectively the commingling of that in a central source. In 1990, disclosure -- this is pre-Edgar -- took data out of financial statements and copyrighted the organization of the data. Not just the meta data, but literally how the column sat within your spreadsheet exactly where they think that value adds to the end-user came from. And what excites me this morning, I found I was listening to a lot of ideas that I had seen in the financial markets occur, but one of the exciting things that I feel being touched on, but my question now is is the peer production somehow and this peer network somehow create value, other than what a traditional aggregator does in cleaning, normalizing, commingling and copyrighting some organization. So is there something where -- I think Scott kind of put it -- where one piece of data -- I produce one piece of data, the guy next to me produces one piece of data, and somehow between those two pieces of data, there's a third piece of value that comes out. That's what strikes me as the holy grail.
MR. H. MORGAN: Certainly Bloomberg did that with analytics. They did it by adding lots of analytics that people didn't have easy access to.
MR. TAHTA: But I view that still as part of the aggregator model. I got all this data. I'm putting it in one place. I'm organizing it. I'm providing some centralized value to it. It's not the producers themselves.
MR. H. MORGAN: But the meta data provides that.
MR. BURNHAM: Jeff.
MR. JARVIS: Example on Del.icio.us is person A tag, person B tag, I have a content, but together, you created more value because you've created (inaudible). Where, in turn, I get the value in other ways. The consumer gets to buy more stuff, but I as the owner of that content get to target better (inaudible). So there's new volume created, or even out of Web 2.1 (inaudible). If I take all the tags and the content associated with that here and here and then analyze, then I can renew this content and say what (inaudible). So there's an exponential growth in the value of it, and it wasn't just the old fax machine. Metcalf's Law, and I won't get into it because I'll get killed in two seconds, but trust did have an impact, because there was a trust issue. There was, in fact, wisdom. It was more than just keying it. It was truly adding intelligence to it.
MR. PINCUS: What I add to that is I think what I heard this morning and the pattern that I think we've seen as all these markets grow is whether it's the financial markets or these new kinds of markets is there's some enormous fixed cost initially to making a market where there was none. Whether all fragmented. Whether it's online dating or the web or ads or whatever, and so someone goes out and creates a market, and for a long time, they get paid a lot of money just because they created that market, but eventually, it gets to the point where anyone can bring that same group together, those commodities, and someone else goes and does it cheaper, openly or whatever, but the one factor that I think -- the formula that runs through all this from its starting close to ending open is consumers are driven by time per lead. So consumers, whether they're going to Craigslist or Hotjobs or Match, they are always going to be viciously following the least time per lead, and Del.icio.us, for a lot of people now, is reducing their time per good lead versus someone else. You have to use a certain community now. Advertisers are driven by dollars per lead, so they're going to viciously go where they can spend the least dollars per good lead they get, and I think the world that we're all entering into is the one where community and identity and trust start to enable new, better models for time per lead and dollars per lead, and I think that it's an inevitable model where the Match.coms and the Hotjobs and the Monsters, they're all going to -- and Ebay will all be put out of business just because of these vicious market phenomena, unless they go and play on the same curve. They're off curve. So Ebay, by making more and more money every quarter, is eventually off the curve, because someone can eventually find -- the advertiser can find a buyer for their good for less money somewhere else.
MR. BURNHAM: So I think the time per lead and the dollars per lead is a very good way to think about the potential for hyper deflation in a flat, connected, you know, information economy. But what does that mean in terms of the ability -- I mean, you almost think about, you know, enterprise value as being associated with trying to create some disruption in that flat economy, some pyramid or some peak in that economy. Is it possible for us to do that?
MR. D. MORGAN: This came up from John at lunch, but if you take a flat, if you commoditize everything, then the brand becomes the differential. Sort of the peanut butter principle where you go into the grocery store, you have six apparently similar, the same chunky peanut butters and all you really have to differentiate them is the brand. Well, what makes the one jump off that the others didn't? And maybe there are some mixed ones that start getting better, but it takes a long time until they move to the center of your attention and can buy the shelf space and all the other things that go with it. So I think that's one of the big things why the wall gardens have a lot of defensibility. Arguably, reason AOL is flat is not just because it didn't open up its content, it's probably because that wasn't the only thing. Maybe that wasn't just -- not just opening up any database, it didn't open up the most important database to open up at the right time, and Ebay figured that out.
MR. WILSON: I wonder if let's take Craigslist for an example, they're right here at the top, Craigslist makes money, could they make more. Today, the Craigslist model is you list for free, except if you're listing for employment jobs and real estate in certain markets; is that correct?
MR. BURNHAM: It's really employment in three markets and they're just beginning to experiment with real estate.
MR. PINCUS: Commercial real estate in New York.
MR. WILSON: And so they really haven't monetized that business very much.
MR. BURNHAM: Well, this is the point that Tim made earlier this morning when he said that many of these businesses, the relationship -- the ratio between the value provided to the consumer and the amount of monetization taking place is different in these businesses.
MR. WILSON: So I want to propose an idea of how Craigslist can become more open and potentially make more money at the same time. I'm curious of what people think of this idea. So if Craig recognizes that eventually these listings are going to get distributed, that whether it's Oodle or Indeed or whoever is going to crawl them and they're going to get outed, and it's too bad Matt Turck's not here, he had to leave, Matt started this service called Word of Blog, which I like, but if Craig simply said okay, I'm going to be the place that people come to create the listing, and on top of that, I'm going to let them provide a rule that stays with that listing, and the rule would be if this listing gets clicked on, I will pay you 50 cents, or whatever that rule needs to be, and then the listings get everywhere, but then any time anybody interacts with that listing, there's a business rule that delivers value, and because Craig was the place that that rule was created, he might also be able to be the place where that rule gets sort of transacts and the value gets distributed, and I wonder if that's where these large brands need to go.
MR. HEIFERMAN: Isn't that AdWords?
MR. WILSON: That is what AdWords is. That's exactly what AdWords is. AdWords to me is the defining model of what an open database web service -- the only problem is that AdWords won't let Marc or Dick take their feed and run it wherever they want. And that's why everybody else in that business is probably going to go and do that to them.
MR. COSTOLO: Right. So I think the definitive answer to your question is yes, Craigslist can make more money doing that, and that gets to -- and I may be putting words in Tim's mouth -- what Tim was talking about this morning in creating just an architecture for providing transformations and APIs around transformations is a much more compelling way of outflanking your competitor, because once Craigslist does that, there becomes no way to out aggregate them. They've just already said anything can aggregate me. I'm just going to provide some APIs or rules for how to, what happens to the concept when it gets aggregated. So it becomes much harder to outflank them when they do that. And they then also defined a market for how they make money.
MR. PAREKH: And pushed out the walls to their garden.
MR. COSTOLO: Right.
MR. O'REILLY: I'm not sure I agree. Here's the issue I see. I think you ultimately, in the long wrong, you get money from doing things that are hard, right, and if you think about the network effects, and that's, you know, that's wonderful. You build this nice, you know, you can get a great network effect doing something that's easy, but to extract money, you have to do something that's hard. And so, for example, take Google AdWords, it's like when you talk about a business rule, the question is, well, there's a lot of in that -- you know, what is that business rule? Well, the business rule of AdWords is I will actually find a relevant advertisement for you, and that's hard, right. It's not just something that they just kind of put a rule that "Thou shalt get thy ads through Google." They're good ads, and it's not just that they happen to aggregate the advertisers to do this sort of self-service opportunity, they've also figured out somehow to track and they've done a lot of stuff that really works. And I guess what I would say is I don't think there's any magic that's going to just -- somebody can just sort of syndicate it out. To me there's network effects in building a big aggregate, and you can definitely -- there's sort of a set of rules for how you can set something like that in motion. As Flickrs demonstrated. As Del.icio.us has demonstrated. And that's I think step one. The web demonstrated it. (Inaudible) Tim figured out the simple things that had network scaling effects. So I think for us, the question as investors has got to be what is hard about what, you know, someone is doing.
MR. BURNHAM: The only thing that doesn't ring through there --
MS. GREEN: You don't think the graphic interface is kind of hard?
MR. O'REILLY: The graphic interface for what? (Inaudible). He didn't actually invent the graphic interface. He made quite a lot of money.
MS. GREEN: But wasn't he actually --
MR. O'REILLY: It is debatable.
MS. GREEN: When he released it as open, right, Tim?
MR. O'REILLY: Tim did. Right.
MS. GREEN: They did with it what they wanted.
MR. O'REILLY: Oh, yes. I guess all I'm saying is the extraction of value, if the barriers are -- well, first of all, I really do believe that you get network effects. You know, if you look at Ebay or Amazon or Google, they're all network effects businesses where they've built some critical mass where the user is adding to the service in some way gives them critical advantage. But I also think that they've each added some things which mean they can't simply be displaced by someone else who just kind of says, "Oh, yeah, this is easy."
MR. BURNHAM: Let's get Tom in here.
MR. EVSLIN: Let me try gaming a little bit. I'm sorry, if we think of Craigslist's strategy, it's a brilliant anti-competitive strategy and I don't mean a value judgment by that. How do you compete with them? If you're not in the areas that they charge for, then you can't make any money because you can't charge for those areas either. So you can say, "Okay, well, I'll go compete in San Francisco by giving the ads away," but you don't have the readers in San Francisco. So what they've done is they've said that I've got the network effects from all the places that I give the ads away that I use to make the ads more valuable in the places that I sell them. And so I don't think it's just negligence that they haven't got around to charging for the other places. It's deliberately focusing the value on where they're extracting the money. It's like the Microsoft strategy of always pricing under Lotus, because you just don't want to leave any room for a competitor. And so here's the brilliant I'll price it free in most places so you can't go around me and get in anywhere, and I'll expect value from a few sweet places which you can't beat me there because the value I'm giving away for all the free places --
MR. O'REILLY: So how does the value in some other city add to the value?
MR. EVSLIN: Because of the job seeker. Remember the network effect, right. So the job seeker is often not as localized. They don't necessarily know they want to go to San Francisco. So there are some ads where it adds no value, obviously, Tim.
MR. BURNHAM: I hear lots of people saying I find my life on Craigslist. So it's not I just found my job. It's I found my apartment. So there's a whole bunch of other services that keep people coming back to Craigslist, and therefore, aggregate the audience around jobs in that market.
MR. EVSLIN: But it's really all about creating that network effect and not really leaving room around the edge.
MR. PINCUS: Why does Craigslist grow in a market like Milwaukee where they don't have crap on there? They still grow. There's something that's missing. Users are not driven. They're not going to Craigslist because it has the most of anything.
MR. HEIFERMAN: No city is a vacuum. People in what did you say, Milwaukee? People in Milwaukee hear about it from their friends in Chicago, but I think we're sort of -- the question on the table is kind of being asked in the world view of sort of the one ring to rule them all in a certain way, when here you've got -- one of the most fascinating things I think in recent history is how much of a weird dude Craig is, and that's how -- and how, you know, there's a giant body pile, a giant body pile of all the MBAs and marketing people that pat Craig on the head in 1997 and '98, '99, 2000, 2001, 2002 that said, "That's nice, Craig. You do your little thing. We're going to build a real business here." And they absolutely didn't expect to come out of Craigslist what it is, and that is because he is such a weird dude. And he's not sort of thinking this way. But really connected to that is --
MR. GOLDSTEIN: Not weird enough.
MR. WILSON: Can you get weirder.
MR. HEIFERMAN: What were we just talking about, because I wanted to make --
MR. JARVIS: That was a weird moment.
MR. BURNHAM: We'll come back. Michael.
MR. PAREKH: Craigslist has been talked about as the poster child. I'd be curious as to what this group thinks what the other poster child is for Wikipedia. How would you --
MR. PINCUS: What, Michael?
MR. PAREKH: The question is Craigslist has become Craigslist because "he was a weird guy." He let it happen for free. Maybe there was tactical elements as Tom mentioned, but there was ultimately a lot more value given out for free essentially for him to build a network, and one of the points that was brought up earlier is you got to let this thing grow to a certain point before you get the network effect before you can reveal the big model.
MR. BURNHAM: Well, Jeff, you were talking at lunch about some of the other initiatives.
MR. JARVIS: He's got Wikibooks, which Wikibooks you got the textbooks competing with the textbook world and maybe (inaudible). So my 13 year old needs to learn how to do more about Linux and stuff and I can go to Wikibooks now and find things for him there.
MR. PAREKH: So there would be Wiki For Dummies on any --
MR. JARVIS: Well, yes. It's about capturing knowledge. And so could you put ads on that? Yes. Could you sell those ads for a price? Yes. It becomes a gathering point of something. And he doesn't make money at that, right. You can instead use that as the beginning of a tutoring network. You know, who knows what the value will come out of it, but I want to hear what Tim thinks about it.
MR. O'REILLY: You know, it's hard to know. I don't know that it depends on which market. My basic thinking about online books is that books don't exist, you know. Books do various jobs, and the answers are very different depending on what jobs we're looking at. Already, you know, Google and the web have wiped out a whole pile of reference books. Other books are about teaching, right, and so he's setting aim at textbooks, and it's certainly true that it's quite possible the product can be developed. That the question, though, for example in the academic system is the distribution system for books is fairly resistant actually to the quality and content from outside the system, because the way that the textbooks, you know, for example in K through 12, they're actually adopted by committee, and it's typically in a few states, and then other states follow suit. In Higher Ed, it's actually one professor who adopts his buddy's book because he expects the same thing to happen in reverse. So there's often information about channels that's kind of missing from. So reference is pretty clear that the web is kicking the butt of traditional publishing. In other categories, entertainment is a third category, and already I tend to think the book has been bypassed in many cases by other forms of online entertainment. And, you know, in the entertainment world, it's pretty clear that while peer production definitely, you know, can play a role, high production values really matter, use that small games, and we're getting back to that with casual gaming, but the costs of now approach the cost of the developing movies. And so people basically find a way to up the anti again.
MR. H. MORGAN: One of the business models, by the way, is hardware. Apple obviously with the iPod shows the way to make money with hardware. And once we get a really effective electronic book platform, that will destroy, you know.
MR. EVSLIN: That suggests a possible monetization if you charge (inaudible).
MR. BURNHAM: At a slightly higher level of abstraction. What we're saying is look for the intersection between hardware and software. Look for the intersection between the information and the transaction for the area in which you're going to begin to develop that value; is that right?
MR. EVSLIN: I think lesson one is don't try to build a business and network at the same time. Don't try to build a business and a network at the same time, and the reason I'm saying that is because of the --
MR. WILSON: Because you didn't take your own advice when you launched your book?
MR. EVSLIN: I'm giving the book away now, right. Let's forget Reed's Law for a moment because Metcalf's Law is steep enough, that everyone knows that Metcalf's Law says use networks that increase value, and everybody forgets the converse, which is small networks have no value. And so what value is there to the first few users in joining a small network. Almost none. And so you can't extract anything in return. You can't put friction in the way of people joining a small network. You have to make it incredibly attractive and easy. That's the secret of Skype's success. They only had distribution experience. Then they used that, and they didn't introduced Skype in and Skype out because it would have been a distraction until they reached critical network mass.
MR. BURNHAM: Let's keep going down the list. Google, I mean, launched in '96. Revenue in 2000 for the first time. Something like that.
MR. PAREKH: In fact, they discounted heavily the distribution deal with AOL and others relative to what they could have monetized that for.
MR. BURNHAM: So there's a long history of people who have successfully built networks and invested. From an investor's perspective, there was an example of someone who was willing to invest for a long time before they took the risk of putting friction in the networks.
MR. EVSLIN: Yossi Vardi, who was the poster child for that with ICQ. I said to him, "Yossi, how are you going to make money?" He said, "I'm probably not going to let myself be distracted by that. That's not what I'm doing."
MR. BURNHAM: And Skype.
MR. EVSLIN: And Skype is another example.
MR. BURNHAM: We didn't hear the joke.
MR. PINCUS: Seth said revenue is a distraction.
MR. GOLDSTEIN: I was quoting Yossi.
MR. PINCUS: I said investor's are the biggest distraction.
MR. WILSON: You better remember who paid for your lunch.
MR. BURNHAM: All right. Let's come back to --
MR. EVSLIN: So you have this enormous network value, but it's not until you get there that you can worry much about how you're going to monetize it. Maybe don't monetize it directly. Maybe do what Skype did. You sell it to somebody else who has an infrastructure for which that network, you know, that that network can add value to, even though it shouldn't have grown to value by itself.
MR. HEIFERMAN: Skype has growing revenue, and I would just say, revenue let's you be confident in sustainability and growth and support of the network.
MR. BURNHAM: But let's challenge that, because if it is a peer produced network, it could actually compromise the sustainability and the growth of it. So if you introduce Google AdSense to Wikipedia, it might in fact compromise that, or if Craig became blatantly commercial, it might in fact compromise that.
MR. D. MORGAN: It would also depend who got the revenues from AdSense or Wikipedia. I don't think it would bother me one bit if I thought it was going to the foundation.
MR. JARVIS: What you're saying too is sometimes the infrastructure has to resist for flowers to bloom on top of it. So the fact that more press being an open source, it exists, and flowers are being let (inaudible). But that's where you try to reach value. We'll all help build it because I know --
MR. BURNHAM: Well, this is a point that Tim was making earlier, which is you may not be the one that capitalizes on the platform you create, but what Tom is suggesting is that Google did create the platform and then capitalized.
MR. JARVIS: It's the advantage of creating the network without revenue that you're the first guy to capitalize on it, and others will too, but you will benefit first and most because of that, not because you created the network.
MR. O'REILLY: Tom's point is really about timing.
MR. EVSLIN: This is just about investment strategy.
MR. BURNHAM: Let's get Michael in here. Then Josh.
MR. FRUMIN: A question I wonder is if as people, as all the peers doing the producing begin to realize that they're creating all this value for other people and they're used to getting some kind of value back for themselves -- you make a web page. You create a value on Google. But once more, the people who are doing the producing are realizing that they're creating economic value for other people, and there's a way to share that economic value more explicitly with the people. Can that contribute to the growth of the network. So not having destruction of growing the network and then creating value that --
MR. O'REILLY: Isn't that what Google does with AdSense?
MR. FRUMIN: Well, it started to, right. But the question is --
MR. BURNHAM: Well, Yochai would make the point that introduce financial incentives into something that is otherwise driven by something else, you can potentially destroy it.
MR. WILSON: Is there an example of that? Yochai, is there an example of that where someone had a nice flowering peer to peer business, they introduced too much commercialness into it and it just blew up?
MR. FRUMIN: But not just commercialness, but commercialness for the producers, for the peers doing the producing, if they can get value.
MR. BENKLER: I mean, Heather said something here on the side that I think is important as a disciplinary measure for this conversation which is Bubble 2.0. this needs to be a standard -- a standard question that we're all asking, because the idea of don't worry about it, let's grab the eyeballs is familiar.
MR. EVSLIN: We've heard that.
MR. BENKLER: So that's important. The other thing that's happening here -- by the way, you'll notice I say lots of things and I won't answer them. The other thing is that there are two design questions now on the table, not one, with regard to the relationship between our one, understanding the source of value being peer production in the social relationship, and number two, understanding this conversation among people who are interested in making money. And one has to do with getting the social cultural framing of the behavior right so that people -- and this goes to Scott's point about Craig being a weird dude. That is to say, having the characteristics of yeah, you know, we'll do it for free, and we'll then figure it out and having being trusted to do that, which has to do with people being willing to put in more, and that's actually where I would say that if you are looking at introducing economic value and people begin to see the monetized value of their contribution, and the only -- well, the only thing that at least comes to my mind at the moment is the difference between the free distributive computing systems and the one or two efforts to create paid per cycle or paid per month of cycle distributive systems, and if you look at all of the successful large scale distributive computing systems, they all give people a sense of what they're for. A sense of social mission. They render reasonably transparently the success of the common project. They provide a particular platform for people to express how much they're giving by saying who's the user of the month, the user of the week. They actually do a lot of mix and match, but when you compare them -- I don't remember their names now -- there are two or three efforts on the web who actually pay people X dollars a month for your -- and all of those have been very small and very flat, even though they're trying to do the same thing. There are some complicated factors, because they're actually not doing exactly the same thing, but that's one place.
MR. BURNHAM: Let's give Martin a chance to get in here.
MR. NISENHOLTZ: I think at some level, what you're trying to do at this meeting, Brad, is reduce risk. I mean, that's what I'm hearing. In other words, what I'm hearing is there's this notion of peer production. Most of the companies that evolved out of this peer production idea have been created almost -- goes back to my lottery point, although it wasn't a perfect metaphor -- but almost out of lightening striking with weird dudes in the middle of it. Brad, what I'm saying is, as investors, what you're trying to do is you're trying to provide a systematic overlay to this conversation in order to reduce your risk, and that's a perfectly credible thing to do. I just don't think it's possible, and I think at the end of the day, the success of this meeting is going to be measured in some part by you or perhaps should be as to whether we as a group can figure out how to identify and then in some ways reduce the risk around investing in these very weird entities.
MR. BURNHAM: Yes. I would say whether it's a personal investment in --
MR. NISENHOLTZ: No. No. I'm not accusing you of anything.
MR. BURNHAM: But one way to think about this is Heather's point of Bubble 2.0, and basically is there something different between Bubble 2.0 and Bubble 1.0, and is there anything that we've learned. It's smaller, right. If we can, let's get Bruce in here.
MR. SPECTOR: I think it needs to be said that in the early days of the "commercial internet," what have now become some of the very most successful players, Yahoo, Ebay, Google, with, of course, some exceptions, like Amazon, were not really started as businesses as all. They just were not. They were started as passionate enterprises.
MR. PAREKH: They were the Craigslists of the time.
MR. SPECTOR: Yes.
MR. NISENHOLTZ: And a ton of them failed.
MR. SPECTOR: They shared as the impetus for their creation the passion of the founders, and that's what they were. They were that and they were only that. It was David and Jerry's directory that became Yahoo, and it was a better search algorithm. It became Google, and it was some passion to create Ebay. But what's interesting is that there appears to be a cycle at work. So you need the innovative passion. You need to be able to simply have creative people, artists if you like, creating whatever the social networks are, creating whatever the network effects are by whatever means, and then at the point that those things are engaged, then the MBAs come in, the investors come in, and various monetization schemers are effectively applied with more or less success. And I think that's the cycle that we've seen in Web 1.0. and I think it's likely we're going to see a variant of that same cycle in this new cycle.
MR. WILSON: But part of the question we want answered is when do you bring the suits in, right?
MR. SPECTOR: I think to a certain degree, we're answering some of those questions, what Yochai has been saying and others, Tom, that the efficacy that the web has brought as this new network environment is in fact that network effect. The ability to, by virtue of aggregation of data, if that's what it is, by virtue of aggregation of eyeballs, if that's what it was, for Yahoo in 1998, then by virtue of that kind of effect, you gather enough interested parties in whatever that network environment is to generate revenue opportunities. And then that's the time presumably to go to your question, "when do the suits come in." It's a little bit like, if you will, it's a little bit like creating Seinfeld. Seinfeld wasn't successful from day one. Seinfeld needed time to gather an audience that became, you know, a very, very large and successful audience. So these things take that kind of time in cycle, and it's hard to know exactly how much time there is.
MR. BURNHAM: There's one thing --
MR. SPECTOR: You're going to see more successful ones, but the notion of so much the environment, I think what it really is is it's a very low barrier of entry to create network effects. That's the cool part. The cool part is it doesn't take a lot of capital to create a quite large network effect relatively quickly. Del.icio.us is a recent example. Passed that, then you need to attach, if you're going to go after money, then you got to attach a variety of really quite well-known models for how you make money. Whether it's an advertising based model. Whether it's a subscription based model. Whether it's a brokerage based model.
MR. BURNHAM: One thing that we've noticed is that a lot of things that we really like are up and running and working on the web, aggregating passionate group of users before the person who created it quits their day job, and that's not something that you saw, for instance, in the enterprise software development world or in the hardware world or the chip world, you couldn't do that. So there is a difference in the capital requirements for taking that first step and building that first passionate audience. The point that you're making, Bruce, is that it may be that taking that from there to some kind of business, then will require a certain amount of investment. It may not be --
MR. JARVIS: But it may be less.
MR. BURNHAM: It may be less.
MR. D. MORGAN: We need to see a different part of strategy, though, which says because the cost of creating networks is so low now, you have all sorts of people creating networks and you can sit back and wait until those few that make it out and you can go pick them, but I don't think that goes against the fact that there may be some of the most powerful networks that we haven't seen because it's still so immature that are commercial from the get go. And, actually, may be much more successful because they recognize that they're building up around an actual customer or problem that people will take on and not the accidental 164 million that can be extraordinary.
MR. BURNHAM: Josh. Then Tim.
MR. STYLMAN: I was actually going to make a similar point to Dave and I think something important to keep in mind is that where the network is built is not always where the revenue is derived from. So we were talking about credit earlier and sort of giving everything away for free. I make the same case about Tim mentioned that Google kicks the reference world's ass -- well, my words, not his. So Google does not make money from reference generally. Even when people are monetizing, not a lot of people can pay a lot of money for reference keywords. However, it gets people coming back to Google, and presumably some percentage are going to be more commercial in nature, thereby building those things.
MR. BURNHAM: Tim.
MR. O'REILLY: Actually, couple points since I first raised my hand. I keep hearing it's easy to build networks, and I would just say it's easy to build networks early in a new market, right. I mean, I built the first commercial web site, GNN. And it was really easy to do, all right, but we sold it because we realized that it was going to take capital to grow it, and I wanted to stay private, so I sold it. And the fact was there comes an inflexion point where you've got the organic growth and the next stage does take money. So there's definitely I think still an opportunity for capital. So the second point that I wanted to bring to this whole line of discussion came off what Yochai said about the distributive computation start-ups and it brings us back to peer production and what drives peer production. And I was part of an event that Macro Media held the other day, and they had a speaker (inaudible). He was talking about why do people pay so much for coffee. Coffee is a commodity. So put that thought in one spot, and then I want you to jump to a comment probably of all the things that happened at the Web 2.0 conference, the thing that struck me the most is a comment that Michael Powell made. He said, "My kids basically think the music should be free. Yet, they have $60 a month ring tone budget. And why is that?" He said, "Because music is something you consume privately. Ring tones is something that you -- they're part of your display. Kids are basically used to paying for this." And there's something in terms of all those thoughts about why do people contribute to folding at home, but not to a paid start-up. Why do kids pay for ring tones but not music. And this whole idea of the experienced economy. What is it that we're increasingly paying for in an economy of abundance?
MR. WILSON: Bling.
MR. O'REILLY: It's basically bling.
MS. GREEN: (Inaudible).
MR. O'REILLY: Absolutely. Yeah. From the provider point of view, you get money because it is closed and that's the only way you can do it, but they've been able to sell it, I think partly because -- well, first of all, you know, they have produced pricing that people thought was reasonable early in the market. But I do think that an interesting concept in that is what do we value in today's economy.
MR. PINCUS: But, Tim, I think it's more basic than that. Anything that can be free will be free. So to Heather's point, the only reason they pay for ring tones is they can't get them for free. Kids paid for music when they had to, and they did, and they'd save up for it, and now they don't, so they won't. I just think anything that can be free will be free, and then there's questions of laws that may stop it and that's it.
MR. BURNHAM: There is a start-up that's like an open source ring tone start-up.
MR. WILSON: We were talking about it at lunch. It's a service that allows you to upload an MP3, create a ring tone and have it SMS'd to you and you use it. And Shana said she thinks that that will get shut down.
MS. FISHER: A lot of labels are looking at it. And that constitutes use for them.
MR. SPECTOR: It doesn't have to be a service. There's software available that allows you to take any MP3 file and turn it into a ring tone.
MR. BURNHAM: Well, it's interesting, because even with the existence of those free services, you know, at three bucks,

you know, from the carrier, I'll do it. So there's a point at which people will pay maybe for display, you know, even at a price.
MR. WILSON: Seth was going to say something about five or 10 minutes ago and you never got the opportunity.
MR. GODIN: Well, the topic went away and I thought it would be selfish to bring it back. Well, I was just going to try to recast some of what we're here for, and say what if we think of this entire topic of the whole day as just another way of talking about marketing technique, but it's a marketing technique that's extremely counterintuitive and that's almost never attempted by anyone who has an alternative. Meaning, that if you got investors or cash, you want to control certain things. You want to spend. You don't want to treat the people who are providing the stuff you are controlling with the same level of respect that you would someone who you're paying. And so what we see is that these very low-funded, garage-type start-ups use these marketing techniques to do all these counterintuitive things and we notice the ones that work, and the act of interacting with Craigslist makes you talk about it. The act of having auctions on Ebay makes you talk about it. People notice them. So, in essence, it's a marketing technique, and the reason to date that traditional companies haven't done it is you'd have to make all your decisions along the way and always do the opposite of what your instincts tell you to make it go. So I want to give Lens Masters as much as I can, but advisors say well, you can get away with paying them less. Well, if you're viewing this as a marketing technique as something that is cheaper than running ads in a newspaper, then it's totally appropriate to pay people's budgets. You treat these people with respect and generosity, they're going to come and not go somewhere else. And the reason why I want to say this to the group is if we're doing this for Brad and Fred, which we are, and they're looking at the world to say, "What should we do and how should we help our existing companies do this better or find a reason for doing it?" I think it would be interesting to look at it from a marketing point.
MR. HEIFERMAN: Actually, I want -- I'll bypass and throw something right back. Why do you use the word co-op and how does that -- Seth is using the word co-op. Why that word, and how does it relate to peer production?
MR. GODIN: When I think of co-op, I think of two places, the Harvard Cou-op and the grocery store where you have to pack your granola in a bag in order to get a discount, and so we're not legally a co-op, and we make it clear to people we're not legally. That's a structure you could have. What we're saying is that if money comes in, money goes out. We're sharing with people who contribute the most to the community, the same way IRI or the Harvard Co-op sends you a check if you shop there a lot. But part of the story we're telling, which says you should join this and contribute because you're going to get something back, which you can give to charity or keep, but it's different than most organizations. Wikipedia says there is no money, and other organizations, Google says, "We're going to keep it."
MR. HEIFERMAN: I just want to respond real quick. You sparked me to really study up -- when I started using that word, you sparked me to study up on co-ops, the co-ops recently, and maybe most people are aware of this, I wasn't aware of this, did you know that, you know, many world famous brands are co-ops? Sunkist, Ocean Spray, AP, Best Western. It really is a fascinating thing to look at in the context of what we're talking about here. These are not hippy commie concepts. But if I could just adjust one quick note about this whole past hour, which is that the, you know, at Meet Up, Fred asked the question can you come up with an example of something that had momentum, a network that had momentum that introduced commercial interest and then failed. I think that Meet Up is out of the woods, but the jury's still out. We had 50 percent organic growth on basically a free service in the first two months of this year, after people thought we would be dead after the election. Then we introduced a mandatory fee, and we lost half of our active Meet Ups. Now we're at about 80 percent of where we were before, but, you know, there is always that nagging thought of if we grew 50 percent in the first two months, what would we be here in November or October if we hadn't introduced the fee in April. But the big point there is is that this notion of like, well, first you build a network and then you introduce the monetization, I think with all respect, that's very conflicting. The point is, for Ebay or for us or for, you know, like the question is how does the monetization make the product better. I would argue that AdWords makes Google a better product, not just is this sort of like you have to sift through the commercial messages.
MR. GODIN: And your fee makes your service better. Also, Ebay and AOL were monetizing from day one.
MR. GOLDSTEIN: Just this point of co-op is I think a useful segue to something which hasn't come up, which is related to monetization. So USAA, anyone belong to USAA? It's military families. One of the largest financial institutions in the world. 80, 90 billion dollars at least. It's a co-op. It's essentially owned and operated by the families insofar as they all -- sort of think about this from an open data frame work perspective. They're aggregating this audience and they're kicking back a lot of the savings to the users. The users get some kind of dividends at the end of the years because you're owners. And it sort of works. So when I think about -- we talk a lot about well, we got to use organic network e-business and you have to figure out monetization, somewhere in between is price. Price discovery. Price discovery is not monetization. Price discovery is what's the data worth. And that comes before it gets sold. You can't go to the selling of the data before you know what it's worth, and I think we're all trying to grapple through A, who owns the data, and B is, what is it worth. What is a click worth. What is a weed worth. What is my click stream worth. What is my attention worth. What is a tag worth. And you feel, at least in the beginning of figuring out a valuation around the consumer as opposed to the publisher. And until we get there, you're going to be flailing for business models. Not to complicate things, but there's a lack of price discovery in Google. They're a price higher, not a price sharer.
MR. O'REILLY: Right. But they figured out the price discovery. There's a big variation in the price -- in fact, that's what I was going to say. One of the things that Google teaches us -- and this is maybe a bit of an answer to Martin as well -- there are different business model for different classes of data. The price of a Google AdWords for Java programming is 11 cents. For some other category where there are products for sale is 50 bucks, and the fact is that, you know, that $50 cost per click in one area will support a very different type of publishing business. I mean, Matt's blog works because he's in a high cost per click area, whereas our stuff in technical information, we have to have a subscription business model because you can't make it on cost per click because the click's not worth very much. And so I think that one of the things that we need to learn, I think your point about price discovery is a good one. It's not one size fits all.
MR. STYLMAN: Well, the network decides the price. So the business model is completely aligned with the participation of the community.
MR. BURNHAM: So this is a peer produced price discovery mechanism that Google has created.
MR. GOLDSTEIN: But there is no price discovery. It's their price. Google's exchange.
MR. EVSLIN: There's a whole industry of search engines.
MR. GOLDSTEIN: The point with Google is there is one broker set up to sell access to that exchange, and it's called AdWords, and you basically either trust the prices you're getting from AdWords, or you trust the price you're getting from AdWords.
MR. BURNHAM: So you're talking about the cost per share to trade, not the actual --
MR. STYLMAN: Well, there's a difference between price of discovery and ultimately --
MR. GOLDSTEIN: But so much less.
MR. KURTNIT: It's just you don't have (inaudible).
MR. H. MORGAN: In Overture, we built a model. We actually showed the price per click on every search and --
MR. BURNHAM: Speak up, Howard.
MR. PINCUS: You're assuming that the end-user cares.
MR. BURNHAM: That's the interesting lesson.
MR. PINCUS: I don't think anyone gives a shit. I don't think they care about the value.
MR. BURNHAM: Wait. Let Howard finish the lesson.
MR. H. MORGAN: The lesson was that the end user didn't care. The advertiser cares, as long as they controlled and could see it. Once Google came in and closed it and Overture (inaudible).
MR. WILSON: Tom, we're going to give you the last word on this, and then I want to change direction a little bit with this.
MR. EVSLIN: Go back to what we were talking about is how Bubble 2.0 is different than Bubble 1.0 it's not different in that prices aren't going to get overinflated. It's not different in that it's not going to collapse because that's what bubbles do. It is just like Bubble 1.0 in that it will raise huge amounts of money in what is essentially a lottery that will fund a discontinuous advance in infrastructure. But one big difference and I think a constructive one when we're talking about investment is in the exit strategy, and the exit strategy, the IPO is the dominant exit strategy was one of the real weaknesses of Bubble 1.0 and a lot of good companies were born like premature babies -- and I'm not making personal statements -- but you can't not go public if your competitors are going public. And so once companies with no profits, with no revenue, with no predictability could go public, the companies themselves were hurt. I'm not worrying about the investors now, but the companies were then unable to perform the way that they could have.
MR. PAREKH: But in today's Web 2.0, the IPO market is not there as much as it was in the first one, but the exit strategy of choice is selling it to an incumbent.
MR. EVSLIN: I absolutely agree with you.
MR. PAREKH: But when they do that, that company then is changed. Myspace is forever changed now that it's part of -- their priorities are going to be changed.
MR. EVSLIN: And some of them will certainly be ruined, but I think that the companies run in expectation of acquisition, because they still can be run privately -- can be run better up to the point where they're acquired, and there's better chance that post-acquisition -- not 100 percent, you know acquisitions fail plenty -- but post-acquisition is probably better than post-IPO, particular for a class company, and Fred, you wrote about some of this, that we don't know if some of the companies that we admire now are really companies or technologies or networks, you know, can they really turn into a business. I question that about Skype. I don't want to reopen that argument, but the answer doesn't matter if there's a market that can fold them into something that's money making.
MR. PINCUS: You're talking about rationality, Tom. You're saying what you think is more rationality driven in this market --
MR. EVSLIN: I think from an entrepreneur's point of view, and I'm not a player this time around, but I think some entrepreneurs will be able to do a better job of running their companies, not that lots of monies are going to be lost, because the IPO was almost victory through death, and this is a better exit strategy both to plan for and to effectuate.
MR. JARVIS: Isn't the exit strategy also to risk high, in that because you start things so light -- I don't know. (Inaudible) I had lunch with Ed Siv yesterday about this topic where you say all right, you're going to start right. You get up while you still have your job. You learn you didn't need anybody. So then you come along and say, "Okay, it's great." Certainly smaller cheaper, but how do you manage many more investments to get to the same level of investment. And how do you find those things. Your entry is almost harder now than your exit.
MR. EVSLIN: I think Jeff has a very good point. You become angels.
MR. BURNHAM: Do we become talent agents?
MR. COSTOLO: Angels.
MR. BURNHAM: Angels.
MR. JARVIS: It's about at some point you have money and you have trust and you know smart people, and that maybe that it's more about trying to bring businesses together.
MR. BURNHAM: One of the possibilities, there's room for small pieces loosely joined -- and this sounds really bubbly -- we can get back to the CGI or CMGI, but is there some room, and obviously there are some lessons learned about the way you build synergies and things like that, but if you think about very like -- so there's -- wait. So there's no synergies. So then you're making argument that there is no value, but I'm trying to follow where you were.
MR. JARVIS: I'm trying to say what happens if you're a VC fund, do you have to come along and say shit, for me to really play, I can't make that bid, just like an add buyer (inaudible), you can't find that one big --
MR. BURNHAM: Right. But one of your proposals was there's a possibility of kind of a social network around the capital that creates value.
MR. JARVIS: You get your fingers into more pies, you share a risk and you share knowledge.
MR. O'REILLY: He's going to send you a T-shirt that says "Fuck you, I don't need a VC."
MS. FISHER: When I was at 2.0, these are young people who learned from Google's success. They said the big thing they took away is that don't work for VCs. It's just not a good thing. I think it's easier to start VC companies that way. I think it goes to something which I was going to tell John that there were -- I feel like there is a lack of heart in data, and I don't know if that's just there's a lot of easy things to be done. It's just judgment. I wonder where the hard problem being solved is.
MR. HEIFERMAN: There is no long tail of household words. Meaning -- like out of the whole boom, there's a few words that 80 percent of the world knows. Ebay and Yahoo and a whole big chunk of them know about dotcom. That's the thing about you look at the list of Web 2.0 companies and projects and sites, and it's all this explosion of words and things and things that no one will ever remember, and no chance that more than one percent of the population will ever be able to digest and make a part of their vernacular. And this isn't about sort of brand and trust, it's literally about how can human beings process this stuff. And this may not sound exactly like it has precisely to do with what this conversation is, but I actually think it does, because there is no long tail for new household words, and household words are like the currency of real value, because you have to be the go to place. Craigslist value is driven by the fact there's a go to place called Craigslist. If it was just that you could go to Google and see this smattering of listings of real estate, that if Google is the ultimate go to place for everything, then the system doesn't function well, but there are these go to places.
MR. PAREKH: But Skype is a virtual peer to peer go to place. There was no Skype --
MR. HEIFERMAN: When I say "go to place," I mean that people are starting to say I'll --
MR. BURNHAM: Let's get Heather in here and wrap this up.
MS. GREEN: I wish Clay was here. I was talking to him yesterday, and I was trying to find out what's going on with kids, right. Myspace. What the heck is going on with Myspace? I went to Clay and I said to Clay, "Clay, you work with kids, do you know what's going on?" He said, "I can't know what's going on." He said, "When you were in like '99, 2000, you could. You could know Ebay. You could know all these companies because there's a finite number of people and you could find stuff." What he talks about now is there's so many people online and there's so many different kinds of places they go, there's not going to be household names in the way we thought about in the past. Craigslist comes up over time. We know it now, but maybe over time that's what happens, but his point is that there's all these different kinds of things that users of the VCs that can probably find a lot faster than anyone else.
MR. BURNHAM: I do want to wrap up this kind of part of the discussion because we have one more topic that we want to touch on, and we are actually rapidly approaching the time that we committed to break, and I would like to keep to that commitment. Get people back to the real world here. I think if I hear what Scott's saying and what Heather's saying and come back to something that Dave Morgan said is that ultimately, that there's a limited amount of attention, and that that attention is going to be directed to a limited set of household names. Those names are going to represent brands, and the tricky part is if we go back to 1.0, everybody spent money building brands. I think we now figured out that particularly when the brand needs to have something to do with authenticity, you can't just buy it. You have to grow it in some other way. So it's not a slam dunk. But I think there probably will be a way of managing attention around concepts that are linked to household names, and that there will be -- back to Martin's 164 million to one scenario -- we will see a very steep curve in terms of what's successful and what's not. The one other thing that we wanted to talk about briefly, and I want to just try to get everybody's ideas on this, is does the current -- this is the last topic in the agenda -- but does the current regulatory frame work help or hurt the evolution of I would say the consumer -- the potential consumer benefit from the services that we're talking about, the peer produced services?
MR. EVSLIN: Did you mean that to be a US centric question?
MR. BURNHAM: My mind is mush if I try to get outside of the US, but I think thinking about -- I think it was Mary early on who introduced the idea of China as being a place that you can't control, therefore, the regulatory frame work is irrelevant, as long as there is a place in the planet where different models are emerging. So I don't mean it to be US centric. Just is there anything that we can be doing? Do we as a body of people who have an interest in this have an obligation to be voicing an opinion about, you know, about the way copyright works in this scenario, the way patents work in this scenario, the way any other legal frame work impacts the development of this market.
MR. WILSON: Brad, let me take a shot at that. Earlier this week I went down to the 11th floor where Joshua runs his business, and the only guy in the office at that time was this guy Dan, who I don't know, how old is Dan, 24, 25?
MR. SCHACTER: I think so.
MR. WILSON: I said, "Hey, how you doing, Dan? He said, "Good." He said, "I got a question for you." He said, "When you post those MP3s on your blog, is that legal?" I said, "No." He said, "Why do you do it?" I said, "It's a political statement." And this is the political statement I'm making. I've posted my policy on how I post MP3s to my blog, and the policy is this, I encode them at 96 kilobits a second. I link to Amazon or some other place that you can purchase that music legally, and I basically do it to Amazon, because through Amazon, I can actually get a report back on actually how many people do link through and buy it. And if the RIAA ever comes after me, I will elevate this as high in the law of the land as possible, because I believe I'm doing the artist a favor by that policy, and I believe I'm selling more music, and I think that's what needs to happen in this country, is people need to take a stand, and fuck what the law says, let's create a new law.
MR. O'REILLY: This is very related to the Google library. I really believe that, you know, opt out that's the only viable option. If you look at the music industry, it said, "We cannot clear the permission." They are waiting to be rescued from the problem of too many property rights. Book publishers are waiting to be rescued. They just don't know. The fact is, we have built so many rights into the system, that you cannot act unless somebody cuts the guardian off.
MR. JARVIS: The big guy at the last lunch -- the last lunch I went to --
MR. WILSON: There may have been some since.
MR. JARVIS: The big guys can't do anything. Somebody's got them on hold. It's either the MSO or the talent agent or the talent. So it's the little guys or the outlaws who are the avengers.
MR. D. MORGAN: I think you can take the risk and have the one out of the 164 million and be socially legally disruptive. The problem is as a business, you can't. I mean, at the point you decide this is going to be the business, I mean, Michael Roberts and others have to say, "Okay, I've got to play by a certain set of rules." And then we'll just have to look to regulatory structure and say it's really problematic, and unfortunately, it's also skewed towards those that don't want to be disruptive, because they control the regulatory structure. They control the legislative structure. The ones with the money can create inertia problems. And I don't know that it necessarily helps our industry to be the ones that are going against the policies all the time, because it actually feeds the forces that are trying to stop it because it plays well in Washington. You can't trust these people because they don't obey laws.
MR. WILSON: But, Dave, the mistake we can make as business owners and business operators is to pay too much attention to those entrenched players in the marketplace and try to serve their interest, because at the end of the day, they'll be an anchor that will sink your ship. What you have to do is you have to at least recognize the only wind in anybody's sail is coming from the people, you know, the ants. This is a revolution of the ants. That's the people that need to take you forward. And the big dinosaurs are going to sink your ship.
MR. JARVIS: When are you running for the Senate?
MR. WILSON: No.
MR. MORGAN: But, Fred, your blog's not a business.
MR. WILSON: I'm not talking about my blog, though. I'm just talking about you as a business might be better off -- I don't want to argue your business -- but you might be better off providing services to me and 1,000 other people like me who have the wind in my sails than Martin and his crew.
MR. D. MORGAN: I don't disagree with that. I just said that I think -- and this is maybe more personal policy thing, which is you get to the point where there are clear laws, and I think that -- I don't think you can say, A, I'm going to directly violate those laws and still say I'm operating a business because I want people -- or to do things where I know that people are going to break laws, but I'm not really going to ask the questions. I do think that, though, that you got to work with the small pieces, because I do think that's where the wind is.
MR. BURNHAM: I think we're talking about something that isn't clear. That isn't black and white. Yochai.
MR. BENKLER: I want to push Fred a little bit in an awkward situation where I will stand as a law professor for the first time in my life. I'm not sure that your marginal contribution is best being putting on the MP3 with a very clear policy as opposed to using the fact that unlike the millions of teenagers who can do exactly what you're doing, you're also a venture capitalist who knows lots of people who are in the business, and can speak a language that the political system understands, which is here's the value we're preserving, a whole 11 billion dollars recording industry, that's the whole recording industry, 11 billion dollars a year, that's it, versus let's take a look at the IT industry's somewhere on the order of an order of magnitude or more larger, and say where's the direct? We're saying the new value is coming from people interacting with and around the information as freely as possible, creating new sources of value, which is where our relative advantage is, and in this regard, I agree that in the sense national is the wrong way perhaps to think about it, but the political economy is such that Congress thinks of US jobs and the EU thinks of European jobs, and an argument about relative advantages of various economies is important, and in this regard, I think, putting together a coalition of people who are investing in this area and intervening in the legislative process, rather than saying the ants will win, is much more powerful, because there are lots of ants that don't have the advantage you have, which is speaking credibly in the name of business.
MS. GREEN: You have to run for Senate.
MR. BENKLER: You don't have to speak.
MR. O'REILLY: I have a comment on that, having done some lobbying in Washington. Unless you're prepared to make that a full-time job, you're spitten to win, because there are people who have that full-time job, and you know the people in Washington know that you're visiting, you're going to be gone.
MR. JARVIS: And we organize an association that pays for --
MR. O'REILLY: Absolutely.
MR. BENKLER: So, today, the Freedom Law Center is being launched by and funded as essentially a standing public interest law firm in favor of the free software with funding from some very big players who understand that that's important. Clearly you have to be there, but let's remember, the recording industry and Hollywood are two industries that are optimized around a particular regulatory system called copyright that they have specialized for close to a century in lobbying around that, and of course they're going to be pulling against, because everything we're talking about kicks them in the business model. And so the point of actually getting together and saying, well, they're using a particular set of regulatory systems to undermine our business, one of the things we need to invest not individually one firm, but to begin to understand that the whole cost of people who are looking at these kinds of technologies need to be able to generate is something like the Freedom Law Center. Something like a standing body that represents the interests of those people who as a business want --
MR. HEIFERMAN: That is both effecting and actually directly about what I think we're talking about here. This concept of peer production, production -- I think production isn't just content and web pages and stuff like that. I'm making one of these long bets that says that one of the biggest impacts of technology over the next 10 years will be something that really doesn't exist yet, which is collective power, collective action. I'm betting that half of the Fortune 500 10 years from now will have their business drastically affected by the ability for consumers and people to organize via the net. And that's going to be, you know, if five million Walmart customers signed an active pledge that they would reduce their Walmart spending by 10 percent if Walmart does not give their employees good health insurance, Walmart will listen. And because they'll listen -- well, they will listen, and you'll say, "Well, why would five million people sign that pledge?" Because the social network of the million Walmart employees is easily five million. Point, bottom line here is that peer production isn't just production, it's peer power, and so this kind of stuff, this ant stuff is really like the net itself is going to fuel an incredible amount of people power that we've never seen before I believe.
MR. WILSON: Marc, this is your soap box, so stand on it for a second.
MR. PINCUS: I tried to start something at the same time (inaudible). And the idea was to create a web-based platform that would enable revolutionary ants, and the idea was that it would be like the people's lobby, and if you had five million people who all agreed to put money or votes into things they cared about, you have the power, and I learned quickly, people care about issues, not big esoteric things like their own power and move on. You've heard of an e-party cabinet. But I do think totally believe it'll happen, but what's interesting me, I was one of the founding investors in Napster and one of the ideas I pushed to Hank Berry or before Hank Berry that they rejected and went the legal route was let's say "fuck you." Instead of trying to work with the RIAA and everybody, why don't you just put something out on the front page of Napster and tell everybody if you want to -- here's all the list of people who are being bought by RIAA, and if you want to contribute to all their competitors click here, and just go to your 80 million people and just ask them to donate money to make what they're doing legal, and I do think if any of these guys now built that in, I think you would get an incredible amount of money, and I do think that money is what buys politics more than votes or anything else. So, I think it would be -- I agree with you, and I'm interested to hear what your long term thought is.
MR. WILSON: Mary, you had something you wanted to say.
MS. HODDER: Oh, I'm building a remix community for video, but anything that can get plugged into video.
MR. JARVIS: She likes VCs.
MS. HODDER: I do like VCs. I can build for 1/10th of what it would cost five years ago, so. Anyway, so I think that there's some place in between what Tim was talking about and Yochai was talking about. You know, Viacom has 14 people full time in Washington, Viacom Digital. We can't compete with that. And I also don't think that -- I mean, I have blogged for years about how all these people with top down contents should be doing something different than what they are doing. Or they shouldn't be suing whoever they are suing. I use MP3s that are copyrighted in the videos that I remix and put on my blog, and probably will remove those before I go live because I don't want to spend two days a week at depositions like Hank Berry did. Bloggers often made their own content, and so many eyeballs followed along behind them, that newspapers had to pay attention. I think we can do this with rich media. So if we make it easy for people to upload their digital cameras and, you know, cell phones to upload snapshots of video to the web and play with each other's stuff under creative commons licensing, so that it's for noncommercial use and figure out the issues around remixing for that kind of content, if enough eyeballs go away from top down content, then those top down producers are going to pay attention, because they're going to realize that they're losing a percentage of their audience to something else, and then they're going to have to figure out how to play with that world.
MS. GREEN: They might do that. (Inaudible).
MS. HODDER: No. No. But the point is, if users are remixing each other's stuff, they're completely out of the loop.
MS. GREEN: I agree with what you're saying. I think you should do what you're doing. I think that's great (inaudible).
MS. HODDER: But, I mean, New York Times started paying attention to blogs when enough people -- when they were having enough traffic driven to them through links, and also, when they realized that enough people were reading other stuff and were just going to simply be lost if they didn't figure out --
MR. WILSON: So, Cyril, you were going to add something.
MR. HOURI: Yes. I was wondering why there hasn't been a way to work both sides of lobbying. We're talking about companies like the 15 lobbyists, and you know, I wonder if people's votes are not more powerful than the lobbyists making cases in Washington. And I haven't seen really that happening. Like people saying to like congressmen or senators or whatever, "Look, we're against this. Even if AOL Viacom tells you that it's a thing you should do, if you do it, we won't vote for you." And I'm wondering why.
MR. BURNHAM: So we need a peer produced lobby.
MR. EVSLIN: I think the answer to that is there isn't any mass movement that thinks Viacom is doing anything wrong, because consumers don't understand the intricacies that we understand. There is a Web 2.0 effect. Every campaign that is getting going now understands the value of the internet. The Bush administration is very effective in using the internet to reach its faithful. So it's not that the technology don't have any affect, but it doesn't split along ideological lines at the level that you're talking about by the people who are the consumers. I think Marc's over optimistic in thinking that Napster could have raised a lot of donations to be channeled to the opponents of the enemy, because people are disassociative, and it has to do with what you believe is credible where. You know, much as Jeff Jarvis' point. So you think Napster's great when it comes to music, but it's not associated with politics. So you see that you ought to contribute to this campaign, and I'm not talking about us now. I'm talking about the downloaders, and you just turn it off. I don't want music. I don't want politics. So I don't think we have as much power there as we think we have.
MR. JARVIS: So given the three choices, competition, votes or lobbyists, what's the best path, those three things?
MR. WILSON: Competition. Get in the market place -- I mean, I'm with Mary. Get in the marketplace and just beat the crap out of them.
MR. EVSLIN: If Chris Anderson were here, he'd point out that he did speak for the Churchill Club where he talked about how he's being competed with by the long tail, and he was the head, and he's got wires and he's got two million subscribers. He's got Cande Nast behind him. And there are amateurs who can do things better than he can do in specific areas. Since there's much less competitiveness now for those amateurs doing their things, and the amateurs doing it out of love or what many of the reasons we talked about, I think what happens in the end, that's why Mary's directly on point, that there's enough talent that was buried by the hit system -- not the copyright system, the hit system -- and the hit system was a function of distribution, and as the hit system disappears and the distribution gets better, then we get effective contribution from the tail, and in the end, the copyright holders can lose by winning. They can win their regulatory battle and be left with a lot of worthless property, if you want to be an optimist, and I am, because Mary makes it possible for the tail to produce the content that people want, and the issue simply doesn't exist anymore.
MR. BURNHAM: Doesn't look like you're optimistic.
MR. BENKLER: No. No. It's not a question about optimist. The sense that these are either/ors as opposed to what are the set of strategies and what are the relevant competencies, right. Clearly, you compete -- you start with competing. That's the existence proof, and that's the source of funding, but then the question is, you know, I don't think that there's no social movement around IT. I think creative commons is first and foremost a social movement and only secondarily a set of licenses and a set of practices. I think what we're beginning to see are the international trade and IP regimes. You're suddenly seeing a very interesting alliance between developing nations, some of the NGOs, and players like Cisco, who are concerned about excessive IP coming together and beginning to lobby. So the question that I thought I was answering was what is the relative advantage of people who are concerned with the business side of the emergence of peer production who understand the excessive -- who understand the drag that the regulatory system called copyright exerts on the whole amount of value. What is your relative advantage in this system. And one is clearly actually competing, but the other is also speaking with a particular voice that the volunteers and the peer producers out there don't have. Sure, they have the potential votes, but what they don't have -- and this plays an important role in DC -- is the credibility of this is a business position as opposed to an anarchist position, and that's an important voice that needs to be articulated. And in the area of spectrum, there are all sorts of activists talking about free spectrum, and then around 2001 or so, Intel and Cisco and Microsoft suddenly stepped in and told the Commission, "We can make a lot of money if you make more unlicensed spectrum available." And as it turns out, that changed the way the Commission saw things, and we got a lot more spectrum and a lot more focus on open spectrum policy. And I think crystalizing that as a pro business deregulation, right. Scaling back copyright from businesses is a pro business form of deregulation would be an enormously important means.
MR. WILSON: That's very good. We need to wrap. I want to just sort of tie this back to where we started. Rikki said that he was looking for the one plus one equals three. In the old model, if you aggregated, you got a certain aggregate set of data. If you had two people producing content, you know, you'd get the same aggregated set of data. Was there some magic that came out of the fact that the content was peer produced versus aggregated in some other way, and Seth Godin said that it's really a marketing technique, and that, you know, engaging the peer producers in the production of the content turns them into advocates of the content, and they then have equity -- he didn't say it this way, but I interpreted it -- they have equity in that product, and they become promoters of it, which is free marketing. He then went on to say that this event was for me and Brad, and I'd hope that he's wrong about that in the same way that I hope that one plus one does equal three, and what I'm hoping is that since this was a peer produced event, that each person who came and contributed walks away with more than what they contributed. I certainly have, and I hope all of you have too.
MR. BURNHAM: We have one more housekeeping detail, and that is to thank our hosts. And I'd like to have Jonah just say, very briefly, who Eyebeam is and why you invited us.
MR. PERETTI: Eyebeam is a nonprofit hardware technology organization, and one of the main reasons we were interested in having this here is because we are launching a new lab, we just got funding, and the idea behind the lab is that there are certain things that businesses can do better than a nonprofit and there's certain things that work better in a nonprofit environment. Instead of just thinking of nonprofit work as being something like, you know, does good things that the market won't take care of, we wanted to develop this idea of doing actual productive work that adds value, and that's partly why this concept of peer production is interesting to us. So this lab through the glass there is the new Eyebeam open lab, and starting in November 4, we have four fellows who will come and they'll develop work for a year, and all the work they will do will be under open licenses. Some of it will be media based. Some of it will be software based. Some of it will be hardware and physical objects. It says reading, printer, laser cutter, and the idea is that the work that happens in this R and D lab will be public domain R and D, so it's something that people can develop. Take with them. Say here, it can be shared. It can be developed and distributed by a network of people beyond Eyebeam. So we're starting to look into these issues and how to do creative R and D.
MR. BURNHAM: Thank you for your space.
MR. GOLDSTEIN: I think it's appropriate that we had it here too, because Del.icio.us is the Web 2 poster child for New York, and I think it's a great contribution to the Union Square's portfolio. I met Josh who I introduced to Fred and Brad through Mike Frumin.
MR. PERETTI: And Mike Frumin is the technical director of the new lab.
MR. BURNHAM: And, also, for those of you who don't know Mike, Mike is the guy that created the election database by Geography and put that up in the rafts, so it's another interface on top of the data set.
MR. WILSON: I've got one more thing. We do want this to be more open than just what happened in this room, and Eddie has been busily transcribing the entire thing, and I've been watching. It's just amazing. I could never do what you do, Eddie. This is going to be on our Wiki at usv.jot.com. All of you, because you were invited to this event, actually have that URL somewhere in your databases. Just click on the public site of the Wiki. There will be a link to Sessions, and everything will be in there. Mark, who's been photographing, we're doing it all digitally. Hopefully by sometime tonight or tomorrow by the latest, we will have taken those photos and put them up on Flickr. They're there? On Flickr? What's the tag? Linked to it from the Union Square Ventures.com web site -- blog, correct. And what we're going to do is if anybody blogs anything about this event, we are going to tag it in Del.icio.us with a tag USV Sessions. There's already a pretty nice group of links there with Yochai's work and some of Tim's stuff and some of Jeff's stuff. Umair, who's been very quiet -- put a plug in for Umair -- he is doing some of the best thinking about this stuff that we've come across. There's a bunch of links to some of Umair's stuff. Tom's debunking of Reed's Law is on there. So anyway, contribute. Blog. Keep the conversation going. We'll try to pull it all together and keep people focused on it, and thanks for coming.
MR. EVSLIN: Can I ask one question, in the interest of opt out, is there anyone who doesn't want anything they said attributed to them?
MR. O'REILLY: I think Scott probably doesn't want to be quoted as saying Craig's a weird guy.
MR. EVSLIN: I'm happy to have anything attributed to me, but just in case somebody doesn't.
MR. WILSON: If anything's on the transcript that you find objectionable, let us know and we will censor. Thank you. (TIME NOTED: 3: 10 P.M.)

CERTIFICATION

I, Edward Leto, a Notary Public in and for the State of New York, do hereby certify: THAT the witness(es) whose testimony is herein before set forth, was duly sworn by me; and THAT the within transcript is a true and accurate record of the testimony given by said witness(es). I further certify that I am not related either by blood or marriage, to any of the parties to this action; and THAT I am in no way interested in the outcome of this matter. IN WITNESS WHEREOF, I have here unto set my hand this 21st day of October, 2005.--------------------------- EDWARD LETO