Interview with Web Guru Tim O'Reilly "We're Moving into a New World"
SPIEGEL ONLINE: You coined the term "Web 2.0" a little over two years ago. Are you beginning to get sick of it yet?
Tim O'Reilly: Yes!
SPIEGEL ONLINE: Why is that?
O'Reilly: Well, yes and no. It's a wonderful feeling to be able to see something and then get other people to see it, too. I wasn't actually the one to coin the term Web 2.0, it was one of my co-workers, trying to think of a new idea for a conference. He was the one who put the name to it, and the name stuck -- to these ideas that I'd been evangelizing for many years.
SPIEGEL ONLINE: And now the term is starting to get on your nerves?
O'Reilly: Recently I was on the Virgin Islands with Richard Branson and his crew, on the way here I had a meeting with the technology planning group of British Petroleum, I just got an invitation to meet up with the CEO of IBM -- and that's wonderful! The downside is: You get tired of hearing yourself talk. It would be much more fun to get out there, learning from other people, and not saying the same thing over and over again.
SPIEGEL ONLINE: These days, the term Web 2.0 seems to be used mostly by people who are fascinated with the idea of making money out of other people's upaid work
O'Reilly: That's part of what keeps me engaged with the idea. A lot of people are trying to bastardize the term and try to turn it into another version of the original dotcom-bubble. I keep saying: "No no no, there's something really important happening here!" A lot of the ideas wrapped up in my framing of Web 2.0 are fairly difficult ideas. It's nice to see them catch on, bit by bit. So that people realize: "Oh, it's not just about mashups, it really is about this idea of harnessing collective intelligence." People are starting to get it. I just try to stay away from the flacks and the froth people.
SPIEGEL ONLINE: Nevertheless, many people are thrilled by the prospect of making people work without paying them. Isn't there an ethical problem? For example the music database service Gracenote is based on data originally provided by its users. Now it's a business and someone is making money out of it.
O'Reilly: Linux is an operating system constructed by its users and people make money out of it. There just has to be exchange of value -- it doesn't have to be money. Open Source software communities create a lot of value -- not just through companies like Red Hat, who sell Linux distributions, but also through the success of companies like Google, Yahoo or Amazon, who all use Linux software. The programmers who've made all of this possible -- are they resentful? No. What did they get out of it? Reputation, better jobs -- now if you're an open source programmer your resume is the work that you've done!
SPIEGEL ONLINE: But that doesn't apply to many of those who contribute to Gracenote ...
O'Reilly: People also recognize that if you make a small contribution to a collective work -- you don't necessarily need to get paid. When I stick in a CD and see that it has unrecognized tracks -- I submit them to Gracenote, I don't have a problem with the fact that they're going to get money out of that -- because they're providing a valuable service.
SPIEGEL ONLINE: Another catchphrase is that of the "wisdom of crowds." Isn't there a risk that large crowds of people are more stupid, and not smarter than individuals?
O'Reilly: A lot depends on how a technique is used. If it was a simple voting algorithm -- yes. Sites like digg.com that let people vote up news are easily kept hostage. But Google has a kind of "wisdom of crowds" elements to it -- it takes into account how many other Web sites link to a particular page. At the same time, they still have people making judgements about the quality of the results. There's an interesting example: There is a national park in Alaska that's called Glacier Bay National Park. Google thought: "We're getting the wrong result, people are typing in Glacier Bay, and we're giving them the name of a company that makes toilets!" They studied the results, then realized that actually more people were looking for the Glacier Bay toilet company! So their human judgement was actually wrong, and the system was smarter.
SPIEGEL ONLINE: But aren't large crowds uncontrollable, maybe even dangerous?
O'Reilly: Spam is an example of the madness of crowds. But it can also be countered by the wisdom of crowds -- individuals can identify spam, and that has become a very powerful anti-spam technique. We're going to move into a world that's not just about people expressing opinion -- it really is about distributed data gathering and real time intelligence.
SPIEGEL ONLINE: A thought experiment on the wisdom of crowds. Say you have a visitor from another planet, an alien, and you want to show him human culture. Would you show him the top ten videos on YouTube? Or a random sample of MySpace pages?
O'Reilly: No. There is an American comedian whose answer to that question was: "This is not really our civilization, it's just something we're playing around with. Come back in 10 years and we'll show you our real civilization." Either that or we'd shoot them and they'd turn around and go back in their spaceship. But seriously: I would show them Google and say: "This is the furthest we've come towards an artificial intelligence, what do you think of that?" I'd show them some of our great works of art. I'd show them some of the things that we're doing wrong and say: "Do you have any ideas about how to fix this?"
SPIEGEL ONLINE: Lately there has been a lot of talk about online communities such as Wikipedia and the news Web site digg, which has a very clued-in, tech-savvy community that digs up really interesting and new stuff. But if digg were overrun by the mainstream, that would soon be over. Is there any way of controlling these kinds of developments?
O'Reilly: Look at the history of Open Source software communities -- they have a long history in that area, so they're a good example. They have a tiered approach -- and Wikipedia has copied a lot of that, which is why they've done a good job. Anybody can make a suggestion, submit a bug report, submit a patch. But it doesn't get used unless someone in some inner circle says: "Oh, that's a good one, I'm going to apply it." And you only get invited to that inner circle after you've sent in enough useful patches.
SPIEGEL ONLINE: So you need a mechanism of exclusion and a hierarchical structure?
O'Reilly: Yes. So you get successively smaller circles. You get down to the point where Linux founding father Linus Torvalds won't even talk to you unless he hears from you from someone who he trusts. We build these trust networks in open software projects that help manage this process of crowd submission. Someone like digg is going to have to figure that out.
SPIEGEL ONLINE: Jason Calacanis, who until recently ran AOL's Netscape division, basically copied digg. But he started paying people to do the same thing they did there for free -- dig up stories and link to them. Is that a good method of ensuring the quality of your community?
O'Reilly: Maybe. In my opinion, a passionate volunteer is actually worth more than a paid person. If you can get them -- if you can't, then paying people is a good second chance. But if you look at the results, they would say it's not been working as well.
SPIEGEL ONLINE: The virtual reality developer Jaron Lanier calls Wikipedia dangerous, because it creates a knowledge monoculture -- and it can be used for character assassination for example. Is he right?
O'Reilly: Everything can be misused. In general, Wikipedia is a useful resource. Have entries that been hijacked? Absolutely. Have people been mistreated? Absolutely. But Wikipedia has developed some internal mechanisms for trying to deal with that. They're not perfect -- but for example they're a heck of a lot better than our political system is right now, at getting things to work correctly. I think people hold the Internet to very high standards when they're asking questions like that. I always like to refer people to a famous quote from science fiction called "Sturgeon's Law." Theodore Sturgeon was a science fiction writer, and someone once remarked to him that 95 percent of science fiction is crap. He replied: "95 percent of everything is crap!" And Wikipedia is pretty darn good! You can get a pretty good, concise explanation of virtually anything there.
SPIEGEL ONLINE: But isn't there some truth in comedian Stephen Colberts idea of "Wikiality?" That if something is so good, it eventually becomes a universal standard that no-one questions anymore, even though it still contains mistakes?
O'Reilly: Go read history! There is an official version of virtually everything -- and it leaves out a lot! It's wrong! That is no more true of Wikipedia than it is of a textbook. Wikipedia is generally less likely to have that systematic bias than a textbook. Anything we do is a selection of reality. Do people forget that? Absolutely. That's a great source of disorder in our societies.
SPIEGEL ONLINE: Calacanis started paying bloggers and "diggers," MySpace and YouTube are being invaded by PR and advertising -- is there a danger of the whole Web 2.0 being taken over by the pros and their own version of reality?
O'Reilly: Yes and no. Web 2.0 is a name that we're attaching to a deep long-term -- everything is becoming connected. The Internet is becoming a glue that ties together everything we touch. Yes, people are going to invade that and try to use it. There are people learning how to drive and manage network effects. But ultimately it's like saying: "Could somebody build a better airline by just being better at PR?" Richard Branson can create an airline that's competitive edge is its PR. But he still has to deliver the final product! You can't have PR and no ability to fly. On the Internet, the service still has to work. You can gain an edge by being good at marketing -- but they can't take it over.
SPIEGEL ONLINE: If people push it to the point where it stops working
O'Reilly: the users will stay away. Eric Schmidt has this great line that he uses internally at Google: "Don't fight the Internet!" When you create new services, think about "where do they really want to go?" The Internet is a little bit like gravity. There's tricks you can use to fight gravity, to fly -- but you really have to do everything right.
SPIEGEL ONLINE: Your name is closely connected to two of the catchphrases of our times: "Open Source" and "Web 2.0." What's the next buzzword you'll invent?
O'Reilly: The next big thing we're working on is wrapped up in a new magazine we're doing, called "Make." We're really focussing on how computing is starting to interact with the physical world, like with custom manufacturing. What we're all seeing is a huge amount of hacker activity around making things. There's a lot of disposable hardware now. People are now on their 3rd or 4th digital camera, and what do they do with the old one? They can re-use it. We also have all sorts of custom manufacturing devices, like laser cutters and 3D-printers. All this stuff is about the same price point that typesetters were in the beginning of the desktop publishing revolution.
SPIEGEL ONLINE: So the focus is shifting back from software to the real world?
O'Reilly: Yes. We're entering an era of custom manufacturing. There's synthetic biology -- where custom manufacturing will extend down to chemical processes and materials as well as electronics and physical goods. Not to mention we have the ability to mass-produce things more easily in some part of the world where there's relatively cheap labor in relatively small units. People are getting increasingly sophisticated tools for simulation and design. Like when people are building things in "Second Life" or in Sketchup on Google Earth. The tools of design of virtual things that can become physical things are getting much more democratized, much more widely spread. For example, there are a couple of services where you can get your avatar from "Second Life" printed out in 3D. I think that's one of the big frontiers.
Interview conducted by Christian Stöcker.