A chat with Jimmy Wales, founder of Wikipedia, on the eve of our own Future of Finance Summit 2019. My questions were designed to get Jimmy to discuss how traditional institutions will survive in the shared economy, given that Wikipedia was essentially an effort by a world full of volunteers. He brilliantly outlined the rules by which the collaborative economy actually works – the right mix of top-down and community approach that is required to be successful in the digital world, rules by which, I thought that APIs and other collaborative platforms should also adhere to.
Following is the unedited transcript:
Creating a Platform Based on Open Collaboration
ED: Very pleased to be able to speak to Jimmy Wales, the founder of Wikipedia, that huge library or encyclopedia which has five million articles now and growing at about 570 articles every day, 270 languages. A whole new business model. Just talk us through a little bit about the early days of Wikipedia what it took to kickstart the idea of a collaborative momentum. Now, I ask this with a little bit of evil in my mind because when we think about Alibaba, for example, Jack Ma says that in order to get Alibaba going, the first 100 transactions were fake. They actually had to buy back the goods that were being displayed on their website. So, there’s this whole thing about starting a collaborative momentous it were. What was it like with Wikipedia?
Jimmy Wales: Well, I mean the first thing that I say, remember it’s 5.8 million in English, but globally there’s more than 50 million entries. And so, we’re in hundreds of languages, and so that’s a big place to be today. But in the beginning, it was very small. People always ask, “What were the first articles in Wikipedia?” And the truth is some of the early article history as lost, and so we’re not exactly sure. But I do know what the first words were. The first words on Wikipedia were, “Hello world.” I know because I wrote them myself. When I had installed the software and I was testing, I just wrote hello world to see if it would work. In the early days, what was interesting about Wikipedia is that we had a prior project which was not successful called Newpedia which was the same vision of a free encyclopedia for everyone, but I didn’t really understand how to build online communities and understand what communities needed. And so, we had designed a very top-down, old-fashioned system that didn’t work. It didn’t allow people to flourish and to do good things. It lacked trust in the users.. So, once I launched Wikipedia, we got more work done in two weeks than in almost two years. And we had a core community of a couple of hundred people who communally turn in and start building and growing things.
ED: Was that core community, was it constructed? Or was it incentive…?
Jimmy Wales: Yeah, there were no incentives – the only incentive was to share knowledge. You know? I think because we – with the Newpedia project we had gathered a group of volunteers for a couple of years to discuss how could we create an encyclopedia and to think through what does it mean to be neutral and so on. So, once we got the software ready, people were really ready to get working. So, that core initial group was a small number of people, but it started to grow very, very quickly. In part, because it was really fulfilling a gap in the marketplace at the time. People really – I mean you can tell now – they love Wikipedia and it’s one of the most popular websites in the world. They really wanted a concise summary of human knowledge. And so, it grows very quickly. IN the early days, I remember I used to create a report that the community could look at, and it was called, What Google Likes, and it would see what Google was sending traffic to to see what was popular on the site. And also, what kind of searches where people searching for and finding us. So, people would find things like, “Oh, wow,” anything related to World War 2 is always a popular topic in history. And they would say, “OH well, look, World War 2 things are popular, let’s build more World War 2 things.” Or they would say, “Oh, we see that you know we have five of the nine planets – well there were nine back then because it’s eight now –
ED: Nine, nine. And back to nine again. So, like really.
Jimmy Wales: And they were eager to grow, and it was interesting because it was a group of volunteers, but they took very seriously the idea that if we wanted to be successful, we’d need to write things that people found interesting.
ED: So, that was one, and how much of it was technology? As in –
Jimmy Wales: Almost none. Almost no technology.
ED: The Wikipedia, the wiki, you know?
Jimmy Wales: So, the concept of wiki, meaning a website that anyone can edit, it had actually already been around for about six years. There’s a great programmer, lovely guy, called Ward Cunningham who invented wiki back in 1995. Then in 2001, is when I started Wikipedia, tried to apply the Wiki idea, this really open editing system, to the encyclopedia. Of course, even today, the amount of technology involved in Wikipedia is really very tiny. We have very minimal use of algorithms and things like that. There’s a little bit, but what little bit we have is mostly supportive and experimental more than anything.
ED: So, what went wrong with Wiki Tribune?
Jimmy Wales: Nothings gone wrong with Wiki Tribune.
ED: Okay, in that here, you were taking it one more notch in that you were covering news and credible news. So, and participatory news, you know, it’s almost like street journalism online. But you did have a setback in that the formal editing process was not able to capture –
Jimmy Wales: Sure, basically, what I talk about is kind of amusing is: Before Wikipedia, I had the project called Newpedia which was very top-down, editor-in-chief, you know, people reviewing things, and nothing being published until it went through this process. And I started Wiki Tribune and I made the same mistake again. I hired a very senior editor. There was a review process. The problem with that model is it is very intimidating for volunteers. You’re going to submit your work to a very senior editor who is going to judge whether it is good or not is very hard for people as opposed to collaborating in the community. But I still think that pure citizen journalism is not enough. That citizens can do a lot in the area of journalism and can produce quality work. We see this at Wikipedia, but you actually do need professional journalism. Although you need them for different things than what a lot of journalists think. A lot of journalists think that a lot of ordinary news judgement. But I think sometimes they have better news judgement because they know what they’re actually interested in, not what takes place. But they’re not good at going out and doing interviews. They may be a little shy about that. They’re really good at desk research. So, we’re feeling our way forward, and right now we’re in a mode of building that community. Giving them that power and going wiki and saying, “Let’s figure this out together.” And it’s working. The community’s growing. We’re continually updating our software.
ED: So, would it grow as quickly as Wikipedia did or…?
Jimmy Wales: It’s very hard to say, very hard to say. Of course, I would dream about that, but I don’t think it’s very likely. Although, remember you know, Wikipedia grew quickly, but not as quickly as certain things like Facebook, which is much younger than Facebook. You know, it is a process of time involving a community. When you have genuine community, where people are meeting each other and learning to trust each other and learning to judge each other’s work, and so on. It takes time for those friendships and those bonds to –
ED: Right, so in the time that you have been building Wikipedia who did you leave behind, who are your peers, and who’s running faster than you? Like how do you pace yourself?
Jimmy Wales: I don’t know. You know, it’s a funny thing because at Wikipedia – like my way of doing things is I don’t think about competition. At Wikipedia, in fact in all of my work, I try to do things that are interesting, and then if other people like them, I think that’s great. I’m happy if they like them, but we don’t think about competitors, we don’t think about being ahead or behind other people, and in some ways, Wikipedia is so unique –
ED: But at the same time, Google stands a chance to overshadow Wikipedia if they structured their content a lot more in all the nice ways, you know, although every Google search brings you back to Wikipedia.
Jimmy Wales: Yeah, I don’t really worry about that. The thing is, to do the thing that we do, it really requires this kind of spirit of community. And it’s very hard for any other organization to replicate that, and, you know, I just don’t think in those terms. And if they do, and if they do a better job than we do. I think they deserve to win, but there seems to be very little chance of that. Google has no interest in trying to do what we do. I mean for one thing, it’s a terrible business, so why not just let us do it?
Connecting with the Community
ED: Let me just zero in on this, the commercial process. Why didn’t it lure you? Why didn’t it tempt you?
Jimmy Wales: Well, I do have a for-profit site, Wikia, which is now being rebranded as Fandom, which is a series of fan wikis, so Game of Thrones wiki, House of Cards wiki, Fortnite wiki, and all of those wikis. And those are advertising supported, and, you know, it’s a very successful company, it’s about number 20 on the internet, and so, people should not worry about me. I’m not gonna starve. But, for Wikipedia I just think it’s a very special place. My vision for Wikipedia has always been for it to be a temple for the mind. A place to go think and learn and reflect. And I think it’s important for our mission that Wikipedia is driven fundamentally by supporting the community to share knowledge rather than fundamentally looking at ad revenue or paid user or metrics like that, which I think would distract us and lead us down a bad path. I mean, there are a lot of internet companies – google and Facebook – who think of themselves as very global, and they are focused on every person on the planet. But in terms of their actual commercial investment, obviously, they make a lot more investment in getting the next million viewers in California than the next million viewers in Nigeria simply because the ad revenue is so much better.
ED: Right, and then it becomes screwed up that way, I mean we it was screwed in the other direction.
Jimmy Wales: Yeah, of course, so, although I think they do a good job for idealistic reasons of supporting smaller languages and that, and so I applaud them and that. It’s not fundamental to their mission in the way that at Wikipedia it is fundamental to our mission.
ED: When I think about businesses that have taken over this collaborative platform. GitHub was bought by Microsoft recently. Is it inevitable that once you have that ecosystem that someone commercial is going to be interested in you?
Jimmy Wales: I don’t know. Maybe? Again, I don’t think so much about commercial versus non-commercial. I don’t find that problematic. I mean, if Microsoft does a good job of running GitHub. GitHub has never been a charity; it has never been an idealistic place. It is a tool – a very useful tool with a big community around it. And that’s fantastic. There’s lots of ways they could screw that up, but it would be foolish if they did. Companies make mistakes all the time. But I think it’s hard to say.
ED: The reason I’m very impressed and very intrigued by what you’ve built is that you stand on one extreme of the continuum from commercialization and community. And today you have a whole range of institutions banking, finance, which is where I come from – large institutions trying to harness the digital to survive. But they have thousands of employees they have thousands of processes. They have an institutionalized approach. And on top of that they are product-centric. They have something to sell, you know? Whereas you are now the benchmark on the other extreme saying that there’s a world where you create a community, you create something that is commercially viable, and yet it’s, you know, it’s collaborative in that sense. And in the middle, all these other elements, like you’ve got APIs for example. Should APIs be institution-centric or community-centric? In other words, it’s driven by the user rather than the producer. It’s user-driven rather than provider driven and, in that sense, where are you in your thinking about commercialization and building an institution? Do you think we are now heading toward a world where the formal institutional structure is not able to carry the digital?
Jimmy Wales: Well, I mean, I think one of the things that a lot of organizations need to think about is transforming the business to be more open can benefit you in ways that can be very difficult to predict, which can be very surprising. So, let’s say that you have some sort of an API to interact with the financial service or something like this. And generally speaking, most banks, for example, they don’t really open up those APIs to third parties in a very minimal way, with very heavy-laden partnerships. But you can imagine if someone said, “Look, here’s the API. Here’s the login protocol.” They become as open as a lot of other traditional internet services and then people can build apps on top of that and build interesting things on top of that. I mean a lot of that is happening, of course, but a lot of times I think it’s happening in a very kind of narrow way. And what this means is we’re not seeing the kind of innovation that we could potentially see, and I do think that some organizations, financial services of all kinds, and then we can look beyond financial services and kinds of organizations. They are beginning to recognize that, “Gee, it’s incredibly powerful to be a platform.” And so rather than thinking of ourselves as a narrow, vertical constitution that does this type of business, and we have partners to say, “Actually, let’s be a platform where all kinds of interesting things can take place on that platform.” That’s a very sort of high-level philosophical. What does that mean in practice? That’s going to vary in context from situations.
ED: I’m sure you get requests for collaborations from institutions. You know, businesses coming to you saying, “Can I create an API with like …?” You know?
Jimmy Wales: Not so much.
ED: Not so much? What sort of collaborations do you have with businesses?
Jimmy Wales: Not very many. So, with Wikipedia we have some collaborations with galleries, libraries, archives, museums, so those kinds of places. So, people who have huge public archives of photos for example. That are in the public domain, they’re not under copyright, and the institution that is managing them is a public service mission. Then many of them are beginning to realize, “Well, if our mission is public service, we need to be on Wikipedia, we need to be partnered with Wikipedia because that is the public which is accessing culture to that.” So, we do have a lot of those kinds of partnerships. But in terms of other sort of, tech partnerships and things like that, it is very minimal.
ED: Where is technology taking you? Where you on PHP before? You know, like, you have you’re right on media wiki, and is it calling technology moving away from the service side to the client side?
A Platform Based on Collaboration
ED: Right, so you mentioned platform and that’s a phase in itself, which is a platform economy. You have the Facebooks and so on. And not so much Wikipedia but your peers – Facebook, Google – there’s been a kind of move towards personalization, increased ownership both in terms of security and curating it for the person in that way. So, as that evolves, do you think that this platform approach is got a lifespan still?
Jimmy Wales: Well, I do, but I think we’re probably seeing a little bit of a backlash against that algorithmic curation, because it brings in a lot of complicated problems, right? So, with Wikipedia we have very little of that. I know they were showing me on the app there’s a little bit of suggesting articles to people based on what other people have read, and we’re looking at that a little. But because our objective is not to maximize page views or maximize time on site, because our business model is so different, and that’s probably worth pausing for a second to explain what I mean. For Wikipedia, we are a charity, but even so, we have a business model, so to speak, and that business model is to ask people to donate money to support Wikipedia. In order to get someone to donate they have to love Wikipedia. They have to say, “This is improving my life.” And it’s a meaningful thing to me. If our business model, or if you’re Facebook for example, your business model is total ad impressions. Ad page use. So, its’ about targeting the ad, it’s about getting people to stay on the site longer. And for that the algorithm is pushing you toward keeping you on the site as long as possible, which means that algorithmically showing you tempting things to keep you going is really important. But also, people seem to be having a bit of a backlash against it. There’s a sense that what keeps me clicking right now is kind of appealing to my reptilian brain, but my higher mind, that’s not how I want to spend my time. I would actually rather be doing something more thoughtful, more reflective. And so, we’re seeing some backlash that those, that curation and personalization isn’t doing something that is helpful for me. It is doing something to maximize my time on the site.
ED: This is very interesting because it seems the Chinese have taken curation to another level, which is curator around your entire life cycle, lifestyle. You know, payments, purchases, bookings, on a daily, and the Western platforms have consciously avoided that for entire cycle, and still avoiding it if they can, but I think Facebook, WhatsApp is coming in step-by-step, trying to put a payment module into it and so on. Do you think that this platform model could have been richer if – and why was it avoided, why would the major players so simplistic in their approach?
Jimmy Wales: It’s a very good question. I don’t know the answer to that question. I think that it’s a very interesting question, and I think some of the problems that we see which come from pure advertising business model, could be alleviated to some extent if there were other sources of revenue. So, for example, Amazon. Amazon’s website is designed with a different goal in mind. The goal of Amazon’s websites is not to keep me there all day long, every day. If I’m just viewing things on the site is useless to them. They only make money if they sell me things. And so, their optimization is about showing me things that I want to buy, which makes sense because it’s a store. But it’s a subtly different thing. Nobody says that Amazon is making me depressed in life because I sit on this site for six hours a day, and I’ve wasted my time. They say, “Wow, Amazon’s great because I get on there and I buy and I check out very quickly, and then the thing comes the next day, and it’s wonderful” Or maybe they’re buying more than they should because they’re showing them things they want to buy, and then there’s there that whole consumerism thing. So, one of the things that I’ve found interesting is to think about some of the surprising impacts where this is your particular business model, but then that drives you on a certain direction that might or might not be were you were going to go.
The role of humans in the age of technology
ED: Right. What is the prospect for AI for example? Is that going to automate a little bit more of your volunteer driven?
Jimmy Wales: I think we’ll see a little bit more AI. It’s impossible to deny that there have been some great advances and as process and power gets cheaper, it gets more feasible to do things. On the other hand, the idea that AI’s going to write Wikipedia entries is very far off. 30 years off. You know, it’s that is one of the highest forms of human activity. Particularly when you really understand it. If you have not done it, you know, people who aren’t maybe journalists or publishers or writers of some kind, they sometimes don’t realize the amount of complexity involved. But once you understand that it isn’t just coughing up some random facts, it’s actually thinking about what does my reader know? What do they understand? What do I need to explain? What are the most important highlights here? What are the things that I need to include or not include? Now, you’re talking very high-level cognition that is far beyond any AI that we have today. Even as we move into things like driverless cars, but if you really think about the process of driving, you know, you’ve got machine vision – that’s an impressive achievement. But it’s basically: here’s the car, don’t crash into anything. It’s sort of a bunch of parameters around that. It’s not easy, I don’t mean to demean that work, but it’s nothing like writing an original Wikipedia entry or even more writing an original news story. So, I think we’re pretty far away. At the same time, I think there’s some really exciting and interesting concepts around the scale of Wikipedia and the kinds of things that might happen. So, for example, machine translation is still pretty terrible, but it’s a lot better than it used to be. And it gets to the point where you can now imagine – and we’re doing some of this and it’s becoming more popular with community because finally it starts to be usable. But if we think about a language like Tamil, which has over 100,000 entries in Tamil Wikipedia. Some of those entries will not be in English Wikipedia. It will be about something local. A local temple. A local politician. And they don’t exist in English Wikipedia. So, now if we can imagine a machine translation from Tamil into English that won’t be perfect but if it’s pretty good and it enables the community to take something and quickly to fix a few small problems, that’s an amazing tool for transferring knowledge. And the reason I go in that direction is everybody thinks, “Oh, now you can translate English into Tamil.” But I think this is too simplistic, right? And it’s also too simplistic because I wouldn’t be in favor of that sort of mass translation because your reader is different, their context, what they need is different. The cultural assumptions and biases from English speakers that they haven’t thought of certain things. It’s all there, and so we need to be a bit more thoughtful on that. But you can also imagine a tool. Let’s say, I would love this tool. Suppose I could take the Arabic Wikipedia entry on Yasar Arafat and the Hebrew Wikipedia entry on Yasar Arafat and translate them both in to English and compare them. But that’s a really, that’s a really interesting. And I could say – but of course right now that translation is such a brute force. It’s raw. It’s not very good. You can’t really understand the nuance. Then the other thing is, I like the idea of with AI of things like: You can easily imagine things like if in one place Wikipedia says, “This person was born in 1934.” And somewhere else it says, “This person was born in 1935.” A machine could easily detect that this is a contradiction. But I think we can go even a bit more than that to say, “Here are two sentences that are in tension with each other. They could be both true, but they don’t seem like they’re compatible.” And a computer cannot understand that. But I think we have a level now of linguistic analysis and that you could flag that up to a human to say, “Gee, we seem to be saying something completely different over here. Let’s take a second look.”
ED: So, that begs the last most intriguing question, the whole question of fake news, right? Yuval Harari says that if 10,000 people believe a piece of news for one week that’s fake news, but if a million people believe a piece of fake news for several generation that becomes a religion. So, the new ones versus fake news, you know?
Jimmy Wales: Yeah, it’s a tough one. And I think again here that algorithms can help. Algorithms can spot certain signifiers that something might be fake news, and I’m not talking about anything super sophisticated. If I’m going to share something on Facebook and if Facebook detects this is a website with only one page on it, it was registered as a domain name a week ago. Yes, it looks like a news site, but a news site with only one article that was registered a week ago? And seven people have flagged it as fake news, that’s enough of a signal to maybe warn me and say, “Ooh, we have some signals that this might be fake news. You might want to check before you share it with your friends.” And I think that’s where algorithms would help, but they can’t really decide nuanced things. I mean, Donald Trump calls any story that he doesn’t like fake news, right? It’s nuanced. And sometimes he calls something fake news, which when I look at it, and I read it, I say, “You know what, he’s wrong. It’s not fake news, but it is biased against him. Like it did misquote him. This wasn’t a great story. I can see why he’s upset about it.” A more traditional president would have said, “Unfortunately, the CNN article missed some of the key elements of my speech and actually portrayed my view incorrectly, and it was biased against me, and here’s my explanation.” And Trump just goes, “Ah, fake news.”
ED: Yeah, so instead self-imposed out of bounds for you that it’s the judgement part that you would avoid getting into or you would want to contribute to?
Jimmy Wales: I think we still need humans at that level, because I don’t think I’ve seen anything that impressive around algorithms being able to be able to judge what’s true and false in the big picture sense. In the place where we’re having problems in society.
ED: The interesting thing about the model created in Wikipedia is that it’s not so much the cloud of the technology, it’s the cloud of the community.
Jimmy Wales: For sure. 100%. I always say – I’ve said for years going way back to the early days – that stories about Wikipedia would frequently appear in the technology section of the newspaper. And I always said, “This isn’t really correct. We should be in the culture section of the newspaper, because, yes of course we use technology and it would be impossible without the internet and databases and modern technology.” But the true story of Wikipedia is a story about people and values. It’s a story about thoughtfulness and kindness and working for a better world.
ED: So, when you think of yourself as the founder of Wikipedia what is the self-perception in terms of achievement? Like, “Wow, I’ve got this role now.” What is that role?
Jimmy Wales:I mean, I think my role is quite symbolic. I try to represent the community in that way. I’ve sometimes joked in the past I’m like the queen in the UK. No actual power but I wave at crowds a lot and say certain things. And so, I really do think for me the most important thing about the whole world of Wikipedia is about the values that we started with and being true to those of openness, collaboration, of quality, of thoughtfulness, neutrality – really striving to do that and making sure that the community doesn’t get overrun by trolls or commercial interests or whatever it might be.
ED: Jimmy Wales, thank you so much for sharing with us this whole scope. The interesting thing is that this interview was really meant for institutional people, and help thinking about their business and the way they’ve shared your story. And in the way that you’ve shared your story there is an institutional angle to it, and there is an opportunity for profit. It’s just that what you’ve done is you’ve kept the profit element out of the community element and maybe that’s…
Jimmy Wales:Yeah, I mean I think some of the principles of openness and collaboration apply everywhere, whether you’re trying to do a non-profit encyclopedia or trying to grow a local wine store, whatever it might be