Loading summary
Transcript131 lines
- [00:02]
A
Ford Bluecruise Hands free Highway driving takes the work out of being behind the wheel, allowing you to relax and reconnect while also staying in control. Enjoy the drive in Bluecruise enabled vehicles like the F150 Explorer and Mustang Mach E. Available feature on equipped vehicles Terms apply. Does not replace safe driving. See Ford.com BlueCruise for more details. And so one of my employees, Jeremy Rosenfeld, showed me the idea of a wiki and I was like, this is very interesting wiki. Like people can just edit and get started. And we were afraid that the academics and people who we had brought in, you know, as volunteers to work on Wikipedia would find this crazy and offensive to have like wide open page, anybody can edit. So we like, oh, we won't open it as Nupedia. We'll do it as a side project and just call it Wikipedia. And it'll be a scratch pad where people can then as articles get good, they can maybe put them into Nupedia. But it was really evident very, very quickly, like we had more work done in a month than we had in almost two years. And I was like, okay, hold on. This is actually the way forward.
- [01:23]
B
Jimmy Wales, thanks for coming on to talk to us today.
- [01:27]
A
Thanks for having me.
- [01:28]
B
Well, we are here because you have a book out called the Seven Rules of Trust, A blueprint for building things that last. I want to start with rule number five. Your mother was right, because I have a suspicion that maybe this is a motivation for why the book. Now, your mother was right is essentially the argument that you feel like maybe we've lost the, that sense of basic politeness in these days. Was that a motivation for the book and for doing it now?
- [02:02]
A
Yeah, definitely. Although I guess I should say I don't think we've forgotten it universally. I think people are still people and basically everybody's pretty nice. Not everybody, but mostly. But yet our public discourse has deteriorated and become quite toxic. So even though people, you know, if you get into an elevator with lots of people, you're going to find they're all basically decent, nice people online there's a lot of hostility and in politics and so forth. So, yeah, let's remind ourselves as someone.
- [02:41]
B
That has dealt with the difference between how people behave online versus how they behave one on one for decades now, have you ever come to a conclusion about what makes people behave differently when they're on the Internet?
- [02:57]
A
Well, I mean, I think some of the, you know, the oldest elements, if the interaction people are having feels very atomistic to them. That is it's a one off encounter and they're anonymous and they're never going to see you again and so on. Then people don't tend to behave in a proper way or they are more likely to behave in an improper way. Whereas if people have a consistent identity over time, it doesn't have to be their real life identity, it's a pseudonym maybe, and they build a reputation, then they tend to behave much better. And so that's one of the elements. Other elements have to do with.
- [03:41]
B
To.
- [03:42]
A
The extent that we see algorithms that are promoting outrage and bad behavior, people are very social and they'll follow the norms of whatever is set before them. And so if they see the way we deal with something here is to scream at each other, then they tend to do that. And if they see, actually if I did that, people are going to not hear me and I'm going to get sort of booted out quite quickly. Then they are like, oh okay, this isn't that kind of place, I'm going to be nice.
- [04:13]
B
So it's a combination of, on the one hand, if I'm mean to you online, I think I can get away with it scot free with no repercussions. But then conversely, the way the incentives work on the Internet, being mean to you online might actually boost my reputation in a way. Online?
- [04:29]
A
Yeah, yeah. I mean I wouldn't say on the Internet or online, I would say in certain spaces online. So definitely if you act like a jerk in Wikipedia, you're probably going to get yourself blocked quite quickly. And if you act like a jerk on some social media platforms, you're going to get a lot of followers quite quickly. So there you are.
- [04:53]
B
I want to go all the way back. I feel like you grew up in a house of books. So I'm curious If you think 10 year old Jimmy would be impressed that Jimmy has written a book that's coming out.
- [05:05]
A
Yeah, probably so. It's actually really interesting. I don't know exactly how old I was when I read this, but I remember reading it when I was young in it was a, in Reader's Digest, I'm pretty sure is where I read it. And it was a quote from someone who said most people who say they want to write a book actually want to have written a book, in other words. And I sort of have always taken that to heart. I from time to time feel like I want to write a book, but actually I probably want to have written a book. So I'm very happy to have written a book because now I can go out and talk about it. And the actual process I did enjoy ultimately, but it was, you know, it's a, it's a pretty heavy thing to sit down and a blank screen and start writing.
- [05:50]
B
You grew up in Alabama?
- [05:51]
A
Yes, I did. Huntsville, Alabama, which is, I think probably most people are vaguely aware, but Huntsville is where the Space and Rocket center is located. So after World War II, they brought over the German rocket scientists to work on the space program. So when I was a little kid, we actually lived close enough to where they would test the Saturn V rockets that sometimes the windows would rattle in the house. So it was all a very exciting technology aspect of growing up that all this was starting to happen.
- [06:28]
B
But what did your parents do? Were they were scientists or.
- [06:32]
A
No, no, no. My, my mother's at that time was a school teacher and my father was a grocery store manager. Mom later went back to school to become a pharmacist. But yeah, when I was growing up, when I was little, they were, you know, school teacher and grocery store manager.
- [06:52]
B
Well, and tell me about house of learning.
- [06:55]
A
House of learning. So my mother and my grandmother set up a small private school which was, I always say it's so sometimes it's reported that it was a Montessori school. It wasn't technically a Montessori school, but they were very influenced by the ideas of Montessori. But it was very similar to what you might call a one room schoolhouse, if you think Abraham Lincoln, that sort of thing. So we had actually two rooms, but we had, you know, first through fourth grade in one room and fifth through eighth in the other room. So that was kind of interesting because there were four kids in my class, but the, the four kids in my age year, but there were 16ish, you know, vary from year to year. And. But what that meant was you, you did hear the lessons of the older kids. And my grandmother had this saying. She would say each one teach one. So actually the older kids would sometimes hold a little lesson for the younger and that sort of thing. So it was a very different kind of learning experience from most.
- [08:01]
B
Not top down, which might be relevant more sort of crowd learning A little bit.
- [08:08]
A
A little bit. I mean, I, I think my grandmother would bristle about it not being top down. She was in charge of things. But yeah, but you know, a lot of it we. Because when the teacher would be working with the, say that the older kids, then the younger kids would be doing our own work. Work and you know, a little bit of free time and a little bit of, you could sort of spend your time doing whatever you wanted to. And there were things like we had our. Our math workbooks and, you know, we would have a workbook for each year. And we came to understand that if you hurried up and got finished with all your math, you could stop a month or two early. So sometimes we would just go crazy and just do math a lot every day. So that was kind of fun.
- [08:54]
B
So this is lore that has been reported many times, so you can correct if you want, but the lore is that your mother bought World Book Encyclopedia, a set from a door to door salesman and you fell in love.
- [09:08]
A
Yeah, yeah, yeah. So we had the. We had the World Book Encyclopedias. And I would say the. The particular edition, I believe was copyright 1966, the year I was born. So she bought it, you know, when I was. Before I could read. And we would. Every year they would send out the annual update and sort of, you would. You get the. In the annual update, sometimes they would just have this and that and the other. But sometimes they would rewrite an entire article and then they would give you these stickers. So you would. I remember next to Moon, you would go and put a sticker in the original encyclopedia saying, oh, there's an Update in the 1974 annual. You know, so we did that as a kid. And I always joke that was my first editing the encyclopedia.
- [09:58]
B
Right. So to be clear, if it was the original edition was copyright 66 by LIKE 1970. The sticker might be, hey, by the way, people landed on the moon. Sort of.
- [10:09]
A
Exactly.
- [10:09]
B
Add that in. Yeah.
- [10:11]
A
And they would rewrite the article. And then, you know, it's like the sticker would say, oh, you better go read the 1970 version.
- [10:18]
B
One more sort of background thing. I love to ask this question. And you can pick one or the other. Either the first computer you remember using or the first computer that you felt like was yours. And I want nerdy make and model services.
- [10:33]
A
Oh, yeah, yeah, yeah, yeah. So my. My uncle had a music store selling guitars and keyboards and what have you. And next door to his store was a Radio Shack. And the Radio shack had the TRS 80. My uncle eventually opened a computer store because he was way into. He was an electrical engineer by training, so he was way into tech. So he opened a computer store. But the first computer we had was a TRS 80, which was quite, quite famous, I would say, from Radio Shack. It's sort of, you know, we had. You could program in basic and he taught me a little bit of programming in BASIC language. First Computer I thought was actually mine. That's actually a very interesting question. Nobody's ever asked me that. I mean we had. After the, you know, after the TRS 80, my uncle sold Commodore computers. We had that at home. He sold Eagle Computers, which was a short lived brand of computers. This is all before the IBM PC came out. I was actually really. That was part of what led me to become an entrepreneur was it was kind of exciting reading about these new companies like Eagle Computers. Unfortunately, the guy, they did an IPO and he accidentally drove his Ferrari off a cliff. So quite a sad ending to that story. But those were all family computers. I didn't feel like they were mine per se. So probably it was when I went to college and I must have gotten some kind of a computer. So I don't really remember now. That's actually very interesting. Probably by that time a PC. IBM PC, Right.
- [12:19]
B
But maybe not a laptop because laptops were like $6,000.
- [12:24]
A
Yeah, yeah, yeah. It wouldn't have been a laptop. It would have been a very basic, simple, you know, sort of Windows 3.1 computer or something like that. I'm going to give this some thought, but I, you know, I don't think I'll be able to think of it while we're on the, on the call here.
- [12:39]
B
Well, but you know, if you're old enough to remember, taking a computer to college meant taking a screen, the box that sat on the floor.
- [12:49]
A
Yeah. Well, I mean, the first, when I was first in college, I would have used, I was living at home, so I would have used the computers we had at home. And actually I remember being quite put off by. So at the University of Alabama in Huntsville where I was going to college, the first computer class I took, which I was quite excited to do, it was an actually absurd thing. We had to learn the Hollerith punch card code.
- [13:18]
B
Oh.
- [13:19]
A
And we had a test on that. And then we actually had to program. We would have to write programs on punch card readers and then submit a batch job. And even at that time, like I'm like, this is ridiculous. Like I have a PC at home. Like, why are they even. I mean they should. It was, it was actually, it was offensive as a student that, you know, like paying all this money for a school and this is clearly not, you know, and people would protest and the professor would be like, no, no, no, if you're going into this as a profession, you have to know this. And I'm like, don't think so. And of course you didn't have to.
- [13:51]
B
Know This, I mean, I'm old enough. I went to film school and they still taught us to cut actual film with razors and stuff. Like. So it's still like, well, this is. You need the basics of computing, even if it's moving beyond it.
- [14:04]
A
Exactly.
- [14:05]
B
What did you go to college to study?
- [14:08]
A
Well, so I. Partly because I found the computer science curriculum to be lacking at the time I chose to study finance, I was actually very interested in the stock market and finance and entrepreneurship and all of that. And so I studied finance and actually went on to get a master's degree in finance from the University of Alabama. And then I also went into a PhD program in finance, which I never finished. But, yeah, I was. I was very much into mathematical finance.
- [14:41]
B
Did the math of markets and like, I don't know, the psychology of incentives or like, did that kind of get lodged in your brain from studying that?
- [14:53]
A
Yeah, I would say a fair amount. I mean, certainly I sort of. In graduate school, I studied game theory quite a lot and was very interested in both, you know, sort of as a mathematical modeling tool. It's quite cool and quite interesting and there's a lot of fun things there. But also just in terms of beginning to think about incentives and, you know, in a more sophisticated way than I probably would have otherwise. And I was very influenced by. Not something that was actually part of the curriculum, but actually there was a paper by Friedrich Hayek in American Economic Review, 1945, so quite an old paper, but where he talked about the flow of information and how the price system is an information system. And he talked about how if you're trying to organize society, and at this time, this was a very live debate as to whether a centrally planned economy could be more efficient. Efficient than a market economy, because instead of the haphazard nonsense of the market, you would have very smart people in a centrally located place who would make all the planning decisions. And he argued that actually the market provides all the information you need and is communicated in a very efficient way. And so nobody has to say, gee, how much bread do we need here and there. It's just the price system says, oh, look, bread's bread prices are higher than the cost of the inputs. Let's make more bread. And so that was influential, and not in a direct analogy, because Wikipedia is not a market in that sense. But that idea of saying, gee, would an encyclopedia be better if we just centralized all the really smart people in one building and sort of broadcast out to the world, or would it be better if the decision making were made at the end points, at the level of the article, and the readers can participate in all of that. So this analogy, it's not really an exact thing, but that type of thinking about decentralization and the power of, you know, information at the end points was quite influential in my thinking.
- [17:04]
B
You mentioned that you did not finish the PhD in finance. You didn't go to Wall Street. Wall street, but essentially you go to finance.
- [17:14]
A
Yeah, yeah, I went to Chicago and worked as a futures and options trader. So it's sort of like going to Wall street. And, and that was quite fun. It was very, very cool stuff that we were doing there. And it was using my math background and programming and all of that to trade. It was good fun. I liked it.
- [17:37]
B
Might come back to that in a second. But let me ask you one similar to the computer question. Do you remember the first time you saw or used either the web or the Internet?
- [17:45]
A
Yeah, so there's a couple of things. So my high school, which was in a very tech oriented town, Huntsville, somebody, a parent I suppose, donated a PDP 11, which is like a mini computer, so like a mini mainframe and about 10 terminals. And so that was the first time that, you know, that kind of networked computing and we could send each other email on the system and to make things and send files to each other and all that kind of stuff. So that was the first experience. It wasn't the wider Internet, obviously, then later I think I was at Auburn University when I think this was before they were hooked up to the Internet or maybe they were, but I just wasn't aware of it. But there was something called Bitnet, which was a store and forward network. So basically you could send emails and then they would go out in a batch job overnight or something and emails would come in. So that was the first time I was like, oh, communicating with people in other places and all of that. So I do remember it very much.
- [18:57]
B
Somebody else that started in finance and then the web happens and they're like, oh my God, I can't miss this excitement is Jeff Bezos. And Bezos big idea is an everything store. Your big idea is everything knowledge. But I'm curious if the same thing happened to you that happened to Jeff, which was, you see all this excitement and energy and you're like, I need to be over there.
- [19:24]
A
Yeah, definitely, definitely. I, I remember. So I, because I'm just a geek. I was the day that Netscape went public and I think it was worth $4.3 billion on the first day of trading. And I was as a hobby. So in. In Chicago, the markets close at two because they follow New York time because it's, you know, and the markets close at 3 in New York or they did back then. And so I would get off fairly early and I would go home and I would spend a few hours. I had no life whatsoever because I was new in town, didn't really know anybody. And so I was writing a web browser because I was very interested in the Internet and technology. And when Netscape went public, I was like, you know what? My homemade web browser is not as good as Netscape 1.0, but it's not $4.3 billion worse. You know, So I thought, actually there's something really interesting going on here and it is very exciting. And so I was very interested to sort of think about, oh, okay, like, actually I should. I should get involved in this.
- [20:33]
B
So does that lead to your first startup?
- [20:38]
A
Essentially? Yeah. Yeah. So basically, a friend and I, Tim Shell and I started sort of noodling over different ideas. And so we. We did something called Loop Lunch, which was, you could order your lunch online. Oh, boy.
- [20:55]
B
You're the first one to have had that idea 30 different times.
- [20:59]
A
Yeah, yeah, exactly. But at the time, it was like I would go into a restaurant to try to convince them to sign up, and they literally had no idea what I was talking about. Like, the Internet. Like, what are you talking about? Is that like a fax machine and so on. And so we were way too early. And actually our technology was very, very simple. You know, it was sort of making web pages and so forth. It was quite terrible.
- [21:22]
B
But.
- [21:24]
A
And then we, you know, we tried this and that. So in order to try to get popular, we were like, okay, we need to try something. And we saw webrings was a. Was a thing that was popular. So we made a webring type of website. It was like a search engine web directory kind of thing, and started doing that. And sort of, I would say the innovation of that, that was possibly the surprising piece that it does foreshadow Wikipedia a little bit, is we thought, oh, well, we can't just sit here and categorize. I mean, I remember we made a web ring about Jupyter or something like that. It's like, this would take forever.
- [22:03]
B
We need people describe more specifically what a webring means.
- [22:08]
A
So it's sort of like basically in our context, what it was is a category page, like if you remember a Yahoo Web directory. So you go to a category, like, you would drill down through a hierarchy and say, oh, Right again.
- [22:26]
B
Interrupt for context. Yahoo. Was not a search engine at first. It was a directory with actual humans.
- [22:32]
A
Yeah, yeah, yeah, Actual humans. Yeah, exactly. So you would go to a category space and then a subdirectory. Planets. Subdirectory Jupiter and there'd be some links about Jupiter. So we did that. But then the idea of a ring is when you clicked on a link, it would open up in a frame and you could go next, next, next, next and go through the, the, the ten links there you would go in a circle so you could quickly look at a bunch of different sites. It was very primitive back then. But what, what we found was that, oh, this is going to take forever. So we opened it up so anybody could create an account and create a webring. So that slightly foreshadows Wikipedia. And it also gave us a little bit of a learning that was helpful with Wikipedia because at the time if you opened the webring about Jupyter, you owned it and it was yours. And if you did a terrible job or if you were a spammer, nobody could correct what you did and so on and so forth. So we thought about, can we open it up? Can we let multiple people do it? We never really got very far with that, but that idea of like that experience. So when with Wikipedia, when people said, oh well, obviously we're going to need to have an editor in chief of this section and who's going to be in charge of this article. I was like, maybe that would be bad because if one person is in charge of anything, they might do a bad job, they might wander off and get bored and so on and so forth. So we let people build whatever they wanted and so that, you know, it was moderately successful.
- [24:07]
B
What was the name of that company, by the way?
- [24:09]
A
Bomas, that are men in suits.
- [24:12]
B
Wasn't there also a. There was a browser component to Bomus. Right.
- [24:19]
A
We may have at some point done something. No, not so much. I mean actually what was actually very interesting was at one point it was Mozilla web browser went open source. And so we did have the idea of taking the browser and adopting it for our purposes and sort of changing it to do whatever, but we never got very far with that.
- [24:49]
B
But Bamis is one of those.com era companies that is basically banner ad supported.
- [24:56]
A
Yeah, yeah, yeah, very simple. You know, we had banner ads and actually that was quite lucrative for a short period of time because we. Well actually we got, we didn't get bought by NBC. So at the, at a certain point in time, NBC Television Network in the US had they thought they were going to become an Internet thing. And so they had NBC, a portal competing with Yahoo and so on and, and search and all that. And they were going to do news and do every, you know, a portal like whatever. And so we, they contracted with us and we moved our domain onto nbci so it was bamas.nbci.com and that meant they got to count our traffic towards theirs. But they didn't buy the website. But we gave them all the ad space and they sold the ad space and they paid us quite well for that. And then they blew up and it was all a big disaster. So we, we're kind of glad we hadn't sold. Although our future wasn't that good either after that because the Internet was changing very, very quickly. And that concept of a web directory was really blown away by search because once search got good, you didn't need people putting links into categories, people just typed into a search engine, you know.
- [26:18]
B
But Bamis was successful enough for a certain period of time that if this encyclopedia idea in the back of your head is still banging around like Bamis can fund noodling on that.
- [26:32]
A
Yeah, yeah, we, so we had, you know, I didn't know anything about venture capital, I didn't know anybody, so I couldn't go into Silicon Valley and raise money. Maybe I should have tried harder, but I just didn't because I was too shy and didn't know what to do. But we, yeah, we, for a while we were, we were profitable enough to, you know, start Nupedia. So Nupedia was the predecessor to Wikipedia and Nupedia was very top down. You know, it was, you had a seven stage review process to get anything published and had an editor in chief and we had, you know, all of these things like a traditional encyclopedia. It wasn't ultimately successful because, and the main reason was it wasn't very fun. So at some point, after getting frustrated at the very slow progress, even though we had, you know, quite a few volunteers came and signed up and were like, oh yes, I want to help with this. It just was taking forever. And so I decided, okay, look, I have, I did all the coursework for the PhD in finance, didn't finish, didn't write the dissertation and I had studied option pricing theory and I knew very well the work of Robert Merton who had just recently won the Nobel Prize in Economics for his work in option pricing theory. And I thought, okay, I could write a short entry about him. And when I went to do it, I was like, it was like massive writer's block. It was very intimidating because they were going to take my article and send it to the most prestigious finance professors they could find. I hadn't been in academia for a few years by this point and it was just like really intimidating, like, I don't know if I can do this. And then I realized this isn't fun. Nobody's going to do this. It's too hard to participate. So then we started looking around for like, how do we make this easier, how do we make it more open? And so forth. And so one of my employees, Jeremy Rosenfeld, showed me the idea of a wiki and I was like, this is very interesting wiki. People can just edit and get started. And we were afraid that the academics and people who we had brought in, you know, as volunteers to work on Wikipedia would find this crazy and offensive to have like wide open page, anybody can edit. So we like, oh, we won't open it as Nupedia. We'll do it as a side project and just call it Wikipedia. And it'll be a scratch pad where people can then as articles get good, they can maybe put them in, into Nupedia. But it was really evident very, very quickly, like we have more work done in a month than we had in almost two years. And I was like, okay, hold on, this is actually the way forward.
- [29:12]
B
Two things on that. First of all, you said it was it wasn't fun and it wasn't fun for someone to contribute. But also I'm sure it wasn't fun to manage. So it's one of those sort of both sides of the market need to, need to be getting something out of it, the juice of doing it.
- [29:32]
A
Yeah, it wasn't fun. And also it became evident that actually despite having this very complex seven stage review process in a very old fashioned way. So one of the things that happened is one of the earliest articles that we published. So we're, you know, to great fanfare within our tiny little community, we're like, ah, it's made it through and it's done. The copy editing has gone through the review and all of that published it. And within a few days somebody said, oh, by the way, this is plagiarized and we had to retract it. And that wasn't caught by academics doing review. It wasn't caught by the copy editing, it just wasn't caught. And it was caught though by, you know, in open source software there's a saying, given enough eyeballs, all bugs are shallow. Meaning if you get a Lot of people looking at something, errors can be spotted and so forth. And so it was like, oh, actually this traditional top down process is deeply flawed. That actually would have been better if we had had a couple hundred people read it and actually scour over it and check it. So that was another impetus to say actually maybe this wiki thing could be better than a traditional process. Right.
- [30:46]
B
Because the second point is scale, where if you are going to be all of human knowledge collected and especially all of human knowledge collected in a reasonable amount of time, the scaling problem couldn't be solved any other way.
- [31:03]
A
Yeah, I mean, definitely. Even when we were working on the Nupedia model, we were struggling with those ideas. It's like, well actually we're going to need a lot of people and how do we organize people? And at the time, of course, we weren't thinking in a very open, fluid wiki way. We were thinking an editor in chief of each section and would they volunteer because that's maybe not a very rewarding job and so on and so forth. And so ultimately we, yeah, we, we decided that was never going to scale and never going to work. And actually reminiscent a little bit of how the web directory didn't scale. Like you can't hire enough people to categorize every website on the Internet. Jack Daniels has proudly served in fine establishments, questionable joints and everywhere in between. So no matter where you go in.
- [31:58]
B
Every bar, you'll always know someone by name.
- [32:02]
A
Jack. Jack and Coke.
- [32:03]
B
Shot of Jack.
- [32:04]
A
Jack Daniels please. Right away. That's what makes Jack Jack.
- [32:09]
B
Please drink responsibly. Responsibility.org Jack Daniels and Old Number 7 are registered trademarks. Tennessee whiskey, 40% alcohol by volume. Jack Daniel Distillery, Lynchburg, Tennessee. Some people think nature is like this, but actually it's like this. That's why Columbia engineers everything we make for anything nature can throw at you. Columbia engineered for whatever. So I think Nupedia launches in March of 2000. If. Is that right?
- [32:38]
A
You think that sounds about right? I'm not sure.
- [32:42]
B
I mean I think we were working.
- [32:44]
A
On it before then.
- [32:45]
B
But yeah, if it is right, I mean that's the exact time when the NASDAQ hits its height and the bubbles bursting. And so my question is sort of extrapolating on the timeline here where Nupedia doesn't work because it doesn't work because it's not structured correctly. But also the bubble's bursting. And so you said you were making good money for a while at bomis, but that might be going away. So how do you. I'd Love to know, sort of the behind the scenes of, hey, this other thing is getting traction, but maybe we're running out of money. How did that all play out?
- [33:21]
A
Yeah, so yeah, I mean, it's actually very, very relevant because I think had we had, had I gone out and raised money or had we made, had some source of money, I think the natural instinct at that time, which is actually still kind of a natural instinct in a lot of social media platforms and so on, is if you have some sort of a social problem on the website, you just need to go out and hire a lot of community managers. And if we had done that, if we had had access to lots of money, we might have said, oh, actually what we need is we need somebody who's the editor in chief of the science area. We need this, we need that. And we would have hired people and probably cut off any possibility of the innovations that happened. But instead, because we had no money and no access to money, we actually had to say, hold on, we actually have to figure out how to empower the volunteers to do quality control, to sort of manage themselves and all of that and all of the details about how the software works and all of that were conditioned on that. Well, we can't just hire community managers to run everything. Like that's never going to scale and we have no money, so we can't scale. And so that was sort of necessity as a mother of invention kind of thing.
- [34:40]
B
Is it Nilay Patel that says for any social network, the actual products. Maybe it's not Neelay, the actual product is the management of the community. And so this is going to start bleeding into what you talk about in the book. So like, what you're describing is you sort of accidentally discover that, hey, this sort of open source crowd, whatever sort of thing is the only way that this idea can function. But then also it still could have failed if you didn't get the core norms right of the community, of things like referencing the book like assume good faith and things like that. So where did that idea come from again? Was it just necessity being the mother of invention or where the key part of your product was the structure around the community?
- [35:36]
A
Yeah, I mean, it's really hard to say because we don't have any alternate histories to review and compare and do an A B test and all of that. But I will definitely say some of the core values definitely played a huge role in making things happen. So things like assume good faith, things like no personal attacks, things like Wikipedia is an encyclopedia, not, you know, anything else. And There were some key decisions made. So as an example, at one point somebody said, oh, well, we shouldn't just have one article on each topic. We, we could have articles from many different points of view. So we might have the, you know, on. On a controversial topic. We might have, say, the Catholic Church view of abortion and, and this and that. It would be competing views and so forth. And I was like, I don't like that. I think we should have one article. I mean, you can have an article about abortion, those points of view, if it's relevant, but there should be one article, and, and that article should address all those points of view in a fair way. And so that was part of the idea of neutrality. You know, as neutrality is really important if you're going to try to do one article about each thing. And it's really important for two reasons. One is the social reason. Like, if you and I have a real disagreement about some fundamental issue. But we're both nice people, we can still work together as long as what we're trying to create is a fair explanation of the debate that we are. We're having. And that works quite well, actually. Better than people might think, given, you know, if you spent too much time on Twitter, you think, well, that sounds impossible. People just scream at each other. But no, actually there's a lot of nice people who disagree and can reasonably disagree and are happy to sort of write things down in a fair way. And then there's the knowledge reason or the epistemological reason, which is, if I go to an encyclopedia, I don't want to be just told one side of the story. I actually want to understand the debate. I want to know what the best arguments are on both sides and so on like that. So that's where neutrality plays another very helpful role.
- [37:56]
B
We skipped over something that I think is important too. It's sort of the chicken and egg situation where, oh, hey, we've got this encyclopedia with all the world's knowledge except the first people to show up. It's not there. And I know that you came from Nupedia, so maybe you were seeding things like it was a tool for Nupedia, but how did you solve at the very earliest days, the, hey, we're doing this, but we might not have an article on X, Y or Z yet.
- [38:24]
A
Yeah. So first thing to know is I think Nupedia had 24 articles, so that wasn't much to seed anything. And they were on pretty random topics. There was no real org. It was like whatever volunteers had been interested in Writing. And so part of it was that an encyclopedia article can be very, very useful even though the whole product isn't complete, you know, so if you did happen to Google, I don't know, Paris, if we had a decent article on Paris, then that's what you needed and that's all you needed. And there's a lot of funny red links that you click on and it says, sorry, we don't have this yet.
- [39:03]
B
Oh, that's interesting.
- [39:04]
A
The main article is useful even before anything else. And actually that was a real challenge and arguably might still be a real challenge for we have a dictionary project called Wiktionary. The thing about a dictionary is it generally doesn't get to be particularly useful. I mean one definition could still be useful to you, but as a place to turn to, unless it's got pretty obscure words. Because you don't turn to a dictionary to find out what dog means. You already know that. So when do you reach for a dictionary? Well, these days nobody reaches for a dictionary. You just Google and see if you can find a definition. Whereas with an encyclopedia it was useful, you know, even when it wasn't very complete. And so that was, that was a piece of it. But also I think at that time the, the Internet was chaotic and full of a bunch of mess. And it turns out oftentimes what people were looking for is just a basic explanation of something. And so an encyclopedia, there was demand for it that wasn't being met by, you know, at that time there were a lot of, well, I'll give an example. There was a really fantastic. On GeoCities, which was like a free web hosting provider that had lots of pop up ads and blink tag banners. It was quite a garish website. But there was a great site that I actually enjoyed. It was about the Beatles and it was pretty comprehensive. I don't know how many pages, probably 50, 100 pages about the Beatles and written by some guy who was super passionate about it. And the problem was he had written quite a lot and then I guess he probably got bored of it and wasn't maintaining it really. But I mean, how much need to maintain the Beatles? But. And it was actually quite common at that time. You would go to a website like that and you would say, hey, sorry, I haven't updated lately. And I always thought it was funny then because I'm like, I don't think people are coming here every day looking for updates. That's not a blog. Like, it's like you're the only one who knows you haven't updated lately, but. But also, like, it wasn't really sufficient. And pages like that might be popular, but then they don't get updated and they're not new. And then there is news about the Beatles, like, occasionally one of them does something in modern times, and so on and so forth, or a new thing is revealed from the past and so forth. And so there was demand that wasn't being met for a constantly updated encyclopedia. And then we did quickly learn that we would see lots and lots of traffic to things that were in the news. And actually, we used to monitor that when we were very, very young. I had a page that I created not on the encyclopedia, but just I would put it up to help the community, where I would look through our access logs and see the traffic coming in from Google as the referrer, which is a lot harder to do these days because everything's encrypted. So you don't know where people are coming from as easily as you did back then. And I would publish a page called what Google Likes, and I would rank the things that Google liked. Now, that was articles we already had, but it was twofold. People could go to those articles and make sure they were gonna be good and improve them. But also you get a general theme. You're like, gee, seems like people are very interested in World War II. Maybe we should write more about World War II. And probably World War I would be interesting too. And the Civil War, that's a popular topic. So people would use that to go, oh, let's fill in the encyclopedia with whatever people are interested in. So that helped grow as well.
- [42:31]
B
Yeah, it's interesting because, you know, YouTube is constantly, like, suggesting, hey, maybe you should make a video on this. Because our logs are saying, you know, so everything old is new. Again, let me ask this in two different ways. So to bring it back to the book, the introduction of the book, you titled it from a Joke to Global Trust. So do you remember a moment or metric that first made you think, oh, Wikipedia can work. But then also was there a moment when you realized that Wikipedia had crossed that trust threshold?
- [43:05]
A
Yeah, I mean, so the moment where I really thought it could work, there were a couple that I remember. I guess it's sort of a constant discovery, you know, But I remember sort of at. At a certain point realizing, like, oh, like, wow, we. We thought. We hoped we would get to 3,000 articles by the end of the year, and we're at whatever, 15,000. I'm like, wow, this is growing quite quickly. And that's very Very interesting. Another was on September 11, 2001 when the, you know, the World Trade center, the terrorist attacks happened. I sort of, it was amazing to watch on Wikipedia as the community was responding by doing lots of research. So a lot of articles just improved a lot that day about, I don't know, World Trade center and the airlines that were involved that day and, you know, all these kinds of different terrorist groups who on day one were mentioned in the news as being possible suspects. People were going and researching and adding information and that was a very different experience from, you know, if you flicked on the tv, all you saw was the planes flying into the buildings over and over and a lot of talking heads speculating. And I thought, oh, this is interesting. Like the community is. The questions that you might have in your mind. The community is actually researching and like, oh, I heard on the news they said Al Qaeda might be involved. I'm like, who the hell's Al Qaeda? The community was also like, who the hell's Al Qaeda? And they would go and write and sort of add. So I thought, oh, that's very interesting. We can react to news events in a way that's different from, you know, what, what else is out there. And then the second question, which was when I trust because. Trust.
- [44:47]
B
Yeah, because if you, I remember, you know, oh, there was a time where it's like, if you are a college student and you use Wikipedia to research your paper, whatever, you're doing it wrong. Do you remember a period of time when all of a sudden, from academia, from just general, hey, this, this is a real thing. Wikipedia is trustworthy.
- [45:10]
A
Not a particular moment. No, because it sort of slowly has evolved. And you know what I say about this is we were never as bad as they thought we were and we're not as good as they think we are. So as a Wikipedian, I'm still like, God, there's so much work to do and there's so much that's not good enough yet and so on and so forth. So we tend to be quite self critical and so we're not like, oh yeah, Wikipedia is great, you know, yeah, use it for anything. We're like, well, hold on a second. Like, you know, there's still a lot that's not good and really needs improving. So. But yeah, I mean, I would say some of the moments, I will say a moment in the theme of what you're saying was the Sigenthaler incident, as we call it. So there was a bad error in Wikipedia. It turned out somebody had put it in as A prank on his friends and so forth, and had written something very negative about a journalist, John Sigenthaler Sr. And he found it, or a friend showed it to him and it basically said that he was briefly suspected of being involved in the Kennedy assassination. And turns out he's like a close family friend of the Kennedys and was a pallbearer at Bobby Kennedy's funeral and you know, like that. So he found it quite offensive of and rang me up. And from the moment I was told about it, I got off the phone and within 10 minutes it was gone. And fortunately it wasn't really linked to from any other articles in Wikipedia. And I checked, I had very little traffic and so on, but still it was quite bad. And that was a moment not when I realized Wikipedia was trusted. Although to some extent it was like he cared more than he would have had. Wikipedia been just some random idiot's webpage, right? It seemed like it should have more quality and more authority, but it was the first. Not the first, but it was a big moment when there was this realization of like, oh, actually we aren't just a bunch of geeks off in a corner, nobody's paying attention, just working on a little project. Like we're having an impact in the world. Like people are reading this and we have a responsibility to get it right. So we made a lot of policy changes in the aftermath of that to say for, particularly for biographies of living people, you can't write anything negative about someone unless you've got a source, a good quality source. That's a rule in Wikipedia. And if you see something that's negative that doesn't have a source, the correct thing to do is just take it out immediately, then go to the talk page and talk about it. Because what could happen back then was somebody would see something negative, you know, say this person, when they were in their teen years, they, you know, were arrested and then they would go to the talk page and say, is that true? I don't see a source and I Google and I don't find anything. And then people would debate about it and like, no, don't debate about it, take it out and then debate about it and figure out what's true and then put it back in if it's true. And so that change was a result of realizing like, oh, actually we are impacting the world and we have ethical responsibilities.
- [48:25]
B
Wikipedia comes out on Nupedia, which is, for lack of a better term, a dot com era startup. So I'm assuming it's maybe around this Time when do you realize you needed to move from what was essentially a founder controlled startup to a community governed whatever Wikipedia was going to become.
- [48:45]
A
Yeah, so. Well, it was. So the first thing we did, you know, like we needed more servers. The site was growing, it was getting sluggish and I had paid for all the servers up to that point, or the company had.
- [49:03]
B
And there were. And there were no ads, right?
- [49:05]
A
There were no ads. It didn't want to have ads and thought, you know what, actually we need more servers. And frankly putting ads up at that time wouldn't have been, wouldn't have made enough money because it was the dot com crash. Like the ad market was in devastation and you know, there was no way to raise money from venture capitalists to give us more Runway. We needed to buy more servers. And so, and, and there was also like quite a lot of demand from the volunteers and a sense of what our mission was. It's quite academic, it's educational. It really should be in a nonprofit. The volunteers were very much in favor of that. I thought, yeah, that actually makes a lot of sense. And so then, you know, created the nonprofit, donated everything into the nonprofit. And then we did our first fundraiser. Our first fundraiser was to buy Brian Fibber, our main lead developer, a volunteer, but to buy him a laptop because his laptop was terrible and he needed one. And I was like, you know what, we need to get this guy a laptop. So we did a little fundraiser for that. But then we had, on Christmas Day, we had three servers, so a database server and two front end servers. And two of the three servers crashed at the same time on the same day. Like they're broken. There was something with the hard drives or whatever. It was like a disaster on Christmas Day and nobody was near the data center to do anything. So I had to like reconfigure the software so everything's running on one server, which was very slow obviously. And I was like, we have to get more hardware. And so we did our first real fundraiser and hope to raise. I set a goal of $20,000 in a month's time, 30 days. And we raised 30,000. And I think that the 30,000 goal was hit within a couple of weeks. And that was the first time it was like, oh, actually this whole nonprofit thing might actually work out. Like maybe we don't ever have to have asked because we didn't say absolutely never to ads. I'm too sensible to sort of make an absolute like that because I was like, well, we need to survive. And even if we're a non profit. A non profit can have ads. There's no rule against that. But, gee, like, that's. Really don't want to do that. And. But I'm not going to say absolutely not, because what do we need to do to survive? But that was the first time. It was like, oh, actually this is going to work. Like, people can. People are responding. Like, the public likes Wikipedia enough that they'll donate some money. So that was great.
- [51:38]
B
Yeah. Have you learned any lessons about that, about being a nonprofit, asking community for support that like other people in the field of nonprofits or whatever? Like, is there, like some sort of golden rule that you've learned about that, that maybe you could share?
- [51:57]
A
I mean, I think, you know, a lot of it is in the book Seven Rules of Trust.
- [52:04]
B
Yes. Seven Rules of Trust is what's still there.
- [52:07]
A
All right. I do think that's important, like, because you have to say, like, if people are going to give you money, they need to trust you. And if I actually, I remember this was. This was a mistake. This was not correct. But it sort of illustrates kind of some of our thinking. One of the volunteers said when we were going to do a fundraiser, they're like, you're completely insane. People aren't going to just give money unless you tell them exactly what it's going to be spent on. And sort of insisted that we make up the hardware order and publish the exact specs of the servers we're going to buy and ask people to donate to fund that. Well, it turned out after some experimentation over time, because we did that, that was fine. But actually, most people didn't care. They trusted us. They're like, I like, whatever. What do I care about? I know nothing about servers. I assume you're going to spend it wisely. And appeals about the vision of the project caused people to donate more. They were like, oh, actually I believe in this free encyclopedia for everybody. That sounds like a good idea. Let's do that. And so people would contribute.
- [53:13]
B
I'm going to pull this right from the book. You describe early 2000s tug of war with a prankster who's posting graphic images on the homepage of Wikipedia.
- [53:21]
A
Yeah.
- [53:22]
B
I think in that section you say that the solution to that was, like, openness. When you adopt a policy of transparency, you can't be selective about it. So is that part of what we're talking about, like, being radically open so that you can gain trust?
- [53:42]
A
Yeah, yeah. As much as possible. We actually still sometimes have discussions in the community about how the board operates. Right. And so the, The. What we try to do is to be overly transparent as much as possible, but there are obviously issues you have to discuss as a board that are not really suitable for public discussion. So we discuss human rights cases of volunteers working in very dangerous places. It's not really appropriate for us to talk, you know, our internal talk about how we're going to help people, that you can't publish that, or, you know, questions about staff and salaries, like that's just not appropriate to publish in a public way. But most things you can. Most things you can be very, very open about. And sometimes it seems a little tedious and annoying, and that's, you know, but it is what it is. But in our culture, it's very, very important. And, you know, the community is deeply involved in the board. So we have a. The structure is, you know, half of the board roughly, is selected by the community, and the other half are selected by the board for expertise. So, you know, we always have a. We need somebody on the board who is an expert in auditing and the financial statements. Because, like, just because you're a great Wikipedia editor who's very popular in the community and, you know, very trusted, that doesn't mean, you know how to, like, supervise the auditors and things like that. But, you know, the majority of the board is community members, and that's been very, very valuable in terms of keeping the foundation really true to the mission.
- [55:25]
B
Sliding doors sort of thing. If you had chosen the other path, do you ever think about this, like, if you had gone for profit, what do you think Wikipedia would look like today?
- [55:37]
A
It's really hard to say. It's really, you know, I don't know, actually. I mean, I think we would have ads. I think we would probably might do some sort of freemium product where you can pay to get extra something. I don't really know, like, because we didn't go down that path. It's just sort of hard to figure out what. Because, you know, everything is evolutionary as you go along. I mean, I can give one example. At some point, a volunteer suggested because Gmail had just launched. Gmail was very popular and we had been around for a while. So Wikipedia is older than Gmail. But somebody said, oh, maybe we should offer free webmail accounts. And I thought, I mean, that's interesting, but no, I don't think so. Like, I knew enough about managing an email server that I was like, that's going to be a lot of distraction and quite hard. And also, what's it got to do with building an encyclopedia. Whereas if you were a for profit, you might evaluate that in a different way. You might say, ah, well, maybe we could put ads on the email to subsidize not having ads on the encyclopedia. And we might do this and that and the other, and you might just go down a certain path and then you might raise money from venture capitalists and then they might want a return and they would be pressuring all the time. They would need a seat on the board. There'd be pressuring all the time. Like, shouldn't we just put ads in the encyclopedia? And you know, like, you could just go to a different place and that might not be good because, you know, like, Wikipedia is awesome. So hello friends. Guess who? That's right, it is I, the replacer. Once again, I've been called on. So you can play the new Call of Duty Black Ops with three expansive modes, 18 multiplayer maps, and the tastiest zombie gameplay you've ever freaking seen.
- [57:32]
B
Available November 14th pre order Call of Duty Black Ops 7 now rated M for mature. That, that, that leads into sort of questions about today and, and Wikipedia and the age of AI. Because if you were for profit, you know, Reddit has turned around and made a lot of money by like, hey, you're scraping this content. Anyway, do I have a question here other than how are you thinking about Wikipedia as a very valuable source of truth and data in this era where data is like oil?
- [58:12]
A
Yeah, so there's a couple of things. So one, we do feel sort of heavier than ever the responsibility of quality and neutrality because we're part of the infrastructure of the world. And you know, I, I wouldn't be so happy about a world in which an emerging AI technology were trained only on Twitter. Like that sounds like a bad idea. So we do think it's important. And then the so part of it is it's not so much about being a nonprofit, it's more like it's open source, so it's freely licensed, so everyone can take the Wikipedia data and use it. And that's part of our core mission and it has huge benefits to us. And it's really the gift that we're giving to the world is that Wikipedia is not proprietary. So even a non profit could say, no, we own the copyright to this and we're not going to let you use it unless you pay us, whatever. At the same time, we do have our Wikimedia enterprise product because one of the things that we see happening, and I only recently came to understand this in a better way, so we're getting crawled a lot by AI bots and, you know, they could just download the database dumps, that would be fine, but they need the latest information for their training and so on, so they like to crawl. And the issue with that really is, so if you go, so let's say something big, a big thing happens in the news. Queen Elizabeth died and we got. I don't know how many millions of people were going to read the page about Queen Elizabeth, because that's what people do when they hear big news. But that page will be cached in memory because so many people are loading it. And to send that page out costs very, very little. It's just some HTML text and we just blast it out to you and it's very fast and very easy. But if you come and crawl a million pages that haven't been seen recently, they're not going to be cached in memory. We're going to have to build the page and send it to you, or we're going to have to invest in caching more pages in memory and dealing with it. So it turns out those crawls cost us disproportionately. And so what we're trying to do is to encourage all of the commercial operators, who are obviously benefiting heavily from their use of Wikipedia, to say, right, you know what you should do? You should sign up for the enterprise product where you can download through the API and you'll pay for that additional service. And we're getting good uptake from that. Like, it's both a practicality. We try to make the experience better. We haven't yet blocked them, but we could. And, you know, cloudflare has got this whole new product where they're blocking AI crawlers unless they pay and all of that. That's kind of interesting.
- [60:58]
B
Are you tempted by that? You're keeping an eye on it.
- [61:00]
A
Yeah, keeping an eye on it. I mean, I'm speaking for myself, not for the board and not for the organization, but, yeah, of course, that's super interesting because, you know, you just have to say, look, if you're costing us a lot of money and you're benefiting a lot, you should chip in. Right, yeah, that seems fair. And I think that argument is carrying the day. But, you know, it's a matter of making all the deals and all of that, but at the same time, you know, we're not interested in saying, no, you can't use Wikipedia. We want all the AIs to read Wikipedia because we want them to be sensible, and Wikipedia is pretty sensible.
- [61:32]
B
So, yeah, you're happy feeding the AI MA at this point?
- [61:39]
A
Yeah, definitely. I mean, I think our whole goal is to share knowledge. And so we're looking at this as yet a new avenue to share knowledge. Now, the one thing that we would say is our free license. You know, like one of the few conditions on the license is attribution. And so attribution from a legal perspective is complicated matter because facts aren't copyrightable. So if you learn something from Wikipedia and you write it in your own words, like you as a person, you learn something from Wikipedia, you write it in your own words, that's not a copyright violation, and it shouldn't be. Like, we definitely are in an era where some of the, for example, academic journal publishers would probably love to make the argument, if you learn something from an academic journal article, you have to pay us, whether you, you know, in order to tell people about it. Like, that's just not good. Like, facts should not be copyrightable. At the same time, from a moral and ethical point of view, obviously AI should be explaining where they got what they got. You know, if they learn something Wikipedia, they say, according to Wikipedia, this and that, that turns out it's quite a hard technical problem because the AIs generally have no idea where they learn something because they're just probability engines for predicting the next word. But they're working on it. And I do think, you know, to the extent that we can persuade them that actually it's quite beneficial, both because it's the morally right thing to do, but also for the trust in what you're publishing. Because one of the problems that AI has right now is the. It's not trustworthy like you, you know, if you ask AI a question about something obscure, you know, who is so and so. Actually, I, I always have the example of my wife, who is Kate Garvey, because she's not a famous person, but she's known enough that the AI thinks it knows who she is and it makes stuff up like, it's unbelievable, it's quite amusing, but it's always plausible, but wrong. And so that is fun, right? But you just say, but yeah, people know that. And then they think, well, I can't trust this. It might just make something up. And that's a problem. So attribution can help them a lot with that, because if they can say, oh, here's where I got the information, then people will go, oh, okay, either I trust that source or I don't, or whatever.
- [63:57]
B
You mentioned the crawlers, and I've seen Articles about strain on infrastructure and stuff, but is. It's not, you're not concerned yet. It's not like an existential threat that you're worrying about all of this crawling?
- [64:09]
A
No, no. I mean, we do worry about, you know, disintermediation, you might say. So if people are getting knowledge from an AI. Yes, but they don't know it came from Wikipedia, then they won't know they should be donating money to Wikipedia in the way they did in the past. Now, so far we haven't seen that. We haven't seen any really material impact on traffic. And in fact, there was a study.
- [64:37]
B
Not at all. Because, I mean, there are a lot of, you know, the whole Internet in the Google era was built on, hey, do a good piece. And. And people are claiming that their traffic is plummeting, but so far you all haven't. If you're watching this video on YouTube, then you'll notice that we maybe are wearing different clothes because we had an Internet connection issue. And Jimmy was kind enough to come back several days later to finish our discussion. Jimmy, thanks for doing that.
- [65:05]
A
Yeah, you know, Brian, what I always say is just think what amazing things we'll be able to do once the Internet doesn't suck. But here we are.
- [65:14]
B
Funny that you say that because I'm going to lead into something. But first of all, what we were discussing was AI. And just real briefly, there's lots of publishers and sites that are complaining about AI crawlers and it's exploding costs for them. I'm sure you've seen costs go up, but it's not existential or a big enough problem yet.
- [65:35]
A
Yeah, no, I mean, we are getting hit pretty hard with crawlers these days. Lots and lots of bots from the major players and also a lot of minor players. What we're encouraging people to do is to sign up for the enterprise product because that's much easier on us. You know, we have a whole system for giving people access in a much more lightweight way. For us, if you're just crawling the website indiscriminately, that's, you know, it adds a lot to cost. I mean, I only recently came to understand this because you think, well, how much can bots cost compared to normal people? And the answer is the traffic patterns from the general public tend to be very spiky. So for example, when, when the Queen died, millions of people came to to read her biography in Wikipedia, which means they were all looking at the same page, which means you construct the page once you have it in memory and just shoot it out to millions of people. If you come and visit an obscure page, then obviously we have to go to the database and we have to build the page and send it to you. It just costs a lot more money, particularly if you're doing it at scale.
- [66:41]
B
Have you seen human usage overall go down? Because again, there are publishers that are reporting, oh, traffic down 30, 70%. Have you seen something similar with humans?
- [66:52]
A
No, no, we've seen no material change in traffic. I mean, it's down a couple of percentage points, but that's within the range of noise. And we actually think it has to do with the use cases of AI versus the use cases of Wikipedia. People initially think, oh, they must be very similar, but actually, once you start using AI, you realize, actually you use it for all kinds of stuff, that you wouldn't use Wikipedia forever. So I use it, for example, to ask for recipes. I love to cook and I always joke, I fancy myself a pretty good cook. And I wouldn't advise you to use ChatGPT for recipes unless you already know how to cook, because if it hallucinates, you might not realize that's way too much salt. But, yeah, so, yeah, but nobody would go to Wikipedia looking for a recipe. That's just not a thing that we do. So, you know, actually it is very interesting, though. A lot of people are reporting traffic declines, certainly. I saw a headline the other day, apparently Stack Overflow, which is a site that programmers use to sort of you, the classic use cases. You have some bug and you get an error message. So you Google the error message and you end up on Stack Overflow, where people are discussing how to fix that. And now people just go straight to ChatGPT and say, oh, I've got this bug. How do I fix it? And it explains it to you. So their traffic is apparently about 10% of what it was a few years back. So that's like, whoa, that's a big deal. But yeah, for us, we haven't seen any really major impacts.
- [68:27]
B
In a broad way. Do you fear that AI is an existential threat to the open web? That people won't go to things like Stack Overflow and maybe won't need to go to websites, period? Do you have a broad fear about that?
- [68:42]
A
Not really. I think we've had that fear for many, many years. I mean, people were like, oh, Facebook's getting so popular and it's a closed garden and it's going to be the death of the open web and HYP still here and doing fine. You know, I think it's, you know, AI or, or even like, you know, never mind a narrow AI player like OpenAI, but a very broad player like Google, which has all kinds of different things. They no one company, no one organization, no small group of organizations can produce everything on the Internet. I think it doesn't really make any sense and it still remains easier than ever to start your own new project on the web. You know, it's like domain names are really cheap and hosting is really cheap and you know, so there is a whole vibrant world out there that I don't think is going away anytime soon. And indeed, I think we're going to see a lot of new things pop up as some old things die. So, yeah, I'm pretty optimistic.
- [69:41]
B
Well, let's wrap up today by coming back to the book and to another one of the rules. Rule number three was Wikipedia is an encyclopedia, which sounds silly, but the point there was to emphasize discipline in terms of the scope of what you're trying to accomplish. Other folks in this era of like TikTok explainers and hot takes on social media, if how should other knowledge platforms like police their own borders in that sense of scope where something that is posted that could be entertainment, but maybe it could be taken as authoritative and the line's really blurry.
- [70:18]
A
Yeah. So I haven't memorized the rules in order, but I think rule number three probably wasn't Wikipedia as an encyclopedia. It's probably purpose. Got it.
- [70:27]
B
I could have gotten it right.
- [70:28]
A
Having a central purpose. Well, yeah, I mean, our example is that Wikipedia encyclopedia is an example of a clear purpose. And so I do think for a lot of these social media websites, it's actually a very hard problem. So for Wikipedia, it's pretty easy. You know, we're here to build an encyclopedia and if you're not helping with that, then you're not welcome here. And, you know, any proposals? You know, some. Once I remember somebody suggested, when Gmail came out, somebody suggested to me, oh, you should probably have free webmail. And I was like, I don't think so, because that's like a whole big product and a thing to manage. And it's not about building an encyclopedia. And so, no, we just stick to our main mission and we do that. And for social media, it's actually kind of hard sometimes to define what the purpose is. And when you don't have a clear purpose, it's hard to define the boundaries of that. So, you know, one of the things that's easier for Wikipedia when we think about moderation questions is we don't have a little box that effectively says what's on your mind. Because if that's the starting point, then people have literally all kinds of things on their mind, some of them quite horrible and not helpful in any way. So I do think, you know, having a purpose that's a little more narrow could be very helpful in clarifying. So one potential purpose for Twitter would be to say we are the. The public square for discussion of public affairs. And, you know, in a wide sense. And if that's the purpose, then you do begin to see, like, actually a lot of trolling actually isn't beneficial to that. It's just like, oh, we're trying to have a town meeting, and the people sort of off to the side, sort of, I don't know, blasting loud music and things so nobody can hear each other. They're not here to have the town meeting, so maybe that doesn't belong here. But it's hard because it's hard to define what that means. And it's also then very hard if your business model just depends on sheer number of viewers. So then you find it really hard to say no to anyone, even if they are problematic, especially if by being problematic, they're causing other people to respond and engagement and more page views. So it's hard.
- [72:53]
B
Finally, this is all about trust. This book, and I believe that you wrote that rebuilding trust in the US should be treated like a national emergency. But let's flip this around. Instead of talking about what platforms can do, what publishers can do, what would be something that someone listening to us right now could do in their personal lives to sort of nudge their own community towards higher trust?
- [73:18]
A
Well, I mean, I think it depends on who you are and what situation you're in. Certainly, I think one of the things that we should all do, as you might say, consumers of politics, that is. I'm not a politician. I have no plans to be a politician. But I'm a voter, and I do maybe sometimes answer polls, and I talk to friends and family and I think you could put on the agenda. Actually, what I really want is someone I can trust. I think that's really super important. And so, you know, that that almost transcends a lot of questions about specific political views. Somebody trustworthy is going to sort of try and balance multiple views and try to find a good, solid middle ground. You know, I remember saying to people, when it was Obama versus McCain, I thought, I'm happy with the situation. I'm, like, proud of the US Like, I'm going to vote one way, and maybe my person will win and maybe they won't, but they're both proper people and we're going to be fine either way. And, you know, like, we should demand that. We should say, actually, I'm. I'm really unhappy if we're offered candidates who are, you know, like, not keeping their word and then, like, being random and all these kinds of things. That's just one piece of it. And that, you know, that can go down to the local level. You know, don't vote for lunatics. Even if they. You agree with their policies in some way. It actually is more damaging to the policies that you love if you've got somebody erratic and sort of untrustworthy representing you.
- [74:57]
B
Well, once again, the book is called the Seven Rules of Trust, A blueprint for building things that last. Jimmy Wales, thanks for coming on twice.
- [75:06]
A
To tell us about it. Thanks for having me. And sorry about the Internet snafu. I apologize on behalf of the Internet.
- [75:13]
B
The worst interview I ever did in terms of sound quality was with the guy that invented the MP3.
- [75:19]
A
Brilliant. That's hilarious.
- [75:20]
B
All right, thanks.
- [75:21]
A
Okay, thank you. Yep.