Russ Roberts

Wales on Wikipedia

EconTalk Episode with Jimmy Wales
Hosted by Russ Roberts
Zywicki on Debt and Bankruptcy... Klein on Truth, Bias, and Disa...

Jimmy Wales, founder of Wikipedia, talks with EconTalk host Russ Roberts about the birth and growth of Wikipedia. He talks about the role of Hayek's insights into the design of Wikipedia, how Wikipedia deals with controversy, the reliability of Wikipedia relative to traditional reference sources and the future possibilities for projects that rely on voluntary contributions of time and creativity.

Size: 19.1 MB
Right-click or Option-click, and select "Save Link/Target As MP3.

Readings and Links related to this podcast

Podcast Readings
About this week's guest: About ideas and people mentioned in this podcast:


Podcast Highlights
0:36Intro. What was the original conception? Came about in 1999, first had idea of free encyclopedia written by thousands of volunteers in many languages. Came up with Nupedia, first effort, ultimately not successful. Top down project, complex review system, people had to apply and prove they were qualified. Not fun for volunteers, intimidating. In 2001, launch of Wikipedia, very open system; got more work done in two weeks than in previous two years. Why bottom up approach? Complaining: Top down approach was costing a lot of money, was very slow, Wales tried to write an article on option pricing theory and Robert Merton, based on his finance work; faced with writer's block. Thought open site would be forced to be locked down in a week or two, though wanted to try to keep it open. Still very open today.
3:47Work of Hayek influenced the work of Wikipedia. How did that come about and is it true? Read Hayek's "The Use of Knowledge in Society" as an undergraduate; deep impact on thinking. Game-theoretic, rule-making within society, how people interact; basic ideas, e.g., simply making a rule doesn't make anything happen, but just changes the costs and benefits. At the time ongoing debate in economics on the efficacy of a centrally planned economy versus the efficacy of a market economy. Hayek's point in the essay was that this is a question of information: is it more effective to communicate all the information inward to a central authority who will use that information to make a decision, or is it more efficient to leave the information where it is and push decision-making out to the endpoints? Breadmaking example: how much bread to make. Parallel question same for encyclopedia: is it better to gather all the world's information into a group of experts who then make editorial decisions, or to push the decision making out to the endpoints, the people who have the information? That analogy came before or after the fact? After the fact, more about understanding what works or doesn't work in Wikipedia and why. Proposals people make: Is this an effort to centralize something that can be left decentralized? Don't want a bottle-neck of failure.
7:37How much centralization in Wikipedia now? Anybody can edit it; what role is there for infrastructure, if any, to oversee the whole process? All community-driven; within the community people are elected to be administrations or to the arbitration committee, though in theory you could be elected to the arbitration committee without being an admin. Wales has a role in the process making decisions on when elections are held and final appointments to arbitration community, though traditionally done by following the vote. What role does the arbitration committee play? Deal with disputes that others have been unable to resolve. Last stop in that process. Hear cases in a similar way to the Supreme Court--no lower courts but there are other dispute resolution processes. Very aware of precedent: can set a precedent, can in their findings put forth a principle people will point to in the future so they don't have to go to the arbitration committee. How big is it? Total of 15, but all volunteers so if people are away, runs 9-11. In English Wikipedia about 1000 active administrators, 1622 total, some not particularly active. When started, knowledge of the Oxford English Dictionary? No, read The Professor and the Madman on the subject only several years later. Crowd-sourcing, thousands of individuals contributed to project. Government structures evolved over time; policy pages changed when a document is not actually being done, descriptive not prescriptive; evolved for historical reasons, not attempting a rationalization. Americans have harder time getting it than Britons--where customs and traditions are not always written down. Hayek: discover norms and customs to make laws and to rule, more British. What's been the biggest challenge? Fund-raising. Never had any serious problems but have to take it seriously.
14:04Out of pocket costs: how many people are using Wikipedia at a point in time? In any given month about 280 million people using the site, staff of only 25 people; everyone else, including Wales, is volunteer. Cost of bandwidth, servers, things like that. Bulk of cost technology related. Is pace of new material on site growing at an increasing or slowing pace? Slowing pace, at least in English. We've figured everything out! How many articles will it eventually have? Don't answer because eventually any answer will sound funny. 2.7 million entries in English, there is a limit, not going to have 35 entries on Colin Powell, will have one. In Wikipedia, haven't checked recently; don't have an article on every member of Parliament in India, about 600 people. Example of something we wish we had but we don't. Long way to go. Do the articles get longer? Yes, over time, to a certain point. Feeling in the community that they shouldn't become too long so that they are not unusable. If article becomes too long, break it out: China, broken with side articles like Human Rights in China. Are insiders, administrators, making those decisions? Yes, but nothing about that process that requires administrators to do. Anyone can create a new article and move some of those things over. What is cultural impact of Wikipedia? Daughter, age 16 wants to know if it can be used for high school history paper? What is acceptability of Wikipedia as a source and is it changing? Don't think it should be used as a cited source. In college, using The Britannica as a source is unacceptable, not the right role for a research project; instead should be going to original sources to write a college-level paper. Separate issue from the reliability of Wikipedia. Younger students, a 10-year old writing for class, we should just be glad the kid is writing and giving credit. Reliability: study done three years ago, out of date, comparing articles from Wikipedia to articles from Britannica. Sent the articles to experts, how many errors? About 3 errors per article in Britannica and about 4 errors per article in Wikipedia. Within striking distance. Surprised people that there were 3 errors per article in Britannica. Doing a good quality reference work is hard. Articles may have been cherry-picked for study. Focused on much more scientific topics where Wikipedia may have had strength. Not a lot of research into the quality of Wikipedia. Would find that Wikipedia's quality is pretty decent. Would love to see research helping understand better where Wikipedia is flawed, where are weaknesses, strengths. Astounding cultural impact, number 3 for most anything you type in; research warranted.
23:59With 280 million users, some people are getting some value. Whole issue of accuracy is wrong issue. Milton Friedman's entry on Wikipedia has an error, suggesting the phrase "Inflation is everywhere and always a monetary phenomenon" came from the Monetary History of the United States but that's probably not true. Came from a lecture he gave in India in 1963. But that kind of error is not really very important. Could get you a red mark on a research paper in high school but the deeper thing is that a written encyclopedia has only one expert's view, so if acolyte praise-worthy article, if antagonist, very insulting. Britannica on economists is limited to facts, very little discussion on why particular economists were important. Wikipedia dwarfs Britannica on that score. Neutrality; in general not going to see an article about an important figure that is extremely biased in one direction or the other. Bigger problem of bias of non-neutrality on less controversial subjects than on more controversial subjects. Friedman an important economist and controversial; therefore have a lot of people who are interested and who either like him or don't like him. Obscure topics end up being written by someone who either really hates somebody or are fans; more difficult time. Because there aren't enough eyeballs on each side? Not just numbers, but something like that. Obscure corners of Japanese anime written and read only by anime fans, so no critical eye. A few years ago problem since resolved: two minority religions in U.S.: Church of Scientologists, and Church of Latter Day Saints, the Mormons. The Scientologists famously have been very litigious, lawsuits with people revealing their secrets. Entries on their articles tend to be very good, vetted by supporters and proponents. Mailing list post: reading about someone in the Mormon religion that said someone was a prophet. People writing those entries weren't intending to be biased, just wrote what they knew. Added clarifications like "according to the Book of Mormon, so and so is a prophet.
30:38Clay Shirky podcast: suggested that Wikipedia is the product of about 100 million hours of human effort; watching TV about 2000 Wikipedias a year. Any way to harness that time for other projects? Happening already in a number of ways. As people turn to the internet, they read and become participants. Expect to see a lot of this. Staggering amount, 200 billion hours of television a year. Nice to imagine if 1% of that time, 20 Wikipedia's worth, could be used productively. Wikipedia-like folks are all a bunch of geeks, but TV-watchers are on average less geeky; more likely to watch sports. A few hours here and there. Can successful crowd-source projects help open people's minds about bottom-up economics rather than top-down? Wikipedia and culture: We've thought of it as commercial culture, like pop music, movies, most television, which has produced fantastic things like the Beatles, as well as rubbish; and also as the fine arts, which we think of as something that has to be subsidized either by the government or wealthy patrons, like the opera. Wikipedia is very different: folk culture, ideas spread in grass roots decentralized way, can be harnessed, coordinated in a voluntary fashion. Not a market system but like a market.
36:08Hard to imagine a better ride, extraordinary achievement. What can you possibly do next? Emotional part. Astounding, hits the most when traveling. Dominican Republic, building technology centers in poorest areas; kids living in shacks use Wikipedia there every day, wanted pictures; really rewarding. Hope that internet will bring learning and knowledge. Now: Wikia--building the rest of the library. About 12,000 communities from political activism to humor to computer gaming; expand citizen participation model. Question and answer site experimentation, tweaking that. Cell phone carriers. Unlike other founders where there is a successor issue, there is no path to follow here. Self-sustaining. Still have role. Colonel Sanders of Kentucky Fried Chicken, real person turned into a cartoon after death--Wales's destiny.

Comments and Sharing

TWITTER: Follow Russ Roberts @EconTalker

COMMENTS (15 to date)
Ajay writes:

While Wikipedia certainly benefits from Hayekian dispersed knowledge, it is more accurately a product of current internet economics. It is built to gather as much data as possible, much of it low-grade, and to involve as many people as possible, who the Wikipedia backers then hope will donate to keep the site running. In this, it most closely resembles the communal model of open source, down to the license they use, the GFDL, which is an offshoot of the GNU project that produced the most widely-used open source license, the GPL. Both projects are built on convincing contributors that they're either taking part in some altruistic mission or to donate intellectual work that they might have created for their own benefit anyway. As a result, both are somewhat successful first experiments but ultimately evolutionary dead-ends, as they cannot compete against a real market, where work is valued and where there are real institutions devoted to evolving quality. With Wikipedia, there is nothing stopping competing encyclopedias from harvesting its facts and sources, since there is essentially no copyright protection for such factual data, and creating better online works, with real organizational processes to pay for work and improve quality. Inevitably, this future will not look like the expert-dominated encyclopedia or creative work of the past: it will be much more inclusive of the work of random contributors but the anarchic Wikipedia model goes too far towards the extreme of inclusivity. The ideal model is somewhere in between the extremes of an expert-dominated, rigid hierarchy and communal anarchy. When that optimal hybrid is implemented, Wikipedia will wither and fade away like all first experiments.

Johan Sigfrids writes:

There was a popularized course in astronomy open to everyone at the university just after Christmas, which I immediately jumped on it. A great chance to bother a real astronomer with all those weird questions that keep me up at night. One interesting thing he said was that if we found a contradiction between Wikipedia and some other book or site on astronomy, believe Wikipedia. Apparently the astronomy pages on Wikipedia are all written by real astronomers and are amongst the most accurate and up-to-date sources for modern astronomy you can find.

Mads Lindstrøm writes:

Ajay, you write:

"As a result, both are somewhat successful first experiments but ultimately evolutionary dead-ends, as they cannot compete against a real market, where work is valued and where there are real institutions devoted to evolving quality."

But both projects are already competing successfully against commercial projects. Wikipedia for about five years, open source software for 10+ years. Why do you think that open source software or Wikipedia, has so far been able to compete successfully against closed source / proprietary players ? Why do think that this will change in the future ?

I have heard the it-is-just-an-experiment argument since the late nineties, and while I disagreed, I could understand why people thought so back then. But now, 2009, still using the it-is-just-an-experiment argument seems quite week.

John Barry writes:

Excellent podcasts, all. I just subscribed (March 9, 2009) and have been listening while at the gym. I teach economics, and this is superb stuff. JB

Ajay writes:

Mads, A social experiment can be prolonged for a relatively long time by the lack of better alternatives or the fervor of its believers, I believe that's what happened here. As for their success, open source isn't that successful on the desktop, it's only somewhat successful on the server. I think that it will be replaced by a better alternative that I came up with. Wikipedia is no doubt a success online, but that's only because it has no real competitors and because people believe in the idea and keep funding it. I used to donate 4-5 years ago but never would now, even though I still use it. I think we will see a working micropayments system deployed in the coming years. Once information can be monetized with micropayments, I think many encyclopedia competitors will spring up and Wikipedia will not be able to compete. Wikipedia is an artifact of the current, incomplete technology market but its organizational process cannot survive in a real market.

John Strong writes:

The final pronunciation of "Wikipedia" may well emerge, in true Hayekian fashion, and become something different from what Jim Wales "planned". However, you should know that it is a common practice in the engineering community to install a "Wiki" on the company LAN with information like work procedures, coding standards, etc., and we pronounce it like Mr. Wales, not like George Mason economists. Just a nit. :-)

Gregory Kohs writes:

How ironic that Russ Roberts and/or Jimmy Wales would label Wales as "founder" of Wikipedia, the viewpoint of the central authority (Wales himself). In contrast, the distributed endpoints (the rest of humanity) view Wales as "co-founder" of Wikipedia, since we haven't unceremoniously stuffed Dr. Larry Sanger down the ol' memory hole.


NormD writes:

Russ, I think you are too naive about Wikipedia. Anytime something becomes an authoritative source of information on a topic, there will be attempts by zealots, political operatives, marketing staff, etc. to control it. The Wikipedia method of controlling such bias is to have zealots on both sides of an issue conduct edit wars making and undoing each other's edits. If one sides zealots are less zealous (or have a life) then the other side "wins".

Among other annoying features is that Wikipedia changes, so you might go there and find some very useful information and then return later and the information is gone for any of number of reasons. You can try to search the history, but that can take some effort.

Don't get me wrong, I use and love Wikipedia for many things, but I am also deeply frustrated by it. I know many people that have simply given up on it.

Adam writes:

My jaw dropped to the floor while he was describing the arbitration process. Wikipedia has evolved into a system of customary law in less than ten years! Extraordinary! Scholars of Law and Economics should be tripping over themselves to document and study this phenomenon! Where are they?

BoscoH writes:

As perhaps the strongest skeptic here of the Eric Raymond open source story, I think many of the above commenters are being a bit harsh on Wikipedia. Perhaps the difference only boils down to humility. Mr. Wales knows the shortcomings of his creation. I thought it was kinda funny at the end of the podcast when Russ was trying to get him just say "yeah, Wikipedia is pretty cool.". Or perhaps the difference is that Raymond comes off as a spokesman for a whole bunch of people who never appointed him.

Or perhaps it's how participatory Wikipedia actually is versus how participatory open source software actually is. I know regular people who use and even edit Wikipedia, just as I know regular people who blog or comment on blogs. And to augment John Strong's example, I've deployed a Wiki for non-technical people and it's been pretty useful for capturing the evolving business knowledge of a small but widely geographically distributed organization. I don't know any regular people who work on open source software. The non-programmers I know who tout open source software (like GIMP) tend to be pretty nerdy.

I'd say if you want to understand the difference between open source software (in the predominant GPL model) and Wikipedia, watch the first few episodes of Dancing with the Stars this season. Think of Steve Wozniak as "open source". Think of everyone else on the show as Wikipedia. Open source software is just a little bit too out of step to be as relevant as it thinks it is.

Seth writes:

Great podcast. I would be interested to hear Mr. Wales response to Eric Raymond's assertion in the January 19 podcast that open source doesn't work so well on something like Wikipedia because the metric for success isn't objective. It seems like 280,000,000 people per month think Wikipedia is good enough.

I like the idea for Wikianswers. I think Wikidebate would be good too.

Regarding Eric Raymond, on my blog I posted why I disagree about needing an objective measure for success for open source to work well, ( I'm interested in thoughts on that.

tw writes:


I share some of SteveD's concerns, especially given the news story earlier this week that pro-Obama zealots are removing anything negative from his Wikipedia entry.

Listening to the podcast, I was struck with the idea that instead of an "I say what the facts are" mentality, Wikipedia has a "We say what the facts are" mentality....where the "we" is still a very small group, though larger than the "I". And along those lines, I'm not sure your analogy to bottom-up economics holds.

In any event, I did enjoy the discussion about the comparison of the number of errors between Wikipedia and Encyclopedia Brittanica entries, and would suggest you post a link to that study, if it's available. That would be fun to read.

Bruce writes:

Although I've never entered an article on Wiki (and given the following, am unlikely to do so), Mr. Wales claims about neutrality and bias sound a bit flat.

Have a look at "Global Warming". There is an obvious political bias to my eye. Look at the talk page; there are folks who try to present the topic in a more balanced manner, but they're overruled by admins.

Every time I hear "the debate is over" (when did it occur?) or there is a consensus (was Galileo wrong too?) despite the fact that the globe has been cooling for ten years as CO2 levels continue to rise (results not predicted by any of the models), it becomes clearer that the issue is about politics and has very little to do with science.

/dev/cpu writes:

BoscoH, you wrote

"Open source software is just a little bit too out
of step to be as relevant as it thinks it is."

Pls checkout the platforms on which most Web 2.0 is built. It is mostly opensource. If you are not a part of the community, you will NOT understand or appreciate it.

Haboush writes:

I have to say that I am a big user of Wikipedia, as many are, and grateful for its existence.

But people should be aware that there are about 200 individuals or so that dominate the editing and administration of the site, and not always in a completely benign and harmless manner. Wikipedia is not as "free" and "open" as it and the 200 or so hardcore editors would have you believe. is related to Patrick Byrne's (CEO of project at and it reveals how Wikipedia, and other social media sites are used and abused by everyone from hedge fund short sellers to nutty, power hungry Wikipedia editor/administrators.

Here's a link to antisocialmedia's stuff on Wikipedia:

Comments for this podcast episode have been closed
Return to top