Jimmy Wales and Wikipedia

Abstract

Wikipedia launched on January 15, 2001, as a side project to support a failing peer-reviewed online encyclopedia. Within its first month it had produced more articles than its parent managed in its entire existence. Within three years, it had surpassed every printed encyclopedia in scale. The English edition passed seven million articles in May 2025, one of over 340 language editions, maintained by unpaid volunteers and reaching more than 1.5 billion unique devices per month. Jimmy Wales co-founded it with Larry Sanger, a partnership that ended in dispute and became one of the most contested founding stories in internet history. What Wikipedia demonstrated, against nearly every prediction, was that collective intelligence on a massive scale could be more reliable than controlled curation.

Jimmy Wales, September 2015. Image: cellanr, CC BY-SA 2.0, via Wikimedia Commons.

Before Wikipedia: Bomis and Nupedia

Jimmy Donal Wales was born in Huntsville, Alabama, on August 8, 1966 (his family long celebrated August 7; Wales confirmed the certificate date in 2011 by sending Britannica a scan of his passport). His mother and grandmother ran a small private school, and his early education was unconventional, less structured than a standard curriculum, with more reading, more discussion, more self-direction. His childhood fascination with encyclopedias was not merely academic; he has described spending hours reading entries in the family’s World Book Encyclopedia, following the threads of interest from one subject to the next, frustrated when the edition was incomplete or out of date. The desire to connect all knowledge was personal before it was organizational.

Wales studied finance at Auburn University and the University of Alabama, then pursued a doctoral program in finance at Indiana University. He left in 1994 without completing his dissertation. The reasons were practical: the internet was becoming visible, and a finance trader in Chicago (where he moved to work at Chicago Options Associates) could see the commercial possibilities more clearly from there than from a graduate seminar. He made money trading interest-rate and currency options and began thinking about what to build.

In 1996, Wales co-founded Bomis with Tim Shell and Michael Davis, a web directory and search company based first in Chicago, then relocating to San Diego. Bomis organized web content into topical directories, ran web rings, and maintained a network of fan sites. It was an early-internet business built on the same category logic as Yahoo. Its most profitable vertical, which Wales has acknowledged without pride, was adult content. Bomis’s “Babe Engine” (a directory of images of women) generated the revenue that funded everything else. It was an unglamorous foundation for an encyclopedia.

Nupedia went online on March 9, 2000, funded out of Bomis revenue: an online encyclopedia modeled on academic peer review. Wales hired Larry Sanger, then finishing a philosophy PhD at Ohio State (awarded in 2000), as editor-in-chief. Nupedia’s editorial model was rigorous: every article went through a seven-stage process of assignment, finding a lead reviewer, lead review, open review, lead copyediting, open copyediting, and final approval and markup. The quality would be high. Only credentialed experts would contribute.

The first full year of operation produced twenty-one articles. Nupedia was, by every practical measure, not working.

The Wiki Suggestion

On January 2, 2001, Sanger had dinner with programmer Ben Kovitz in San Diego. The conversation turned to the web and collaborative tools. Kovitz described WikiWikiWeb, the original wiki, created by software engineer Ward Cunningham in 1995 for the Portland Pattern Repository, a collaborative website for software design patterns. Cunningham’s key insight had been radical in its simplicity: if you make editing a page as easy as reading it, and if you trust that a community of interested participants will fix errors faster than they accumulate, you can produce useful collaborative documents without any formal editorial process.

Sanger called Wales the next morning. He proposed using wiki software as a feeder mechanism for Nupedia: an informal, open space where articles could be drafted quickly by anyone, and the best of them submitted to Nupedia’s formal review process after reaching acceptable quality.

Wales agreed. Wikipedia launched on January 15, 2001. Sanger picked the name off a long list: a portmanteau of “wiki” (a word Cunningham had taken from the Hawaiian for “quick”) and “encyclopedia.” He later called it “a silly name for what was at first a very silly project.” The article count passed one thousand on February 12, 2001, four weeks after launch, and ten thousand by September 7. Every week brought more contributors, more articles, more cross-links.

Info

Ward Cunningham created WikiWikiWeb in 1995 as a tool for documenting software design patterns, not for encyclopedias, not for the general public, not at scale. His original software was spare: pages could be created and edited by anyone who visited the site, changes were logged, and the most recent version was what you saw. Cunningham called it “the simplest online database that could possibly work.” Wales and Sanger recognized that this minimal mechanism could scale in ways Cunningham had not attempted. Cunningham himself became an active Wikipedia contributor and has been consistently generous in crediting what the project made of his invention.

Exponential Growth

The two projects coexisted for about a year, and the comparison was not kind to Nupedia. By the end of December 2001, Wikipedia held roughly nineteen thousand English articles. Nupedia would finish its whole run with twenty-four approved ones. The academic peer review model, with its requirement for expert engagement at every stage, could not compete with open editing’s ability to aggregate small contributions from many people. A retired teacher who knew a great deal about the history of railroads in rural England could write a stub article in twenty minutes. Under Nupedia’s process, that article would have required the recruitment of a credentialed expert and months of review.

Nupedia ended without ceremony: its server crashed on September 26, 2003, and nobody restored it. Twenty-four approved articles and about 150 drafts were left; the finished ones went to Wikipedia. By that point, Wikipedia had passed one hundred thousand articles (January 22, 2003) and the outcome was not in doubt.

The growth curve continued. The English Wikipedia passed one million articles on March 1, 2006, with an entry on Jordanhill railway station. By 2007, it was the largest reference work ever assembled in any language in any medium. The German, French, Japanese, and Polish editions were each larger than the Encyclopaedia Britannica. The scope expanded to include topics that print encyclopedias had never had space for: every minor character in every television series, every small municipality in every country, every notable (by Wikipedia’s own contested definition of “notable”) event and object and person. The first commercial casualty was not the print sets but Microsoft’s CD-ROM encyclopedia Encarta, which held about 62,000 curated articles and could not answer a free encyclopedia that had millions and grew by the hour; Microsoft shut it down in 2009.

The Wikimedia Foundation and the Economics of Free

On June 20, 2003, Wales incorporated the Wikimedia Foundation as a nonprofit in St. Petersburg, Florida, to own and operate Wikipedia’s infrastructure. The governance decision was critical: by placing Wikipedia inside a nonprofit rather than Bomis or any other commercial entity, Wales ensured the encyclopedia could not be sold, monetized through advertising, or restructured by investors. Wikipedia would be funded by donations or not at all.

The annual fundraising campaigns (featuring banners displaying Wales’s own photograph and handwritten-style appeals for small donations) became one of the internet’s most recognizable recurring phenomena. The campaigns reliably raised enough to cover server infrastructure and a staff of several hundred. Foundation revenue was $154.7 million in fiscal 2021/22 and $185.4 million in 2023/24, against roughly 650 employees. By comparison, the company that Wales co-founded alongside it, Wikia (later renamed Fandom), became a for-profit wiki-hosting platform and grew substantially through advertising revenue.

The decision not to carry advertising on Wikipedia has been consistently cited by researchers and readers as central to the site’s credibility. A source with no financial incentive to manipulate your beliefs is, all else equal, more trustworthy than one with such an incentive. All else was not equal (Wikipedia has other reliability challenges) but the absence of advertising was a structural choice that shaped how the site was perceived.

Dead End: Nupedia and the Expert-Verification Model

Warning

Nupedia’s failure (and the success of its amateur alternative) established a counterintuitive principle in collaborative information systems: rigorous curation at the point of creation is outcompeted, at scale, by mass editing with post-publication correction. This is not because expert knowledge is useless; Wikipedia’s best technical articles frequently match or exceed print encyclopedia quality precisely because domain experts do edit them, unburdened by a formal process. An expert can fix a specific paragraph in Wikipedia in five minutes without committing to a seven-stage review. The same expert, under Nupedia’s model, was asked to own an entire article through every stage of production. Wikipedia harvested expert knowledge incidentally and iteratively; Nupedia required it explicitly and upfront. The requirement made it impossible. The lesson applies broadly: participation systems that minimize the marginal cost of contribution, even at the expense of some curation, outproduce systems that maximize individual contribution quality.

The Founding Dispute

Bomis, squeezed by the dot-com bust, laid Sanger off; he resigned his Nupedia and Wikipedia roles on March 1, 2002. His subsequent accounts of the project’s founding diverged sharply from Wales’s. Wales’s public narrative described Wikipedia primarily as his own idea; Sanger appeared as the hired executor of someone else’s vision. Sanger, who had independently proposed the wiki mechanism, maintained that he was a co-founder whose central contribution had been systematically downplayed.

The dispute has produced a substantial secondary literature and has never been resolved. The Wikipedia article on its own history acknowledges Sanger as co-founder; Wales’s public statements have varied. The difference matters partly as a question of historical accuracy and partly because the founding story shaped how the project was understood: whether Wikipedia was a community-generated phenomenon or the extension of a single founder’s vision.

Sanger announced Citizendium at the Wizards of OS conference in September 2006 and opened it publicly in March 2007, a rival encyclopedia requiring real-name registration and giving credentialed experts editorial authority over their subjects of expertise. Citizendium attracted a small community of contributors and produced a fraction of Wikipedia’s content. It demonstrated that the peer-review model that had failed as Nupedia did not become viable simply by lowering the barriers to contribution somewhat.

The Reliability Question

Wikipedia’s open editing model generated persistent reliability problems that the community developed sophisticated mechanisms to manage.

Vandalism (deliberate insertion of false or offensive information) is a constant problem on high-traffic pages. Bots that automatically revert obvious vandalism can act within seconds. Semi-protection, requiring account registration to edit certain pages, limits anonymous vandalism on the most-targeted topics. But determined vandalism by registered accounts, and subtle manipulation by those with apparent credentials, has been harder to catch.

The 2005 Seigenthaler incident crystallized the problem for mainstream audiences. John Seigenthaler Sr., a journalist who had served as administrative assistant to Attorney General Robert F. Kennedy, discovered that his Wikipedia article claimed he had been “thought to have been directly involved in the Kennedy assassinations” and had moved to the Soviet Union in 1971. Both were false. The text was inserted on May 26, 2005, and stayed up until October 5, more than four months. The author turned out to be Brian Chase, a Nashville delivery-company manager who said he had meant it as a joke on a coworker.

Wales responded on December 5, 2005, by barring unregistered users from creating new articles, and on December 17 Wikipedia adopted its biographies of living persons policy, which demands sourcing for contentious claims about the living and allows immediate removal of unsourced ones. The episode did not reduce Wikipedia’s usage (if anything, media coverage of the controversy brought more traffic) but it established that the project took errors seriously and would modify its processes to reduce them.

Systematic bias remained harder to address. Studies consistently showed Wikipedia’s editing community to be disproportionately male. The Wikimedia Foundation’s 2011 editor survey put female participation at 9 percent; a 2018 survey across twelve language editions found 8.8 percent overall and 13.6 percent among English Wikipedia contributors, while other English-focused studies have estimated up to 20 percent. The community skews young, and concentrated in wealthy English-speaking countries. Technology companies received detailed coverage; women scientists from periods before Wikipedia’s founding were often stubs. The Global South was systematically underrepresented. Various initiatives (Wikipedia Education Program, Wiki Loves campaigns, outreach to libraries and universities in underrepresented regions) addressed the margin of the problem without changing its structural character.

Wales himself was involved in controversies. In December 2005, writer Rogers Cadenhead documented that Wales had edited his own Wikipedia biography, including edits that trimmed references to Sanger as co-founder; Wales apologized for a practice Wikipedia discourages. In 2008, a romantic relationship between him and a Wikipedia subject whose article he had edited was reported by the press, raising conflict-of-interest questions about how the project’s own conduct standards applied to its founder.

The Scale of What Was Built

Whatever the controversies, the scale of what Wikipedia represents in the history of knowledge production is difficult to overstate. Before 2001, encyclopedic knowledge was controlled by institutions, publishers, universities, national academies. The gatekeeping was sometimes valuable and sometimes arbitrary. Access was unequal: encyclopedias cost money, and libraries in rich countries had far better collections than libraries in poor ones.

Wikipedia made the sum of human knowledge (imperfect, biased, contested, and constantly revised) freely available to anyone with an internet connection. A student in rural Nigeria, researching a science topic at midnight, has access to something closer to a research library than any individual student in 1990 could have accessed without physical travel to a major institution. The access inequality is not eliminated, but it is reduced.

Wales’s vision of the encyclopedia as a project of radical accessibility (knowledge for everyone, produced by everyone) turned out to be achievable in ways that his contemporaries in academic publishing could not have predicted and did not welcome. The encyclopedia his mother bought from a door-to-door salesman was updated once a year with a packet of stickers he helped paste onto the pages. The thing he built updates itself continuously, in more than 340 languages.