What this covers
Jimmy Wales sits down with Tyler Cowen to defend Wikipedia's durability in an era of polarization. The conversation centers on how incentive structures shape the character of information platforms: advertising-driven social media optimizes for engagement and agitation, while Wikipedia's nonprofit donation model creates no such pressure toward sensationalism. Wales argues that Neutral Point of View is achievable objectively, but increasingly hard to maintain recognition for as people abandon the ideal of objectivity itself. He frames Wikipedia as an enlightenment project—one whose dry, matter-of-fact style stands quietly radical against today's outrage-driven media landscape.
Beyond Wikipedia's mechanics, Wales ranges across business models, political economy, and technology policy. He examines how incentive-caused bias corrupts judgment when someone's paycheck depends on a particular outcome, why selection bias makes niche topics skew positive, and what makes Facebook's content moderation an impossible bind. He touches Ayn Rand's epistemology as underrated, defends Gradualism over radical policy shifts, and critiques both the European right to be forgotten and the bipartisan push to repeal Section 230—each position rooted in unintended consequences and first-amendment reasoning. On education, he argues against rigid testing regimes, having himself benefited from unstructured curiosity and encyclopedia reading. Throughout, Wales emphasizes that Wikipedia's real durability rests not on heroic leadership but on community governance systems and the Wikimedia Foundation's ability to maintain that partnership without becoming a moderation monolith.
Wales argues that Wikipedia's neutrality and quality flow primarily from its nonprofit, non-advertising business model and its open, community-governed processes, and that this incentive structure—not heroic leadership—is what lets it remain a dry, enlightenment-style reference work in a polarized media culture.
- Business model drives outcomes: ad-driven platforms optimize addictive engagement while Wikipedia's donation model optimizes meaningfulness
- Open, public community processes and rules, though bureaucratic, are the hard and important part of sustaining quality
- Neutrality is achievable objectively but increasingly hard to have recognized as people abandon the ideal of objectivity
Wikipedia's nonprofit, non-advertising model prevents the engagement-maximization corruption that afflicts for-profit social media platforms.
- Business model drives outcomes in often unanticipated ways: advertising-driven social media is incentivized to maximize time-on-site, which pushes it to build addictive products and prioritize agitation and argumentation over education and thoughtfulness, whereas Wikipedia's non-advertising donation model gives the Wikimedia Foundation no incentive to optimize engagement.
“the real incentive is to show you as many ads as possible, so they really obsess over time on-site, and it’s driven them to create addictive products”
- If Wikipedia applied basic machine learning to show promotional sidebar links to maximize time-on-site, it would probably end up recommending hot, off-topic content like Kim Kardashian or Donald Trump rather than related concepts, which is not what users want from an encyclopedia—illustrating why optimizing engagement would corrupt the product.
“the best thing to do is, when you go to look up Tyler Cowen, to show you on the sidebar links to Kim Kardashian, Donald Trump, whatever the hot topic of the day is”
- Wikipedia was highly contingent: a publicly-written wiki encyclopedia was not inevitable, and the choice of a nonprofit, non-advertising model was a conscious decision rather than a foregone conclusion—one could imagine a Craigslist-style for-profit hobby with retained control and minimal monetization instead.
“was it inevitable that we would choose a nonprofit, nonadvertising business model? Definitely, that was not inevitable. That was a set of conscious choices that we made.”
- Wikipedia is primarily an enlightenment project, and its dry, matter-of-fact style is quietly radical in today's outrage-driven media culture because it insists people understand what something actually is—neutrally and clearly—before deciding whether they are for or against it.
“before we know if we’re for it or against it, maybe shouldn’t we actually find out what it is?”
Maintaining community health, trust, and procedural integrity is Wikipedia's central ongoing challenge, more vital than finances or technology.
- Although flattering to imagine Wikipedia depends on his sage guidance, Wales doubts it, holding that the durable challenge is designing community systems, models, and procedures that maintain Wikipedia's values long-term; the genuinely unsettled hard problem is the interface between the community and the Wikimedia Foundation, ensuring the Foundation can act without taking on an unscalable Twitter/Facebook-style moderation role.
“without my sage wisdom and guidance and calm ceremonial neutrality, the whole thing would fall apart, but frankly, I really doubt that.”
- The chief ongoing threat to Wikipedia is community health—keeping the volunteer community happy, productive, thoughtful, kind, and free of trolls—rather than finances, which are stable thanks to conservative management and successful fundraising; technological shifts like voice assistants extracting Wikipedia content without bringing donors are a secondary risk.
“it’s about community health. How do we make sure that the community is happy, productive, not full of trolls, thoughtful, kind”
- Wikipedia's accumulated history, rules, and procedures—each arguably doing more good than harm—have made it bureaucratic in a way that makes it hard for new people to get involved or understand things, as illustrated by a famous author whose self-correction was met with a request for a source because he emailed from an unverifiable Gmail address.
“we’ve become very bureaucratic in a certain kind of way. We’ve got a lot of history. We’ve got a lot of rules.”
- Sock puppets are prohibited not because of multiple identities per se—legitimate alternate accounts can be declared to the arbitration committee—but because Wikipedia relies on trust and uses straw-poll 'not voting'; pretending to be multiple people skews the perceived balance of opinion and can conceal a conflict of interest, as when a PR firm created a fake retired car-collector persona.
“bulking up your impact by double-voting on something, by pretending to be different people, is super problematic.”
Wikipedia can objectively maintain neutrality, though polarization and rejection of objectivity itself make that neutrality harder to recognize.
- Wikipedia can objectively maintain a neutral stand, but as politics polarizes along educational lines and respect for traditional journalistic institutions declines, it becomes very hard for that neutrality to be universally recognized—especially among people who have philosophically abandoned the concepts of objectivity and neutrality, a stance found on both the right and in postmodern critiques on the left.
““Can Wikipedia maintain a neutral stand?” Yes. That, I think, yes. Objectively speaking, yes. But that would be recognized as such... that could be quite hard.”
- Highly contested articles like George W. Bush become high quality through an enormous expenditure of human hours of open, adversarial debate—far more than four or five thoughtful people would need—but the open process of everything being subject to challenge has independent value.
“how many human hours were spent hammering away at making this what it is today?” It’s anenormousamount of time”
- Even articles on seemingly stable old topics like Hamlet or Julius Caesar receive continual edits because scholarship continues, new perspectives are generated, and the contemporary impact of the subject keeps evolving—evident in how much was learned since the 1911 Britannica articles Wikipedia imported.
“since 1911, what have we really learned about Julius Caesar?” Well, it turns out, a lot.”
Paid advocacy editing corrupts neutral judgment through systematic bias, while legitimate compensated contribution based on genuine knowledge is acceptable.
- Per the old saying, it is very hard to get someone to understand something when their paycheck depends on them not understanding it, so a person paid to improve their boss's Wikipedia page will, through advocacy bias, hold a substantially different view of what is neutral than the community does.
“it’s very hard to get someone to understand something when their paycheck depends on themnotunderstanding it”
- Wikipedia distinguishes paid editing from paid advocacy editing: a professor paid by a university to contribute knowledge is fine, while someone paid by a PR department to make a subject look good is the real problem, because advocacy bias systematically distorts one's sense of what counts as neutral 'filling in gaps.'
“we try to make a distinction between paid editing and paid advocacy editing.”
- The highest marginal return for a Wikipedia contributor depends on their skill set: contributing unique knowledge on an obscure topic, or helping build communities in smaller languages in the developing world, yields far more impact than being the 11-millionth editor watching a high-traffic article like Donald Trump's.
“people who have the largest impact for the amount of time they’re spending are people who are working in smaller languages in the developing world”
Business models, organizational structures, and leadership incentives drive outcomes in ways that shape social outcomes and moral responsibility.
- Wales describes his own role at Wikipedia as ceremonial—he gives advice and people listen, so he has influence but not power—and he is content with that, contrasting people who expected power and instead hold only influence and are therefore unhappy.
“People do listen to me, so I have influence, but not power. I’m very happy about that.”
- Because Mark Zuckerberg holds one of the strongest CEO positions of any company and effectively cannot be ousted, he can make decisions that reduce engagement and near-term revenue without fearing shareholder revolt, so he—unlike CEOs who must panic about quarterly results—can take a long view and should be held morally responsible for doing so.
“Mark Zuckerberg has probably one of the strongest CEO positions of any company. It would be very difficult for anybody to oust him as CEO.”
- Money does not buy happiness but it buys freedom, letting people pursue their interests; fame, by contrast, is often net-negative—as with Sean Parker, who would give up fame over wealth because the grief it brings (over things like his extravagant wedding) is unaligned with the cancer-research and medical interests that actually make him happy.
“I don’t think money buys happiness, but the truth is, it buys freedom, so people then can pursue interests”
- Facebook faces a contradictory bind on disinformation: ten years ago, if Facebook had announced it would shape public conversation by deemphasizing low-quality news, people would have called it Orwellian, yet now the public demands Facebook 'do something' about disinformation—putting it in a tough spot where it neither wants to be the arbiter of truth nor can avoid taking responsibility.
“We would all have freaked out and said, “Oh my God, it’s Orwellian. Facebook is now trying to decide for us what’s true”
- Ayn Rand's objectivist epistemology is underrated: separate from her controversial politics and ethics, her work contains genuine, under-appreciated insights about concept formation, which are overlooked by people who dismiss her as merely a ranting novelist—and her core value of 'chewing on ideas' (thoroughly thinking things through) is undermined by the unkind, flamboyant way she presented her conclusions.
“there’re some real insights, interesting insights there about concept formation that have not been appreciated”
- Huntsville, Alabama is a surprisingly high-tech, underrated place because NASA scientists are headquartered there at the Space & Rocket Center, giving it apparently the highest number of PhDs per capita of any place in the US during the early 1960s.
“for a time in the early ’60s, Huntsville apparently had the highest number of PhDs per capita of any place in the US”
- Black-Scholes option pricing has swung from overrated—treated as the 'Second Coming' early on—to underrated today as a 'little toy model,' when in fact it was a genius innovation that revolutionized finance and enabled derivatives, despite real weaknesses as a description of reality that prompted more sophisticated successors.
“it’s a genius sort of innovation, a very, very clever model. Actually revolutionized a lot about how people thought about finance”
- Wikipedia editing is still overwhelmingly a desktop activity because the small-phone form factor makes editing intrinsically difficult—not merely a matter of more investment—so there is a divide between mobile reading (which is fine, even improved by shorter scan distance) and desktop writing, weakening the connection between mobile's rise and article content.
“Wikipedia editing is still overwhelmingly a desktop phenomenon. Part of that is that mobile editing experience isn’t great... the issue is it’s a small-phone factor.”
- Wikipedia can be biased without intent because some topics attract only their fans: articles on Japanese anime tend to be positive because the only people who write about anime are people who love it, unlike contested political topics where opposing sides take positions—making certain niche articles insufficiently critical though usually harmlessly so.
“they tend to be quite positive because nobody actually cares about Japanese anime except for people who love Japanese anime”
- Because of the genus-species structure of definitions—clicking the first link of any article eventually leads to philosophy—Wikipedia presents concepts like beauty as the history of philosophy rather than as live philosophical inquiry, effectively making history more central to how we read about ideas.
“what you would be reading about, in a sense, is the history of philosophy”
- Wikipedia's oldest 'ignore all rules' rule—if a rule prevents you from making Wikipedia better, ignore it—is the most frequently ignored rule, and exists because rules and procedures, while reasonable and there for a reason, sometimes produce bad outcomes like keeping a notable candidate off the site through bureaucratic process.
“if some rule is preventing you from making Wikipedia better, then ignore that rule.”
- Conquest's Second Law—that non-explicitly-right-wing organizations drift left over time—is not necessarily correct, because an organization like the Catholic Church has maintained a steady set of values for a very long period that does not map cleanly onto contemporary left-right categories, showing durable non-partisan value maintenance is possible.
“can an organization maintain an explicit set of values over a very long period of time that is neither left- nor right-wing in its nature? I think so.”
Section 230 legal protection is essential to Wikipedia's existence and faces serious bipartisan political threat.
- Section 230 of the Communications Decency Act is essential to Wikipedia's existence, and the bipartisan push to repeal it—from Trump on the right and from those worried about big tech's power on the left—poses a serious threat to anything open on the internet.
“There’s a lot of noise these days about repealingSection 230 of the Communications Decency Act, without which Wikipedia couldn’t exist.”
- The European right to be forgotten as currently enshrined is deeply flawed because Google can be forced—without any judge involved—to delist perfectly legal content that remains on a publisher's website, which by American First Amendment logic is like banning a legal magazine from being sold in stores; if content is egregious enough to censor, a judge, not a company, should decide.
“Google can be forced to take down links to content that is perfectly legal on a publisher’s website... and there’s no judge involved.”
- Facebook's moderation approaches both fail to scale—using people in sweatshop conditions to make content judgments, or AI that is not ready to judge the nuances of human conversation—so what it really needs is to put more power into the hands of the community, the model that works for Wikipedia, though what that looks like for Facebook is an unsolved hard problem.
“Hiring people in sweatshop conditions to make content judgementsortrying to get AI — which is not really ready to judge the nuances of human conversation — to do it, neither of those is scaling very well.”
Government and institutional interventions require strict caution because they frequently produce unintended harms that markets might have avoided.
- Government interventions frequently fail to work the way most people expect and can cause more harm than good—as with raising the minimum wage to $40 an hour, which would not help poor people much—so interventions warrant strict-scrutiny-like caution; over his life Wales moved from a radical anti-state stance to gradualism, holding that abruptly eliminating zoning laws would not reach the same outcome a market-oriented system would have because zoning roughly replicates what markets might have produced.
“we know very well from economics that many interventions don’t work the way most people think they work, and they actually can cause a lot more harm than good”
- Traditional education has a big downside in that, if not handled carefully, it becomes a rigid process of sitting in rows and meeting testing standards that is bad for creativity; an unusually open upbringing with free time to follow curiosity (Wales spent much of it reading the encyclopedia) is enormously valuable, and—while standards and tests have a place—creativity must be emphasized because solving future problems will require innovation.
“there is a big downside to traditional education in terms of, if you aren’t careful you can get into a very rigid process that is not good for creativity”
Wikipedia is primarily an enlightenment project insisting on clear, neutral understanding before judgment in an outrage-driven media landscape.
- The wave of cheap news stories about Wikipedia vandalism—like the immediate flood of calls when a celebrity's page was vandalized on the day of her death—reflected something wrong with journalism rather than a deep flaw in the Wikipedia model, and that fashion faded the way 'someone's selling X on eBay' stories did once people realized such content gets removed.
“It shows something deeply wrong with what’s going on in journalism, that this celebrity dies andIget eight phone calls about vandalism in Wikipedia.”
- Wales is now focused on WT.Social, a pilot project rethinking social media with a different business model, which has about 500,000 sign-ups though many are inactive and a core community is doing the work.
“WT.Social is rethinking social media, trying to create a new social media platform with a different business model.”