Information wants to be unfree
Prediction markets and the recommodification of journalism.
I.
One way to think about prediction markets is as an efficient way to transfer wealth from suckers to inside traders. (You may have read news stories about anonymous bettors on the crypto-powered Polymarket placing suspiciously well timed wagers predicting Taylor Swift’s engagement announcement.)
From a 2025 white paper by Felix Reichenbach and Martin Walther:
In this study, we use a dataset from the leading platform Polymarket that contains more than 124 million trades by nearly one million traders with a total volume of USD 48 billion. … We find that fewer than 30% of traders earn positive net profits and that the distribution of profits and losses is highly skewed, which indicates that a large number of traders systematically lose money to a small minority of skilled participants.
If a significant majority of the people who wager lose money, that meets my definition of a losing game. Yet prediction markets like Polymarket and Kalshi will be a greater feature in our lives in 2026 than 2025, including through the growing financialization of the news people get. This raises the pressing question for the news industry of whether the world benefits when candy gets taken from babies, and how long that arrangement can last.
Last week, Polymarket launched a partnership with Dow Jones & Company to provide real-time “betting odds for economic, political, and cultural topics” across titles like the Wall Street Journal. The announcement came a month after prediction market Kalshi’s agreement with CNN, touting that “journalists can more easily surface credible information to their audiences about the real-time probabilities of future cultural and political events.” Polymarket had previously in 2024 hired as consultant the stats journalist and off-putting social media personality Nate Silver, who wrote last November:
As an advisor to Polymarket, I’m obviously happy to see all the attention to prediction markets. As a journalist who would like to see more statistical literacy, I also think it’s basically good when people are more exposed to probabilities and they become more normalized. Some journalists like to claim that they don’t make predictions, but journalism is chock-full of speculative statements about the future.
A basic problem with, I don’t know, life, is that humans are bad at making predictions. We all have our emotions and biases, our rigid and narrow way of looking at things. We’re bad at guessing how frequently the unpredictable happens. Making things worse is that the modern information ecosystem has created new attentional rewards for making up bullshit: Just log on to Elon Musk’s X, the everything app.
As Polymarket founder and CEO Shayne Coplan put it in an early interview:
“The idea behind it is that we’re in this phase as a society where there’s a fundamental misalignment of incentives in the social news cycle, and that kind of manifests itself where the news and information that’s distributed to us is very skewed, because all the platforms that distribute it are optimizing for engagement, getting people glued into their sites instead of actually distributing the truth.”
Coplan added, “We are missionaries for the concepts of market-based journalism.” We should investigate what, exactly, those concepts are.
II.
The theory of something like Polymarket in our attention economy is that, unlike social media attention-hackers or friendly but hallucinating AI agents, prediction markets disincentivize bullshit by putting real skin in the game. Money on the line rewards the opinion-haver who does more research, organizes their thoughts more clearly, and punishes the bullshitters who don’t. The wagers are a performance incentive, the same way CEO pay packages are larded with equity awards to align individual effort with the value of the broader enterprise. Behavior modification is the point.
We should be real, though. Part of what makes prediction markets “accurate most of the time,” as Reichenbach and Walther report, is the presence of bettors with inside information whose wagers are reflected in the odds. (When a mystery bettor netted more than $50,000 by suspiciously wagering on María Corina Machado to win the Nobel Peace Prize 12 hours before the announcement, it reportedly caused her Polymarket odds to go from 5% to 70%: vastly closer to accuracy.) This is not an accident. Prediction markets aren’t just parlor games, like a fantasy sports league. They’re information-elicitation machines: The piston powering the contraption into motion is presence of dumb money and the opportunity to separate it from its owners. That’s the whole incentive for quality-information-havers to risk getting in the game, bringing along more accurate “signals” into the public domain that might otherwise stay obscured. The bigger the morons and the hype, the bigger the reward for accuracy, and the greater the chance that accuracy shows up. Inside trading and the business of sucker exploitation may violate our sense of fair play, in addition to probably betraying confidences and breaking laws in some cases. (Americans aren’t legally allowed to be on Polymarket, ha ha ha.) But if the point of a prediction market is to predict, bettors with inside information help a “prediction” market do exactly what the name says on the tin.
Naturally I am wary of all this, while recognizing that the rise of traditional journalism was itself an earlier practice of commoditizing and disseminating private information to the public. Many journalists are frequently in the position of coaxing secrets out of people otherwise expected to keep their mouths shut, with the goal of socializing that information to bring the public in closer alignment with factual reality. This is the magic industrial process by which wages paid to a reporter are converted into what economists call social “welfare.” This blockbuster investigation shows a mayoral candidate might be a Ponzi schemer! Better mark somebody else on your ballot.
Prediction markets linked to news outlets go further in securitizing the old process of newsgathering, with a critical and novel distinction: Ethical journalists can’t pay leakers, but prediction markets can. Traditionally, reporters’ sources are financially uncompensated and must have another motive for taking the risk of leaking, whether it’s to pursue the civic good or, say, the chance to knock a rival down a peg. Unfortunately, the nonmonetary incentives for those sources to share nonpublic information with journalists have deteriorated in recent decades: News outlets have less leverage with their subjects (who now have their own means of publication and message control on social media), a diminished public mindshare (lessening the utilitarian impact of exposés), and less trust than before (thus less of a civic incentive to cooperate). Worse, information itself has become increasingly assetized and privatized in the United States and made riskier to possess, via the proliferation of nondisclosure agreements with coercive financial penalties for sharing nonpublic information. It is dangerous to be a news source today. Maybe a bounty for leaks pushes a source’s calculation back in a public-facing — and profit-seeking — direction.
III.
I’m just not sure prediction markets offer a stable future for news outlets to rely on, given the markets’ central contradiction: They need a steady stream of suckers to entice the information-havers to play. If the suckers figure out they’re suckers and stop betting, their disappearance may paradoxically drive down the usefulness of prediction markets by causing the smart money to go somewhere else. I didn’t even mention problematic incentives for news subjects to change their behavior to exploit pending bets, or the enormous cynicism that the monetization of all phenomena reliably engenders in everyone.
Journalists had resisted cynicism long ago by adopting professional codes of ethics that reject gifts, pocketbook journalism and the incentive mindset: Truth ought to be its own reward. But truth is a value, and values struggle to pay their share of the rent these days. Information is a commodity. And like all commodities, information keeps looking for ways to be unfree.
Sometime in the not distant future, we're also going to learn that people wagered large sums on the prediction markets to shape the future. Press coverage of the prediction markets during elections definitely influences the election, and having a huge amount of money wagered against a candidate can increase the chances that candidate wins. I think we've likely already seen this happen in the last election cycle to some extent.
Both from the news consumers perspective and that of the journalists, it further encourages the bad practices of horse race coverage -- but on steroids as it is effectively polls on everything, continuously available, without any of the depth or nuance that actually can make well-done polls insightful.
There was already way to much focus on predicting outcomes rather than substance, but people betting on it will only accentuate those trends.