Generated by Codex with GPT-5

Techmeme surfaced the Wall Street Journal investigation They Looked Like They Were Getting Rich on Polymarket—but None of It Was Real, published June 21. The article examines how a product built around publicly observable wagers marketed itself through videos of trades that never happened.

The Journal reviewed more than 1,100 videos, creator instructions, and interviews with people who worked in Polymarket’s paid content program. It found dozens of mostly college-age creators filming staged trades and, in some cases, fabricated wins. The videos looked like personal accounts of successful betting, but the activity shown did not correspond to transactions on Polymarket’s actual market.

One creator, college student George Makihara, appeared in his videos to place 145 wagers worth almost \$410,000 between January and mid-May. A particularly striking clip showed him winning \$100,000 by betting that President Trump would publicly say “McDonald’s” during January. The Journal found that the trade was not real. More than 50 accounts did place the actual wager, and all of them lost.

A demonstration presented as evidence

There is a meaningful difference between dramatizing how a product works and inventing a customer outcome. A normal demonstration makes its simulated status clear. These videos instead borrowed the visual language of an authentic trading diary: a creator chooses an unlikely contract, risks money, and records an apparently enormous payoff. The implied message was not merely that Polymarket was easy to use, but that people like the viewer were already getting rich on it.

That message matters because a prediction market’s appeal rests heavily on credibility. Polymarket presents its prices as an information signal produced by participants putting real money behind their beliefs. Its blockchain infrastructure also gives the platform a useful transparency story: genuine trades leave records that others can inspect. Fabricated creator trades reverse that advantage. They use the appearance of a verifiable market while withholding the one fact that would let a viewer verify the promotion—that no such trade occurred.

The staged wins also distort the basic economics of the product. Every winning position has a counterparty, and spectacular payouts are rare rather than representative. A Washington Post analysis of Polymarket trading data through March 29 found about 1.7 million participants had lost money, compared with roughly 765,000 who had won. Marketing built around invented jackpots replaces that distribution with a much more seductive story: unusual foresight reliably turns into effortless money.

Growth incentives outran disclosure

The Journal’s findings fit a broader pattern in prediction-market promotion. Earlier June reporting found that Polymarket paid creators to amplify its markets and corporate announcements, sometimes without clear sponsorship disclosures. Separate NPR reporting said Polymarket and rival Kalshi offered creators as much as \$500 per post and gave affiliates wide latitude as long as they promoted the markets. Some paid posts then helped spread false claims about election results.

These are not isolated content-moderation mistakes. The affiliate model rewards attention, while prediction markets supply an endless stream of provocative odds that can be framed as breaking news. A creator earns by producing a dramatic claim; the platform gains reach and trading activity; and neither has much incentive to emphasize uncertainty, losses, sponsorship, or the difference between market odds and evidence.

The regulatory distinction is also important. The Federal Trade Commission requires material relationships between advertisers and endorsers to be disclosed clearly. A sponsorship label can identify who paid for a post, but it does not cure a fabricated transaction. Disclosure answers “who funded this?” The Journal’s investigation raises the more fundamental question of whether the event depicted happened at all.

The larger risk is that prediction markets are trying to become both trading venues and information products. Their prices increasingly appear alongside political commentary, news coverage, and social-media discussion. If the same company manufactures viral proof of easy profits, it weakens trust not only in its advertising but in the market signal it wants others to treat as informative.

The practical takeaway is simple: transparency cannot stop at the ledger. It must extend to distribution. A platform that asks the public to believe its prices should require paid creators to show real positions, real outcomes, and conspicuous sponsorship—or label simulations as simulations. Otherwise, the marketing layer turns an ostensibly auditable market into another feed where the most persuasive evidence may be synthetic.