When Elon Musk took over Twitter, he discovered that promoting free speech was harder than allowing all unfiltered thoughts to be published.
A functional public square requires structure to keep online conversations from devolving into trolling and death threats.
Kalshi, a regulated exchange where users can trade on the outcome of real-world events, recently introduced Kalshi Ideas. It’s a feed that allows people who’ve staked positions in Kalshi markets to write why they made their trades. Kalshi Ideas includes terms and conditions like any other social media platform. However, it’s innovative in two ways:
- Allows users to self-select into conversations they’re knowledgeable about
- Requires users to stake money behind their opinions
Each post has like and comment buttons, but there’s also an option to buy the same position as the person who wrote the post. Kalshi Ideas is structured to encourage posts that defend financial positions instead of seeking engagement through outrage.
“That creates a significantly more objective debate and structured debate and also filters the cesspool of people just saying random things and making the current conversation uncivil,” said Kalshi Founder and CEO Tarek Mansour.
Kalshi created a new type of futures contract to reward users for correctly making predictions. Its markets include trades on whether interest rates will hit a certain level by the end of the year. Traders can also trade on rainfall amounts or whether the Center for Disease Control will identify a new COVID variant of high consequence.
Although these markets can take advantage of the insurance function derivatives can play, most traders will use Kalshi to speculate on high-risk, high-reward contracts.
What early chatrooms can teach modern social media platforms
Before social media, the mid-’80s Internet saw the rise of small chatrooms. One of them, the WELL, was what author Gal Beckerman called a “conversational exchange.” In his book, The Quiet Before, he described the philosophy behind the WELL, an early chatroom where anyone could join a chat and post:
“What Tex [a WELL founder] brought to the WELL was a faith that communication itself could be redemptive. He believed that this was the key to self-government, to making this new virtual community work.”
“But he’d also learned what almost anyone learns when they dabble in such a human experiment: that success rests on the fragile balance between the needs of the individual and those of the collective, a balance that had to be monitored, calibrated, and recalibrated daily … But without that vigilance, without those rules, without a structure that pointed people toward productive deliberation, things could quickly go off the rails.”
That’s not so different from modern social media struggles. The difference in the mid-’80s was that user bases were measured in the hundreds or thousands instead of millions. WELL included a key role missing from modern social media sites: the host.
How Kalshi revived the host on its social feed
Originally called the “fair witness,” the host was a key component to WELL’s functioning. The host would be in charge of moderating discussions, including rewarding interesting posts that moved discussions forward or banning users who were bad for the conversation or only there to troll.
With a user base of just a few thousand people, this could be done by a thinking person who was knowledgeable about the nuances of balancing impolite and blunt remarks with the welcoming environment necessary to attract new users to the conversation.
That middle layer doesn’t scale and can’t be performed by AI, which can still make content moderation mistakes as Facebook’s flagging program did when it labeled part of the Declaration of Independence as “[going] against our [Facebook’s] standards on hate speech.” Facebook restored the post and offending account shortly after human review.
Kalshi doesn’t need to rely solely on human hosts to promote productive conversations. Instead, its markets attract people with money to stake on their predictions. People who put their money where their mouths are can then defend those decisions. Further, elite traders with large gains are featured at the top of the feed, rewarding users who make successful predictions rather than provocative noises.
Skin in the game and selecting for expertise
AI programs can accomplish impressive feats of data analysis, but by themselves, they can’t screen each user for expertise. As a proxy for expertise, Kalshi designed its social feed to reward users with skin in the game.
“It’s an incredible selection mechanism because if you have enough skin in the game, you’ve done enough research to actually put your money into this market,” Mansour said. “I think you’ve bought yourself a ticket or a seat at the table to discuss that topic. And so I think that’s a very clean markets-driven selection mechanism to foster pretty high intellectual debate.”
This sentiment echoes a theme that Nassim Taleb sounds late in his book Antifragile:
“Reality removes the uncertainty, the imprecision, the vagueness, the self-serving mental biases that make us appear more intelligent. Mistakes are costly, no longer free, but being right brings actual rewards.”
Kalshi’s markets are already a conversation. The price changes reflect traders’ ideas and predictions about the future. Kalshi Ideas allows its users to put the prices they’re willing to buy and sell at into words.
Bringing content boxes to social media
Having a social media platform where anyone can post anything for free has value. It allows people on the front lines of breaking news to post and stream the news as it unfolds. However, the dark side of social media comes from the guardrails that don’t surround content. Terms and conditions from the top down and the block button at the individual level are insufficient ways to prevent trolling, doxing, bots, and other sources of online derangement.
Kalshi is a comparatively small platform that’s able to limit conversations to its markets, and Kalshi Ideas is a natural extension of the existing information ecosystem. In contrast, social media sites are a sprawling morass of many information ecosystems, which can have their own facts and logic. Companies like Facebook and X aren’t designed to create structural guardrails the way that Kalshi is.
However, social media sites could explore ways to add the missing middle layer of content moderation at scale. Taleb’s maxim that he “want[s] predictors to have visible scares on their body from prediction errors, not distribute these errors to society” could be a useful guide for social media companies as much as it is a distillation of everything right about Kalshi Ideas.