The recent Manifest 2025 conference that took place in Berkeley, California was far more than a gathering of prediction market enthusiasts. Intellectuals with diverse interests from sex work to calculating a precise P(Doom), come together to converse about their deeply reasoned philosophies that answer many of the world’s most profound and difficult questions.
Then they have the great joy of arguing about them.
Prediction markets were a small but crucial part of the conference. The attendees had to grapple with the explosive post-election growth of the event contract industry and how to balance the forecasting potential with problem gambling harms.
Here are my five takeaways from Manifest 2025.
Prediction markets fit in a larger picture
Diverse professionals share an interest in prediction markets’ abilities to forecast outcomes. Effective Altruists, entrepreneurs, and even a bureaucrat from the Defense Department came to better understand how prediction markets can aid in predicting the future.
There was plenty of hype for not only Kalshi but also for other prediction markets like Futuur and Bayes Market. With topics like elections and possibly sports being fair game for exchanges, many new players are considering moving into the space.
Manifold co-founder Stephen Grugett told Prediction News in a closing interview that he was “very bullish” on the prediction market industry.
“There’s orders of magnitude more volume now than a year ago,” Grugett said. “Everything has totally changed. “Unfortunately,” he acknowledged with bemusement, “this is mostly due to the efforts of our competitors.”
The effects of Kalshi’s election victory and aggressive post-election marketing have sent ripples through the once-niche industry.
Manifest attendees largely believe in prediction markets’ value
One of the social goods Manifesters hoped to extract from prediction markets was a way to make better choices in an uncertain world. Probabilistic thinking is viewed as a way to correct for one’s own biases and make better choices. For those who subscribe to Robin Hanson’s vision for a futarchy, a government run by prediction markets, trading could lead to better government policies.
Prediction markets can aggregate current information into a price that reflects the odds of an event occurring. However, the information must be relevant to the market’s outcome and available to the market’s traders to produce a useful forecast.
On the other hand, many markets seem to exist just because they’re fun to bet on.
Manifesters concerned about gambling on everything
I participated in a panel with Jeremiah Johnson and Isaac Rose-Berman on Sunday titled Has trading and gambling gone too far in the US? It covered the similarities between prediction markets and gambling.
Some of the differences are regulatory or academic. Financial futures have additional uses beyond the contracts, while sports wagers have arguably little use beyond their entertainment value. The question over whether the potential benefits of prediction markets outweigh the concrete harms of increased gambling troubled many attendees. And the discussion couldn’t have come at a better time.
Sports contracts, which come in a format familiar to the general population, work as customer acquisition tools that expose new traders to Kalshi’s other markets. However, state gaming regulators and Kalshi are engaged in ongoing court battles over whether sports outcomes are fair game for federally regulated derivatives markets. The answer to that question remains to be seen, but we could have more clarity on that in the coming months.
Political forecasting remains a hot topic
One of the best panels was a conversation between Democratic pollster David Shor and political trader Jesse Richardson, better known as @politicalkiwi on X. They discussed how they were able to see Trump’s 2024 victory coming and how so many intelligent people missed changes among both parties’ voters.
Among those changes was the difference that voter turnout has on both parties’ fortunes. Democrats now benefit from low turnout, and Republicans perform better when turnout increases.
Since Democratic organizations run many “turn out the vote” programs, this puts the Democrats in the awkward position of helping their political opponents. Republicans’ endorsements of policies like voter ID laws make it harder for some of their own voters to vote, reducing turnout among their own voters.
“Democrats…and Republicans are both kind of stuck in the mindset of thinking of their coalition from 10 years ago, and they haven’t actually realized that things have changed,” Richardson told Prediction News.
The overarching lesson from the event’s political traders was to rely on data describing what voters actually believe instead of proxy variables like voter registration.
AI concerns loom large
While prediction markets were a cornerstone of Manifest 2025, one of the dominant concerns remained the promise and threat of advanced AI.
Attendees directly involved in the debate had pins on their lanyards listing their P(Doom), the probability that AI will end humanity. (The University of Chicago student I discussed the Central Asian Enlightenment with thought it was below 10%.)
But AI also has important implications for prediction markets. Large Language Models (LLMs) can already assist traders by aggregating information from news sources and recommending which trades to place. AI bots aren’t populating real-money platforms like Kalshi, but Manifold and Metaculus have experimented with allowing AI bots to place play-money trades or make forecasts without currency.
The question of whether to allow AI bots or continue requiring human traders at the helm could press real-money platforms soon. The Forecasting Research Institute published a study in February 2025 estimating that LLMs could perform as well as superforecasters within the next four years.
The future seems to be barreling toward the prediction market industry as quickly as it is the rest of us. And they appear to be on a rapid trajectory into the mainstream.
