Stephen Grugett (Manifold Markets Founder) - Predictions Markets & Revolutionizing Governance

Dwarkesh introduces Stephen Grugett as the founder of Manifold Markets, which he says has received a grant from Scott Alexander and raised a $2 million seed round. Grugett, one of three co-founders, describes the product as a platform for user-created prediction markets. Anyone can post a question about something they care about, and friends and other users bet on it. In his telling, the betting mechanism is what produces the best and most calibrated probabilities available.

Dwarkesh asks about the mechanism, assuming users buy Manifold dollars with real money, and wants to know why anyone would care about their balance. Grugett corrects the premise: Manifold uses a play-money currency, and a new user receives a thousand Manifold dollars to bet as they see fit. His hypothesis about human nature is that status and competition motivate people more than money does. Manifold sees itself as a social game in which users sharpen their forecasting and then show others, through an objective track record built from their betting history, that they know what they are talking about.

people are driven more by status and competitiveness than greed

Dwarkesh relays a tongue-in-cheek objection from Tyler Cowen: if prediction-market users were really good at forecasting, hedge funds would be trying to hire them constantly, and the fact that they are not suggests this is only a hobby. He asks whether Wall Street will eventually recruit people from the top of the leaderboard. Grugett, admitting he is biased, says something like that is possible and that Manifold's top predictors are quite good. He mentions that the current top-ranked user did well on a market about whether Russia would invade Ukraine in February, although the transcript of this sentence is garbled (as transcribed).

Grugett says he does not entirely accept Cowen's premise. In his account, the forecasting community overlaps considerably with financial professionals who manage money or trade for a living. As a play-money site, Manifold also attracts hobbyists without a finance background who are still quite good. He believes the top predictors represent untapped potential and that sites like his can identify them.

Asked what makes a top predictor, Grugett puts experience first. Many of Manifold's best forecasters came from other sites and have spent a lot of time thinking about forecasting in general. He identifies the first hurdle to improving as clarifying and crystallizing one's views about the world. People easily misremember whether they got a political event right, so without numerical predictions that are actively recorded it is very difficult to improve. Many of Manifold's users had already cleared this step before joining.

Dwarkesh then turns to news organizations, citing a Scott Alexander post. An investigative reporter who discovers serious government wrongdoing could, for example, short the market on that government's GDP a year out, publish the story, and profit from the journalism. Grugett calls this a clear case of prosocial benefit: a market mechanism supplying a public good that society would otherwise fail to produce. He considers accurate information about corporate malpractice very valuable and says anything that encourages it is good.

Dwarkesh asks whether skill at prediction could eventually be rewarded with money, through crypto or offshoring or some other route. Grugett says Manifold plans to host tournaments with real cash prizes, which it can offer despite being a play-money product. The company began as a crypto idea and then pivoted back to Web 2, which he jokes made it one of the first firms to move from Web 3 back to Web 2, for usability reasons. A real-money crypto offering is on the back burner and under active thought, and it would probably be a separate product rather than an addition to the current one.