Mantic Monday: Let Me Google That For You

Astral Codex Ten Podcast

https://astralcodexten.substack.com/p/mantic-monday-let-me-google-that

Let Me Google That For You

New from Google this month: Creating A Prediction Market On Google Cloud. Google announces that they’ve been running an internal prediction market for the past year, with “over 175,000 predictions from over 10,000 Google employees”. 1 Predictive analytics.jpg

Most of it’s classified because they’re predicting stuff about Google’s corporate secrets, but some friendly Googlers were at least willing to walk me through the article and clarify pieces I didn’t understand.

The market, called Gleangen, is actually the second prediction market Google’s tried. The first, in 2007, was called Prophit - the team included occasional ACX commenter Patri Friedman, who’s since moved into the charter city space. (source)

Prophit wound down because the founders left and nobody really knew what to do with; you can read about some of their findings here. In 2020, with all the uncertainty around coronavirus, some Googlers decided to try again. Gleangen is the result.

Unlike most prediction markets, anybody can create a question on Gleangen. This usually goes badly: most people are terrible at writing questions with objective resolutions. Google manages by having a dedicated team of moderators who go over everything and amend it when needed. The market pays out in play money and the right to be on a leaderboard.

So far it’s not doing much else. The Googlers I talked to saw no evidence that company executives were paying much attention to it when making decisions. Why not? Hal Varian, Google’s chief economist, said in a Conversation with Tyler Cowen:

COWEN: Why doesn’t business use more prediction markets? They would seem to make sense, right? Bet on ideas. Aggregate information. We’ve all read Hayek.

VARIAN: Right. And we had a prediction market [referring to Prophit in 2007]. I’ll tell you the problem with it. The problem is, the things that we really wanted to get a probability assessment on were things that were so sensitive that we thought we would violate the SEC rules on insider knowledge because, if a small group of people knows about some acquisition or something like that, there is a secret among this small group.