A Counterexample to the Bayesian Truth Serum
The Bayesian Truth Serum looks like one of the best possible ways to get judgments that avoid the problem of “groupthink.” There's only one minor problem: Groupthink isn't the only human failure mode.
According to the Bayesian Truth Serum, the most reliable judgments are those that are more common than people expect. (The expected common judgments are those caused by groupthink.) The most recent cases of that are the Trump movement and Brexit. I'm not sure about Brexit, but the Trump movement is an example of something recommended by the Bayesian Truth Serum that does not reflect good judgment.