Using Prediction Markets to Guide Government Policy — LessWrong