Testability and values shape scientific debates more than we often assume.
Forecasting is the wrong test of real expertise.
Distinguishing real credences from convenient but empty statements.