This obviously doesn't help right here and now, but I would like to point out that CfAR is in a good position to investigate this question experimentally. We'll have to wait awhile to be sure, but it looks like they have developed decent debiasing procedures and life outcome measures. I'm also guessing that they can't train people against every bias in a single retreat. Thus the can include different biases in different curriculums and compare their practical effects.