All its work would be advances, I think

I dispute your claim.

I'm well disposed towards your viewpoint on that.

attempting to remove emotions from decisionmaking is what causes the "valley of rationality."

I disagree with the implication of this. I think the main causes are misusing tools like Bayesian updating and considering what a rationalist would do, and trying to do that.

Insofar as poorly calibrated emotions are part of the problem, one must subtract the problems that would have been caused by them under non-aspiring rationalist conditions from those under aspiring-rationalist conditions to calculate what the aspiring rationalism is responsible for. I don't think this usually leaves much left over, positive or negative.