Maybe I should wait for the canonical article, but is your argument that false beliefs are not part of a first-best approach to rationality, even though they might be part of a second-best approach? Or is it something stronger than that?

I, for one, am interested whether there are convenient self-deceptions that promise instrumental benefits, short-term or otherwise. If nothing else, this will help me adequately assess the potential costs of rationality, rather than taking its benefits as a matter of faith.