things that help the typical human are less likely to help you as you become less like the typical human.

Yeah, that's true, and I suppose I didn't adequately address it in the grandparent. It doesn't strike me as likely to cause serious problems, though, except perhaps in the social realm; the body of empirically supported self-improvement research that I'm currently aware of seems broad but shallow, not containing many long causal chains of the kind that could be disrupted by small changes in their early dependencies. It's conceivable that completely eliminating a deeply rooted bias could render broad swathes of traditional self-improvement literature irrelevant at a stroke, but I'm not sure that's plausible in the art's current state.

In the future, perhaps, if LW or a similar community manages to come up with some seriously impressive results, but I don't think that's enough to merit a disclaimer.