I agree in the specific case that the state of the art is in a local maxima, but saying "the best way to approach a problem is by not being aware of disregarding previous attempts" ignores the much more frequent banal work of iterative improvement. Leaping out of a local maxima is rare and sexy and gets articles written about you and is important, but the work of slowly iterating up to a nearby peak is also important.

I think progress needs both individual achievements who break out of preconceived notions and the communal work of improving within the notions we currently have.