This isn't true for people who approach their undergraduate degree as a kind of classic liberal arts education or intellectual orientation. Many professional trainings are offered as graduate degrees, without the expectation that an applicant's undergrad training will necessarily get applied. One good example is the Harvard Graduate School of Design, where people with all kinds of backgrounds train to be architects at the graduate level.

Whether you think undergraduate degrees should be vocational (whether they currently are or not) or, in contrast, really just an intellectual taster for topics that students won't necessarily pursue, it would be a mistake to rule out significant "course corrections" at graduate level. Even at the level of PhD admissions, people change fields—especially strong people.

The idea that "genius" is always specialised in a particular academic category, and always knows in advance exactly where it can contribute, is old-fashioned. GH Hardy probably believed it about mathematics, and (for example) the Cambridge math degree probably endorses it implicitly by being so tough. But this kind of macho thinking excludes the possibility of transdisciplinary work, and discriminates against people whose talents aren't easily categorized. It also makes the disciplines which it effects markedly less diverse.

I disagree pretty strongly with both, your undergraduate does not determine what breakthroughs you'll make. There are plenty of examples of businesses built by people that did not complete an undergrad.

As for Elon Musk, though I don't claim to know what he's done to be successful, I doubt very that he has not asked for advice.