No more than you do the other way around. That method works fine for my genealogical work, and has so for most Scandinavian genealogists for a long time. It is even approved by several history scolars, and this is the way new genealogists are tought.
But maybe we have a differnet definition of what a source is. To us, a source is whatever and whoever you get the information from, therefore they have to be classified into caathegories according to their trustworthyness. It's not a problem in my neck of the woods, since most scholars in fact do agree about a source quality and trustworthyness. At least the ones most commonly used.
I use the method of if it is not provable by primary or several good secondary sources, it's not going to be linked in my personal software, and I think some sort of that approach needs to be done on Geni to. We can't just link persons just because some online tree without sources have it written down. (And that online tree is to me a source, but a tertiary one.)
When it comes to the time before year 1000 almost all of them agree that the genealogies in the nordic stories about the persons living in the Nordic countries are untrustworthy and should not be used in our genealogy as proof of family relationship, and I do believe them.
Not even the Norwegian line of kings are provable genealogically longer than back to king Sverre born around 1150. Almost all genealogists and historians say it is impossible to prove any family relationship back to king Harald "Fairhair" Halvdanson, the first king of Norway, even though there are plenty that tries. I have never understood the fascination with a relationship with kings or nobles.