"Not really sure how this can actually be better??"
Well for one thing, when you create an in-place family, you clutter up the project by creating all kinds of stuff like reference planes and other datums and views that you don't need after you finish the in-place family. The model can be a confusing enough, without adding all that junk and then later going in and trying to clean it all out. But in the family editor, all that stuff was never part of the project, and it goes away when you finish the family. So creating the family in the family editor is a very clean process for whole project. And for another thing, in-place families were never really what Revit is supposed to be about. Revit is about a lot of independent elements, all with parameters, coming together to create a model. When all those elements and families are independent, they can be moved in and out of the project, then system families can be moved more easily "revised instantly", (Revit). You can even take all your non-system families out of the project and start the whole project over from scratch without even spending as much time doing it as you might think.
I had to learn to get away from in-place familes myself when I moved into Revit LT. And slowly I'm learning more and more reasons why I never wanted in-place families in the first place. But again, I'm not overlooking the fact that in-place family creation is quick and convenient because you are using the model itself as a reference, instead of lugging dimension into the family editor. So maybe sometimes you might want that in-place capability. And of course this is available in the full version of Revit.
This is all just my opinion from my own experiences, I'm not making any claims that anyone can dispute. But I would certainly welcome any differing opinions. This is a topic that MANY Revit users would like to understand more about.