The possible applications for yaml frontmatter data are endless. In the video you will see another area of notes and how obsidian is even able to present yaml in a friendlier and less technical way (that alone is a value of its own in regard of writing and reading):
I don’t think that it would collide IF the way in which the user can retrieve the data would offer a useful addition to the tags. By supplement I mean that although it would be quite possible to save some tags (why not for the sake of a more tidy tag tree?), the data from the yaml could be used to realize queries that are not so easy with tags. In this context, take a look at the “cooked” field in the video. It is obvious that for good reasons the data from yaml never ever should appear in the tag tree to avoid feature duplication and therefore the usage should totally be seperated from the feature to write nested tags.
I think that some search syntax like [ingredients:eggplant] would be nice but would not exploit the potential of yaml data by far. Obsidian has implemented that new visual representation of yaml as preparation for the upcoming datacore/dataview feature that will allow inline querying of all data and presentation as tables, cards, boards, calendars combined with filter and sorting fields. That, as far as the visual aspect is concerned, is reminiscent of Notion but doesn’t go that far to become a database-centric application.The notes are still the focus. The “Dataview” then offers a flexible display of the notes, a “super map of content”, so to speak, with many possibilities, as you can see in the following screenshots (taken from notion as datacore is yet not released for obsidian).
Would that be something for bear 3.0 or even better for panda 2.0? 
