Gerard's point that the items are typically found in Wikimedia Commons I think is key. If the item is part of a sister project to Wikidata then it has a corresponding place on Wikidata. Unless I misunderstand the interoperability and mission of Wikidata.
Also: I am not a fan of deleting content -- especially content that is curated, focused, captures a time and place, incorporates hard work on projects especially as it relates to a Wikimedia project. To me this is not defensible. So deleting entries seems similar to my experience with Wikipedia editors hostile to added content focused on deletionism -- of course to a notable women's page where I as an editor am trying to establish said notability -- who characterize the information as yes, "too encyclopedic." #Ridiculous I wish this wasn't true.
So I agree with Gerard and others here.
There are several things at play. First, Wikimania and its talks will find its place in Wikidata because typically much of the papers, presentations and registrations will be found in Commons. So they will be registered anyway. Second, this thread is also about the way our policies are maintained. This is done in an arbitrary way and consequently much of the arguments based on policies have lost much of their validity. Third, the number of items involved is so low that it not even registers. When other conferences like TED find their way, it is not a problem so why should granularity be a problem now?
When people want to know about how we think about what we do, the Wikimania talks is a prime resource. Papers are published about what we do. We could easily refer people to Wikimania and other talks. We could and should because it makes sense to do so. In the end it is our history.