Very true. I do expect that the tools will advance and that in a decade many of us will be supporting multiple databases. Perhaps that is some different "version" of SQL Server (a NoSQL edition?), but likely I expect that we will find NoSQL type solutions just solve some problems very well, and they are worth implementing.
I also suspect that much of the workforce that comes along in the last (and next) 5 years will be more comfortable working with multiple technologies and won't get bogged down into using this vendor or that one, or this language or that.