One big use case for Boost is to provide alternative implementation for standard containers and sometimes even language features that whatever toolchain you need to use doesn't have yet. Being bound to an ancient Boost version as well kind of limits the use you can get out of that.