Persistent memory is a new technology that straddles the concepts of memory and storage. It can be used to serve as either, but to get the most out of it, you need to adapt a good part of your software. And the kernel is too slow to help you much this time.
As these concepts are not widely known yet, and the software stacks are not quite mature, this talk will mostly cover what you can use persistent memory for, and basics of programming it, but it is time to start thinking how to integrate pmem support in a distribution.