My main question is this - I heard Richard Dawkins say in a video that after 1000s of years, any given individual alive today will be either an ancestor to ALL of the humans (in that future time) or none of them. What is the rationale behind this? I'm unable to find the link to that video. I'll keep searching and add it once I find it.
But my fundamental question is perhaps this - is it necessary that a given species must have at least one common ancestor? I understand that there are multiple common ancestors to humans, the most recent one being MCRA. But could it have been otherwise? I think there's some underlying logic that I'm missing.
I'd appreciate it if someone could answer in a non-technical manner since I know very little about genetics and how all this works.
Update:
Consider the following example. Here's there's heavy inbreeding and as you can see no matter how many generations pass, the population are never going to share a common ancestor from current generation or later. (They could still have a common ancestor from before). But this case violates Dawkins' statement.
But this is an unrealistic example with no mixing and cross breeding. Let's consider another example as shown below. I want to see how the individual A from current generation can be related to all the population in future. As shown in the diagram, although A is related to everyone in some future generation, he is not a common ancestor to all of them. The actual common ancestor belongs to some earlier generation.
For any given no. of generations, theoretically, I could conceive similar lineages by which someone from current generation need not necessarily be a common ancestor to all of the future population (or none). So I don't understand how Dawkins' statement is inevitable.