Hi, Lamden.
Welcome to EvC!
I want to point out that I have a lot of skepticism toward any attempts to assign hard numbers to these things. In my opinion, not only is it a difficult answer to get, but it's also a rather meaningless answer to get.
Evolution isn't an entirely linear process. At any given time, there are many lineages living, reproducing and mutating simultaneously, and there are many ways in which those separate lineages can intermix their genetic material. So, you effectively have multiple "generations" happening at the same time. So, we'll need a parameter to describe how many lineages are likely to have been going at once throughout Earth's history. We'll also need a parameter to describe what proportion of the accumulated mutation load of a given lineage is likely to be transferred into another lineage. Then, we'll also need some parameter to describe how many of those lineages will go extinct, irrevocably removing their accumulated mutation loads from the equation.
To me, the game of parameterizing a model like this feels like sitting in front of a huge panel covered in dials, and arbitrarily dialing up and down until the numbers look reasonable to me. It's probable that any number I might fancy could be achieved by a variety of different combinations of the dials.
So, even if this thread resulted in some number that we could take home, it would not satisfy me in the least. As much as I would love to have you convinced of the reality of evolution, I see no point in encouraging you to accept it because of some silly numbers game that we could only fail if we grew bored of constantly turning dials.
-Blue Jay, Ph.D.*
*Yeah, it's real
Darwin loves you.