The number of atoms on earth in the past isn't relevant because we were not close to exhausting the earth's resources in the past. The author is demonstrating that if this rate of growth were continue in the future, eventually we would run up into physical limitations; the idea that we could sustain a whole present day earth's worth of activity for each individual atom in the galaxy seems more than a little unlikely. Of course, there would certainly be other bottlenecks hit first; looking at atoms in the galaxy is just a clear-cut way to show that there is almost certainly a limit to growth somewhere, and that even in that extreme case we would hit it in a short time frame, relatively speaking.
