You should really have a look at the book. The point is that you can pick any set of events you like and you always get the same curve. That is not because there is no objective meaning - it is because technology accelerates, it grows exponentially rather than linearly. That is the point. He's describing the Moore's Law of Technology.
OK, I pick the events called out in the History section of this article on clocks. The data points are these
Years ago | Years delta | Event |
5006 | 2913 | clepsydra |
2093 | 937 | Antikythera mechanism |
1156 | 326 | mechanical timekeeper |
830 | 74 | Sens Cathedral installed a horologe |
756 | 184 | earliest reasonably accurate clocks |
572 | 223 | first self-striking clock |
349 | 13 | pendulum clock. |
336 | XX | anchor escapement, |
Here are the points for log-log plot
3.699490845 | 3.464340485 |
3.320769228 | 2.971739591 |
3.062957834 | 2.5132176 |
2.919078092 | 1.86923172 |
2.878521796 | 2.264817823 |
2.757396029 | 2.348304863 |
2.542825427 | 1.113943352 |
If you plot them you don't get a line. Instead it starts out curved then hooks up and then back down. I wish I could post it somehow.
Do you get the point now?