But there WERE more than two points; surely you know that. Q4 was not the first quarter that Model 3s were produced. Q3 of 2017 is. So that's three points. To

But there WERE more than two points; surely you know that. Q4 was not the first quarter that Model 3s were produced. Q3 of 2017 is. So that's three points. To use your example: T-1: 2 vehicles T0: 10 vehicles T1: 40 vehicles That is exponential. If it was only a linear ramp-up, that'd imply negative 20 Model 3s were produced in Quarter 3 of 2017. You might think this is being pedantic, or that "exponential" is just a buzzword, but it's not: the point about exponential ramp-up is that if you're off by a factor of, say, 4x in production, that may mean you're only behind a few months. A relatively small time change corresponds to a very large difference in production. Exponentials don't last forever; eventually a plateau is reached. But as you're bringing automation/production up, you're still in the exponential phase (as the data shows), and so small delays can make a big difference in production without telling you much about how many you'll produce once production is fully up and running.