Speaking of comparing Berkshire to the S&P, some folks might find this entertaining.
How have sales grown?
www.stonewellfunds.com/BerkshireAndSpyRealRevenue.png

"Sales per share" isn't a great metric of value when comparing two companies, but in this case it's not too terrible.

For the S&P 500 it's quite a useful metric, since it's such a broad set of firms. Over the very long run profits can't rise any more quickly than sales can. Shorter term variation comes mainly from changes in tax rates and the costs of borrowing. Plus a shorter-cycle component for the business cycle, affecting operating margin, giving the one big advantage of using sales versus profits: sales are very much steadier through the business cycle.

For Berkshire, there are two flaws in looking at revenue per share as a proxy for value.
The big obvious one is that revenue doesn't include "look through" revenue of the stocks we hold. (The only mitigation here is that the value of the investments has been a remarkably steady fraction of the value of a share for a very long time, so the rise in revenue on the subs side is probably, by coincidence, not a horrible measure of the progress of the investments side too).
The less obvious is that sales per share will jump around each time there is a transformative acquisition or consolidated investment that is consolidated.
In this stretch, however, there are only two of real note: BNSF in 2009, and the kink in the last period from the consolidation of Pilot since February.
So, overall, I think it still gives some information.

Boring footnotes about the graph:
The data points are monthly, but the data are only quarterly for the S&P 500 and annually for Berkshire, with linear interpolation for each month end in both cases. Revenue for each is inflation adjusted, then plotted as a log value so a constant growth rate would be a straight line. The two are scaled arbitrarily to align them at a baseline at the start of 2004.

Jim