There are quite a few different approaches to the "two column" valuation.

Mr Buffett's is very clean and strict.
There is no provision for underwriting profit or buy-low-sell-high profit on the equity portfolio.

My own approach is to split the firm into those things that are marked
to market regularly and have an obvious value, and everything else.

The "everything else" couldn't exist by itself, but it isn't by itself, so that problem can be ignored.
This includes the facts that float and current assets are needed, and
the fact that the "everything else" part has a whole lot of debt.
Given that view you can value the two sides separately and just add them.

The first one is easy: value all investments per share at market value.
Apply a cyclical adjustment to that if you like, and if you like you can apply a small
discount for an extra level of dividend taxation and the perpetually high cash allocation.

The second can be valued pretty simply based on some plausible multiple
of the on-trend sustainable after-tax income on that, which of course does not
include the dividend and interest income from the marked-to-market half, but could include
a reasonable conservative estimate of cyclically adjusted underwriting profit.

You might use .9 times investments per share plus 9 times pretax EPS on the rest,
or .8 times investments per share plus 10 times pretax EPS on the rest.
I haven't checked recently, but both of those would likely give figures in the $170k-180k range today.
Knock off maybe $15k if you buy my notion that the broad US market is
pretty overvalued these days, and that you conservatively assume that
that issue applies equally to Berkshire's current equity portfolio.

FWIW
For those who want to estimate the long run average pretax underwriting profit,
the average cost of float since and including a horrible 1999-2001 for
Gen Re cleanup has been 0.57% of float, but 2003-2011 it was 3.48%.
It's up to you to decide which is the sustainable level based on
whether you think the Gen Re losses were just rare things that will happen
over the long cycle of reinsurance or truly one-time things resulting
from a badly reserved portfolio of already-written business that was purchased.
0.57% of the current $71.1bn of float is $245/year pretax EPS.
3.48% of the current $71.1bn of float is $1498/year pretax EPS.
1% probably wouldn't be a crazy number to pencil in = $430 pretax EPS.
I think Tilson assumes $1bn/year which would be about 1.4%.

Jim