The next Berkshire CIO will do spectacularly better by just buying the top 1000 market cap stocks in equal amounts and holding them.

It would be hard for Berkshire to buy that portfolio in meaningful quantities, as the 1000th biggest firm has a market cap around 1/500th the size of the largest by market cap.
There is a case perhaps to be made for "rank weight".
$1 worth of the 1000th biggest company, $2 worth of the 999th biggest, $3 worth of the 998th biggest, up to weight 1000 for the biggest.
(or maybe add a small constant: using weights 10 through 1010 gets you about the same ratio of 10th-largest firm to smallest firm by both market cap and this rank weight)

But I agree it would not be such a bad investment strategy*, provided (and only in the case) that it was rebalanced to the target weight from time to time. If you let it be for too long, it gradually becomes more like cap weight strategy, which has a huge disadvantage: whenever anything is overvalued, you have a disproportionate number of dollars allocated to it, and vice versa. And Berkshire would never be in a position to rebalance, neither in practical terns nor philosophically, so maybe that's out.

Jim

* The S&P 500 has done extremely well in the last few years, so it has been a hard target to beat.
But for whatever it's worth, imagine someone buying all the Value Line stocks in equal weights every January, then holding them all for a year and repeating the process. Requiring only that each stock have a domicile in the US, and that it have known market cap and ROE figures. (The last two checks eliminate investment funds). This gives an average of 1279 stocks. Performance including dividends and after trading costs but before tax would have beat the S&P by 2.41%/year in the last 27 years. That's probably a very slight underestimate, as this one year hold test assumes that any bought-out or delisted company is held as cash until the next anniversary.

Taxes for this are quite a bit lower than you might guess. Things get sold for two reasons: no longer in the eligible list, which on average means they have not done well, meaning often a loss. And things that have done well and you're merely trimming the positions size. The gains and losses from that churn cancel to a certain extent.