So if I'm not mistaken we have been mandated here to discount Berkshire based on the over-valuation of Apple now for some time, trillions of market value ago. Yet the Apple itself was probably the largest segment of the intrinsic value gain for all the time we were required to discount it...because of course it was surely headed far lower based on extreme over-valuation.

I sure hope you don't think I said those things about Apple overvaluation!

FWIW, my estimate of the value of Apple is up 80% in the last three years.
Same rise and same trajectory as the (smooth) earnings per share, which makes more sense to me than imagining that fickle market prices mean much.
The fair market cap implied by a constant multiple of (smooth) earnings, as I do it, is up by over $1 trillion in that time.
The earning power is up a lot. Not as much as the stock price, though.

Their earnings are only very slightly cyclical.
Yet seven years ago the market thought the right multiple was under 10 times earnings, and 2.5 years ago thought the right number was well over 40.
I suspect neither figure was correct, and I also suspect the correct multiple won't have changed much.
These days, the difference between those two equates to over $3 trillion in market cap changes just from short term price movements... pretty large error bars for those who use market cap.
To me, that's too much statistical noise to get a meaningful signal.

Instead, it makes sense to value Apple the same way as I value the railroad or the utilities: adjust after-tax earnings for cyclicality if needed, then apply a fixed multiple.
Different multiples for different kinds of businesses, but each one constant through time.

Jim