It's not revisionist. The problem is identifying which ones will survive ex ante. Given 100 assets to invest in, if you knew that one of them would become amazo

It's not revisionist. The problem is identifying which ones will survive ex ante. Given 100 assets to invest in, if you knew that one of them would become amazon in 20 years, but the other 99 would die, how do you value them? Well, you value all 100 at 1/100th the expected value of Amazon. Now, maybe you look at each individual company and try to pick a winner to put all your arrows behind, but assuming you can't reliably do that (which is a reasonable assumption), you would want to value all 100 at quite a bit more than they currently appear to be worth.