> While I find the 17% number interesting to think about, "most" is far less interesting. Multiplication doesn't care about order so you're instantly cutting 2^64 possibilities down to about 2^63. That's a hair's breadth away from "most" already
It's much worse than that. It's difficult for a 64-bit product to have the high bit set if the multiplicands are both no larger than 32 bits.
I wouldn't say difficult. To set the high bit, the geometric mean of the two multiplicands has to be at least 3 billion. 32 bit numbers go up to 4.3 billion so that's quite common.
If I did the correct integral (the area inside the unit square above y=.5/x), then 15.34% of products should have the high bit set. That's much less than half but it's still happening constantly.