There are around 6 billion humans, so this would make the total water consumption equal to around 16 times the human population.

I started out writing this post on how this number is way too low, but it's actually more reasonable than you might think. I do think it's off by an order of magnitude (but I thought it was off by at least 5).

Assume water consumption is roughly proportional to mass. The rest is based on wiki Biomass article. From the "domestic biomass = 700 mill tonnes = 1% of earths biomass" estimate a total biomass of 7*10^10 tonnes. Say a third of that is from animals (there are a lot of small critters in the sea, so maybe), so we have around 2*10^10 tonnes animal biomass. 6 billion humans = 1*10^8 tonnes biomass, so we're talking the equivalent of 1200 billion humans, ie, around 10 times your estimate. So that would make the two quantities even closer.

The interleaving is assumed random. I believe that the deck is randomly divided into 2 stacks, and then we sequentially take cards from each of the stacks, choosing the stacks with probability in proportion to the number of cards in each. I would actually take a look at the paper if you're interested - the result is not that esoteric. What's nice about the proof is the very clever idea (something about certain sequences of cards), which makes the proof itself pretty accessible.

Her mom was an 'analyst', so she probably wasn't incapable or ignorant. I agree, though, that children are the future, which is why my children will all use Macs.