Our theories of the nature of the universe give us good reason to suppose that there is nothing particularly special about the fundamental constants nor any reason to suspect that they could have been quite different. Imagine each value as a single number corresponding to a coordinate. As there are ~20 constants, we'll say there are 20 values, so there are 20 coordinates, requiring 20th dimensional (Euclidean) space. This is the probability space, in that every point is a set of possible values for the constants. This space extends infinitely in 20 different "directions" (or along 20 different dimensions). We would expect that, given the enormous number of possible values, there must exist a wide range for them such that large variations wouldn't really effect much. That is to say, we would expect that the probability space of universes in which we could exist out of the total probability space would be a large region.
In other words, we expect to find what we do in your example of whales and salt water: yes, wales require salt water, but you can change an huge amount of variables by (comparatively) incredibly vast amounts and not really do much, including the salinity of the ocean. Instead we find that rather the probability space for a universe in which we could exist isn't some huge region, but more like a single point.
Not only that, but if we assume that the universe was designed for us, we can (and have) made successful predictions about its nature. Not only that, but the other alternative in that paper which mentions the two popular explanations for fine-tuning, the anthropic principle, suggests we take as fundamental not fundamental forces & particles but us. It isn't much of a leap from here to suggest that if we should do this, and it works as it has, then our status is equivalent to that of fundamental particles/forces, and as we weren't around to be fundamental anything until long after the universe began, that it was designed for us.
In essence, the ways to explain the precision of the values for these fine-tunings have been to propose that our universe is one of a vast number, to suppose that it was designed, or to suppose that we can treat it like it was designed.
Same here. The universe has the constants.
The fine-tunings are not. They are given only once the universe is a given, or alternatively that granted this universe we live in has the fine-tunings it does, it has them. But it need not have them, just as you need not have one the lottery. And we have reason to suppose that it should not have them.
Why do physicists call fine-tuning a problem?