Short answer: Some people just like anthropomorphizing more than others do.
The comment made by user "a human" on the original question inspired the following train of thought:
Take two people, Alfred and Zack. Alfred is a "people person". Zack is a mathematician. Your satisfaction with this answer will likely depend on your tendency to support the following intuitions about Alfred and Zack.
Alfred likes to think of the world in terms of what people intend and want. Key word: people. So when confronted with more abstract concepts, like chemistry, Alfred handles things better when he can assign them names and talk about their behavior in terms of human behavior. This is called Anthropomorphizing. When dealing with people, it works really well. When dealing with abstractions, it's kind of a hack; it gets him by, but doesn't provide him a platform with which to excel at, say, vector calculus.
Zack is less socially apt than Alfred. Maybe a little nerdy, withdrawn. Maybe he has Asperger's. But, he's very good at abstract thought. He doesn't "get" people like Alfred does, but he can do proofs like a boss. Calculus? No sweat. Physics? Easy.
(If you're Zack, you might wish I had just talked about "people person" and "not-people person" in the abstract, rather than creating concrete examples and giving them names.)
I'm going to venture a guess that Alfred finds it harder to believe in a Godless universe than Zack. And that these perspectives aren't likely to change much with time.
So, what makes Alfred and Zack different? Is it neurostructural, neurochemical, or simply habitual? If it's neurostructural, they're doomed to disagree, but maybe their kids aren't. If it's neurochemical, then maybe one or the other will go on Prozac (or something stronger) and temporarily come to Jesus (or run away from him), then go off Prozac and revert back to his old self. If it's habitual, then maybe they'll switch places down the line due to some personal experiences.
EDIT: In light of the above post (about empathizers and systemizers:) the Alfred I described is an extreme empathizer, while the Zack I described is an extreme systemizer.
In the field of neural networks, the number of layers in a network is sometimes said to have to do with the "amount of abstraction." To me, this suggests that different human brains might also have different network depth. So, a 'deep' human brain might be more abstraction-focused and, perhaps, systematic (Zack) while a 'shallow' one might be more feeling-focused and empathetic (Alfred.) (I apologize for the slightly stigmatic nature of the words "deep" and "shallow" -- maybe someone can suggest alternative terms.)
That would be the "neurostructural" explanation. But it seems clear that neurochemistry also plays a role in empathy (we're starting to think that even Tylenol might have an effect on empathy, not to mention stronger substances like cocaine) so clearly the issue is more complex than the above duality would like to suggest.