That’s essentially because evolutionary biology, physiology, and ecology are three different fields, with three different sets of experts, and three different sets of papers. If I want to know about the physiology of likely intermediate states, I would ask an animal physiologist, not an evolutionary biologist or systematist.
The size change between those is trivial. There are members of individual genera (like Paleoloxodon) that differ in mass by factors of over 100, and seem to share common ancestors within 3 million years. Birds tend to vary much less, but as there are rather smaller birds known from the late Jurassic (crow to jay-sized), there’s not that much of a jump there.
It didn’t:
Is a perfectly valid guess for how that might have worked. I see no impossibilities in that sequence, and it seems to fit with what fossils we have.
There are fish intermediate between the two.
But there are a lot that they can, and evolutionary mechanisms are one of those.
To take a related example from my actual field of expertise, the number of people that I would trust to accurately identify eulimids in Neogene deposits in the Southeastern United States to species without looking at a reference is 0. I look at my own publication and others to identify them. The number of people whose “turrid” identifications I would trust (living or dead) for southeastern US Neogene deposits is maybe about 10. Because those are the only people who have studied them enough to do so.