Besides, the rarity of predation on giant herbivores seems to be
relatively morphology independent - there might be the occasional
big-game specialist in the Mesozoic, but I don't see any reason to
think that the rarity of predator attacks on much larger prey across
multiple clades and environments today should not also apply to
Mesozoic systems for the most part.
In fact, I'll pose the reverse question: why should we suppose that
giant herbivores in the Mesozoic were predated any more readily than
giant herbivores today? (noting, of course, that "giant" is
relative, and so we really mean the size ratio of predator:prey).
The Mesozoic ecologies didn't have to work like modern ones, but
given how widespread the size advantage trend is, I think we need
more than what we have at present to overturn the null hypothesis
that the Mesozoic size ratios had the same effect as present ones.