If we could compare the curvature of the Earth to the expansion of the universe then it may make sense. Over short distances the expansion isn't that noticeable, but over larger distances it is noticeable. Is this what you meant?
I'm not entirely sure either.
But my guess is that your take-away here is slightly off.
The curvature of the Earth analogy wasn't meant to be specifically applied to the expansion of the universe.
The curvature of the Earth analogy was meant to broadly suggest that there are times where we "think things are as we understand them" (early science thinking the world is flat) but they eventually turn out to not be true (science eventually uncovering that the Earth is round).
I think this curvature of the Earth analogy was meant to be applied to the broad sense that our knowledge-of-the-universe is at the level where we "think things are as we understand them" but cannot guarantee that this will not "eventually turn out to not be true."
"Eventually turn out to not be true" does not apply to anything specific.. like expansion of the universe or current-verifications-of-the-speed-of-light. It only applies to the very general "we don't know as much about the universe as we know about the surface of the Earth" and therefore, there is (significant?) room for a "oh, hey... the earth is round" moment in our understanding of the universe.
It seems unfair that the point is general and vague. However, the point necessarily must be general and vague. If it had specifics... well, then we would be on the road to beginning to understand if those specifics will lead us to an "oh, hey..." moment.
The entire point is that our knowledge-of-the-universe is at such a level where there is easily enough room to fit in such general, vague and historically-precedent-ed issue-possibilities. That is, science-history is filled with "oh, hey... " moments where we thought everything was understood this way, until we also uncovered information that shows it's not "this way" but actually "that way."
If that is, indeed, the point. I would offer that science is well aware of this. It is not a new idea by any means. And that the only way to make progress into it... is to carry on exactly as we have been:
-making predictions based on things as we "currently know" them to be. -act as if everything acts as we "currently know" them to be. -when (if?) anything ever contradicts these assumptions... then study the area for replication/verification and a growth in understanding to update what we "currently know."