That the flash translation layer is probably not writing those logical blocks back in the same physical location is pretty much a given, for reasons of simple impatience - doing a block erase to write back in the same place would simply take too long.
So a fairly implicit migration of data around the actual storage is a reasonable assumption.
IMHO, the proper question is not “does the translation layer level wear” - it is “can you trust the translation layer not to get confused about where it put things when a metadata block fails, and fail the entire card”.
What’s particularly insidious about that issue is that not only is the translation layer secret and proprietary rather than many-eyes-audited code, nothing you do at operating system level (short of duplicating data to a distinct physical device) can protect against it. If the FTL metadata gets corrupted, OS-level concepts like partitions are not meaningful firewalls against loss.
And as the memory cells reach the point where read life limitation requires the FTL to periodically move and refresh data just to preserve it, even treating the entire card as being logically write protected in operation is not really enough.
