I think the answer is clear. They don't care about accuracy in the same way you clearly do. And them not caring about accuracy to the same level as others should be OK; just as OK as caring a lot about it.
If you are building to make a hyper-accurate scale representation of a specific ship, errors will probably bug you. And if a person falls into that category, and likes correcting errors or making things micro-millimeter perfect, that's awesome, because it means they are building models to the standard that brings them joy.
If you are building for any other reason, then your level of give a damn for accuracy could be low. If you fall into this category, and moving a hatch 1/4" from where it should be in life is your thing, that's awesome, because it means they are building models to the standard that brings them joy.
Personal preference is, well, personal, and shouldn't really matter to anyone other than the person with the preference. At the end of the day, as long as the person likes what they are doing, it's awesome because it brings them joy.
Neat how people who do things differently can still be happy doing it, huh?
I'm not good enough at my hobby to be able to build for the world so I build to make myself happy, which was the original aim when I took up scale modeling at the age of eight.
Also time periods matter, details that are incorrect for one time period may be correct for another.
A few days ago I watched a YOUTUBE video of a guy using a hyperdetailing set for the 1:200 Missouri. Having been on three of the Iowas, a number of the details did not look right. I checked my photos and saw a lot of discrepancies.
Jim Slade sent me some of his measured drawings that clearly showed the discrepancies. When I pointed them out to Jim, he sent excerpts of blueprints showing that they clearly were discrepancies.
This makes me wonder, what is the point of hyperdetailing in error? I'd think that you'd be better off with no details than wrong details.