Sucks to be Boeing, I want them to be successful. Hell, I even liked that big goofy looking X-32 which was a JSF competitor. People die when their planes fail, it's nothing like a PCB as mentioned by the grandparent. They f-ed around with the 737Max and found out. Honestly though, if your loved ones died on a 737Max and then you found out that they weren't building planes to the specs that they defined, what would you think?
Bob: You've called for a 5mm hole here, but you don't have a precision specified?
Alice: I really only need the sort of hole a 5mm drill bit would produce.
Bob: Do you know what level of precision that is? Have you done any calculations to confirm that's the appropriate level of precision?
Alice: Not really, in my judgement this doesn't rise to the level of needing such calculations.
Bob: Well, our drawing quality standards require a precision to be specified. Would ±0.5mm be OK? If not, how precise does it need to be?
Alice: How precise is the laser cutter we're cutting this out on?
Bob: The spec sheet claims ±0.05mm
Alice: The required precision is ±0.05mm
Not every hole needs a tight tolerance, depending on the loading and assembly methods. And manufacturing prefers a standard tolerance range to be specified to enable consistency in manufacturing operations, rather than some weird lopsided super precise tolerance callout. So you determine the largest tolerances that will meet your design requirements, then select a fit class from the manufacturing standard that meets your tolerance requirement.
Sometimes you just need a specification, because you have to tell your mfgr something[0]. Notably 5 thou is a pretty standard idgaf-tolerance, but in this specific case might be important.
[0] e.g. a spec of 5" is meaningless, 5.00000 +/-.000001 is insane, and 5.000 +/- 0.005" is (generally) readily achievable and good enough.
The way I see it is, if there's a spec. that specifies such a small tolerance (in an airplane of all things), it must be there for good reason.