I don’t know if what I have to say really is all that interesting or anything, but I guess I’ll give my opinion anyway.
It seems like a lot of people either do tons of SRS or nearly none, but I think the most efficient way to do it is to balance things in a way that makes sense and addresses your weaknesses.
When it comes to reading, that would mean using some SRS if you come across a decent amount of words you don’t know (but not to the point where it’s the only thing you do, because that just seems pointless) and spending more time reading instead if you only come across a few words you don’t know.
That’s obviously just the way that seems best for me though. Regarding the debate about learning through exposure, I know it’s possible because I learned English that way, but it also took a really long time before I really could consider myself fluent at a level at least close to a native. I honestly feel like I’m learning Japanese faster than I ever learned English despite how it supposedly is much harder. Part of that may be because I have more experience with learning languages in general, but I’d like to think at least part of it is because the methods I’m using are more efficient.
To summarise: I think both SRS and exposure help a lot. Exposure is more important, but using SRS as well is more efficient for me, judging by how fast I’ve learned Japanese so far compared to how long it took me to learn English and German without using any SRS at all.