I know what you mean, but it’s important to realize that kind of thinking is a sunk-cost fallacy. Based on the GitHub discussions about filesystem performance,

I know what you mean, but it’s important to realize that kind of thinking is a sunk-cost fallacy. Based on the GitHub discussions about filesystem performance, it seems like they had really hit a wall and there was no other way to squeeze more performance out without a radical re-architecture.