True. But I'm also worried about the "removes ALL boundaries" claim. I don't know btrfs or the algorithms well enough to do the math, but the claim is, essentially, that btrfs may, in the pathological case, consume an infinite amount of space, for actually storing nothing.
That is unlikely to happen for real workloads, but should still be impossible. Especially since pathological workloads can often be provoked deliberately by an attacker. (typical for race-conditions, for example, even race-conditions that would "almost never" happen in normal situations, become a problem if an attacker can deliberately create them)