I think the point of the parent was that systemd will be much harder to replace than other init mechanisms once something better comes along, because it grows into every part of a Linux setup. And if that's true, then systemd actually hinders progress and innovation, because the cost of replacement will be too high.