To them, the computer is a tool, not an environment, and to use it is to press buttons to get some tangible result you want from it, and the buttons you have are not buttons you are going to know how to change.Your general point is correct, but the analogy you're using is incorrect.
It's not tool/environment, but rather, it's tool/appliance.
The fundamental cultural shift going on between many anti-systemd and pro-systemd people is that the former want the computer to be a tool, something that with proper discipline and skill can yield great things, whereas the latter want it to be an appliance - something that you fire and forget, and is meant to be inflexible and that "just works".
The problems, as always, are that tools have learning curves and often need to be composed into more complex configurations to yield more powerful results. Whereas appliances, although they "just work", they work only for predictable and formulaic configurations that the appliance explicitly expects, and proceed to break down and reveal their opaque innards when used for more esoteric or unusual situations.
In summary, the anti-systemd side wants to do things with their computer, and the pro-systemd side wants their computer to do things for them.
This is why there will never be a consensus. It's also why systemd opponents will often resort to stupidity like defending the messy hairball of SysV initscripts, while proponents will just casually look at unit file configuration being outwardly simpler and not think of the bigger picture.