I think the ultimate problem here is one about accountability and quality control. As a game designer myself, I absolutely understand and agree with having next to no democracy for many design decisions because game design, however complex or simple people think it is, ultimately has wiggle room. There are a few no nos that are universally agreed on, and there are many identifiable objective bad design choices, but these often are mired by people arguing with fallacies or require a bit of study and dissection.
This is where I feel the point of quality control/accountability comes in. A problem I've seen excessively of late is many merges are introduced that are either hilariously buggy, simply do not work, or quite frankly giant pieces of shit. Pubby and Donut SM merges were tire fires visible from space. The Russian Surplus PR was a complete shitshow of balance that not a single maintainer cared about. Fireman carry flat out doesn't work on the ones you'd expect it to. Even the stamina crit debacle was insane. All of these are actually recent and just off the top of my head.
Meanwhile, we see PRs being closed for either completely arbitrary reasons or because the maintainers usually want those coders to go above and beyond what the scope of the PR was. There was a PR to move mech rockets into security lathe and the response was effectively "come up with a completely new system just because;" Specifically, a way to have different access to different lathes. At this point, any new coder is immediately shut down because this may be so far out of scope for their skill that it doesn't have any chance to get done. Meanwhile, maintainers can push whatever they want fundamentally ignoring the restrictions they impose on others.
This brings up an additional problem in that maintainers can arbitrarily decide if or when something is good. Nominally, in any game design or large term project, you establish a design document or overreaching policy to help guide people to the goal they want to receive. These could be somewhere between simple to extensive, but gives contributors factors they can understand and limits they can experiment with to really enhance the open source portion of the platform. Saying something as simple as "Security should control production of all weapons on the station" is a very simple design goal with a stated goal. Now you can make pulls that police cargo null crates requiring (say) a security card to open. Mech weapons are pushed to security or robotics is made a sec department. This is a very simple design document, and is just an example, but it creates an understanding that others can follow.
Our current maintainers do not seem to be interested in doing anything even remotely like that.
This just goes back to the above problems of accountability and quality control. There are no maintainer elections or coder elections. This means unless they screw up so bad you can see from the moon, they will likely receive no need for accountability. If they do not change, however, this does mean that a global design document can exist (and should exist) because it allows open source contributors to push forward to it.
I think the policy going forward should be one of bringing coderbus back down into the realm of accountability or at least making them require certain goals or thought processes beyond "I feel like it." This can be done with some measure of elections or just by forcing them to do documents with certain restrictions then punishing them when they feel otherwise. It's getting scarily close to some things getting merged for no reason whatsoever while the person atop the INTERNAL DISCUSSIONS tower mocks the flunkies below.
As for democracy, an important part of game design is also responding to community wants and needs. Fighting game combos are probably the universal example of this as they were never intended, but once they came out the game designers realized its popularity and started making them baseline. For a most esoteric example, Wavedashing is another great one as the designers acknowledged that it allowed for higher level play. They disagreed with making it baseline, and would eventually evolve it into Perfect Pivot.
Players may be terrible judges of making good design decisions, but many can and readily will tell you when something
feels fair, interesting, or engaging. This is also why metabreakers tend to come either with small choices that expand out dramatically (Brigitte from Overwatch) to complete reworks (pick a jungle rework from League of Legends). The difference is, most of those changes tend to be a little more accepted because you take the time to explain your thought process and you engage in the community. Alternatively, you could just do the Blizzard WoW aspect of "We know better than you." Hint: That almost never works because statistically, there are more people willing to show you why it doesn't. Related Link (Massive Irony Expected):
https://1d4chan.org/wiki//tg/_gets_shit_done