> If their percentage of society becomes dangerously large then people will vote what to do or the system will naturally collapse and be replaced with something else.

i.e., if one of your foundational assumptions is wrong then the system won't work.

Put another way: the system won't work unless your underlying assumptions about human behavior turn out to be true. That is, unfortunately, a key flaw that has doomed many utopian visions in the past, and will likely continue to do so in the future.

One of the most important features a socioeconomic system can have is resilience to human misbehavior. A great many systems would work if humans behaved in just the right pro-social ways, but history has shown us that it's very naive to expect that from large (10M+) spread-out (country+) societies. One of the great successes of modern systems such as democracy and regulated capitalism are that they are fairly resilient in the face of bad actors.

Take, for example, democracy vs. dictatorship. At its best, a dictatorship can be amazing -- decisions are made quickly, efficient solutions are deployed, waste is minimized -- but in reality bad behavior tends towards the "Dictator's Trap" of fear leading to poor information leading to poor decisions leading to poor outcomes; at its worst, dictatorship results in genocide and collapse. Democracy, by contrast, has much less variance -- it can never be as efficient as an enlightened dictatorship, but it will also never sink to the depths of a corrupt or murderous dictatorship, and history has shown that on average the democracy will tend to give better results.

So while it's certainly interesting to consider alternative social, political, and economic systems -- and while there are almost certainly better ones out there that we haven't tried yet -- those systems have to be resilient to human misbehavior to be even remotely realistic.