The issue here is that saying a thing on twitter with the potential audience of the whole world is an inherently different thing than e.g. saying that thing to your blokes in the pub.

Imagine it is 1980 and you went into a disco with a certain percentage of people of $ethnicity: What is your expectation for what would happen if you started yelling that people of $ethnicity have $negative_attributes and no right to exist? That they come to you and tell you that you have a point there? Of course the group will not ignore you, you are directly criticising their existence from an outside position.

And this is the issue here: expectations. When speaking on a global online platform with near instant feedback your expectation shouldn't be that you can just voice a controversial opinion and have nobody disagree like it would happen when you talk to your likeminded friends in a small circle. Your expectation should be that you are in close vicinity of all the people you are talking about, because you are.

And this can also be a good thing, because assuming a ton of things about others that you could just directly talk to is a sounds like a receipe for disaster. Imagine that instead of yelling stuff in the disco you were to talk to people and getting to know their point of view. The issue is that you can't do that on social media because you are always in the spotlight. And in the spotlight it is hard not to wave the flag of your team and have a human conversation instead.

Social media is a constant spotlight, global, instant and huge public space and the behaviour you discribe is a direct result of the fact that people treat it as their living room conversation instead.

Shitstorms are real and often they can get quite bad, but paradoxically you won't get rid of them with less moderation ("more free speech") but by interfering with the spread of the word.