The danger of social media is that it fosters radicalization by connecting extremists far more efficiently than is possible offline, and that in creating communities of extremists who give themselves psychological permission to act out their beliefs. The consequences of this spill out into the real world in ways that society is not fully able to manage.
People using social media to call each other shitheads for supporting (or not supporting) Joe Biden isn't a problem. No one outside of those involved in the conversation will have any reason to care and the people involved will get over it.
Social media connecting unhinged people, however, and helping them create a parallel reality where burning down 5G towers becomes the right thing to do to protect people from autism, is very much a problem. This kind of activity spills over into the real world and does have direct consequences.
Regulations of social media platforms ought to be far, far more concerned with mitigating the risks of the latter problem than with maintaining civility.
It won't, though, because the best feed algorithm for increasing engagement and stoking outrage and making money off ads is not the one that is best for fostering calm discussion and building a community.
Twitter etc. want you to use the engagement one, but in order to actually fix the social problems created by their platform, people have to use something else that will make them less money.
I personally think that's fine, but I doubt Twitter will be happy with that. And I'm not sure how you legislate a requirement that your platform isn't psychologically tuned to bring out the worst in people.