Discourse quality is 99.9999% a trust/safety/moderation policy and tooling outcome. Content can be problematic on any of dozens of dimensions (tone/energy/relevance down to actual criminal consequences) at any size. Whether conversations retain signal and communities develop and grow trust entirely depends on what decisions moderation does or does not make.
Did shorter content chunks simplify the moderation problem, in the context of an org that wanted to build communities? Absolutely. Does Twitter post Elon have completely (completely) different goals and values? Of course, yes, opposite end of the spectrum. What relative percentage of resources does Elon twitter devote to moderation? Maybe 1%. So is it obvious that conversations are going to be completely different and lower "quality"? Yes, yes it is. That is a goal.