If a journalistic institution's goal is to advocate for a side, they are motivated to report on when the other side is doing something radical and not when they're doing something sensible. The opposite for the side you're supporting.

I don't know if it's more or less charitable to suggest that they're primarily driven by revenue. The incentives are basically the same in that model, except the lean is audience-generated rather than centrally controlled. You'd expect more extreme positions on both sides, which is consistent with what you're saying.

A combination of both best fits the evidence, I think. You write controversial articles about both sides to get clicks, and write about the center of your side to get political support. So you get people on both sides with very skewed views (the far) of the other, and who think of their own side as idealogically diverse (the far) but grounded in reality (the center). If the media landscape is primarily liberal, you'd expect it to look the way you're describing it.

I really like that conclusion, because it's perfectly consistent with my experience. I don't really have evidence to support this, though, and I'm not sure how to go about getting any.

That's a great perspective, and seems both logically consistent and supported by current reality.

The "your centrist, but no other" political motivation (even if not economic) is a piece I missed.

Right-wing support for Trump and his far right policies dominates the GOP. Moderate or even conservative Republicans, who even say that the election was legitimate, are cast out and can't win primaries.