Prosaically, all of your examples are moderation. And as a private space that a user must choose to access, I'd argue that's great.

There is (or should be, in any case) a difference between moderation and recommendation.

There is no difference. Both are editorial choices and protected 1A activity.

> What part of deliberately showing political content to people algorithmically expected to agree with it, constitutes "moderation"?

Well, maybe it's just me, but only showing political content that doesn't include "kill all the (insert minority here)", and expecting users to not object to that standard, is a pretty typical aspect of moderation for discussion sites.

> What part of deliberately suppressing or promoting political content based on the opinions of those in charge of the platform, constitutes "moderation"?

Again, deliberately suppressing support for literal and obvious facism, based on the opinions of those in charge of the platform, is a kind of moderation so typical that it's noteworthy when it doesn't happen (e.g. Stormfront).

> What part of suppressing "misinformation" on the basis of what's said in "reliable sources" (rather than any independent investigation - but really the point would still stand), constitutes "moderation"?

Literally all of Wikipedia, where the whole point of the reliable sources policy is that the people running it don't have to be experts to have a decently objective standard for what can be published.