Face recognition would only matter if the adversary was targeting someone specific. There are simpler use cases than that: who was present at a protest for example. Activists could find themselves in a situation where it makes more sense not to get tested in order to avoid correlating their identity to their movements.
Just because you chose to sacrifice your privacy for convenience doesn't mean it should become a baseline for what we expect. Especially nor for government services.
I don't think those are good examples. The communities you described formed because they were unwelcome in the mainstream, so it's only logical that they would devolve into radicalization as that's the very essence of their existence.
I don't think social platforms should be non-moderated. Calling for harmful action should lead to a ban. Believing unrealistic things shouldn't. In such a context, most non mainstream communities would have no need to exist in a bubble. In the same vein downvotes and upvotes are fundamentally antithetic to free speech as they only serve to amplify the dominant echo and push minorities to their own bubbles. Instead people should be able to downvote locally, informing their own future interactions without affecting the community as a whole.
This is a disingenuous argument. When all mainstream platforms decide that your opinion is not welcome, your only real alternative becomes either silence or small echo chambers where you're more likely to enter into radical feedback loops. This is not healthy for our societies.
Historically, public spaces have been used to confront ideas. Free speech isn't a luxury, it's a necessity that ensures our societies don't devolve into radicalized micro bubbles that are more likely to resort to violence. The issue that we're facing is that the public space is being replaced by the virtual private space, which doesn't have such guarantees. We're already seeing the effect of such divisiveness already, and it's only going to get worse. So should we just accept that free speech is dead? I don't think that's the right direction.
By this logic Twitter should also ban all Muslims because there are extremists who call for the death of infidels. Instead it's far more reasonable to just crack down on hate speech and harassment in general, which I would have assumed they already did.
You don't watch YouTube videos for free. You pay by relinquishing information to Google with the implicit understanding that they can monetize that information. If you are okay with that, or aren't fully aware of the implications then that's a different matter.
In any case there are people who have decided that such transactions ran against their own interests and prefer to pay content creators directly. Now if Google offered a platform to discover content creators, and if I trusted Google not to try to profile me, I would gladly pay a premium to access it. But that's not the case, so we rely on services like Invidious and NewPipe to access the content while remunerating the creators through different channels.
In the long term we hope that people are going to take the issue of privacy seriously enough to either cause Google to change its model or to at least cause creators to publish their content on privacy-friendly platforms.
In case you care about these things you should be aware that Google can use your account to track you, assemble a profile on you and target you even if you don't explicitly hand them any data.