If the tech companies want to short-circuit the right's ability to organize from now until the 21st, that's kind of reasonable, given recent events. Past that, though? They need to be very careful about continuing this. If they do, they need rationales and policies that are clearly stated, and clearly seen to be applied in an unbiased manner.
Things might be different at national level, though. States have to deal with more nuts-and-bolts practical issues, so that grounds the politicians somewhat. National issues tend to be more idealistic, so one-party rule could become extreme.
I know several conservatives who voted for Trump not because they love the guy, but because in the two-party system he's more closely aligned with their personal views on things like immigration, taxes, military, etc. I'd also argue that for the average American, the Trump of 2016 wasn't exactly an extremist hoping to overthrow the government.
The American political climate has polarized both sides. It's really unfortunate that in some circles, someone can't say or express conservative views without being lumped into the radical right-wing Trump Patriots. Similarly, some can't express liberal views without being lumped with radical left-wingers. And as a result of this, de-platforming the right-wing extremist is commented on as "silencing half the country" which just isn't true.
[edit] - as an nuclear engineer - that study you cited reminds me of perturbation theory. You make a small tweak to the system and develop a model to fit how such small changes affect the larger system. But perturbation theory conclusions don't scale accurately up to "large" perturbations. If you pull the control rods a tiny bit in a subcritical reactor - you get higher neutron-multiplication factor for your existing/present neutron sources. So you might think that if you pull the rods 10 times as far, you get 10 times the multiplication - but you'd be wrong. If you pull them 10 times as far you go super critical and then you get like 1,000,000,000 times as much multiplication.