Scientific American hasn't seemed very healthy after the 80s. In the decades before, it was an unusual labor of love by one or two chief editors (I don't remember specifically).
Who is actually being cancelled and for saying what?
This is what I find a little frustrating. There's very little censorship and when it does happen it's usually not against those that most loudly cry about censorship.
For example, did you know you can no longer use the Futurama Farnsworth quote on Facebook "we did in fact evolve from filthy monkey men"? Meanwhile, I've reported and had the report rejected nutters I know literally calling for the stoning of gay people using Bible quotes. (Lev 20;13).
FWIW the moment I started wondering if we were losing liberal norms actually was reading Dawkins in the 00s calling for scientists to coordinate against debating creationists. Like I was with him in being convinced even "scientific" creationism is powered by Christianity and not any good evidence from nature, and I guess I need to say I had absolutely no problem with any scientist choosing not to engage with any creationist. But there's something anti-science in a campaign to expel a belief from public debate, by a means other than better arguments. That can conceivably be a good thing in some case; but it's the opposite of science.
Relying on Facebook is a bad idea because it's a corporation operating under different pressures than healthy discourse, further trying to direct your attention in its own interest, applying resources it gains this way to modeling you. You can try to improve its moderation but besides the trouble you bring up, probably any success you can get that way will just seed a competing platform. I prefer to give my energy to an open protocol such as Bluesky's (admittedly I haven't looked at its protocol spec) -- unless you can take away everyone's personal computers, everyone's not going to live under your favorite monitor. An open protocol is compatible with choosing among competing moderators. (BTW the pre-web Xanadu vision included open-ended moderator choice, and how different system designs could have different social effects, and the importance of getting it right.)