That is a terrifying argument.
I’ve only heard two arguments- let Twitter do what they want (viva Capitalism!) or have government regulate (viva Socialism!).
I wish there was more thought given by Twitter, and others, to just self apply free speech rights and stick to using the current legal system to restricting speech. There’s a ton of law to protect people from damaging speech (harassment, slander/libel, dangerous threats, etc). It’s quite possible that Twitter can just use a more rational standard for their process.
What the Government can't do is tell Twitter what statements it is or is not allowed to publish (authoritarianism).
And unless you are okay with your bank providing unfettered access to your ID, account number and PIN, I'd imagine that you are not in favor of general unfettered access either.
Democracy may require access to information but civilization requires that information not be universally shared.
The difference between WSJ and Twitter is that the WSJ has competitors (NYT, Politico, even Fox News and crazies like Infowars) whereas Twitter does not. We live in a world (unfortunately) where the President of the United States and his administration heavily relies on Twitter - it's no longer someones backyard social group. For Twitter to play the censor is no longer acceptable, although we still expect twitter to keep its users safe from harassment and mobs.
So yes, it's unacceptable for Twitter to play the censor or ban people from public service announcements or policy statements by their elected officials. If Twitter is going to facilitate administration statements and my local government agencies letting me know critical information, they are no longer an entirely private platform.
I agree that having restrictions on information access and communication is a slippery slope in the non-fallacious sense, at least with the solutions we usually discuss.
But I’m optimistic that the failure conditions might be more specific than they seem to the current state of affairs — self curated social graphs of unlimited size and bandwidth. Perhaps dramatically limiting scope, connectedness, or agency over whose views you have access to (think true random, not AI) would lead to less or different issues. I don’t know, because these experiments would take a lot of investment, but the point is that it may not be as black and white as you make it out to be.