@101101000 The thing is that Twitter, as a corporation, is a legal person with its own free speech rights. When you see a tweet from Trump, what you're actually seeing is not Trump's speech but Twitter's - it's not Trump saying "covfefe"; it's Twitter saying "Trump said 'covfefe'". If Twitter stops publishing Trump's content, it's exercising its own free speech rights (specifically the right against being compelled to utter speech against its wishes), not impeding Trump's.

Mechanically, Trump could craft and send whatever HTTP payload is used to submit a tweet for publication, even if the website frontend no longer offers him an input. That's his free speech right. But he has no right to make Twitter do anything in particular with that content once it's received.

That's why I think that a serious free speech platform has to be pay-to-use. Once you enter a *contract* with someone to publish your tweets, you have contractual rights (still not free speech rights) to see him honour the deal. Sort of the reverse of a non-disclosure agreement - if you sign an NDA you can be sued for certain speech even though your free speech rights would ordinarily protect you in doing so; what we would want is the mirror image where the provider can be sued for failing to utter certain speech even though his free speech rights normally shield him from having to do so.