What are you proposing exactly? I'm trying to steelman your position, but the best I can come up with is "Only the rich should be able to spread their opinions (which are obviously correct, since they are successful)" and "You can have your free speech in your designated free speech zone (which has bars on the outside, and is surrounded by armed guards)".
This is a clear false dichotomy. The platforms that monetize outrage and disinformation by rewarding it with free reach are at issue here. As to solutions - The only way to filter the volume of information reaching us into meaningful, useful, true narratives is through trusted filters that prioritize attempting to be truthful and operate in good faith over time. Fact checking creates a history of this which should feature into algorithms. This is not to say that people are prevented from speaking, but amplification should come with responsibility.
Trust in the facts of reality can only really be established as a chain - this is how the sciences work. That’s why independent experimental verification of results is so important. In the information ecosystem, fact checkers perform this function. We should establish structural means of support for these functions.
> Fact checking creates a history of this which should feature into algorithms.
At first glance, it looks like you've just re-defined the problem from "platforms should decide which posts are true" to "fact checkers should decide which posts are true", but actually I think you might be onto something.
Platforms could use their existing user-reporting features, and measures of audience size, to inform fact checkers about which posts are the most controversial; then different fact checkers could (reactively) investigate the claims and offer rulings that would appear as a banner over the relevant post.
The missing pieces of the puzzle are that individual users should be able to opt in to one or more fact checker, and platforms should give users a default set of fact checkers, and the fact checkers should be funded by a 1% tax on platforms' revenue (split in proportion to the popularity of those fact checkers among users).
This could actually be a valuable and viable system, and maybe the sciences could also benefit from a system like this for flagging suspect papers and paying for researchers to attempt to reproduce results. It could even be gamified, with researchers winning "points" for correctly guessing which studies have non-reproducible results (as long as these points didn't have financial incentives attached which could corrupt them).
