What we're seeing is essentially the Death of the Free Flow of Information on the Internet;coming concomitant with a pervasive social slide towards Elitism when the whole point of the original revolution was to break down the siloes of power aggregated into gatekeeper hands.
In the Assault on Reason, AL Gore talked about the ability to craft systems that could enable people to create and engage with content in meaningful ways. No doubt viewing YouTube as an antidote to taking people out of a passive TV watching role and towards an active, civically engaged citizen.
The problem is, when anybody can say anything, Everyone will say Everything. YouTube's Really Dumb Idea here is to "mitigate the problem by eliminating discussion." We see how well that worked for the Catholic Church and sexual abuse allegations, the Silent Generation and LGBTQIA rights, etc.
Right after the web exploded in the early 2000s and we were all chanting "Information Wants to Be Free", a large number of us enthusiastically believed in Democracy and Reason. We believed in rational argumentation, with point, counter-point, analysis of biases and nuances, leading towards an ultimately acceptable conclusion.
Instead we got Flat Earth, Anti-vax, and Donald Trump. It's not hard to see why the enthusiasm is failing.
The problem is, it's not The Big Dumb Idiot Public's Fault. It's Ours, as technologists building these platforms.
A +/- button and free-text fields in 2021 is exactly the same technology we were using to talk in 2001, now with far more robots and malicious actors adding noise.
An Obvious Question that has been around since the advent of these technologies, that is in fact inherent in information itself, is "how do we identify Signal from Noise?"
Treating all Anti-Vax view, and even Donald Trump's banned accounts, as just "Noise" are bad precedents.
Rather, we should be leveraging Machine Learning and all the obtrusive user data we're gathering from ad farms to Actually Do Useful Things For Humanity like:
* identifying a user's background and their "conceptual distance" from the topic at hand. Aside from pointing someone to an arcane Wikipedia article likely last edited by an expert into incomprehensible oblivion, how can we determine a "learning path" series of links that can take someone from "completely ignorant" to "common knowledge educated?"
People love anonymity on the internet, but there are times when they don't, as well. (see Facebook's success)
On a random Hot Topic thread in a non-anonymous platform (FB, Twitter), the commenter's background could be probabilistically calculated to inform readers about how likely they are to know what the fuck they're talking about. I. E. If a random computer scientist blowjob like me is mouthing off about how inefficient the vaccine process has been and someone with a PhD in Molecular Biology from a reputable institution that's actively working in a lab developing in a proximal space replies to my comments with counters, that level of expertise needs to be highlighted.
"water flows uphill" and things that are just counter-factual can be detected and auto-flagged.
Actively elevating the discourse by both teaching people And giving easy tools to identify sources, evaluate their credibility, learn the argument space by highlighting the authorities and hubs of different views related to a topic, and ultimately lead to better discussions is Hard... But an infinitely better direction than draconian choices/policies that directly undermine trust in public discourse --which is a core tenet of democracy--.