Another answer to this is "Yes, with proper rules and moderation, you can keep your cake and eat it too".
First off, have a rule that every created thread in this forum need to have some sort of "basis for discussion". Any thread that doesn't, gets closed.
Secondly, have a rule that states that any posts in the threads need to be at least somehow related to the topic, and posts that are not nonsense or off-topic.
Thirdly, have a moderation team that is well aligned, and favors quality posts, removing the rest of the shit.
The end result is a forum where you can discuss any topic, as long as your own posts also add to the conversation, not just try to shut others down or post off-topic/nonsense.
Sure, it will contain opinions you disagree with, but it won't be a cesspool of just casual racism.
(These ideas are all based on a real forum online today where essentially any topic can be discussed without fear of facing real-life consequence for the opinions you hold, assuming your anonymous)
They want 0 limits on their own speech. None. And that includes people blocking them. They want to have their say, no matter how noxious or unpleasant anyone else finds it. They want to "just ask questions" and "just have an open conversation."
But they absolutely want other people to be moderated. There's no amount of moderation of their that's small enough.
Basically, they're (internet speech) libertarians; they want to reap the benefits of a structured, cultivated society and community without be subject to any of the constraints that make it actually work.
The problem with many of these social media platforms (X, Threads, Insta, etc) is that they are actively disincentivized to enforce the first requirement of those two. Strict KYC is (rightly) viewed by many as overtly invasive of privacy, it would hurt the volume of content on these platforms (because 90% of it is AI generated engagement bait), and it would hurt their user signup and retention metrics (being ignorant of how many bots and alts are on your platform is vastly better than knowing and doing something about it).
This would actively lower the freedom of expression. There are opinions people hold that they're only comfortable sharing when they know it won't come back to them in real life, as holding such opinions can be dangerous. These people should still be allowed to share their views and discuss with others that disagree with them/they disagree with. This could be one of the only ways for them to even challenge their own world view.
If you're optimizing freedom of expression, then anonymity is a requirement.
I understand the anonymity argument, even from the perspective that sometimes you don't want even the service operator to know who you are. I'll be blunt on my take on this: I think this is millenial/genx idealism, tossing coins into a wishing well for an internet that does not exist anymore and will never again. Generative AI is out of the bag, and social network service providers can either grow up and recognize that their Paramount Number 1 service they can offer is some reasonable guarantee of protection against generative AI, or they can stay addicted to their juiced engagement numbers and play dumb when it comes out that 99% of tweets are from bots trying to build rapport, sell products, and influence elections (bye bye advertisers!). Its their call. Twitter is making the wrong one. Meta is up next. Traditionally they've all done bot detection; this doesn't work. Bots are indistinguishable from humans now. You need human-detection; KYC, meat-space verification. If you don't like the anti-privacy angle, then don't participate; no one is forcing you, and you're welcome to start your own mastodon server out in the generative AI wildlands (and, by the way, you should always have that right; i'm not prescribing how the world should be ran, just how e.g. Threads should be).