Re: Banning stuff
I largely agree. The recent US trials have nailed at least one thing - the problem isn't the content. The problem is the method. Toxic content is merely the unavoidable consequence of how social media is currently designed.
Toxic content has always existed and always will exist. It's never been a big problem, because there never was very much money in promoting it, which meant it didn't get promoted, which meant not very many people got exposed to it. And because there wasn't much money in it, not very many people produced it.
Today's social media platform have come up with an unchecked feedback loop. Machine learning will figure out what triggers you, and then push that to you. It doesn't really matter what the content is - and, indeed, the algorithms don't really even know or care. That methodology is what we need to ban.
So: don't ban poorly-defined "toxic content", and don't ban based on user age.
Instead, ban pushing non-subscribed content to users, and/or ban user profiling. Either will work, and neither requires censorship or moderation (beyond the bare minimum which nobody but the most rabid free-speecher would object to). Yes, you might miss out on some nice content as a result of these changes - but guess what, believing that you need to see all of the nice content is also a distortion. Weaponized FOMO.
I frankly do not believe that there is a way to salvage Facebook, Instagram, YouTube, TikTok and the like. Their entire business model is broken beyond fixing. But there absolutely are ways to build content services that don't work like that. They would never be as profitable, of course, but they could be viable if they didn't have to compete against literal drug dealers.
Also, newsflash: kids are not a different species. Yes, marketing tobacco to kids is especially effective and worthy of an explicit ban - but tobacco is very much not only harmful to kids. Same thing for self-reinforcing content feeds.