[eropple]

Moderation on the client side doesn't work and never has. The Usenet killfile model was awful and was a big part of why huge swaths of it were an unnavigable mess for most users--because unless you wanted to expend the effort to build one you didn't have one and got the whole sewer pipe spraying straight into your face. And, oh boy, client-side moderation happens on Twitter, it works so well? Then why do female friends whose posts get some attention have a bunch of dudes in their mentions saying how they want to have sex with (consensually or not) and/or kill them no matter how many they block? Could it be because it's an abject failure of a model due to the low stakes of spinning up a new account--so low that Twitter stopped using egg avatars because they destroyed interaction rates with new users because most of them were troglodytes?

I do wonder.

I'm trying to assume good faith in your reply so let me point something out to you: the people who are wringing their hands about this used to bleat about shared block lists being used to get some semblance of value out of "client-side moderation" (which is not to say that it is successful at dealing with shitty actors, just that it's less unsuccessful, when you use them), just as they're bleating about this. Because the real beef here is not with Twitter Deciding, it's with Anyone Deciding that they are not entitled to the attention of anyone they demand it from.

But they lost re: shared block lists, and this is a retrenchment to attempt to keep relevance when they can, and should, lose this one too and be shown the door. If they want "unmoderated", there are platforms like Parler and Gab where they can have that. Relatedly, those platforms are full of literal-not-figurative fascists. I wonder why.

[stale2002]

> Moderation on the client side doesn't work and never has.

Are you claiming that the phone system does not work? Or that ISPs do not work?

I think that they work pretty well, even though there really is not much top down moderation.

> If they want "unmoderated"

Why can't we consider the phone network, or ISPs, which are successful examples of the kind of unmoderated approach, that people are asking for?

[jakelazaroff]

Because phone companies and ISPs only facilitate private peer–to–peer connections between a small number of parties, whereas the platforms we’re discussing host often-public interactions involving millions of people.

It’s a fundamentally different problem. If people use my phone company to have racist companies, I don’t have to know or care. Whereas if I had to deal with people saying racist shit every time I went on Twitter, I’d stop using it.

Consider also that most people’s biggest complaint with the phone system is a consequence of them not moderating, i.e. spam calls.

[triceratops]

Who said ISPs and the phones are "unmoderated"? Try running a server with significant traffic from your home. Or using blue boxes (? I don't know, whatever Woz used) to get free long-distance calls. They absolutely perform moderation, only at the network layer, so most people never see it.

[stale2002]

> Who said ISPs and the phones are "unmoderated"?

ISPs have to follow common carrier laws.

If social media platforms were subject to these same common carrier laws, then that is good enough for me.

That is what I mean by unmoderated. I mean, that they are following things such as common carrier laws, which have very significant requirements.

[triceratops]

> ISPs have to follow common carrier laws.

Firstly, you're wrong about that. This same FCC voted to remove ISPs from common carrier classification.

Also, what do you think common carrier means? It doesn't mean "completely unmoderated" as you appear to believe. As I already demonstrated, even common carriers perform network-layer moderation, since they provide the network layer. Social media provides the application layer, so why shouldn't they be allowed to perform moderation at the application layer?

[stale2002]

> Also, what do you think common carrier means? It doesn't mean "completely unmoderated"

Systems such as the phone network are pretty darn unmoderated. It is mostly unmoderated, even if common carriers are allowed to do a small amount, of highly restricted moderation actions.

> as you appear to believe

I have just clarified. It means that there are very strong regulations, that ban many forms of moderation.

> so why shouldn't they be allowed to perform moderation at the application layer?

I am saying that I would be fine if they were banned from doing most moderation actions, as is the case for common carriers.

And I am saying that I would be OK with them engaging in the very small amount of highly regulated moderation, that common carriers are allowed to do.

Common carriers have large restrictions on the amount of moderation that they are allowed to do. I want those same restrictions to apply to other media platforms.

[triceratops]

> Common carriers have large restrictions on the amount of moderation that they are allowed to do

Yeah, I don't think you know what common carrier actually means.[1]

The moderation that telecoms and ISPs perform relates to traffic that can negatively affect other users of those systems. Social media platforms can easily argue that their moderation already follows similar principles. It just happens that they deal with content, not raw bits, so they have to moderate content.

You also didn't address my other point, which is that ISPs aren't common carriers at all and this same FCC voted to make it so. So this seems pretty hypocritical behavior on their part.

1. https://www.law.cornell.edu/uscode/text/47/153

[eropple]

ISPs aren't a remotely relevant comparison. Neither are point-to-point phone calls, though as an aside the phone system certainly doesn't work. I get elderly-preying spam calls at a rate of eight to twelve a day and the Caller ID gets spoofed so numbers I block are useless. And because I use the phone for my job, I can't go contacts-only.

[eternalban]

"Moderation on the client side doesn't work and never has. The Usenet killfile model was awful and was a big part of why huge swaths of it were an unnavigable mess for most users--because unless you wanted to expend the effort to build one you didn't have one and got the whole sewer pipe spraying straight into your face. And, oh boy, client-side moderation happens on Twitter, it works so well? Then why do female friends whose posts get some attention have a bunch of dudes in their mentions saying how they want to have sex with (consensually or not) and/or kill them no matter how many they block? Could it be because it's an abject failure of a model due to the low stakes of spinning up a new account--so low that Twitter stopped using egg avatars because they destroyed interaction rates with new users because most of them were troglodytes?"

Re client side moderation, it's possible that users can be educated about effective approaches. Just like they started warming the non-geeks about email with hollywood films (You've Got Mail, 1998), and I'm sure there must have been similar sort of concerted media efforts for older technology. Automobiles, for example. I -bet- we can dig up content (old films) that basically were teaching the American public how to use cars using movies.

Given the inherently democratic aspect of client-side filterings, and serious inherent political concerns with centralized moderation, it seems wrong to give up on the idea simply because the avant garde generation on internet (using mostly antique kludgy UI/UX, btw) had problems with it.

We can fix client side moderation with education. It is a transitory issue, only.

Centralized moderation is inherently problematic for public platforms. (I have no issues with private, members only, platforms opting for centralized moderation.)

[eropple]

"Just educate users" didn't work when there were a few hundred to a few thousand users. It's one of the reasons, though granted not the only one, why Usenet was a hole and people migrated to saner pastures when they were available. What makes you think it's any more feasible when it's "now you just need to go administer your ration of racism and transphobia and personal threats" on a millions-to-billions scale?

If you're being legit with your post, you're falling for a tremendous okeydoke here on the part of people who wish to do civil society harm. Literally all this approach accomplishes, literally and exclusively and without exception all it does, is cede the public square to people who want to export misery or deal with the exporters of misery--a vanishingly small fraction of the decent people out there. We can and must be better and that means that no, we are under no obligation to waste time or resources on people who were never socially potty-trained.

[eternalban]

Re. the scale of the effort required on the client side by the user, I think you missed the "kludgy UI/UX" bit. Obviously, we need to address the kludgy kill-file approach.

> you're falling for ..

You seem to think that you know a lot about me. Let me gently suggest that possibly your technical imagination is failing to show you alternatives that solves all these issues.

So, actually, it is not a case of falling for anything, rather waiting since 1996 for someone else to start discussing having social models for networked virtual spaces.

> If you're being legit with your post

And why would I be otherwise? I found that quite offensive, and entirely un-necessary.

[eropple]

Again, the bad actors in question will scream about tools made to help the general public deal with them, too, because it's not about some ethical thing, it is that they demand to have their sewage spewed at all comers and you are bad if you don't let them. There is a level of malice here that I feel you are definitely dismissing while charging end users with a level of responsibility that it is impractical to handle regardless of what UI/UX affordances you want to put on it.

It's not a UI/UX problem. It's a "you are offloading moderation to the end user" problem--it is a fundamental failure of the approach. If you are going to put all of the work of moderation on the end user, regardless of whether you dress it up as "shared moderation" or whatever, the end user is going to leave and cede the floor to those who want to dominate the public discourse with sewage and crud and threats because the end user has a life to live and things to do with their day. There is a fundamental human cost to being exposed to the kind of garbage that Twitter et al filter out on the regular. It takes a toll on the psyche. It will drive people away. And that's not a "but muh engagement" for whoever's metrics and KPIs are involved--it is a realization that this is now the public square and ceding it to actively anti-civil movements is bad for liberalized society in general.

If there's a failing of technical imagination on my part, fine. There strongly seems to be a failing of appreciation of how modern supremacist movements operate on yours. This is not a technical problem, this is a human problem. It's a question of whether the psychic damage inflicted upon bystanders by constant and unending exposure (because that is, genuinely, what the end state here is) is worth whatever hypothetical benefit is realized by giving these parties a place from which to abuse others until they definitely, yessir, are blocked by all of them and never, ever seek to get around it to continue their abuse.

> And why would I be otherwise?

Speaking frankly? Because this discussion is dominated by bad-faith types whose axe to grind is really that there's a movement against their favorite flavors of supremacism and they generally use arguments like the one you're putting forth as retrenchment now that the tide has turned against them. If that isn't your bag, I appreciate that; I think you're incorrect, but that's separate from disingenuous.

[eternalban]

The lack of imagination on your part was an allusion to the inability to see a structural solution to the conflated issues of identity; access; location; and social affinity. Believe it or not, it is possible to have a social network where it is impossible to spam.

> There strongly seems to be a failing of appreciation of how modern supremacist movements operate on yours.

Well, the internet gifted to us, with its flat namespace and the inevitable conflation of user identity with device identity, [empowers] bad actors. I alluded to this.

[eropple]

OK. So where's your white paper? I genuinely want to read it.

[dexen]

>people who wish to do civil society harm >a vanishingly small fraction >never socially potty-trained

A facade of politeness & shaming, a reality of censorious rules. The encroach was slow, almost imperceptible, but here we are at last. The current state of the Twitter rules is impacting investigative journalism. Internationally recognized journalist Glenn Greenwald of The Intercept explains:

Look carefully at what Twitter is saying to justify censoring the Biden story. If applied consistently, it’d mean that some of history’s most consequential journalism — the Pentagon Papers, WikiLeaks’ war logs, Snowden docs, Panama Papers, our Brazil Archive — would be banned. [1]

Ceding the full power to the platforms has failed us.

--

[1] https://twitter.com/ggreenwald/status/1316720230450724864