Recent searches
Search options
Administered by:
Hey remember how for years anyone interested in or working on decentralization was told "yeah but it's not moderatable, you need centralization for that"?
Isn't it funny how the two big centralized social media websites are now both "neo-nazis-are-cool-here" websites
Isn't it just a whole barrel of laughs
*Narrator voice:* But it was not, in fact, a barrel of laughs.
@cwebber one rotten laugh spoils the barrel
@randomgeek @cwebber This feels related to the parable of the Nazi bar: If you let one stay, it's a Nazi bar.
@cwebber well at least we've had to put up with 30 years of anonymous harassment and defamation in order to give those companies a "safe harbor" for content moderation, that was certainly a Section 230 well spent
@cwebber it is hilarious. the only way to moderate content is in a decentralized model where the communities in the ecosystem can regulate themselves. for facebook to even think that they can fact check all political comments and figure out who's good and who is bad ... all with a team of employees at facebook. If the team were 10,000 people, maybe it would work, but I bet it was less than 10,000 people.
@PersistentDreamer yeah, i heard, that's what i'm talking about. it was stupid for them to even think they could do this.
@arrrg oh gotcha. Personally, I think at least attempting fact checking is a good thing, but you have to trust the fact-checker to begin with and I have had zero trust if Facebook for some time. But with them blatantly tearing off the mask, well it's just going to make things veer even more right wing.
@PersistentDreamer
Fact check:
FB has not stopped using external independent fact checkers in all geographical areas yet.
Wow. In a way it goes to show how spineless and amoral zuck is. His political beliefs are: whatever you want them to be so long as you're using his service and making him $. What a fucking coward.
@cwebber Thinking about that Youtube algorithmic moderation apologism talk Tom Scott made. 
@swift @cwebber I might have to rewatch it, the takeaway I remember is that algorithms are a necessary part of moderation at the scale of Youtube, with a passing mention of alternatives like Mastodon at the end, and not much consideration given to whether something of Youtube's scale should even exist. He didn't really talk about how alternatives could work or what the incentives are.
@csepp @cwebber ah, yeah. I should've said the conclusion is that you can't moderate with *just* algorithms. But then, in the fuzzier between space, I think I'm in favour of algorithmic support for human moderation: I think algorithms going "hey, this might be something maybe take a look" is fine, I think algorithms going "we're automatically shutting this down unless a human goes out of their way to intervene" is less so.
And yeah, the question of what should exist is a fair one, but one I'm unsurprised maybe didn't occur to him to question too much 
I agree with the spirit of this, but I think some nuance is in order. An #algorithm is just a sequential, unambiguous series of steps. If I write instructions down & have someone else follow those step-by-step instructions, they are using an algorithm. A human moderating with a set of guidelines uses an algorithm.
The thing that humans bring to the table is their experiences. Moderators of marginalized communities bring their unique experiences.
If you want to delve more into the phenomenology, humans have a sense of judgment that emerges from tacit aspects of their experience.
Tacit aspects are things that are hard to encode into language, so there are certain aspects artificial intelligence cannot capture because it relies on language in ways that human judgment does not.
Humans are still carrying out algorithmic tasks. One that we all do here is curate our timeline. That is an algorithm.
@Cerulean_Code_Wyrm it isn't clear to me that pushing to the formal definition of an algorithm is useful when it's apparent from context that we're using the colloquial definition to refer to the distinction between automated vs human processes. Changing definition is going off topic.
@Cerulean_Code_Wyrm I concur on the value of human insight in this process, and that that's the heart of the issue. Feel free to replace "algorithm" with "automation" in the prior discussion :)
@swift You've gone down a path of a red herring. The crux of my argument is not whether we use an algorithm but rather who is using the algorithm & what experiences or lack of experience they have.
The issue is not the definition; the issue is who is implementing the algorithm. Using the term "automation" doesn't change anything.
You acknowledge that in your first sentence, & then you diverge from that point with a red herring. I'm going to say you are not presenting a good faith argument.
@Cerulean_Code_Wyrm I'm saying that I don't think it's constructive to jump into an existing discussion to mandate particular definitions or terms in use when the existing participants mutually understood the terms as they were using them. If we had been talking past each other due to misaligned definitions that would be a different matter, but I don't believe we were.
@swift I'm not reading this. I told you to stop, you did not. You get blocked.
@Cerulean_Code_Wyrm I sent that before you replied 
@swift That's a Red Herring. From what I gather, we are all on the same academic level and you are using spcefic allusions to Computer Science, so the context is such we can use formal language. This is a Red Herring. The OP is literally the co-author of the ActivityPub protocol.
@Cerulean_Code_Wyrm sure, but she hasn't actually engaged in the thread of conversation that was discussing algorithmic (or automated) moderation, and context indicates that csepp and I weren't using formal definitions, regardless of our ability to do so.
Just to be clear, I agree with the conclusions and spirit, and I am absolutely disgusted by #X and #Meta. I am clarifying the terms and the problem.
#LLMs lack the tacit aspects that inform human intuition in ambiguous cases. Human judgment stems from tacit aspects of experience beyond language, like perceiving the "blueness" of blue. Perception of colors varies among people.
@cwebber it’s almost as if these massive social media sites only make money when they allow the trolls free reign.
When you don’t have a profit making imperative then miraculously the need to cater for trolls immediately disappears.
@cwebber Yeah, I think the real issue here is "too big to moderate" and they aren't saying that out loud so as not to become just another moderator of some small Podunk BBS; but really it's what we all need instead.
@cwebber I also heard a lot of "How could I trust a small instance with my data its just some guy it doesn't even have a GDPR policy"
Anyway Instagram is putting your face in ads with AI now and TikTok is tracking phones to catch whisteblowers
@cwebber the same is true for statism and anarchism. Saying how the state helps protecting the order and fighting criminals is cool only before it is led by nazis.
@ligne @cwebber
I had someone unironically ask me, without the state, how you would stop a bunch of racists with guns from doing whatever they want.
And I was just like... you're aware that the state is straight up giving racists guns and the license to do whatever they want, right? So even a system that just, you know, *doesn't* do that would already be a huge improvement.
yeah but it's not moderatableWe know the benefits of decentralization. Are there any disadvantages?
@cwebber
And they were absolutely right. Without centralization the big tech corps wouldn't have been able to moderate their networks into neonazi-amplifiers Imagine the fascists would have to face the consequences of defederation, that would be totally unfair to them /slightly sarcastic voice/
@cwebber We are in the midst of a giant User Experience Research project.
Terms: testing to destruction of user willingness to endure Nazi abuse
Cohort: entire planet Earth
Budget: total revenue streams for Twitter, Facebook, Instagram & Threads
@cwebber
Today I looked up what power distance is and coincidentally high power distance overlaps with everything I dislike of those platforms
@anarcho_capableism thanks for posting! it applies to the OP but what I’m taking from it is a useful framework for understanding communication differences at work
For meet this is dependent upon if the leader has a screw loose or not.
Interesting to see it's a categorical thing too.
@cwebber we need to discount most "wisdom" that has come out of the past decades of so called "big tech". They are the result of specific political conditions, in a country that is almost tearing itself apart in dysfunction.
This doesn't help with figuring out a path for a sane digital universe for the rest of us. It means most of the phase space is still unexplored, the real difficulties of herding humans being obscured by the dominance of business models that should never have existed.
The people making those claims don't know what Internet Relay Chat is, or an even better example E-mail. They absolutely have no idea what they're talking about.
Also Moderable. 
@cwebber bad actors love fewer points of failure, especially when it's paired with undeserved positive reputations.
@cwebber It doesn't even make sense. Moderation is always hard, as I know from my days moderating on forums. Even when it's a small community. You always end up upsetting somebody. But it gets way harder at the bigger scale. Leading to attempts to automate it, which don't work well. Or to needing a huge team of people, which will still not be big enough, and which will always play it safe in one way or another, whether being too heavy handed, or too lenient, depending on the prevailing attitude.
(not a vampire) @cwebber @beecycling kill all the instances with more than 100 monthly active users
@gabboman @beecycling @cwebber
Taken to it's logical conclusion, every app is an instance..
Instead of signing up for social.coop or Mastodon.social, one creates singleuser@username.Mastodon.social
Let at the instances moderate themselves!
@cwebber Something I highlight a lot on my newsletter is that the moderator-to-user ratio is orders of magnitude better on the fediverse.
There are not only significantly more eyeballs on accounts, but also those eyeballs belong to volunteers who are motivated by community building, not just for a paycheck.
There's also lot of nuance with good moderation, and when the person doing it is also member of said community they can be much more effective.
@JustinH It's the classic carrot and stick motivation that doesn't work scenario.
Trying to motivate people with a paycheck to care about communities that they're not part of is never going to work. And the scale of the effort to police millions and millions of messages with a team of people who work for a company is impossible. They would need to hire 1000s of people to do it effectively, and it still wouldn't work. The overhead of managing all of these people and trying to craft corporate policy around policing ideas would be overwhelming.
@cwebber I wasn’t active in that time frame but if I had been I would’ve pointed them at Usenet and then laughed at them.
@cwebber yet, they keep calling them "town squares." Yeah. Ok. So, go to the center of your town square, set yourself up in the middle with a nice mobile amp, and start spewing. You'll get yourself moderated alright!