Recent searches
Search options
When Elon was losing class action lawsuits for his blatant, egregious, and widespread racism against Black people, and Black folk were decrying his Alt-right antics and the harm that they cause, everyone was fine. Tesla drivers weren't all the devil. People still cheered SpaceX launches.
Now Elon's Alt-Right antics have come for a much wider circle. Now everyone that stays on Twitter is the devil, and an irredeemably bad person? 
This one-way solidarity ain't it.
Telling people to leave Twitter is silly anyway. The math ain't mathing for Elon. He may be losing hundreds of millions a month. Advertisers ain't coming back. The Eng and policy teams are at unsustainably low levels. This doesn't end well.
But I'll repeat what I said in my 1st few months on Mastodon: if you want communities to move over to Mastodon, make Mastodon better for those communities. Don't just yell "Elon is bad!" Many of the people you're telling, knew he was bad way before you did.
If a person says they want to see Mastodon scale to support 100 million users, a lot of Mastodon users will jump on that person's neck. 
They'll be accused of having "VC brain", or being obsessed with hyper growth.
It's not about hyper growth for growth's sake. If you truly believe that Twitter is bad(tm) and Mastodon is good(tm), and that people should leave Twitter, then where exactly do you think those people should go? There's hundreds of millions of them.
@mekkaokereke How about “I believe that modern VC-driven business models are irredeemably flawed, to the extent that that lifeboat comes with an unpluggable leak”.
Personally working on trying to prove that an instance based on co-op principles can be sustainable.
@timbray Same!
But the National Science Foundation is a kind of VC firm LP'd and GP'd by science loving US government employees.
And a crypto DAO is a kind of co-op run by crypto bros. 
A lot of VC assets are controlled by alt-right adjacent dudes. If 95% of all co-ops were DAOs, I think many folk would similarly assume that most co-ops don't do any good in the world, are usually harmful, and are often run by sociopathic alt-right degenerates, intent on destroying the world for profit.
@mekkaokereke Can't disagree with any of that. I have this optimistic vision of a future where the Fediverse is a large collection of small lifestyle businesses and member-owned co-operatives.
I think I'm less worried about tech & scaling than you are. There'll be painful spots but there are enough people around with the development and ops mojo to get us through.
I expect GCP & AWS to offer instance-as-a-service and there's nothing wrong with that.
@timbray @mekkaokereke the tech and hosting parts will likely be solved, there are already some managed offerings.
The hard part about running an instance is moderation and I don't see that changing
@Paxxi @timbray Yeah, but I've become radicalized with the idea that good moderation of smaller communities is *relatively* easy, and of larger communities approaches impossible. As in, moderating for 10 friends is relatively easy, 100 is hard, 1000 is about the limit of an untrained pro, 10K requires a group of trained pros, and I don't know (m)any well moderated 100K+ communities.
Not sure if this is more because it forces a high mod:user ratio, or because mods can understand context easier.
@mekkaokereke @timbray I'm thinking the smaller the community the more likely that you're intimately familiar with the culture, in-jokes and so on making moderation easier.
I'm guessing that a lot of potential moderators have no clue about "it's okay to be white" and for every minority community there are likely to be similar "stealth" hate content that's intended to fool moderators.
@Paxxi @mekkaokereke @timbray I help run/mod a small forum and can hold enough culture and history of it in my head to make that relatively easy. And the handful of mods agree on enough thus more so.
The "C-word" is one of the most offensive words in the US. It's often used in the vilest, most misogynistic contexts. But it's less offensive in Australia? And when combined with other words and contexts, the meaning changes a lot.
When an AUS friend told someone "Mekka's a hard c-word!" that was in reference to me being our rugby team's enforcer. When they said "Oh you're a sick c-word now!" that means I lost weight but stayed muscular. Both intended as compliments.
Imagine trying to moderate abusive language for a Mastodon community server.
"Australian semi-pro rugby players talking about their Nigerian rugby teammate" is just one micro-context. It's not just "anything goes in the in-group!", because there are definitely things my AUS teammates could say that would violate. Either misogyny, racism, transphobia, etc. It's complicated.
How many of these microcontexts can exist in a community of 10 people? 100 people? 50K? 1MM?
Now imagine you think you've gotten it right and *not* moderated the use of C-word between rugby playing Australian friends.
Then a woman replies: "Hi. I'm Australian. I play rugby. I play on the Australian women's national team. That word is *not* less offensive in Australia, and I don't want to see you use it again. Grow up."
Did the mods make a mistake? Or did community standards evolve?
It's complicated. But easier to adjust on the fly with smaller communities.
@mekkaokereke @DamonHD @Paxxi @timbray it does sometimes feel like the men (and diaspora) of Australia held a referendum on that word but forgot to invite the women
@mia @mekkaokereke @DamonHD @Paxxi @timbray A big part of this is because context creates meaning, this is how reclaiming words like "queer" works. It's why some words are okay to use with your friends and in-group but are offensive when someone in an out group uses them. This is also why you can't use machines to moderate, they don't understand context and meaning (this is also where machine translation fails and why literary translators will continue to be necessary...and why literary translation can be very intense and hard work in the meaning mines).
*also, part of moderation work ideally would be helping people understand each other and/or understand why something they said was problematic (if it's not intentional, of course, but that stuff tends to be obvious).
@mekkaokereke @Paxxi @timbray definitely not always easy, though our problems have been more religion, guns, and science-does-not-apply-to-me stuff. Being sweary would be straight out
@mekkaokereke @Paxxi @timbray I kinda wonder if the solution there isn't so much to lift up the fediverse as it is to devolve the internet back towards multiple communities (with the Fediverse/ActivityPub serving as a kind of connective tissue/discovery mechanism to help people find the right communities for them).
Bring back the indieweb, with its webrings, forums and blogs. The primary purpose of cramming everyone into a single site was to resell their attention to advertisers and we have no need to care about that.
@mekkaokereke @Paxxi @timbray in addition to the great conversation below, I wonder if part of the scaling issue is the same one we see in other organizations -- communication limits end up requiring hierarchies and summarization in order to coordinate, and each layer of that is less efficient.
E.g. an instance of 100 maybe needs 4-8 mods (1 per 20), who all know each other and can communicate directly. If 10k users need 500 mods, then you also need at least 25 mod-coordinators. 1/2
@mekkaokereke @Paxxi @timbray in addition, the work ratio for mods changes if you have to coordinate with a mod-of-mods as well as peers, so maybe you can only support 1:15 users:mods, which means more like 800 mods for 10k users. A 1:20 ratio probably also isn't going to work for mod:mod-of-mods, probably 1:10 or so for middle layers (or less!). That means 80+ first-layer coordinators, and a 6-8 person second-level coordination team.
That makes moderation 2x as expensive at 100x scale...