I look forward to the day that my instance (snouts.online) ends up muting (or just outright blocking) activitypub.twitter.com. (They already mute mastodon.social, which means that only people who you actively follow on that instance have their toots pulled over; Reply Guys are just shouting into the void.)
The long and the short of it is that all the Nazis will either move to gab.ai where they can scream into the void at each other while the rest of the federated world blocks their traffic, or they’ll stay on Twitter, which will result in most of the fediverse either muting or outright blocking Twitter. Neither of which would be terribly good for The Engagement, as Twitter currently defines it.
This seems like such a bizarre move for Twitter. The biggest reason they siloed themselves and took away access to features through their API is because it’s harder to monetize your platform if other tools can come along and just skip over the ads while providing better content-complete experiences. Making Twitter interoperable with other platforms removes all of its inherent value for advertisers, especially since it also eliminates the friction of choosing to seek out a less Nazi-laden social media service. If I can leave Twitter('s instance) for the Mastodon instance of my choice without losing all of the incoming material from people I care about, and I can even continue to interact with those people who choose to stay on Twitter, why wouldn’t I leave? I just don’t see how Twitter survives as an ad-driven platform when other ad-free and also free-to-use/Patreon-supported Twitter-interoperable instances arise.
This is absolutely about content moderation; Jack even says so in his tweetstorm:
centralized enforcement of global policy to address abuse and misleading information is unlikely to scale over the long-term without placing far too much burden on people.
Of course, “unlikely to scale over the long-term” is code for “we’ve completely failed to do anything even remotely competent in this area”. I just don’t see how Twitter fixes its moderation policies without also collapsing the house of cards that is its ad-revenue service. They could build a protocol that requires participants to display all federated content to the recipient in order to continue pushing ads from their centralized ad-serving house, but that would just mean that abuse and misinformation would continue to flow out from Twitter into the broader ecosystem unabated. Or they could use an existing protocol like ActivityPub that would empower other instances to take their own administrative action against Twitter’s bad actors, which would also inevitably mean that ads would get caught up in the mix of content being blocked.
Given the two, I can tell you which option Twitter would prefer, but it won’t do anything to address the healthiness of social media as a service.
If they can be blocked or de-federated, then they can have their own little universe to play in. (And infiltration by the feds or groups like Unicorn Riot always works.)
I guess, if you take for granted that Twitter (a) matters and (b) is a unique technological achievement on par with the jet engine or the hydrogen bomb. Which Jack Dorsey would certainly very much like you to.
It tends to bounce off people when I say this, but: reifying Twitter and Facebook as if they’re the Catholic Church or the Imperial Family of Japan is the diametric opposite of pushing back on them. Of course they want their trivial chat service to be the subject of a grandiose UN standards process. Preferably before people start to notice that the whole time-wasting bile tornado is no more permanent or substantial than Nickelback fandom or saying “whassuuuup?”
Twitter’s essence is not the breakthrough invention of mass text messages. It’s the idea of reading large numbers of strangers’ vapid blurts, in exchange for them reading whatever pops into your head. To the extent that anyone really needs that, it will never be technically difficult to arrange.
If it’s anything like what I’ve seen on the ActivityPub Fediverse that powers Mastodon, it may lead to slightly more, but more easily ignorable, nazis. Any good small independently-run server will kick nazis off on sight, while the “free speech above all else” servers will become a haven for abusive bigots and all the other servers will block them, leaving them to just chat with each other.
You get my “Curmudgeon of the Day” award, and I also agree.
When Trump isn’t leader of the ‘free’ world anymore and the next president simply issues press releases, what happens to all of that swirling Twitter traffic?
Mastodon is a toy compared to the amount of traffic, moderation, and overall scale Twitter is talking about here.
I doubt any Mastodon instance in existence could handle the Twitter garden hose (1% sampling of all twitter traffic) let alone the full firehose, in large part because the protocols were designed to be easy to code to, not focused on stupid levels of messaging (which is the case with many api’s, frameworks and the like).
It’s sort of the evolution of HTTP vs SPDY (now HTTP/2) and soon to be QUIC (HTTP/3). Each one became progressively more difficult to implement, but each has huuuuuge gains for moving large amounts of traffic quickly (and in the case of QUIC, for true anonymous, encrypted transport of data of all types) - and each of these standards started as in-house projects from the big providers who were tacking problems no one else had the visibility or experience to handle due to sheer scale.
It reminds me of one of my visits to the USENIX LISA (Large Installation Systems Administration) conference. I was in a storage group roundtable talking about handling petabytes of data and storage management, as were many others. Then we got to the guy in the youtube sweater who was talking about how they have to deal with that much data… per hour. The scale is almost unimaginable and the tools required vastly different.
If we want decentralized to “go big” and get popular, then we need robust protocols and designs that are prepared to handle the worlds social messaging volume. This is a great first step to getting serious about this stuff.
This seems to be all about content moderation. I always get the feeling that Dorsey is stuck between wanting the old libertarian, free speech view point of view of the web and the fact that he owns a company that has to own up to hosting a lot of repellent speech that advertisers hate. If he can make a Tweet protocol, making a Tweet an individual entity like an email that anyone can grab or publish, then he can wash his hands of any offensive Tweets. If anyone can use anything to publish a Tweet, then Twitter(company) doesn’t have to be responsible for it anymore. Then Twitter(company)/Jack Dorsey become responsible for creating both an open commons and a walled garden. Tweets are free flowing, anyone can grab them out of cyberspace if they want to and the protocol allows anyone to create them. Twitter(company) is a walled garden that curates those Tweets in a completely Disney-fied way to make it safe for consumers and more importantly advertisers. Twitter(company) now doesn’t have to answer to critics that say Twitter(company) is limiting their speech, because of course they are, they have ads to sell. Nazi McFuckface can Tweet all he wants, anyone can use any tool to post a Tweet, but now Twitter(company) doesn’t have to publish it.
If you want content by Nazi McFuckface, it’s out there. It’s now an open protocol, Nazi McFuckface can Tweet into the open commons as much as they want. If you want to read that sort of content you will have to get something besides Twitter(company) to get it. Tweet clients that allow free-for-all content or Tweet clients that offer specialized content (Nazi, porn, sex work, furry, anything that upsets advertisers.)
Twitter(company) of course has a metric shit-ton of experience with this. If you want to make a Stormfront Tweet client, Twitter(company) has tools to help you with the content curation that it can sell you for a price. Twitter(company) sells those tools to anyone, they are not responsible for how people use them.
I suppose that this is all assuming competence on the part of Twitter(company) and Dorsey. But, wasn’t there something about how they had tools to block hate speech, but using them also blocked Republicans? This strikes me as a way to profit off of those tools.
Yeah, but that wasn’t really a bug. If you made an impossible tool that perfectly separated hate speech from non hate speech it would block many tweets by Republican lawmakers. But it is not at all in Twitter’s interest to be the one to tell that to Republican lawmakers. Though I don’t have any sympathy for Twitter’s plight, I do see how existing as a speech platform in a society that is in a cold civil war between racist fascists and people who find racism intolerable is a predicament.
Thankfully we already know the solution to this: all of the existing instances will rightly block twitter.com because people don’t want that garbage on the fediverse
If you made an impossible tool that perfectly separated hate speech from non hate speech it would block many tweets by Republican lawmakers.
My argument is that a perfect tool WOULD separate out hate speech from Republican tweets. The tool Twitter has now is imperfect, IMO because Republican Tweets are impossible to separate out from hate speech. Making Tweets independent of Twitter(company) allows them to handle moderation however they want. If Tweets are independent from Twitter, publishing Tweets from the POTUS (any POTUS) is different from publishing Tweets from Milo and Glenn Beck. If you want Tweets from those two, get a different Tweet client, Twitter(company) doesn’t want to publish that.
Technically? No. In every other way, that is, user base, user activity, revenue, etc.? Well - good luck. The fact that that none of the Twitter (and Facebook) inspired attempts like Mastodon, Diaspora, etc. ever took off should be a warning sign.
You can be dismissive of Twitter all you want (I’m certainly no fan of it myself and only use it to follow some artists who don’t post elsewhere) the fact is that it would appear people do want to read large numbers of strangers’ and parasocial “friends” vapid blurts, and more importantly, want to add their own vapid blurts to the choir. It’s about interaction or at least the illusion of it, the illusion that your opinion counts, the idea that by being one voice in a crowd you can influence things. And to a certain extent and in certain ways - it works.
Sure, I assume some people use Twitter to communicate with friends and family. For those people Mastodon and the like are a good alternative. But those who use it to because this way they can have the largest possible reach for their ideas, not so much.
I think what Twitter actually wants here is something where users can continue doing what they do now, except Twitter can go “hey, don’t look at me! It’s not my responsibility!”
If every single node in a distributed social medium has to carry all the traffic, then it’d be doomed from the start. Only node-sites with Twitter-equivalent tech could support it.
Starting from the bottom end, my own personal node only has to carry the traffic that I’m interested in. Back in the day, a full Usenet feed (far too much for one person to read) was ~50M/day, excluding binaries, a much less scary number.
The problem is to be able to access the entire message space, but only pull down (and relay) the parts that I’m interested in. A distributed hash table should be able to handle indexing the message space, i.e. which nodes carry which message areas/newsgroups, and then use an NNTP-like protocol to sync the actual messages with them.
Could it handle the traffic? Eh, good question. A DHT can handle a large volume of data, but if there are too many localized hotspots in the DHT that concentrate the traffic on a small number of nodes, it could rapidly Slashdot itself out of existence. (As each node failed, the DHT would re-align on the next closest node. Bang!)
I mention Usenet, because I think trying to replace Twitter with instantaneous posting across a distributed net isn’t realistic. (Besides, I think that type of medium is bad for people.)
Wouldn’t it make it worse for the company?
Now they at least have the free-speech and the problem of doing moderation at scale as excuses to justify the bad actors in the platform.
If they make a disney-fied version of themselves, any violation becomes much more serious as they are supposed to curate it.
I also cannot see how it wouldn’t make them look more like a publisher, and therefore more responsible for what they host.
Youtube kids is a good example of this, there is no free-speech and a big restriction of what can be aired.
But, every video that slip through the filters gets much more attention and outrage than the average offensive youtube video even if the former are not as horrible than the latter.
That mostly sounds like a good thing to me. Nobody gets censored, but everyone has tools to control their own online experience.
Nazi trolls can be pushed into ghettos where they can scream at each other all day, and, when the Republican politicians who love them star looking for who to complain to about it, they find that it’s distributed, so they can’t just go and threaten one company or grill its CEO in congress.
I’m kind of with @RickMycroft, no Mastodon server would likely ever need to handle the Twitter garden hose, much less the firehose. Even if the landscape remains fixed and absolutely zero additional people leave Twitter after a hypothetical occasion where they implement ActivityPub, the average number of Twitter users any given person follows, times the number of people on a given Mastodon instance who would import their Twitter follows, would probably not create that much higher of a traffic load. If Twitter were to try and somehow create distributed aggregate data on things like hashtag trends, that would probably add some degree of overhead, but still nothing approaching Twitter’s raw throughput, I would think. Down the line, it’s possible that the number of instances proliferates to such a degree that simply aggregating data from that many sources would itself become problematic, but I don’t know precisely how likely that is either. If folks start self-sorting, it seems like for the most part you’d have pockets of high inter-connection between handfuls of servers, with sporadic links to the broader network-at-large. Mastodon’s own software might have to stop being a Rails app and really lean into being a hardcore service application built for high throughput, but the ActivityPub protocol itself seems pretty well-designed from what I’ve read of it.
Of course, if the goal is to perfectly replicate Twitter’s current UX and feature set when it comes to things like full-text search, indexing, and algorithmic “who to follow” recommendations, then yeah, every instance wold need a copy of every post (or at least a way to crawl for them) but I think that’s a scalability concern that exists separate and apart from ActivityPub itself. A sane implementation of ActivityPub on Twitter would allow other instances running on Mastodon, Plemora, etc. to take only what they need, and give back only what they want. Communities (and even individual users) would be just as “discoverable” as they choose to be.
I’d be more interested to see what concessions/improvements Twitter is willing to make to their own service if they decide to interoperate with the likes of Mastodon. Would they adopt content warnings? (Please, Twitter, adopt content warnings. I’m begging you. It’s so nice not having to abandon social media after big movies come out because I don’t want to read or post spoilers, depending on when I see it.) Would they eliminate the quote-tweet? (Technically speaking, you can QT people on Mastodon because there’s nothing to actually stop you from manually copying a link to a toot into your own post, it just doesn’t embed the tweet you’re linking to the way Twitter does, and it counts against your character limit, so it’s harder to use for quick public dunkings.)
That, again, is a framing that Twitter would like to encourage, because it positions Twitter as the only serious player. But it’s no more true for would-be Twitter competitors today than it was true for Twitter when they started.
If you create a new microblogging app using Node on a single $5/month instance, no, of course it won’t handle Twitter’s volume. But it can probably handle the couple of hundred users you might actually have to begin with. And if you can get revenue from them, you can turn it into something that will handle thousands of users, and so ad infinitum. That’s how Twitter started, and it’s how the thing that makes Twitter obsolete will start, which is why they want to set the rules. It’s like T-Rexes ”working with” mammals to standardise a protocol for surviving a nuclear winter while being a giant cold-blooded predator.
