Bluesky Begins To Make Its Decentralized Vision Real

Innovation

from the nothing-but-blue-skies dept

For semi-obvious reasons, I’ve been following developments at Bluesky closely, given that my Protocols, not Platforms paper was originally part of the reason Jack Dorsey decided to create Bluesky. I have no official association with the organization, though I did help Twitter review some of the early Bluesky proposals and spoke with a few of the candidates they looked at to lead the company (including Jay Graber, whom Jack eventually tabbed to run it).

While Dorsey has since soured on the approach that Bluesky is taking, preferring the nostr protocol’s approach (and deleting his Bluesky account entirely), I continue to believe that Bluesky is the most interesting and most promising of the various attempts at building a better social media system out there. I explained many of the reasons why a few weeks ago when Bluesky finally dropped its “private beta/invite-only” setup and opened to the public.

And yet, as many people pointed out to me, Bluesky still wasn’t really decentralized in any real way. It remained entirely centralized, as the company worked to build up both the new protocol for it, ATProtocol, and the Bluesky reference app on top of the protocol.

However, last week, Bluesky took that next step and opened up the ability to federate.

Today, we’re excited to announce that the Bluesky network is federating and opening up in a way that allows you to host your own data. What does this mean?

Your data, such as your posts, likes, and follows, needs to be stored somewhere. With traditional social media, your data is stored by the social media company whose services you’ve signed up for. If you ever want to stop using that company’s services, you can do that—but you would have to leave that social network and lose your existing connections.

It doesn’t have to be this way! An alternative model is how the internet itself works. Anyone can put up a website on the internet. You can choose from one of many companies to host your site (or even host it yourself), and you can always change your mind about this later. If you move to another hosting provider, your visitors won’t even notice. No matter where your site’s data is managed and stored, your visitors can find your site simply by typing the name of the website or by clicking a link.

We think social media should work the same way. When you register on Bluesky, by default we’ll suggest that Bluesky will store your data. But if you’d like to let another company store it, or even store it yourself, you can do that. You’ll also be able to change your mind at any point, moving your data to another provider without losing any of your existing posts, likes, or follows. From your followers’ perspective, your profile is always available at your handle—no matter where your information is actually stored, or how many times it has been moved.

It’s currently limited to smaller situations, of people who basically want to self-host their own Personal Data Servers. While things get settled, there are rate limits and guardrails for these PDS’s (so, things like only 10 user accounts for the time being). If you want to understand this even more (even if you’re not technical), Bluesky’s more “technical” explanation is still highly readable.

I know that some people hear “federation” and immediately think of Mastodon. However, Bluesky’s entire setup is very different and designed to be much more user friendly in multiple ways (once again, this is one of the reasons that Bluesky chose to create the ATProtocol, rather than going with ActivityPub).

ActivityPub federation has both pros and cons. When you sign up for an instance, you’re basically wholly reliant on whoever runs that instance. Rather than being part of a big centralized network, like Facebook, you’re part of a small centralized server that interconnects with lots of others. But whoever runs your server has pretty much ultimate control. That can work out great if they’re committed to it. But it can also unleash some problems.

Mastodon and related ActivityPub systems have put a lot of effort into minimizing some of the downsides of this. For example, threats of “defederation” are a fascinating incentivizing structure to keep ActivityPub instance admins from going totally rogue, while still allowing for there to be experimentation and differences among servers.

But in the end, you’ve still gone from a big centralized system to a little one, where someone else is in control.

With the Bluesky approach, there are many more layers involved, and federation is less about putting your entire social experience in the hands of one instance admin. Rather, it’s just about where your data/account information gets stored. As Bluesky explains:

A summary of some ways Bluesky differs from Mastodon:

• A focus on the global conversation: On Mastodon, your “instance”, or server, determines your community, so your experience depends on which server you join. An instance can send and receive posts from other instances, but it doesn’t try to offer a global view of the network. Your Mastodon server is part of your username, and becomes part of your identity. On Bluesky, your experience is based on what feeds and accounts you follow, and you can always participate in the global conversation (e.g. breaking news, viral posts, and algorithmic feeds). You can use your own domain name as your username, and continue participating from anywhere your account is hosted.
• Composable moderation: Moderation on Bluesky is not tied to your server, like it is on Mastodon. Defederation, a way of addressing moderation issues in Mastodon by disconnecting servers, is not as relevant on Bluesky because there are other layers to the system. Server operators can set rules for what content they will host, but tools like blocklists and moderation services are what help communities self-organize around moderation preferences. We’ve already integrated block and mute lists, and the tooling for independent moderation services is coming soon.
• Composable feeds: We designed your timeline on Bluesky so that it’s not tied to your server. Anyone can build a feed, and there are currently over 40,000 algorithmic feeds to choose from. Your Mastodon timeline is only made up of posts from accounts you follow, and does not pull together posts from the whole network like Bluesky’s custom feeds.
• Account portability: We designed federated hosting on Bluesky so that you can move servers easily. Moving hosting services should be like changing your cell phone provider — you should be able to keep your identity and data. Changing servers on Bluesky doesn’t disrupt your username, friends, or posts.

This is important, though there are still some details to be worked out, especially around the third-party moderation efforts. But, on the whole, having the ability to still interact with the wider Bluesky community while keeping your personal data server somewhere else that you control is a big step forward in realizing how a more decentralized social media could (and I’d argue, should) work. It brings us back towards the world of an open web, and away from locked-in silos.

Now, again, there are still some parts of the system that people are worried about, in particular how they could be open to centralized capture. The thing is, there is always going to be some risk of this on basically any system. To make things work properly, you tend to need certain parts of the stack to either work together seamlessly, or it just ends up that a very small number of giant players end up dominating the otherwise “open” system anyway.

This is a concern worth watching. However, it’s also been one that the Bluesky team has repeatedly and readily acknowledged, along with their ideas and thinking on how to guarantee that future Bluesky (or anyone else) is effectively incentivized against enshittification. That’s not to say it will all work out, but so far I’ve seen no reason not to believe that the team has been building with this in mind. Its last few major announcements have all shown continued movement in this direction.

At the end of this tunnel, there is a very powerful vision, one that is partially (though not entirely) laid out in the Protocols, Not Platforms paper. In this vision, people can either self-host their own data servers or find a trusted third party to do so, with the ability to move if the current host turns out to be a problem. It’s one where there are many different tools to allow people to craft their own experience (though composable moderation and algorithmic choice within the system) and the moderation layer is separate and extracted from the data server, the app, and the hosting company.

There will be services that combine them all (like Bluesky today), but also we’re increasingly moving towards the world in which people will be able to adjust things to their own liking. And that can be powerful in its own way. No, most users won’t want to get down into the weeds and tweak things themselves. But that’s where there’s an opportunity for organizations to step up and provide a comprehensive solution themselves, whether it’s Bluesky itself, or others.

But, just the fact that users can modify basically everything, and that third parties have free ability to build apps and services (and custom feeds) on top of this core, has an added advantage, even for those who don’t want to tweak the details and fiddle the knobs themselves. The very fact that it’s possible (or that it’s possible to jump to other providers) creates a strong anti-enshittification incentive structure.

One of the big reasons that enshittification occurs is because users are locked-in. There’s no easy way to leave, without a massive hassle. And part of that hassle is losing access to friends and family. The exciting part of Bluesky with federation is that there is no lock-in, which means there’s much less temptation for enshittification and rent extraction from users with nowhere else to go.

This move towards federation is a small move towards that larger vision, but it’s an important one.

Filed Under: decentralization, federation, protocols, protocols not platforms, social media
Companies: bluesky

Bluesky Opens Up

Innovation

from the the-skies-are-blue dept

Bluesky is now open to anyone without an invite. And a bunch of other exciting things are coming soon.

As many of you know, I’ve been pretty excited about where Bluesky is going as a social media offering, not just because of the people who have been using it (who have mostly been great, making it a fun place to hang out these days), but because of the concepts behind it.

Bluesky was originally seeded by Jack Dorsey in response to his reading my Protocols, Not Platforms paper. The project moved along (somewhat slowly) for a few years. While Dorsey funded it, and the idea was that Twitter would eventually adopt the protocol, it was created as a wholly independent company, which initially had a contract with Twitter. The whole concept was finally picking up steam, just as Elon bought Twitter and cancelled the company’s relationship with Bluesky.

From there, the company quickly pivoted to release a reference app of its own, to give people a sense of how you could build a social media network that wasn’t awful and wasn’t confusing. But, because of the rush to set up their own network, and the numerous features it didn’t yet have ready (e.g., in the early days there was no “block” feature at all), it was setup as a closed beta, where you needed an invite to use the system.

Even with that invite system, Bluesky grew to over 3.2 million users (not all of whom have stuck around, but the network keeps growing). Over the past year, Bluesky has built out a number of new features, both ones to reach parity with what’s expected of most social networks, as well as some unique (and important) ones.

For example, the company has added some (still early) features that give users much more control over their experience: composable moderation and algorithmic choice. Composable moderation lets users set some of their own preferences for what they want to encounter on social media, rather than leaving it entirely up to a central provider. Some people are more willing to see sexual content, for example.

But, the algorithmic choice is perhaps even more powerful. Currently, people talk a lot about “the algorithm” and now most social networks give you one single algorithm of what they think you’ll want to see. There is often a debate among people about “what’s better: a chronological feed or the algorithmically generated feed” from the company. But that’s always been thinking too small.

With Bluesky’s algorithmic choice, anyone can make or share their own algorithms and users can choose what algorithms they want to use. In my Bluesky, for example, I have a few different algorithms that I can choose to recommend interesting stuff to me. One of them, developed by an outside developer (i.e., not Bluesky), Skygaze, is a “For You” feed that… is actually good? Unlike centralized social media, Skygaze’s goal with its feed is not to improve engagement numbers for Bluesky.

I also have feeds showing me “quiet posters” (calling attention to posts from users who don’t post all that often) or posts that are “popular with friends.” I have a few different topic-focused algorithms as well, including one highlighting breaking stories from journalists, and others highlighting posts from folks interested in tech law and policy.

In other words, rather than letting Bluesky curate my experience (or leaving it up to the whims of a chronological feed), I get to curate the experience myself, with help for anyone else who is creating and releasing their own feed algorithms.

And all of that is about to get even better. Because Bluesky also announced that they’re opening up their moderation system as well, to enable a similar sort of feature for moderation:

In the coming weeks, we’re excited to release the labeling services which will allow users to stack more options on top of their existing moderation preferences. This will allow other organizations and people to run their own moderation services that can account for industry-specific knowledge or specific cultural norms, among other preferences.

One potential use case for labeling is fact-checking. For example, a fact-checking organization can run a labeling service and mark posts as “partially false,” “misleading,” or other categories. Then, users who trust this organization can subscribe to their labels. As the user scrolls through posts in the app, any labels that the fact-checking organization publishes will be visible on the post itself. This helps in the effective distribution of the fact-check and keeps users better informed.

I expect that we’ll begin to see a lot of innovation there as well.

In addition, the company has said that it is finally rolling out its long awaited federation features. While Bluesky and its underlying ATProtocol was always designed to be a federated network, to date, the only real way to use Bluesky was to rely on Bluesky’s servers. There are some amazing third party clients (Deck.blue is an astoundingly great Tweetdeck-like multi-column client), but they’re still just showing you what’s on Bluesky’s servers.

But that’s changing:

This month, we’ll be rolling out an experimental early version of “federation,” or the feature that makes the network so open and customizable. On Bluesky, you’ll have the freedom to choose (and the right to leave) instead of being held to the whims of private companies or black box algorithms. And wherever you go, your friends and relationships can go with you.

I know that a lot of people hear “federation” and worry that it will be confusing and complex, as it often feels on something like Mastodon (though, Mastodon has put a lot of effort into making that experience better). But Bluesky is building from the ground up with a plan to make the federation aspect as seamless as possible.

All of this is pretty exciting. Yesterday, I spoke to Will Oremus at the Washington Post as he was working on an article about Bluesky opening up, and I said something to him (which didn’t make his article) but I think is important. I mentioned that I’ve always believed that there were two ways to make a “protocols” approach to social media work: (1) convince a big company to move away from a centralized system or (2) have someone use a protocol based system to build something that was just, fundamentally, at its core better.

Both approaches have challenges to succeed. But I think it’s fascinating that Bluesky started as (1), but has very much moved to (2) (and, of course, it’s notable that I never included “have a narcissist billionaire ruin one major platform that people kinda liked” as a third option for how this might work).

It’s still a long way to go to see if Bluesky succeeds, and there are oh so many ways it could go wrong. But the inclusion of composable moderation and algorithmic feeds already gives me a way better experience than any other social media platform, and it does so not in the service of any billionaire, but rather in service of me, the user. And that is incredibly encouraging as a start.

And, given the open nature of ATProtocol, it also means that if Bluesky fucks it up, and doesn’t actually continue to build in this direction, others have the ability to make it better for them (and for everyone).

To celebrate opening up, Bluesky teamed up with artist Davis BIckford to create a lovely comic explaining why Bluesky is different, and why it matters. You can see the whole thing in Bluesky’s post on opening up, but here’s just a snippet.

I know lots of people like to crap on social media. And I’ve heard a bunch of people insist that Bluesky is too late to the party, or that Threads will kill it or some-such. And, hey, that may be true. But right now, it’s a place that offers a fantastic user experience, which puts you in control more than any other. And, once federation opens up you don’t even have to worry about it being in service to a single company or a single billionaire.

Filed Under: algorithmic choice, atprotocol, bluesky, composable moderation, federation, protocols, protocols not platforms, social media, user empowerment
Companies: bluesky

Content Moderation At Scale Is Impossible To Do Well, Child Safety Edition

Content Moderation

from the child-safety-isn't-just-about-flipping-a-switch dept

Last week, as you likely heard, the Senate had a big hearing on “child safety” where they grandstanded in front of a semi-random collection of tech CEOs, with zero interest in actually learning about the actual challenges of child safety online, or what the companies had done that worked, or where they might need help. The companies, of course, insisted they were working hard on the problem, and the Senators could just keep shouting “not enough,” without getting into any of the details.

But, of course, the reality is that this isn’t an easy problem to solve. At all. I’ve talked about Masnick’s Impossibility Theorem over the years, that content moderation is impossible to do well at scale, and that applies to child safety material as well.

Part of the problem is that much of it is a demand-side problem, not a supply side problem. If people are demanding certain types of content, they will go to great lengths to get it, and that means doing what they can to hide from the platforms trying to stop them. We’ve talked about this in the context of eating disorder content. Multiple studies found that as sites tried to crack down on that content, it didn’t work, because users demanded it. So they would keep coming up with new ways to talk about the content that the site kept trying to block. So, there’s always the demand side part of the equation to keep in mind.

But also, there are all sorts of false positives, where content is declared to violate child safety policies, when it clearly doesn’t. Indeed, the day after the hearing I saw two examples of social media sites blocking content which they claimed were child sexual abuse material, when it is clear that neither one actually was.

The first came from Alex Macgillivray, former General Counsel at Twitter and former deputy CTO for the US government. He was using Meta’s Threads app, and wanted to see what people thought of a recent article in the NY Times raising concerns about AI generated CSAM. But, when he searched for the URL of the article, which contains the string “ai-child-sex-abuse,” Meta warned him that he was violating its policies:

In response to his search on the NY Times URL, Threads popped up a message saying:

Child sexual abuse is illegal

We think that your search might be associated with child sexual abuse. Child sexual abuse or viewing sexual imagery of children can lead to imprisonment and other severe personal consequences. This abuse causes extreme harm to children and searching and viewing such material adds to that harm. To get confidential help or learn how to report any content as inappropriate, visit our Help Center.

So, first off, this does show that Meta, obviously, is trying to prevent people from finding such material (contrary to what various Senators have claimed), but it also shows that false positives are a very real issue.

The second example comes from Bluesky, which is a much smaller platform, and has been (misleadingly…) accused of not caring about trust and safety issues over its approximate one year since opening up as a private beta. There, journalist Helen Kennedy said she tried post about the ridiculous situation in which the group Moms For Liberty were apparently scandalized by the classic children’s book “The Night Kitchen” by Maurice Sendak, which includes some drawings of a naked child in a very non-sexual manner.

Apparently, Moms For Liberty has been drawing underpants on the protagonist of that book. Kennedy tried to post side by side images of the kid with underpants and the original drawing… and got dinged by Bluesky’s content moderators.

Again, there, the moderation effort falsely claims that Kennedy was trying to post “underage nudity or sexual content, which is in violation of our Community Guidelines.”

And, immediately, you might spot the issue. This is posting “underage nudity,” but it is clearly not sexual in nature, nor is it sexual abuse material. This is one of those “speed run” lessons that all trust and safety teams learn eventually. Facebook dealt with the same issue when it banned the famous Terror of War photo, sometimes known as “the “Napalm Girl” photo taken during the Vietnam War.

Obviously, it’s good that companies are taking this issue seriously, and trying to stop the distribution of CSAM. But one of the reasons why this is so difficult is that there are false positives like the two above. They happen all the time. And one of the issues in getting “stricter” about blocking content that your systems flag as CSAM, is that you get more such false positives, which doesn’t help anyone.

A useful and productive Senate hearing might have explored the actual challenges that the companies face in trying to stop CSAM. But we don’t have a Congress that is even remotely interested in useful and productive.

Filed Under: child safety, csam, false positives, masnick's impossibility theorem
Companies: bluesky, meta

We Teamed Up With Bluesky To Tell The Supreme Court How State Social Media Laws Don’t Take Into Account User Empowerment

Content Moderation

from the empowering-users-is-important dept

As you know, the Supreme Court is now considering the NetChoice/CCIA cases challenging two similar (but not identical) state laws regarding social media moderation. The laws in Florida and Texas came about around the same time, and were clearly written to target ideological speech. Both of them put restrictions on how certain social media apps can moderate or even recommend certain speech.

As you’ll recall, district courts in both states found the laws obviously unconstitutional attacks on the 1st Amendment. On appeal, things went differently. The 11th Circuit agreed that most of the Florida law was unconstitutional (we think they’re wrong about the part they weren’t concerned with too). But the 5th Circuit went rogue and said that of course states can set whatever moderation laws they want (which is in conflict with later 5th Circuit rulings regarding state pressure on moderation, but I digress).

Anyway, you can follow along on the docket for the case at the Supreme Court, where NetChoice/CCIA filed their brief recently. This week, a bunch of amicus briefs are being filed, some of which are really interesting.

I wanted to focus on one brief in particular in this post: our own, written by Cathy Gellis. We teamed up with Bluesky (the alternative microblogging service that is building a federated protocol for social media) and Chris Riley (in his personal capacity as the operator of a Mastodon instance) to make some points that we don’t think other amici are likely to make.

The key point we tried to make is that so much of the arguments being thrown back and forth are really about who it is that gets to determine how a website moderates: should it be the government or should it be the website? If those are the only two options there are, then it already does seem obvious that it should be the website, not the government.

But, the key to our brief is pointing out that this assumes, falsely, that this is the only possible model out there. Instead, we highlight, that it is possible to envision a world in which users themselves get to decide, and any ruling that says the government gets to decide would fundamentally make that kind of user freedom and empowerment impossible.

A fundamental part of my Protocols, Not Platforms article (which, in part, helped inspire Bluesky) was that it would empower users to have more control themselves, or at least let them choose which intermediaries they trust to help them with algorithms and moderation, rather than relying on the same platform that they use for the hosting of the content itself.

For example, Bluesky has an amazingly useful feature called “custom feeds” and it has created a marketplace of algorithms so that you can create your own algorithms and share them with others, or you can just decide to use someone else’s algorithm (or even adapt it further yourself). That is, rather than relying on a company like Twitter or Facebook to decide what you should see, on Bluesky, you get to make those decisions yourself, or hand them off to someone you trust.

But, in that architecture, it means that Bluesky often won’t even know what algorithms people are using (the algorithms don’t have to live on Bluesky’s servers, indeed, Bluesky itself might never even be aware of them). But should these laws (or laws like them) apply to Bluesky, that kind of ecosystem basically would be effectively barred, because the law would limit what kinds of algorithms could work on Bluesky, and Bluesky itself would have no way to control those third party algorithms.

This ecosystem of platforms is necessary in order for there to be meaningful choices in what expression Internet users experience online. Platform choice, and the customization algorithmic choice enables, are what helps realize the expression-promoting value of the Internet and ensures it captures a diversity of expression by putting the choices of what expression to be exposed to in the hands of users. It is not for the government to take away this choice, creating a platform or algorithmic monoculture, which is what the Florida and Texas laws threaten.

Similarly, we used the example of how comments here at Techdirt are moderated, in which much of the moderation is actually handled by community votes, and how there’s no way to comply with these laws for such community moderation practices either:

But while the Copia Institute’s moderation practices can be described in broad strokes, they cannot be articulated with the specificity that the Texas law would require. For instance, the law requires that platforms disclose their moderation standards. See, e.g., TEX. BUS. & COM. CODE § 120.051. And it also puts limits on how platforms can do this moderation. See, e.g., TEX. CIV. PRAC. & REM. CODE § 143A.002 (banning certain moderation decisions, including those based on the “viewpoint” of the user expression being moderated). But even if the Copia Institute wanted to comply with the Texas law, it could not. For instance, it could not disclose its moderation policy because its moderation system is primarily community-driven and subject to the community’s whims and values of the moment. Which also means that it could not guarantee that moderation always comported with a preannounced “Acceptable Use Policy,” which the Texas law also requires. TEX. BUS. & COM. CODE § 120.052. It would also be infeasible to meet any of the Texas law’s additional burdensome demands, including to provide notice to any affected user, TEX. BUS. & COM. CODE § 120.103, maintain a complaint system, TEX. BUS. & COM. CODE § 120.101,24 or offer a process for appeal,25 TEX. BUS. & COM. CODE § 120.103. None of these faculties are features the Copia Institute has the resources or infrastructure to support. In other words, the Texas law sets up a situation where if the Copia Institute cannot host user-provided content exactly the way Texas demands, it effectively does not get to host any user-provided content at all. Or, potentially even worse, it would leave Techdirt in the position of having to host odious content, including content threatening to it, its staff, or others in its reader community, in order to satisfy Texas’s moderation requirements.

There are all sorts of other important 1st Amendment reasons why these laws are deeply problematic. But we assumed (almost certainly correctly) that the briefs from NetChoice/CCIA and other amici will cover all of that.

Our brief was more focused on highlighting how these issues go beyond just the 1st Amendment concerns of big websites, but how they might impact a new generation of social media platforms, like Bluesky and Mastodon, whose very models and infrastructure are fundamentally different from the “giant silos” of today’s major social media platforms.

Too much of the discussion assumes that there are only two parties who might have a say in the moderation of social media: governments and the platforms. But we want to make the court aware that a new generation of services are focused on enabling the users themselves to make that choice, and if these laws are allowed, it could wipe out that possibility.

Filed Under: community moderation, content moderation, decentralized, federation, fediverse, florida, protocols, social media, supreme court, texas
Companies: bluesky, ccia, copia institute, netchoice

One Year In: Elon Has Inspired New Approaches To Social Media. Just Not His Own

Too Much Free Time

from the a-retrospective dept

Today is the official one year anniversary of Elon getting control over what used to be called Twitter, and now is simply exTwitter. It was supposed to be tomorrow, but in a sign of what was to come, Elon and his buddies maneuvered to close the deal in the afternoon a day early, just to maximize their assholish tendencies.

The closing of the Twitter deal had been scheduled for that Friday. An orderly transition had been scripted for the opening of the stock market that morning. The money would transfer, the stock would be delisted, and Musk would be in control. That would permit Agrawal and his top Twitter deputies to collect severance and have their stock options vest.

But Musk decided that he did not want that. On the afternoon before the scheduled close he methodically planned a jiu-jitsu maneuver: He would force a fast close that night. If his lawyers and bankers timed everything right, he could fire Agrawal and other top Twitter executives “for cause” before their stock options could vest.

It was audacious, even ruthless. But it was justified in Musk’s mind because of his conviction that Twitter’s management had misled him. “There’s a 200-million differential in the cookie jar between closing tonight and doing it tomorrow morning,” he told me late Thursday afternoon in the war room as the plan unfolded.

Of course, it was never true that Twitter management misled Elon. What is true is that Elon didn’t bother to do even the most basic due diligence (and, in fact, waived the right to do so), and signed a contract which basically everyone admits was a massive overpay, that also saddled the company with significant debt.

If we take stock of how things are looking one year in, it can be summed up simply by saying “not great, Bob.” Musk told bankers that he had a clear plan to get the company to be worth $250 billion before long, and the bankers bought it. But so far, basically none of his plan worked. The pitchdeck claimed that he would quadruple revenue to $26.4 billion by 2028. Instead, he’s cut it by at least 60%.

He said he’d more than triple users to 900 million by 2028. Instead, user numbers have been dropping. Indeed, the Wall Street Journal got access to some more data (beyond what we had in the previous post) showing that exTwitter seems alone in losing users, as other sites are gaining them:

Also, you see how it was a gradual decline, and then a steeper cliff after July? Want to know why that is? It coincides, almost exactly, with Musk’s big “rebrand” to “X.”

These aren’t issues of “market conditions” or things that were screwed up through outside forces. Every one of the problems stems directly from Elon Musk having no fucking clue what he’s doing.

He claimed he’d increase average revenue per user by $5.39, and his big idea there seems to have been to co-opt Twitter Blue (which had the kernel of a good idea, but wasn’t marketed very well by old Twitter) and turn it into X Premium (while also, ridiculously, getting rid of actual verification and pretending that X Premium was verification). And that program cannot be described as anything but a colossal failure, with even those who were interested in paying gradually losing interest in continuing. The value just isn’t there.

The best estimate I’ve seen for how many people are paying for X premium is somewhere around a million people. This is well less than half a percent of Twitter’s claimed user base. Now, converting people to premium offerings is always harder than people think, but less than half a percent is embarrassing. Those are the kinds of results that gets people fired. It also means that Elon gave up something in the range of $2.5 to $3 billion in ad revenue… to get back about $100 million in subscription numbers. That’s… bad.

So, no, it does not look like the ARPU numbers are going up to $5.39, and I doubt the plan to charge $1/year for everyone is going to help.

Of course, the banks who lent Elon all this money are now pissed off. They’ve been unable to unload the loans like they’d planned, and are recognizing just how much money they’ve lost on this deal. Apparently, the banks own investors are asking how the banks could have been so fucking stupid to loan Elon money for Twitter:

The X deal should have been a fee bonanza for the banks, who stood to earn tens of millions of dollars on the debt. Instead, their inability to resell it has been an albatross on their lending businesses and prompted questions from their own investors.  

Banks limit how much risk they take on at any given time, so holding X’s debt has taken up loan-book capacity that their deal makers would prefer to allocate elsewhere.

But really, the banks have no one to blame but themselves. Too many people fell for the myth of Elon having the Midas Touch, and insisted that he was some sort of ultra genius who could turn any company he touched to gold. But it was fucking obvious from the jump that no matter how much he might have (or might not have) contributed to his other companies, he never had the slightest fucking clue how social media works.

And the astounding thing is that one year in, it’s clear he’s still learned nothing.

Most social media startup CEOs end up going through the learning curve. Eventually, they figure things out. Sometimes it’s too late. But, by the end, they start to understand the basics. The incredible thing with Elon is that he doesn’t appear to have learned anything from his mistakes here.

Perhaps more tragic is that he’s basically destroyed what had been the best place to go for rapid breaking news coverage and analysis. While Twitter was always smaller than the other platforms, it made up for it in being the best “real time” source of news when something big was happening. But, as we’ve learned over the past few weeks, going to exTwitter to find out what’s happening with, say, Israel/Palestine or with the shooting in Maine is an utter disaster. Musk himself initially promoted two grifter accounts with a history of posting completely false nonsense to follow about the Israel/Palestine situation, and given that many more respected/trustworthy news purveyors have reasonably abandoned or limited their use of the platform, exTwitter’s greatest value is basically gone.

That said, he has inspired some unique experimentation in the social media space. At the six month anniversary of the takeover, I wrote about what appeared to be the three big “contenders” to take over the void that Twitter had left open for real time news. Six months later, and there are still tons of interesting things happening. I feel like Mastodon lost a ton of its early momentum by being effectively hostile to people who wanted a new Twitter-like space. I think it’s clear that the team behind Mastodon has realized it needed to adapt, but it feels a little like Mastodon is going to be saddled with being the “Linux” of short form social media: never quite going mainstream, no matter how often its many fans (and I remain one) insist that it’s not as hard to use as you’ve heard.

The new entrant since that post six months ago is Threads, from Meta/Instagram. That obviously hit the market with a huge splash and tons of hype, and then… lost a lot of its momentum, and has been trying to rebuild a space for itself in the market. There have been reports lately that suggest maybe it’s finding its footing, but it still seems a bit shakey. Some of the problem is that it still doesn’t seem to know what it wants to be. At times it acts like it wants to be the Twitter replacement, but then the people behind it keep saying that they’re downplaying the discussion of news on the platform, mainly because they know it gets impossible to moderate. But that also makes people who do want to discuss the news feels somewhat unwelcome.

The big unknown, of course, is if Meta ever actually lives up to its stated plan to federate Threads with ActivityPub, enabling Mastodon users (and users of other Mastodon-compatible ActivityPub implementations) to communicate with people on Threads. People at Meta insist it’s still a part of the plan, and that could be really interesting depending on how it’s implemented. But we just don’t know the details yet.

Six months ago, I had also mentioned nostr, which remains a fun project in its complete embrace of openness, but the project’s leaders seem so naive about what it takes to set up a social media protocol that it’s driving away basically anyone who doesn’t want to just talk crypto all day. I’m still hopeful for interesting projects to come out of nostr, because it has some advantages over the alternatives, but for now it’s just not getting usage outside of a niche.

And then, there’s Bluesky. This remains the one that I’m most hopeful about, and where I’m spending more and more of my time. Six months ago it only had around 100k users, and now it just surpassed 1.75 million, with over 1 million users having posted on the platform at least once. And that’s with it still being gated by an invitation system. I’m unaware of any invite-only app that has gotten that big.

It’s also the one that feels the most like early Twitter.

And also, it’s the one that is clearly thinking about how to actually function as a mainstream platform in the real world, while still setting itself up to be decentralized and not just beholden to whoever controls the company. A few months ago, someone from Bluesky explained that they saw their own future company as a potential threat, and were designing accordingly. Of course, it’s one thing to say that, and it’s another to do it. So far, they’ve been talking the talk, but at some point they need to start walking the walk too. I’m still confident they will, but I know some are quite skeptical. Once Bluesky finally goes federated, we’ll see if they can really find the right balance.

And so, one year in, it’s pretty safe to say that Elon Musk has been truly great at inspiring new and creative ides for better social media platforms. Just not his own.

Filed Under: ads, business models, elon musk, social media, speech, users
Companies: bluesky, meta, threads, twitter, x

Trust, Safety, And Communication

Content Moderation

from the your-comms-strategy-is-your-trust-strategy dept

One thing that should be evident by now is that every online community eventually learns the need for some level of “trust & safety” or basic moderation to take place. And they quickly find that things are a lot more complex than they seem from the outside. Just try to moderate a medium sized Facebook Group if you want an example. Or play our Moderator Mayhem game.

It’s always interesting to see the points at which community organizers realize this and try to figure out how they’re going to handle this issue or that issue — and begin to realize what an impossible task it is. And while some people think that it’s now been long enough that any new community should have “the basics” figured out, it’s important to recognize that (1) there are always new problems, and (2) the “simple” problems are often a lot more complex than they seem. On top of that, there are thousands of things that any new community “should” have, and at some point the people building them need to weigh “releasing something” against “having every feature in place.” You can disagree with where the line is drawn, but everyone has to draw a line somewhere.

I was thinking about all this over the last week or so as there was some discussion when the (still invite-only beta) Bluesky ran into some issues regarding a username filter (specifically, the filter allowed users to sign up with slurs as their user names). This is, obviously, not good.

The debate on Bluesky morphed over the course of a few days from criticism regarding the pretty major omissions on the filter list, to the lack of communication from the company and its (normally communicative) employees. Basically the entire company went silent, followed eventually by some more bland “corporate” sounding responses that went against the “poasting” style the team had embraced earlier. Indeed, the sudden silence from the team stood out even more given their normal willingness to engage in all sorts of ways on just about everything else. Going from super talkative to silent at the moment of notable controversy is, perhaps, the opposite of a compelling communications strategy.

Still, it’s somewhat understandable when looked at in context. The teams have repeatedly talked about how much they need to accomplish in building both a protocol (which may change the nature of some of these issues) and their own platform as a reference app of that protocol. The service is still in beta for a reason. And when there are thousands of trust & safety things you need to set up in addition to building the platform and the service, raising money, finding a business model, and everything else, it can get a little overwhelming. And that’s especially true when the company had made earlier moves and statements suggesting that they took these issues seriously and were working on solutions. So, when things blow up because they missed some things, it can feel like an attack. The team believes their heart is in the right place, and they’re trying to balance the variety of things they need to do, and yet… they’re still getting yelled at.

But, alas, this is the general rule when you run any sort of online community: you will get yelled at, and at some point you need to decide what issues to deal with and what to focus on. Getting yelled at sucks. And often makes people clam up. Of course, the obvious (and very true!) counter to this is that having to deal with hate, abuse, and racism also sucks. And also makes people want to clam up. So if you’re taking a job to build a social network, you’re signing up for this specific kind of abuse, and you need to be ready for it in order to protect others from abuse.

Back in May, I had written a thing about social media Nazi bars, tradeoffs, and the impossibility of content moderation at scale, which I think remains quite relevant here. There are always tradeoffs, and unlike, say, Substack (which is much larger and much more well resourced), I’ve seen no indication that the Bluesky team is simply abdicating its responsibilities here, but rather prioritizing as best it can, meaning some things that everyone agrees are important won’t get put in place as quickly as some would hope.

For example, regarding the filter list, while it was an obvious failing in how the system was set up, any sort of brute filter list runs into problems over time. If you don’t want to deal with a “Scunthorpe” problem, you need a more sophisticated solution, and more sophisticated solutions require more time and thought, and we’re right back to the line-drawing exercise I mentioned above, where the long list of thousands of things you need to accomplish is at least one item longer (and more complex).

The real difference here seemed to be how much the communications problem exacerbated the more classic trust & safety failing.

And it made me start to think about how communications itself is a strategic trust & safety tool, though rarely considered as such. Some of the communications issue was, as many people noted, the failure of the company to come out and say they were sorry for the errors. And, yeah, it seems like this is a case where company leadership should have done so. But sorry only goes so far. Mark Zuckerberg has to keep going on apology tours, and it’s not clear that it’s really helpful.

Instead, I think the failure might be in the lack of clear communication on the larger roadmap from Bluesky. This applies to lots of other online communities as well, but I’m focused on Bluesky to make this point (although arguably, it applies even more to others). Many people (I think, falsely) focused on this one error regarding to the username filter list, insisting that it showed the company “didn’t care.” That struck me as unlikely, given earlier statements and actions by the Bluesky team, which seemed to indicate not just that they cared about this, but they cared deeply, to the point that they wanted more thoughtful, serious, and comprehensive approaches to dealing with it, rather than slapdash duct tape fixes.

But, unless you’re paying close attention, you might miss all of that. And there’s no clearly laid out roadmap that people might have pointed to to alleviate the concerns of others.

Things might have gone a bit differently if Bluesky had a page with a roadmap regarding its plans for federation, composable moderation, trust & safety tooling, trust & safety hiring, and the like. If such a roadmap existed, that showed exactly how the team was thinking about these things, and made it clear that the team was working towards them deliberately, including at least some public explanation of the tradeoffs of various approaches, it would be more difficult for users to fill in the void with “they just don’t care.”

On top of that, it would similarly give the team breathing room to keep working on that roadmap, rather than having to respond to every emergency (some emergencies will still require emergency reactions, but not every emergency will grind everything else to a halt). This isn’t the answer to everything, of course. Nothing is.

But having clear communications, especially regarding a project that is designed to be decentralized and is being designed for the public benefit, is a key element of building trust, which I guess would be somewhere around 50% of the point of building out trust & safety.

Creating such a roadmap is quite a process in and of itself. As far as I can tell, no one else has done it either. And I’ve already been talking about how the team likely already has too much on its plate. But it does strike me that spending a bit more time on this at this early stage might help prevent some of the problems going forward, both in allowing users to point out some areas where the roadmap may need to be adjusted, or in simply having a better understanding of not just where Bluesky is today, but where it’s heading in the future.

I honestly think this understanding of the communications element of trust & safety could help many other communities as well. Many of the complaints and problems come from a mismatch between expectations and how a company actually makes decisions. And one way to deal with that is to better align the expectations. I think the last decade might have gone differently if Twitter, Facebook and others had been more public and upfront with some of their internal trust & safety discussions as well, so this is hardly unique to Bluesky.

But, at the very least, I think it’s important to start considering the role of communications as a part of a trust & safety strategy.

Filed Under: communications, content moderation, roadmap, trust and safety
Companies: bluesky

Bluesky Continues To Explore More Creative Moderation Plans Openly

Content Moderation

from the the-company-is-a-future-adversary-to-its-current-plans dept

I continue to be fascinated in watching how the various decentralized protocol-based social media systems are evolving — in particular how they’re dealing with the challenges of content moderation. There was an interesting discussion a recently on nostr over whether or not moderation should be best handled by relays or clients*.

ActivityPub has, of course, continued to move forward with its systems of moderation handled at each instance level, combined with the threat of “defederation” being used to keep “bad” instances in line (or cut off from parts of the network). That’s worked surprisingly well in some cases, but is also facing a few challenges, as there have been complaints about some of the largest instances, and now that Meta is planning to release an ActivityPub-compatible offering, there’s a weird push to make some instances promise to defederate from any Meta offering immediately.

And then, there was the big “Scaling Trust” report that we talked about last week (disclaimer: I was on the task force that helped with the report), which included an annex all about decentralized trust & safety.

But, again, Bluesky may be where the most interesting discussions on decentralized trust & safety and moderation are happening. A few months ago, we wrote about their plans for decentralized composable moderation, and recently they released some thoughts on how you can handle moderation in a public commons.

The goal of Bluesky is to turn social media into a shared public commons. We don’t want to own people’s social graphs or communities. We want to be a tool that helps communities own and govern themselves.

The reason we focus on communities is that for an open commons to work, there needs to be some sort of structure that protects the people who participate. Safety can’t just be left up to each individual to deal with on their own. The burden this puts on people — especially those who are most vulnerable to online abuse and harassment — is too high. It also doesn’t mirror how things work in the real world: we form groups and communities so that we can help each other. The tooling we’re building for moderation tries to take into consideration how social spaces are formed and shaped through communities.

Somewhat importantly, they make it clear that they don’t have all the answers (no one does!), but it’s really interesting to see them discussing this openly, and publicly, and asking for thoughts and feedback as they move forward. To me, the thing that stands out is that the ideas that are presented obviously involved a lot of thought (to the point that I haven’t fully wrapped my head around some of the different proposals, some of which seem clever, while others may need a bit more baking before they fully make sense).

Historically, trust & moderation decisions come in two forms: formed on high in a centralized system in which little is discussed publicly, and people are left trying to sort through what’s actually happening, or in an entirely distributed manner in which things often spring up ad hoc out of need (see: Usenet killfiles), which often run into problems later on.

The Bluesky folks are trying to think about something that is a more hybrid approach, in which the system itself is design to enable communities to better manage things, not just one giant opaque centralized control bunker, and not putting all the weight on users which is unfair to many (especially the targets of abuse and harassment).

I think this kind of vision seems exactly the right one for an organization like Bluesky to have:

A company is an efficient structure for building out a cohesive vision of how things should work, but locking users into our systems would be antithetical to our mission. An open commons can’t be governed at the sole discretion of one global company. We offer services like professional moderators so that we can help protect people and provide a good experience, but we shouldn’t exert total control over everyone’s experience, for all time, with no alternative. Users should be able to walk away from us without walking away from their social lives.

The reason we’re building in decentralization is because we observed that business interests and the open web have a habit of coming into conflict. Third-party developers often get locked out. Moderation policies come into conflict with the diverse interests and needs of different groups of users. Ads push towards algorithms that optimize for engagement. It’s a systemic problem that keeps playing out as centralized social media companies rise and fall.

On Bluesky itself, the lead developer, Paul Frazee noted that they view the future company as a potential adversary, and are designing accordingly. That, alone, is a fascinating perspective to have on things, and one that certainly makes sense in the age of enshittification. And, unlike the way many companies that start on the open web, and later come into conflict with it, as they seek to pull up the ladder behind them to protect a moat, Bluesky is trying to design its systems in a way that protects the system from their own future attempts at enshittification:

Even when things are working correctly on social platforms, there are weird dynamics caused by people’s relationships being mediated by a single company. The Internet is pretty obviously real life in the sense that its management has real-world consequences. When these places control our identities and our ability to connect and to make money, having no way out from the founding company is a precarious situation. The power difference is daunting.

The goal of Bluesky is to rebuild social networking so that there’s not a lock-in to the founding company, which is us. We can try to provide a cohesive, enjoyable experience, but there’s always an exit. Users can move their accounts to other providers. Developers can run their own connected infrastructure. Creators can keep access to their audiences. We hope this helps break the cycle of social media companies coming into conflict with the open web.

Now, some users point to the complex onboarding of Mastodon, or the “WTF how does any of this work?” nature of nostr, and worry that any decentralized/federated system has to be confusing. And that user unfriendliness, in some weird way, acts as a moderation tool in its own right, by keeping communities somewhat smaller. But it also keeps communities… smaller. So Bluesky has a different vision. A surprisingly refreshing and honest one:

A great experience should be simple to use. It shouldn’t be overly complex, and there should be sensible defaults and well-run entry points. If things are going well, the average user shouldn’t have to notice what parts are decentralized, or how many layers have come together to determine what they see. However, if conflict arises, there should be easy levers for individuals and communities to pull so that they can reconfigure their experience.

A great experience should recognize that toxicity is not driven only by bad actors. Good intentions can create runaway social behaviors that then create needless conflict. The network should include ways to downregulate behaviors – not just amplify them.

A great experience should respect the burden that community management can place on people. Someone who sets out to help protect others can quickly find themselves responsible for a number of difficult choices. The tooling that’s provided should take into account ways to help avoid burnout.

A great experience should find a balance between creating friendly spaces and over-policing each other. The impulse to protect can sometimes degrade into nitpicking. We should drive towards norms that feel natural and easy to observe.

A great experience should reflect the diversity of views within the network. Decisions that are subjective should be configurable. Moderation should not force the network into a monoculture.

Finally, a great experience should remember that social networking can be pleasant one day and harsh the next. There should be ways to react to sudden events or shifts in your mood. Sometimes you need a way to be online but not be 100% available.

There is no perfect content moderation solution out there. There is no whiz bang simple technical solution to the messiness that is human beings. As I’ve said many times, so many trust & safety dilemmas are really societal problems that we think are new or need to be solved by internet companies because they’re appearing through screens over the internet.

And, of course, nothing that Bluesky is working on may turn out to work, or matter. It’s still a small operation, and some of these ideas are completely untested. But, at the very least, it is presenting some pretty thoughtful ideas in an open way, and trying to think through the real consequences of what it’s creating here. And that, alone, is incredibly refreshing.

* The creator of nostr apparently does not believe moderation should happen at the client level, but when I asked him how relay operators could express their moderation rules suggested it didn’t matter since relays weren’t moderating anyway. Of course, since then I’ve noticed that nostr is being overrun with cryptocurrency spam, so at some point people there are going to realize that something needs to be done.

Filed Under: activitypub, at protocol, content moderation, decentralized social media, nostr, public commons, trust & safety
Companies: bluesky