Leaflet stream

Building a new social internet on open protocols is by its vary nature a highly political project. That social media has meaningfully shaped our society in the recent decade is such a true-ism that it barely feels worth stating anymore. I'm not saying something particularly shocking either by pointing out that the state of Big Tech platforms has political impacts. But the implication is that politics is a highly important lens to understand the new open social protocols. So politics will continue to play a main role in how I write about the ATmosphere, the fediverse, and everything around it.

The Netherlands had an election yesterday, with a clear win for centrist party D66. Further results of the election can be interpreted with whatever kind of story you want to tell. It's clear that the left lost, but to what extend you want to tell stories on whether the right and far-right lost depends on how you frame individual parties like VVD.

My interest is in which political parties actually use Bluesky, and how that relates to political parties and open social networks in other countries.

I'm not gonna do a very exact analysis, just a high-level overview to give some indication of what some of the political parties are, if and how they are using Bluesky, and what that says on politics and the open social web. My comparisons for parties in other countries are inexact, because real politics is complicated and local. Also, for context, there are 150 seats in Dutch parliament.

• The biggest winner is D66, a centrist party (9=>26 seats). Theres some flavors of US Democrats in here, combined with old-school European hardcore liberalism. All the top 5 members of D66 are active on Bluesky. The account of leader Rob Jetten (most likely the new PM) has 7.5k followers, but it is not a personal account. It's managed by his campaign team, and he does not respond personally.
• The PVV is the main far-right party of the Netherlands, with the singular official party member Geert Wilders the face of the party. The party lost significantly in the election (37=>26 seats), and is now tied with D66 for being the biggest party. None of top 5 members (nor anyone lower, as far as I can tell) have a Bluesky account. Wilders is obviously a highly active X user.
• VVD, the main right-wing party of the Netherlands has some analogies to the UK Tories, and stayed roughly the same (24=>22) The VVD itself has a Bluesky account (4.5k) followers. It has posted a few times per month for the last year but apparently the social media team forgot about it during election season and it has not posted in the last 3 weeks, which is kinda funny. Party leader Dilan Yeşilgöz-Zegerius does not have a Bluesky account. Only the nr 3 on the list, Ruben Brekelmans, has a Bluesky account, who is also currently Minister of Defence and he posts polished PR updates regularly.
• GroenLinks-PVDA is the main large-test left party, and lost, from 25 to 20 seats, and leader Frans Timmermans steps down as leader as a result. The party is highly active on Bluesky. The main account of the party posts multiple times per day, and there are dozens of party accounts for regional departments of the GLPVDA. Furthermore, party members are highly active on Bluesky as well, the top 10 candidates all have a Bluesky account, and many lower on the list as well. Furthermore, many of the members use Bluesky as a personal account, responding and actively engaging with the community. For examples, see the accounts of Laura Bromet, Lisa Westerveld, or (my personal vote) Barbara Kathmann.
• CDA, another centrist party on the more social-conservative side, also saw a big win, from 5 to 18 seats. All top 5 candidates have a Bluesky account, only the top 3 actually use their account. Usage is mainly polished PR, not particularly personal.
• The next three parties (JA21, FVD and BBB) are various flavours of far-right parties, that overall gained significant votes, combined from 10 to 21. None of the party members are active on Bluesky.
• Finally there are 7 small parties, all with 3 seats or less, across a wide range of the spectrum. The left PVDD, the animal party, is highly active on Bluesky, with leader Esther Ouwehand being a prime example. The CU, Christian Union, is fairly active, but leader Mirjam Bikker stopped using Bluesky a few months ago.
• A special mention for Volt, the European-focused party. This party lost, from 2 seats to one. What stands out to me is that Volts highly active on Bluesky, see leader Laurens Dassen and nr 2 Marieke Koekkoek. They are both active users of Bluesky, and get high engagement on Bluesky, especially compared to the election results.

By looking at which Dutch politicians use Bluesky, and who is popular on Bluesky, you can get a sense of the political leanings of the Dutch Bluesky community. The main (not hugely surprising) conclusion is that Dutch Bluesky strongly leans left, with all varieties of left parties active users in Bluesky. Centrist parties do use Bluesky, but it is clearly part of a larger social media strategy. Rightwing parties barely use Bluesky, and far-right parties are not present on Bluesky at all.

The biggest outlier is Volt, which is active on Bluesky and much more popular on Bluesky compared to the wider electorate. Volt is a leftwing party with a strong focus on European integration, and more technocratic in their solutions compared to other Dutch left parties. That combination of being leftwing on social issues while also more focused on federation and technology compared to other leftwing parties makes Volt a fairly good analogue in ideology for Bluesky and atproto. But the Dutch election result also indicate that this vision only resonates with a small part of the wider population.

Worth noting: a lot of the most active adopters of Bluesky regarding Dutch politics are women, see Amsterdam mayor Femke Halsema as another example.

Some US politics news that is only indirectly related to Bluesky: Kat Abughazaleh, a Democratic congress candidate for Illinois, announced that she has been "charged in a federal indictment sought by the Department of Justice." Why I'm reporting on Abughazaleh here is because she launched her campaign as an explicitly Bluesky-first operation:

A second piece of US news: Graham Platner is a Democratic congress candidate for Maine, and he turned out to have a nazi tattoo. The Platner campaign is also active on Bluesky, and Bluesky posts regarding Platner were largely supportive. This shifted abruptly when the news of his tattoo came out:

What's especially interesting here is that sentiment on Twitter from Democratic voices seem much milder on Platner, compared to how Bluesky is highly critical of him.

I've written before about the perception of Bluesky as a 'left' space (here, here), but the difference in reactions from Democratic pundits on X versus Bluesky is an indication that Bluesky is indeed a 'left' space, and that this differs from the communities on X that are affiliated with the Democratic party.

The larger context of this is that how the Trump administration is relating itself to Bluesky. For further context on that, see this recent article by Erin Kissane.

The main implication here is that an authoritarian regimes are incompatible with spaces with their political opponents can freely gather, organise and communicate. That the DOJ goes after Democrats who have a high profile online (such as Kat Abughazaleh) is an indication that the Trump administration is aware of which politicians pose a relatively large risk for the regime.

That Abughazaleh has publicly declared that the base of her online presence is Bluesky, while the administration has already taken an adversarial position regarding Bluesky (for now limited to 'trolling'), should put Bluesky on alert that the current equilibrium the platform is in regarding the Trump administration is not particularly stable.

Leaflet has added polls! The feature is missing on Bluesky, but its cool another platform has it now. Leaflet says they're hoping other platforms will add polls as well, and that shared systems can emerge from that: "One of the cool things about ATProto is that we can coordinate gradually over time, and all the data is still out there, in a simple format, for us to build on!"

PiPup is a writing and publishing platform on ATProto. The platform also has a reader feed that shows all long-form articles published across the network over a variety of platforms, supporting Leaflet and WhiteWind, next to PiPup itself.

Bluesky is expanding their team, they've hired for Strategy and Operations, as well as DevRel, and have a variety of jobs open. The newest job they're hiring for is Technical Recruiter.

Alex is the new Bluesky DevRel, and he wrote a blog post on why Bluesky loves DNS.

Bluesky also published an introduction to OAuth for developers:

Lycan is a search tool for your Bluesky likes, made by Kuba Suder.

GreenEarth is a new project to build personalised recommendation algorithms for atproto. The project has gotten 300k USD funding from Project Liberty, and was founded by Renee DiResta. The project has 3 goals:

• build a "prosocial" feed, that demonstrates their earlier research that the right algorithms can "make people hate each other less"
• create open recommender infrastructure for others to work with
• support atproto startups, and they are already working with Graze and Skylight

Having freedom over which custom algorithm to use is one of atproto's most powerful features, and it's interesting to see that being reflected in which startups on atproto actually are able to attract funding.

Speaking about algorithms, the For You feed is currently the hot custom algorithm for Bluesky. The latest update from Bluesky makes the 'post seen' status available for all feeds, not just Bluesky's own Discover feed. This is now implemented in For You as well (and this is how I'm finding out that Leaflet supports quotes in quotes in embedded Bluesky posts)

For You also has an interesting problem, in that the algorithm is highly effective in surfacing people's alt account unintentionally. The current solution from the creator is for people to opt-out of the For You feed via a DM, but that hardly seems like a scalable solution.

Parallel to this problem is the longstanding wish of Bluesky users to be able to opt out of the Discover feed as well.

It feels like Bluesky is struggling with transitioning their mindset from building something resembling Twitter to something truly different. Opting your posts out of the main algorithmic recommendation engine does not fit well with the Big Tech approach to algorithms on platforms. But it is clearly the way forward for open social protocols, and its time for Bluesky to catch up to that and let people be able to opt out of having their posts reach the Discover feed.

Tangled talks about the vision behind the project:

plcbundle is a new way to sync the entire PLC directory in a verifyable manner.

That's all for this week, thanks for reading! You can subscribe right here on Leaflet to keep up to date, and if you want to read more related to the fediverse and the open social web, you can find more of my writing on connectedplaces.online

And if you want to know more about the challenges with coordination for building better moderation systems, check out yesterday's article:

A short report with some thoughts on two articles about Bluesky and atproto this week, as well as some additional links that stood out to me.

Roomy dev Erlend Sogge Heggen writes about how he would use 200M USD to fund the developments of atproto. He gets the 200M number by comparing it what the municipality of Oslo has spend on a the renovation of a single swimming pool, also 200M USD.

It points to the increasingly strange situation that we're in. It is getting close to mainstream opinion now in Europe to say that the EU needs its own social media infra. This is clearly illustrated by the recent speech by President Macron:

Greens MEP Alexandra Geese:

The thing is, once you agree that the Europeans have been naive in entrusting our digital spaces to a few American Big Tech companies and TikTok, then there are basically two options:

• you build a European version of Meta
• you work with open protocols

Nobody is even remotely interested in option 1 (for good reason), so that really leaves only the option of working with open protocols. And once you accept that option, you really only have two current realistic options: ActivityPub and atproto. Considering what Geese points out above, this is a short term need, which realistically rules out the option to start from scratch on an open protocol.

Both of these protocols are operating on a shoestring budget, ActivityPub even more so than atproto.

So we end up in the situation where the EU is trending towards a dominant opinion that this is a (supra)nationstate problem, but the most realistic solutions operate on fractions of nationstate budgets.

My main point here is that we're currently far from something that looks like an equilibrium state from social networks.The moment funding for open protocols approaches even a small fraction of what the political rhetoric suggests this problem deserves, the entire ecosystem of the open social web looks completely different.

The second thought regarding nation-state funding for open social protocols is that might trigger some funny and strange competitive dynamics. One major driver of why nations talk about sovereign social media is the loss of control that resulted in depending on foreign tech companies. In the hypothetical situation that Norway decided to fund their own social networking infrastructure (whether that would be a NorSky, something on AP, or something else) I'm not actually sure that other states would look and that and think: "sounds cool lets all join NorSky now".

It seems much more likely to me that politicians would think "oohh shiny new thing that lets me say the sovereignty buzzwords, gimme that too". Case in point: Last week I gave a presentation to a bunch of media execs about bluesky, atproto and eurosky. The first question I got was "does this mean we can instead build a DutchSky?"

I have no idea what such a competitive/collaborative dynamic would look like, except for the obvious point that this would drastically change the open social web ecosystem.

Finally: these musings assume that EU politicians will translate their rhetoric into action, which is historically far from a safe bet to make. I would like this to happen, but I'm far from certain that it will actually happen.

Erin Kissane has written an excellent article on arrival of the White House on Bluesky last week, that is highly worth reading.

The core challenge is that people want and need digital social places that have 'high-context moderation', and that Bluesky is unlikely to going to provide such forms of moderation. Digital places with high-context moderation exist, both on atproto and even more so on ActivityPub. These spaces are great and do deserve all the support they can get, but my worry is that these spaces might not be able to grow at a speed that the descent into various fascist movements around the world necessasitate.

For more on the White House joining Bluesky, see also my post from earlier this week:

And a few links that grabbed my attention this week:

There is a lot of discourse about moderation (which I contribute to I guess), but there is a lot of work also being done on how labeling systems work in practice. The Skywatch labeler update points to how the labeler can facilitate abuse detection in realtime, both with new spam systems abuse as well as for astroturf campaigns.

A recap of Tangled, and its kinda crazy that the git collaboration platform is only 6 months old.

Bluesky is apparently great for sports engagement. It seems to me there are a ton of opportunities to build dedicated apps for sports-based social networking platforms on atproto, but so far I've seen surprisingly few efforts being made in that direction.

atproto-powered streaming software Streamplace is at the forefront of what's possible to build with atproto. It is now possible to self-host Streamplace. Another cool new feature is the embedded metadata, allowing streamers to set how the stream can be distributed, content rights and content warnings.

As video moves more and more in the direction of re-using other videos (whether it is stitches on TikTok, or co-streams for esports or reaction videos on YouTube), having a system that determines distribution and content rights is highly important.

Free Our Feeds with a quick update:

That's all for this week, thanks for reading!

The Trump administration functionally rules via decrees on social media, and extensively relies on social media for validation of their policies and proposals. The actions of the regime are done to a significant extent precisely because of how it will be perceived on social media. Like other fascist regimes, terror is a core part of how the regime projects power. Violence itself only becomes terror when it can be packaged into a message to distribute to the population. US government agencies heavily rely on X for the distribution of their messages of terror, from creating ASMR videos of deportation to sharing the videos of missile strikes on fishing boats in Venezuela. As X radicalises more and more into a place for regime supporters, the platform starts to lose effectiveness for projecting violence. It becomes more important for the administration to find the (digital) places where their political opponents gather, to spread their message of hate.

In this context, it is little surprise that dozens of US government agencies all joined Bluesky at the same time, including the White House, the Department of State, and more. When over a month ago, pundits like Noah Smith and Nate Silver started writing about Bluesky and 'Blueskyism', they shaped the opinion of what Bluesky is. To them, Bluesky is a place for leftists scolds. I wrote about their articles:

We're now seeing the direct impact of this: Bluesky is seen as 'left' space, a place where the political opponents of the Trump administration gather. And that is the perfect destination for a regime that rules via social media trolling.

By and large, people on Bluesky understands the game that the administration is playing. Their goal for joining Bluesky is to spread the message of terror and fear, as illustrated by the Department of State posting that "this is a great place to research visa revocations 👀". The best way to counter this strategy is to starve the accounts of oxygen, and limit their attention and reach as much as possible. The calls on the network where loud and wide to 'block and move on', to prevent the US regime accounts from getting the attention they crave. The White House is already one of the most blocked accounts on the network after a few days (behind vice president JD Vance, who joined a few months earlier), and blocking and starving trolls from attention works to a certain extent, but there are problems that go beyond that.

By joining Bluesky, the government agencies are actively contributing to the perception that Bluesky is not a 'left' place. The logic is fairly simple, it is harder to be perceived as a 'left' place when a fascist government is actively posting on the platform.

For queer people, the Black community and other communities who shaped Bluesky's culture, the platform's identity isn't abstract concept to write Pundit Thought Pieces about. It's the difference between a space that's for them and their community, and one where they have to share a digital place with a government that wants many of them dead or deported. They joined Bluesky as a place they could joyfully be themselves. Bluesky as a politically neutral place isn't what they're interested in. This goes beyond the direct impact on safety that the presence of various government agencies creates (although that's important too). The core point is not wanting to share the same digital place with this administration. There are clear parallels with how people on Bluesky are upset with the decision not to ban Jesse Singal: the argument is just as much about not wanting to share a digital place as it is about his specific wrongdoings on the platform.

At the same time, Bluesky presenting itself as a "Twitter-alternative" works much better if a wide variety of political governments are present on the platform. Bluesky's current user base is indeed largely people on the left of the political spectrum. The presence of the White House on Bluesky provides a good counter for the company to point to, when potential groups are hesitant to join the platform because they feel that it is a niche platform for a specific political group only. It fits in with the larger viewpoint of Bluesky PBC. In an interview with Wired, interviewer Kate Knibbs asked Jay Graber point-blank: "Would you welcome President Trump?" Graber's answer is clear:

The presence of the Trump administration on Bluesky creates a precarious situation for Bluesky however: The risk of political action taken against Bluesky is real, and can take various forms. There is the risk of a crackdown on places that are associated with Democrats and the left, a risk that Trump might lash out against anything he decides to get mad about that day, and a head of the FCC that seems excited to be crack down on what he calls the "censorship cartel", where he has made clear his intention to use whatever regulatory tools available to pressure platforms he believes are suppressing conservative viewpoints.

The administration's accounts are clearly daring Bluesky to take action against them, likely hoping to play the victim in whatever drama follows. It feels like only a matter of time before an account breaks the rules, and when that happens, Bluesky faces an impossible choice: take action and risk political retaliation, or do nothing and face the backlash from a user base that already feels insufficiently protected. Bluesky's relationship with its users over past moderation decisions has been tumultuous, to put it mildly. If government accounts get a pass when they break the rules, that tension will only intensify.

Bluesky board member Mike Masnick has written extensively how content moderation at scale is impossible to do well, calling it the Masnick's Impossibility Theorem. Masnick's argument focuses on moderation being an inherently subjective practice, and at social media scale, that subjectivity makes doing it well impossible. The presence of an authoritarian regime on the platform with the purpose of trolling only makes the Impossiblity Theorem all the more impossible.

Programming note: I've barely written about atproto in the last month: some issues with my shoulder severely limited the amount of hours I can spend at a computer each day, and I've had to prioritise some paid work in the limited time. Things are finally going better now, and I'm slowly catching up on the backlog. I have been keeping fairly close track of all the news shared in the last months anyway. So for the next few updates, I'll be going over the backlog and share some of the news and links that stands out to me. See it as a way to find out what news you've missed over the last month or so.

This also means I won't be doing a deep dive analysis on the 'waffles' situation and the factors that lead to this all. This kinda bothers me since I think its probably the most important story about atproto over the last month. However, I'm still restricted in time, and have to prioritise paid work. So for now it'll be some shorter Leaflet notes while I get up to speed again. Deep-dive analysis will definitely be resumed again, I do very much enjoy writing those. For now, let's dive in!

The For You feed is a custom algorithmic feed that has gotten popular recently. The algorithm looks at posts you like, and finds posts that are also liked by people who like the same posts as yours. It is a surprisingly effective algorithm, and I see a lot of people recommend the feed, over Bluesky's Discover feed as well. I can also definitely recommend the feed, one of my favourite ways to scroll Bluesky. The For You feed also shows how powerful and cool it is to have algorithmic feeds that are fully decypherable and open: The creator spacecowboy has an entire For You feed playground website that allows you to tweak with the settings. Another cool aspect: see what other people's algorithmic feed is like.

Seeing A/B tests on this feed and how demoting popular posts also impacts usage is interesting:

Adoption of new social networks often happen on a community-by-community base, and less on an individual basis. (this is why adoption of both fedi and bluesky happens largely happened in waves). In that context I find it interesting to see what fandom communities are up to:

Leaflet wrote about feedback they've gotten how the publishing platform can be more suited for fandom communities

haetae has been building towards a fandom archive project on atproto, with a blog explaining the process here:

Ms Boba is regularly streaming her atproto dev work for building a guestbook on atproto, and she's also written an extensive guide on what an AppView is and what it does:

teal.fm is a music scrobblr (tracks your music listening, similar to last.fm) on atproto that is currently being developed. Developer Matt wrote about the current state of development:

Teal.fm consists of multiple parts:

• the frontend app
• the appview/api
• the music tracker, called Piper

What's interesting about teal.fm is that the music tracker and accompanying lexicon already work. This means you can already use teal.fm, even though it is technically still in development. In fact, I've been doing that since the last week or so, and I'm impressed by how well this all already works. (shoutout to Bailey for letting me use his Piper instance).

So if you want start tracking your music listening habits on atproto already, you can:

install piper, and see your music stats here:

SkyBeMoreBlue is a platform for introducing people to each other on ATProto. The idea is that you give an introduction and description of other accounts in your network. This introduction is then shown to other people in your network as a recommendation to follow. I think it is an interesting contrast with how platforms like Bluesky (and Mastodon) handle Starter Packs, which are focused on getting people to follow large number of accounts at once. SkyBeMoreBlue is has some of the same final goals (help people find more interesting accounts to follow), but with a very different approach to get there. It is made by the same people who build the ATProto audio spaces app Bluecast, and seems to be mainly in use by the Japanese side of the ATProto community. I'm, uhhhh, unsure on the naming, but I do think it's cool to see more experimentation and new ideas for building healthy connections between people on social networks.

(a short bonus post because I need an example of a Leaflet post with an image embedding as a heading, and figured I might as well write something)

Sebastian Vogelsang, whose the developer of Flashes and working on the commons moderation service for Eurosky posted a very interesting question this week which got me thinking.

The idea behind the Eurosky commons moderation service is as follows:

• Developers can register their apps with Eurosky.
• Eurosky acts as the endpoint where participating apps can send content reports.
• Eurosky then provides moderation decisions that can be shared across apps using the same lexicon.

So far, so good. But the problem is that not all apps are the same. Vogelsang makes a distinction between consumer apps and producer apps. The distinction is simple:

• Producer apps are predominantly used to create content.
• Consumer apps are predominantly used to consume content.

For Vogelsang and the Eurosky team the question is: how should the costs of running the commons moderation layer be split across all participating apps, and should there be a distinction between the different types of apps?

I think thats a good question, and I don't yet have a clear answer to that.

But for me, this distinction between producer and consumer apps has been stuck in my mind for the last few days.

The current social networking era of the Big Tech platforms sees everyone use the same app to access a platform. Musk uses the same X app when he posts 150 times per day as any rando who just wants to scroll their feed. Same with TikTok, Insta, etc, the power users who are responsible for creating the content use the same app to access the network as the lurker who follows 3 people and scrolls for 5 minutes a week.

Social media is well known to follow a power law curve regarding how content on the network is created and consumed. As a rule of thumb, most networks follow a 1-9-90 split, with 1% of people creating original content, 9% occasionally contributing and 90% lurking.

It is one of the quirks of how the current Big Tech platforms work that these very different groups of people all use the same app to access the platform. It is clear that their experience, use case and needs are extremely different. For example, notification management is an absolute must if you are a creator with a big account (and even then it's often a mess), but for a lurker this matters much less. A lurker prioritises good feeds that gives them the content they are looking for. Also important for some creators, but other big accounts often don't have the time to read through feeds anyway so they only need the most relevant highlights.

The cool part of open social networks is that we get to move away from the paradigm of building the same app for very different types of people to access a network. Instead we can now build different apps that respect the needs of different groups of people.

However, what interests me about Vogelsang's post is that this also creates new challenges as a result. Should producer apps and consumer apps be treated different, when it comes to shared moderation costs, for example? Currently ATProto lexicon records do not indicate with which app a post was made. Apps could start adding a custom field to posts to indicate the source. Is that desirable? I don't know!

Only thing I do know is that you give people the freedom to build their own tools/apps/software, they might actually do that, and over time a network like atproto ends up much more diverse, stranger and unique as a result.

I was chatting with Sebastian Vogelsang (who's behind Flashes, Skeets and Eurosky), and he mentioned that he had recently given a presentation about atproto for media people. I asked him what his method was for explaining atproto. He said that compared atproto to wire services, and that this was a comparison that landed really well with the audience.

I've never heard of explaining atproto by comparing it to wire services before, and I think its actually a pretty smart way of explaining atproto and its value for media. It's probably an explanation best suited for people who are familiar with media, but I wanted to share this with, since I think more people might use this comparison in the right context (and shoutout to Sebastian for this!)

The idea is pretty simple: wire services, like Associated Press, Reuters and Agence France-Presse, work by creating news articles by their own journalists, which they then sell in bulk to other news organisations. A wire services aggregates news reporting, they have reporters around the world so they can cover all the news. These news articles get aggregated by the wire services into a continuous stream of news. Others news organisations usually do not have staff around the globe, but do want to fill their newspaper/tvshow/website/whatever with all global news. As a result, many news organisations buy a subscription from one of the wire services, and they get access to a continuous stream of news articles they can use.

Atproto has some quite similar dynamics, but instead of news articles by Reuters journalists, it is now user generated content (posts/videos/blogs/etc). The aggregation and relay is done by a relay, and anyone can plug into this stream of user generated content to build any social media platform they want. Obvious differences is that connecting to a relay is free, running a relay is cheap, and you cannot stop people from posting even if you wanted to.

One of the challenges with the open social web is in story telling: how do you explain all this complicated technical stuff to people? Often analogues to other internet protocol get made: activitypub is like email, atproto is like websites, etc etc. Comparing atproto to wire services takes the analogy away from the internet, to a world that a specific target audience knows well, in a comparison that holds up reasonably well on a high level.

Bonus advantage is that this analogy explains what the best place for economic value creation is for media people. Most people in media have (understandably) no interest in being a wire service: you want to buy from a wire service, place the article in your own branding, and provide enough value that people will use your news platform over any of the other news platforms.

This is sorta similar with atproto: all the low level infrastructure stuff is just that: infrastructure. There is no real money in running relays (or any other supportive infrastructure). The value is in building the app that people will use to view the content on the entire network. Currently Bluesky PBC has captured virtually everyone. But there is space for a whole lot more competitors to acquire some of this market share as well