marc0s
boosted
Protocols for Publishers
and 2 others
boosted
Fediverse Report – #148 – On Protocol Governance
The W3C has announced a new Social Web Working Group, starting January 15, 2026, to maintain and update the ActivityPub protoocol. The group will be chaired by Darius Kazemi, who created Hometown Mastodon fork and the Fediverse Schema Observatory. The aim of the Working Group is to release updates to ActivityPub, and its specifications such as Activity Streams and Activity Vocabulary. Most of the work on the protocol is scheduled to be done by Q3 2026, with the Working Group running until January 2028.
To understand why this matters, some context on how the W3C operates is necessary. The standards organisation distinguishes between Community Groups and Working Groups. Community Groups are open to anyone and serve as incubation spaces for ideas. Since 2018, the Social Web Incubator Community Group (SocialCG) has been the steward of ActivityPub. While Community Groups serve as a grass-roots place, they are very limited in publishing official documentations and formal updates to protocol standards. Working Groups, by contrast, are the bodies that can actually publish official W3C Recommendations, meaning formal standards. Participation in Working Groups is restricted to representatives of W3C member organisations (which pay membership fees on a sliding scale) and invited experts approved by the chair.
ActivityPub became a W3C Recommendation in January 2018, and the protocol work was done by a Working Group. After ActivityPub became an official W3C specification, this Working Group disbanded, and the SocialCG was formed. Since then, the specification has not been formally updated, despite significant implementation experience revealing ambiguities and missing features. The SocialCG has maintained an errata document and developed extensions through the Fediverse Enhancement Proposal process, but these carry no official W3C status. The new Working Group changes this by providing a formal path to update the core specification.
Fediverse advocates regularly point to ActivityPub being a W3C standard as a mark of legitimacy, but for the past seven years the organisational structure that created the protocol has also prevented necessary updates to it. The W3C has done a massive service to the community by holding space for the creation of the protocol in 2018. But since then, the same organisational structure that allowed the protocol to be created also slowed down necessary further work on ActivityPub. This shows up both in errata documents not becoming part of the formal documentation, but also larger work on the Client-To-Server part of the ActivityPub needing more work in order to be suitable for larger adoption.
The new Working Group for ActivityPub changes this situation, and there now a formal path to update the core specification, incorporate errata, and potentially advance new work like LOLA (Live Online Account Portability) to official status. LOLA is a proposal for server-to-server account migration that would allow users to move between ActivityPub servers while retaining both their posts and their social graph. Unlike the current Move activity that only migrates followers, LOLA would enable full content portability. The charter includes LOLA as a potential deliverable, which means it could become an official W3C specification rather than remaining a community proposal.
The are some major complicating factor however, and that is about who actually gets to make decisions. The Community Group lacked power to make official chances to ActivityPub, but it did provide an open place for anyone to participate. In contrast, the Working Group requires participants to either be a paid W3C member or to be an Invited Expert. There are only two organisations that are active in the fediverse that are a paid member of the W3C: Meta and the Social Web Foundation. With the Social Web Foundation also receiving funding from Meta, the company that built Threads now has more institutional standing in ActivityPub governance than any of the organisations actually building open fediverse software. Mastodon gGmbH, Framasoft, and others are not W3C members and cannot participate in the Working Group unless they are invited.
This is by all accounts an extremely funny outcome for a network that aims to be independent of Big Tech’s power.
A few nuances to this. It is unclear if Meta will actually participate in the Working Group, and considering they recently put their Threads<>fediverse integration on maintenance mode, there is a good change that Meta has no interest in actually participating. The Working Group also has yet to communicate who the Invited Experts will be. It could theoretically be that Meta is absent from the group, while Mastodon and Framasoft employees are invited to be part of the Working Group.
Another challenge for the W3C Working Group is that there has long been a disconnect between ActivityPub protocol development and the people creating ActivityPub software. While the above makes it sound that fediverse developers are excluded from the protocol development process, the practical reality is also that the developers of the main fediverse platforms like Mastodon, PeerTube and Lemmy have shown very little interest in engaging with the process when it was openly accessible under the SocialCG. This is illustrated by the meeting last year in which the SocialCG voted to charter a Working Group, where no member of any of the fediverse platform developers was actually present. There has long been a disconnect between the people who develop ActivityPub software and the people who maintain the ActivityPub protocol, with only a few notable exceptions.
This matters because of how W3C standards work. The charter’s success criteria states that updating the Recommendation requires “at least two independent implementations of every feature defined in the specification, where interoperability can be verified by passing open test suites.” The Working Group can propose whatever changes it wants, but for those proposals to become part of the official ActivityPub standard, they need to be implemented in actual software.
LOLA, the proposal to improve account portability, is a clear example of this challenge. Already in fall 2024, Lisa Dusseault, the author of the proposal, said that the specification was ready for developers to start testing implementations. The main bottleneck since then has been getting organisations like Mastodon interested in actually building it. The protocol work is largely done, but what remains is the persuasion and coordination to get implementers interested in using it.
The importance of protocol maintenance and further development of ActivityPub points towards responsibilities for software implementors, especially Mastodon as the dominant ActivityPub implementation. Mastodon’s choices become de facto standards whether or not the project engages with formal standardisation processes. The most clear example is how the Mastodon API has effectively taken over from ActivityPub’s Client-to-Server as the dominant protocol that other softwares have to implement. That position comes with obligations, and when Mastodon doesn’t participate in protocol governance, it creates a vacuum where the largest implementer (in this case also Mastodon) is able to set standards for the rest of the network, but without the governance or formal documentation. When protocol development and maintenance in the Working Group happens disconnected from the largest implementations, the specifications that may not reflect implementation realities.
What this situation reveals is that using network architecture to solve issues of power distribution simply shifts bottlenecks rather than eliminating them. A decentralised protocol does not automatically produce decentralised governance, it also moves power to different, less visible places. The W3C membership structure concentrates formal power in ways that don’t reflect the fediverse’s values, while the implementers who could counterbalance that power have largely opted out of the process. The new Working Group creates an opportunity to address both problems, but who gets to shape the specifications of ActivityPub depends on both who is allowed to participate, as well as who is willing show up and do the work.
https://connectedplaces.online/reports/fediverse-report-148-on-protocol-governance/
Fediverse Report – #148 – On Protocol Governance
The W3C has announced a new Social Web Working Group, starting January 15, 2026, to maintain and update the ActivityPub protoocol. The group will be chaired by Darius Kazemi, who created Hometown Mastodon fork and the Fediverse Schema Observatory. The aim of the Working Group is to release updates to ActivityPub, and its specifications such as Activity Streams and Activity Vocabulary. Most of the work on the protocol is scheduled to be done by Q3 2026, with the Working Group running until January 2028.
To understand why this matters, some context on how the W3C operates is necessary. The standards organisation distinguishes between Community Groups and Working Groups. Community Groups are open to anyone and serve as incubation spaces for ideas. Since 2018, the Social Web Incubator Community Group (SocialCG) has been the steward of ActivityPub. While Community Groups serve as a grass-roots place, they are very limited in publishing official documentations and formal updates to protocol standards. Working Groups, by contrast, are the bodies that can actually publish official W3C Recommendations, meaning formal standards. Participation in Working Groups is restricted to representatives of W3C member organisations (which pay membership fees on a sliding scale) and invited experts approved by the chair.
ActivityPub became a W3C Recommendation in January 2018, and the protocol work was done by a Working Group. After ActivityPub became an official W3C specification, this Working Group disbanded, and the SocialCG was formed. Since then, the specification has not been formally updated, despite significant implementation experience revealing ambiguities and missing features. The SocialCG has maintained an errata document and developed extensions through the Fediverse Enhancement Proposal process, but these carry no official W3C status. The new Working Group changes this by providing a formal path to update the core specification.
Fediverse advocates regularly point to ActivityPub being a W3C standard as a mark of legitimacy, but for the past seven years the organisational structure that created the protocol has also prevented necessary updates to it. The W3C has done a massive service to the community by holding space for the creation of the protocol in 2018. But since then, the same organisational structure that allowed the protocol to be created also slowed down necessary further work on ActivityPub. This shows up both in errata documents not becoming part of the formal documentation, but also larger work on the Client-To-Server part of the ActivityPub needing more work in order to be suitable for larger adoption.
The new Working Group for ActivityPub changes this situation, and there now a formal path to update the core specification, incorporate errata, and potentially advance new work like LOLA (Live Online Account Portability) to official status. LOLA is a proposal for server-to-server account migration that would allow users to move between ActivityPub servers while retaining both their posts and their social graph. Unlike the current Move activity that only migrates followers, LOLA would enable full content portability. The charter includes LOLA as a potential deliverable, which means it could become an official W3C specification rather than remaining a community proposal.
The are some major complicating factor however, and that is about who actually gets to make decisions. The Community Group lacked power to make official chances to ActivityPub, but it did provide an open place for anyone to participate. In contrast, the Working Group requires participants to either be a paid W3C member or to be an Invited Expert. There are only two organisations that are active in the fediverse that are a paid member of the W3C: Meta and the Social Web Foundation. With the Social Web Foundation also receiving funding from Meta, the company that built Threads now has more institutional standing in ActivityPub governance than any of the organisations actually building open fediverse software. Mastodon gGmbH, Framasoft, and others are not W3C members and cannot participate in the Working Group unless they are invited.
This is by all accounts an extremely funny outcome for a network that aims to be independent of Big Tech’s power.
A few nuances to this. It is unclear if Meta will actually participate in the Working Group, and considering they recently put their Threads<>fediverse integration on maintenance mode, there is a good change that Meta has no interest in actually participating. The Working Group also has yet to communicate who the Invited Experts will be. It could theoretically be that Meta is absent from the group, while Mastodon and Framasoft employees are invited to be part of the Working Group.
Another challenge for the W3C Working Group is that there has long been a disconnect between ActivityPub protocol development and the people creating ActivityPub software. While the above makes it sound that fediverse developers are excluded from the protocol development process, the practical reality is also that the developers of the main fediverse platforms like Mastodon, PeerTube and Lemmy have shown very little interest in engaging with the process when it was openly accessible under the SocialCG. This is illustrated by the meeting last year in which the SocialCG voted to charter a Working Group, where no member of any of the fediverse platform developers was actually present. There has long been a disconnect between the people who develop ActivityPub software and the people who maintain the ActivityPub protocol, with only a few notable exceptions.
This matters because of how W3C standards work. The charter’s success criteria states that updating the Recommendation requires “at least two independent implementations of every feature defined in the specification, where interoperability can be verified by passing open test suites.” The Working Group can propose whatever changes it wants, but for those proposals to become part of the official ActivityPub standard, they need to be implemented in actual software.
LOLA, the proposal to improve account portability, is a clear example of this challenge. Already in fall 2024, Lisa Dusseault, the author of the proposal, said that the specification was ready for developers to start testing implementations. The main bottleneck since then has been getting organisations like Mastodon interested in actually building it. The protocol work is largely done, but what remains is the persuasion and coordination to get implementers interested in using it.
The importance of protocol maintenance and further development of ActivityPub points towards responsibilities for software implementors, especially Mastodon as the dominant ActivityPub implementation. Mastodon’s choices become de facto standards whether or not the project engages with formal standardisation processes. The most clear example is how the Mastodon API has effectively taken over from ActivityPub’s Client-to-Server as the dominant protocol that other softwares have to implement. That position comes with obligations, and when Mastodon doesn’t participate in protocol governance, it creates a vacuum where the largest implementer (in this case also Mastodon) is able to set standards for the rest of the network, but without the governance or formal documentation. When protocol development and maintenance in the Working Group happens disconnected from the largest implementations, the specifications that may not reflect implementation realities.
What this situation reveals is that using network architecture to solve issues of power distribution simply shifts bottlenecks rather than eliminating them. A decentralised protocol does not automatically produce decentralised governance, it also moves power to different, less visible places. The W3C membership structure concentrates formal power in ways that don’t reflect the fediverse’s values, while the implementers who could counterbalance that power have largely opted out of the process. The new Working Group creates an opportunity to address both problems, but who gets to shape the specifications of ActivityPub depends on both who is allowed to participate, as well as who is willing show up and do the work.
https://connectedplaces.online/reports/fediverse-report-148-on-protocol-governance/
alcinnz
boosted
Tim Chambers
and 2 others
boosted
A Power Problem, Not a Platform Problem
Happy 2026 everyone. The world’s richest man runs a subscription service to remove the clothing from photographs of children, and I don’t know how to write about it. It’s been just over a week and the global order has already drastically changed, in ways that affect everything, including how open social protocols understand themselves.
Remember 2023 and 2024? When every time Musk did something bad, people got excited because that would lead to Elon Musk Events, with signup waves to Mastodon and Bluesky? And when ‘bad’ was understood be frivolous things like DMs not working? Now in 2026, when ‘bad’ means generating CSAM and NCII on-demand at industrial scale, it’s crickets, and there is no initiative at all to leave the platform anymore. Society is going through the motions of vocal condemnation, pointing at agencies who should enforce something, but then not enforcing anything. It is clear now that actually leaving X is a step too far and unthinkable.
On January 3rd, US forces bombed Caracas, captured Maduro and his wife, and killed at least 80 people. Trump posted photos from a makeshift situation room at Mar-a-Lago showing the raid being monitored in real time. In the situation room photos, visible on the big screen behind Defense Secretary Hegseth, was an X feed showing search results for ‘Venezuela’. Another photo showed the OSINTdefender account on the screen. Few things illustrate the current role of X in our society as well as the heads of the most powerful military in the world, monitoring the X account of OSINTdefender while the CIA director sat next to them in the room.
Grok’s latest update allows people to generate sexualized images of women and children on demand, at industrial scale, generating multiple such images per second. Everyone is aware that this is happening, and continues to happen, as nobody is willing to stop it. Musk’s attitude is to use the crying-laughing emoji on complaints is an indication of how serious he takes the issue. Politicians have universally condemned it in words, calling it unacceptable. What is so maddening, however, is that their actions say otherwise. Governments all around the world are very clearly afraid of picking a fight with Elon Musk and a belligerent US regime.
This fear by politicians is further accentuated by the raid on Venezuela and Maduro’s kidnapping, which shows that the US is now a rogue state that does not care in the slightest about adhering to any forms of law. With realistic further threats being made to annex Greenland by the US, it is in fact understandable why politicians are afraid to take actions against the richest man in the world. When the UK said it might potentially think about enforcing its own laws against X, a US congresswoman threatened to sanction not just Starmer but ‘Britain as a whole,’ calling enforcement ‘a political war against Elon Musk.’ X is both protected by US state power, as well as being a source of US state power.
This widespread societal resignation of ‘guess our government communications now happens on a deepfake porn site’ is maddening, but also points to a deeper issue. We’re used to describing X as a platform, and analyse X accordingly. The photos of X being on the big screen while coordinating the Maduro raid is an indication that X acts as the infrastructure for power, the glue that connects the neo-royalty.
But the refusal of governments around the world to do anything about the CSAM and NCII generation machine, and how other countries get bombed and their leaders kidnapped because it creates content for X, shows that X is about power, and less about a platform.
When Elon Musk took over Twitter late 2022, alternative open social networks gained prominence, and platforms like Mastodon and Bluesky presented themself as alternatives to Twitter’s function as the digital public square. Both networks understood that there was an issue with the idea of having a single platform as a public square, and thus did not copy Twitter fully, but put their own spin on it.
The fediverse and Mastodon focused on there not being a single public square at all, but instead many digital places and communities that interconnect with each other. For Bluesky the solutions were aimed at giving users more control, with composable moderation and custom algorithmic feeds.
What this framing missed however is that Twitter also stopped being the digital public town square, in the years since Musk’s acquisition. Instead, it became the internal coordination space for a political faction that now controls state power. People still treat X as if it is still 2015, pretending it is the town square and using it for everything from talking about sports news to keeping up to date on pop culture. This lends validation and legitimation to X’s new role, facilitating the power of the neo-royalty.
The neo-royalty is the small network of political, capital, and tech elites centered on Trump, and X is their coordination infrastructure. Because they are rapidly gaining power around the world, and are in full control of the world’s military hegemon, you cannot separate yourself from this power infrastructure that X has become. Leaving X does not insulate or protect yourself from the warping effects it has on global power. This goes for both the large geopolitical aspects (see Greenland), and the local impacts, as the Somali families in Minnesota who lost their childcare funding because of a viral X video weren’t on X.
The implicit theory behind the alternative social web was that platform quality would determine outcomes. Build something that’s better, and in combination with the incumbent getting worse, this would lead to such difference in quality, user experience and safety that at some point people would switch from X to alternatives like Mastodon or Bluesky. This theory held up for a while in 2023 and 2024. In 2025 it started to falter, as Musk aligned himself with Trump, the signup waves to the alternative platforms effectively stopped. In early 2026, this theory is now really over, because X has fundamentally changed. Mastodon and Bluesky are not in competition anymore with the platform X, because X has changed. It changed from being a platform to the power structure for the neo-royalty, with the public square shambling along as a zombie, animated by everyone who still treats X like it’s 2015.
You cannot out-compete ‘where the ruling faction radicalizes and coordinates’ by having better moderation policies or algorithmic choice. X is not a platform problem anymore, it is a power problem, and building a different platform does not solve the power problem.
Other countries will need to leave the platform to untangle themselves from this dependency, and reduce its legitimacy. But the functioning of the neo-royalty is such that other governments taking actions against X will be taken as an offensive action by the US regime, that will likely trigger extensive retaliation. No country seems to be willing to be the first one to move to take action and thus take the brunt of the counter-offense of the regime.
We’re now at a strange stand-off, where it is extremely clear it is unacceptable what X is doing, and governments make a lot of noise about how upset they are, without daring to pull the trigger on taking action. Everyone is waiting on everyone else to take the first move.
This leads to three possible outcomes:
- no government dares to take action, and they keep to calling things “completely unacceptable” while accepting the actual situation. Things stay as they currently are, and the world keeps sliding into a more dangerous and harmful place.
- One government takes action against X, and the US regime retaliates so hard that no other government will dare to do meaningful enforcement against the massive harms created by X.
- One government takes enforcement action against X, creating a permission structure for other governments to also take actions.
All three options have a meaningful impact on the open social web. For the first two outcomes, it further cements X as the place of power for the neo-royalty, further cementing its dominance in the political sphere. This position of power is also what prevents the alternative open networks to become a place of political power in it’s own right. The third option, of mass enforcement, is what creates an opening for open social networks to not just be an alternative, but to be a source of political power as well.
I do not know what the outcome will be, and with how rapidly the world has been changing I do not know which option is likely either. It’s easy to be highly cynical, and that point of view has been extensively validated over the years, but I do choose to hold to hope that we can build a better, more ethical, social internet out of the toxic waste ground of the current state of the internet.
https://connectedplaces.online/reports/a-power-problem-not-a-platform-problem/
𝓼𝓮𝓻𝓪𝓹𝓪𝓽𝓱【ツ】☮(📍🇺🇸)
boosted
Chewie
boosted
The Digital Services Act and Theories of Power
Right in the heart of the European quarters in Brussels is the Parlamentarium, the visitors centre for the European Parliament. When I was in Brussels this summer, and I had an entire Sunday to myself to wander around the city, I found myself in front of the building, and decided to enter the Parlamentarium on a whim. I was curious how the European Union understands itself, and the visitor’s centre for the Parliament itself seemed like a good place to learn a bit more about what stories the EU deems important to tell about itself. One thing that was immediately clear is that the EU thinks it is very important to make the stories about the EU accessible to everyone. There are clear and proud advertisements that the Parliamentarium can be experienced in any of the EU’s 24 languages, and on the website the EU can only wait until the second sentence to tell you that the Parlamentarium “can be experienced in any of the European Union’s 24 official languages.”
It seems like the EU thinks of themselves as an institute that cares deeply about making their information available to all of their citizens. Think about it, everything is translated to 24 languages!
When you enter the exposition, it starts on the ground floor with a history of the European Union, from the world wars, to the creation of the European Coal and Steel Community, to the growth of the Union after the fall of the Wall, the creation of the euro, up until the present. The diversity of language and communication is in full display in this exhibition, with walls covered with photos of historical moments, accompanied by explanatory texts in continuously changing languages.
After the short history, you move down a floor, where an extensive set of displays explains how the current EU system works, what the different parties are, and who the members of parliament (MEP) of your country are. Another set of displays and computers shows you all the MEPs, and invites you to click on the MEPs of your country, to learn about their background and what they stand for. Helpfully, this interface also shows you what your MEP has been saying about politics. If you go into this menu, you get presented with a big button labeled X, that opens up the MEP’s feed on X and shows their most recent posts on X. Besides the option to email the MEP, no other option to know more about what each MEP is saying is presented.
The EU is clear in their communications: If you speak any of the 23 official languages of the European Union that is not English, you’ll have to travel to Brussels to get information in your language. Otherwise, it all happens on X, the everything app, including the official and non-official communications by all of Europe’s elected officials.
Its all fine
While the EU is using their visitors centre to tell stories how utterly dependent the government is on X for their communications, the European Commission (EC) is getting more annoyed with the platform. After 2 years of deliberation, the European Commission issued its first DSA non-compliance fine, with a penalty of €120 million for X. While people have many types of misgivings with X, mainly focused on the types of hate speech that are flourishing on the platform, the EC focused on one specific part of the DSA: its transparency obligations. The problems that the EC found are specific to that context. X deceived users by selling “verified” badges to anyone willing to pay, making it impossible to distinguish authentic accounts from potential scammers. The platform’s advertising repository was deliberately crippled with design features and access barriers that made meaningful research impossible. And X blocked researchers from accessing public data, prohibiting scraping while charging prohibitive fees for API access. The maximum possible fine under the DSA is 6% of global turnover, and with an estimated revenue for X of 2.5B USD for 2024 it means that the EC chose a penalty fairly close to the maximum of 150M USD.
The same day, the Commission announced it had accepted binding commitments from TikTok on advertising transparency, with the platform agreeing to provide complete ad transparency. This means that TikTok’s advertisement repository, which provides insight in the ads that are running on the platform, will provide better insight to researchers, with some fairly technical improvements: updating the repository within 24h, showing the full ad with urls, targeting criteria and demographic data, as well as additional search tools.
That the EC made these both announcements on the same day is a deliberate strategy, illustrated by how EC spokesperson Thomas Regnier explicitly frames the fine for X in contrast with the compliance of TikTok in the announcement post. The EC is clearly very worried about accusations of censorship, and via the juxtapositioning of TikTok and X they hoped to show what the two possible outcomes are of the DSA: cooperate like TikTok, pledge some fairly technical updates and achieve a settlement, or go for confrontation and get close to the maximum possible fine.
It’s worth noticing what the Commission did not fine here, with their investigations that opened over two years ago still ongoing regarding the “dissemination of illegal content” and the “effectiveness of measures taken to combat information manipulation on the platform”. The EC says that on this subject “the investigation continues.” It seems that the first enforcement addressing procedural failures serves as a testing bed to figure out how platforms like X, but also the US government, will respond to an enforcement action, before moving to the more contentious points of illegal content and misinformation on the platform.
Henna Virkkunen, the Commission’s executive vice-president for tech sovereignty, further shows how careful and hesitant the EC is being here, saying that “This decision is about transparency” and “nothing to do with censorship”. The fine was “proportionate,” she said, and “we are not here to impose the highest fines.” Her language use shows that the Commission is aware of how this decision will be framed, and is trying to control the narrative. Meanwhile, Musk and Trump are not particularly convinced by this, with Musk calling the EU the Fourth Reich, retweeting a picture of a nazi swastika underneath the EU flag, and calls for abolishing the EU. Trump meanwhile called Europeans ‘impotent’ regarding political deals on Ukraine, and it’s not entirely clear that couching the announcement of a fine in such careful language is sending a message of political strength to Trump.
The timeline shows how the DSA’s regulatory approach operates on a different speed than political events. Formal investigation into X opened in December 2023, with preliminary findings issued July 2024, and a fine announced December 2025, for a total time for transparency violations of two years. This is only for the transparency violations, with the more complex investigations into content manipulation and election interference are still ongoing with no completion date.
Making sense of sense-making tools
Embedded in the DSA is a theory about what X actually is. It treats platforms like X as communications infrastructure where speech happens, and the platform is conceptualised as a singular place, mostly neutral, with certain obligations for moderation and transparency attached. It views platforms as companies that are capitalistic in a textbook understanding of capitalistic companies: entities with the goal profit maximalisation, that are responsive to legal and economic incentives. This place can be regulated properly via transparency and via a set of complex process requirements. The platform companies that run these places will then implement these requirements as they are incentivised to do so via legal and economic pressures. The DSA’s approach follows from this understanding: establish transparency requirements, ensure researcher access, and prohibit deceptive design practices.
Under this framework, the platform is regulable through procedural obligations. The DSA mandates transparent ad repositories so researchers can detect coordinated campaigns, requires platforms to provide data access so systemic risks can be studied, and prohibits deceptive design like selling “verification” to anyone without actually verifying identity. These requirements target the infrastructure layer. They don’t directly regulate what content appears or how algorithms amplify it. Instead, they create transparency and access that should enable identification of harms, which then triggers other enforcement mechanisms. The procedural requirements are the foundation on which content-level interventions could eventually rest. This is a highly procedural approach, with the idea that if the procedural infrastructure is right, then either the platforms will naturally trend towards ‘good’ outcomes for speech, or the infrastructure yields proof of the platform misbehaving, which can then be punished via other regulations.
X under Musk’s leadership operates in a manner that the architects of the DSA clearly did not account for however. Musk’s X is a major platform operated primarily as a political vehicle rather than a profit-maximizing business. There is little in Musk’s behaviour that indicates that commercial optimisation of X is his priority, from haphazardly gutting the workforce to alienating major advertisers, and altering the algorithm to favour right-wing speech. This resulted in cratering ad revenue for X. Musk’s choices do not make sense from the perspective of a business textbook, but do make sense if you understand the platform as Musk’s personal project, with changes made haphazardly and erratically to suit his needs. That his demands and ideas can instantly switch between petty and personal to strategic promotion of Musk’s affiliation with nazi ideology, is a further indication that the incentives of the platform are vastly different than what the DSA assumes.
This is why enforcement of DSA regulations fails when it comes into contact with platforms like X, as the DSA regulates companies by threatening revenue and market access. But X is an organisation that sees its platform as a political instrument, and that happens to have something resembling a corporate structure attached to it. Musk has the personal wealth to subsidise operational losses indefinitely. Furthermore, Tesla’s stock price is at an all time high, and more than triple in price since his takeover in fall 2022. This shows that Musk is generating exorbitant amounts of wealth by treating X this way. X functions implicitly as the marketing outlet for Musk’s personal empire, as financial commentator Matt Levine says: “It is sometimes useful to think of all of Elon Musk’s ventures as one big company.” He wrote that in context of the DSA threatening fines against X in October 2024, but also says, while it makes sense to think of all of Musk’s ventures as a single company, there is a risk against being an actual single company: “See, you’re not really supposed to do that: X is its own company, with its own corporate structure and owners; 6% of X’s revenue is 6% of X’s revenue, not 6% of the revenue of Musk’s other companies. But if everyone thinks of the Musk Mars Conglomerate as a single company, then there’s a risk that it will be treated that way.”
Levine’s analysis proved to be prescient, and the EC treated X as a separate entity, and not part of Musk’s total conglomerate, over a year later X indeed got a fine close to 6% of Musk’s revenue, for 120M EUR. But considering Musk’s personal wealth has grown from around 130 billion to 470 billion dollars, a staggering tripling in three years for a total growth of 340 billion dollars, a 120 million euro penalty seems like an hilariously good deal in favour of Musk.
Where the EC treats X as a communications network, Musk understands intuitively that X is something more than that, although he does not spell it out explicitly. Social networking platforms are collective sense making tools. Social networking platforms, whether that’s X, Instagram or TikTok, are platforms that we use to shape our common knowledge, and to determine which political opinions are currently in-vogue. These platforms are used to create a shared reality. This goes from how TikTok and Instagram influencers can push Dubai Chocolate into a global hype, to how the conversations on X shape what’s inside the political Overton window. The algorithmic feeds actively shape which voices get amplified, which narratives spread, and which facts feel established. Henry Farrell summarises the problem as: “The fundamental problem, as I see it, is not that social media misinforms individuals about what is true or untrue but that it creates publics with malformed collective understandings.” The fundamental power of platforms like X comes from its ownership over the tools to shape the collective understandings of the public, and allows them to be malformed in favour of fascism.
Viewing platforms like X exclusively through the lens of a communications network, without taking into account how the platform affects collective knowledge, leads to two problems, both on the individual level and on the political level. This misunderstanding operates at both the individual and regulatory level.
As Robin Berjon explains clearly in a recent article how staying on X have “a completely faulty understanding of how power works in the digital sphere and operate according to mental models of social media that simply do not match reality.” They imagine X as a public square that can be occupied or reclaimed, or as a neutral venue where audiences can be addressed. But as Berjon points out, platforms like X function like a personalised newspaper, not as a public space. The platform exerts editorial control over what everyone sees. Some platforms make this editorial decisions based on business reasons, while platforms like X make these decisions much more on political reasons. The DSA is also written from the same perspective: it imagines a neutral infrastructure that could be made transparent and accountable, rather than an editorial system whose owner has no interest in neutrality and is insulated from the economic pressures that regulation assumes.
What does digital sovereignty mean anyway
Jon Worth recently created an extensive wall of shame, posting on a per-commissioner-basis whether they were still posting on X, and found that 25 out of 27 European Commissioners have posted on X in the last week, and 24 out of 27 generally post every week, and none of them have stopped posting since the takeover of X by Musk.
It is indicative of how the EC places itself in a position of weakness. After two years of investigation, coalition politics, legal proceedings, and all the EC has to show for it is a measly €120 million fine and the hope for incremental improvement in transparency practices.
The behaviour of the Commissioners demonstrate how utterly dependent the EU has become on X for their communications. It seems that they can only conceive of power via proximity, and that’s why they feel the need to be on X, as that’s where the social power is. There does not seem to be any understanding by any of the European politicians that they have agency and can take power over social infrastructure themselves. By their own actions they have relegated themselves to a place of dependency, where they cling on to shibboleths of transparency, in the vain hope to regain a sliver of control.
Digital sovereignty has become a buzzword in Europe, but how the EU and the EC treat X, both via the DSA’s regulation as well as with their daily usage, showcase that digital sovereignty is not much more than just that, a buzzword. What does digital sovereignty even mean when you are negotiating terms of platform transparency with a platform who’s owner explicitly calls for the abolishment of the EU, and who is backed by the US government who just released a National Security Strategy that calls European regulation illegitimate censorship? How does a better ad transparency system on X lead in any sort of way to digital sovereignty for Europe?
Sovereignty means having actual power, and this recent DSA fine makes it crystal clear that the EC wields a painfully small amount of it. What’s worse, the continuing presence of virtually all Commissars on X showcases that the European leaders both lack the awareness of how digital power gets constructed, as well as how their continuous presence on X creates and legitimises the power of the people who are actively out in the open working to undermine the existence of the European Union itself.
Taking power
Social networks built on open protocols, like ATProto and ActivityPub, are designed around an entirely different premise than the DSA. Where the DSA’s theory of power comes from a top-down regulatory control over platform behaviour, the theory of open protocols is to make the type of power that the centralised platforms currently have structurally impossible. There is no single entity that controls the algorithm that needs to be regulated, because there is no single algorithm. Nor can a single entity block the EC’s ads because there is no single entity to do the blocking.
In a recent blog post, Mastodon calls for “social sovereignty”, as a response to how X can retaliate against government institutions. Mastodon understands social sovereignty here as public institutions taking control of their social media presence, mainly by running their own social networking servers on software like Mastodon. They mention explicitly that the EC already has their own Mastodon server, at ec.social-network.europa.eu, and invite other organisations to follow suit. That the EC already has their social sovereign presence, but only uses it for press releases without any of the Commissioners using the platform, further accentuates the large gap between the rhetoric and behaviour. Still, the infrastructure for alternative ways for the EC to take power already exists. Initiatives like Eurosky further indicate that the tools for the EC to shift power structures away from the platforms they’re trying to regulate are available.
Open social networking protocols offer a fundamentally different model of sovereignty from what the DSA tries to accomplish however. The DSA positions the EC as regulator sitting above platforms, using fines and transparency requirements to shape the behaviour of the platforms. Open protocols position the EC as simply one node among many equals.
This indicates a real loss of power for the EC: from regulating platforms with billions of users, to simply running a server as equals, with no meaningful power over the network. But the honest question is if this power over X actually truly existed in a meaningful sense, or if this power always was a legal fiction. That the DSA can only level a fine of 120 million euro against a platform owner who used that platform to gain over 340 billion dollars in wealth, and is using that power to actively influence politics to abolish the existence of the European Union, suggests the latter.
The Parlamentarium tells visitors that the EU makes itself accessible to every citizen. But when you want to know what your MEP actually thinks, there is only one platform to go. Until that changes, digital sovereignty remains nothing more than a story that the EU tells about itself.
https://connectedplaces.online/the-digital-services-act-and-theories-of-power/
A Power Problem, Not a Platform Problem
Happy 2026 everyone. The world’s richest man runs a subscription service to remove the clothing from photographs of children, and I don’t know how to write about it. It’s been just over a week and the global order has already drastically changed, in ways that affect everything, including how open social protocols understand themselves.
Remember 2023 and 2024? When every time Musk did something bad, people got excited because that would lead to Elon Musk Events, with signup waves to Mastodon and Bluesky? And when ‘bad’ was understood be frivolous things like DMs not working? Now in 2026, when ‘bad’ means generating CSAM and NCII on-demand at industrial scale, it’s crickets, and there is no initiative at all to leave the platform anymore. Society is going through the motions of vocal condemnation, pointing at agencies who should enforce something, but then not enforcing anything. It is clear now that actually leaving X is a step too far and unthinkable.
On January 3rd, US forces bombed Caracas, captured Maduro and his wife, and killed at least 80 people. Trump posted photos from a makeshift situation room at Mar-a-Lago showing the raid being monitored in real time. In the situation room photos, visible on the big screen behind Defense Secretary Hegseth, was an X feed showing search results for ‘Venezuela’. Another photo showed the OSINTdefender account on the screen. Few things illustrate the current role of X in our society as well as the heads of the most powerful military in the world, monitoring the X account of OSINTdefender while the CIA director sat next to them in the room.
Grok’s latest update allows people to generate sexualized images of women and children on demand, at industrial scale, generating multiple such images per second. Everyone is aware that this is happening, and continues to happen, as nobody is willing to stop it. Musk’s attitude is to use the crying-laughing emoji on complaints is an indication of how serious he takes the issue. Politicians have universally condemned it in words, calling it unacceptable. What is so maddening, however, is that their actions say otherwise. Governments all around the world are very clearly afraid of picking a fight with Elon Musk and a belligerent US regime.
This fear by politicians is further accentuated by the raid on Venezuela and Maduro’s kidnapping, which shows that the US is now a rogue state that does not care in the slightest about adhering to any forms of law. With realistic further threats being made to annex Greenland by the US, it is in fact understandable why politicians are afraid to take actions against the richest man in the world. When the UK said it might potentially think about enforcing its own laws against X, a US congresswoman threatened to sanction not just Starmer but ‘Britain as a whole,’ calling enforcement ‘a political war against Elon Musk.’ X is both protected by US state power, as well as being a source of US state power.
This widespread societal resignation of ‘guess our government communications now happens on a deepfake porn site’ is maddening, but also points to a deeper issue. We’re used to describing X as a platform, and analyse X accordingly. The photos of X being on the big screen while coordinating the Maduro raid is an indication that X acts as the infrastructure for power, the glue that connects the neo-royalty.
But the refusal of governments around the world to do anything about the CSAM and NCII generation machine, and how other countries get bombed and their leaders kidnapped because it creates content for X, shows that X is about power, and less about a platform.
When Elon Musk took over Twitter late 2022, alternative open social networks gained prominence, and platforms like Mastodon and Bluesky presented themself as alternatives to Twitter’s function as the digital public square. Both networks understood that there was an issue with the idea of having a single platform as a public square, and thus did not copy Twitter fully, but put their own spin on it.
The fediverse and Mastodon focused on there not being a single public square at all, but instead many digital places and communities that interconnect with each other. For Bluesky the solutions were aimed at giving users more control, with composable moderation and custom algorithmic feeds.
What this framing missed however is that Twitter also stopped being the digital public town square, in the years since Musk’s acquisition. Instead, it became the internal coordination space for a political faction that now controls state power. People still treat X as if it is still 2015, pretending it is the town square and using it for everything from talking about sports news to keeping up to date on pop culture. This lends validation and legitimation to X’s new role, facilitating the power of the neo-royalty.
The neo-royalty is the small network of political, capital, and tech elites centered on Trump, and X is their coordination infrastructure. Because they are rapidly gaining power around the world, and are in full control of the world’s military hegemon, you cannot separate yourself from this power infrastructure that X has become. Leaving X does not insulate or protect yourself from the warping effects it has on global power. This goes for both the large geopolitical aspects (see Greenland), and the local impacts, as the Somali families in Minnesota who lost their childcare funding because of a viral X video weren’t on X.
The implicit theory behind the alternative social web was that platform quality would determine outcomes. Build something that’s better, and in combination with the incumbent getting worse, this would lead to such difference in quality, user experience and safety that at some point people would switch from X to alternatives like Mastodon or Bluesky. This theory held up for a while in 2023 and 2024. In 2025 it started to falter, as Musk aligned himself with Trump, the signup waves to the alternative platforms effectively stopped. In early 2026, this theory is now really over, because X has fundamentally changed. Mastodon and Bluesky are not in competition anymore with the platform X, because X has changed. It changed from being a platform to the power structure for the neo-royalty, with the public square shambling along as a zombie, animated by everyone who still treats X like it’s 2015.
You cannot out-compete ‘where the ruling faction radicalizes and coordinates’ by having better moderation policies or algorithmic choice. X is not a platform problem anymore, it is a power problem, and building a different platform does not solve the power problem.
Other countries will need to leave the platform to untangle themselves from this dependency, and reduce its legitimacy. But the functioning of the neo-royalty is such that other governments taking actions against X will be taken as an offensive action by the US regime, that will likely trigger extensive retaliation. No country seems to be willing to be the first one to move to take action and thus take the brunt of the counter-offense of the regime.
We’re now at a strange stand-off, where it is extremely clear it is unacceptable what X is doing, and governments make a lot of noise about how upset they are, without daring to pull the trigger on taking action. Everyone is waiting on everyone else to take the first move.
This leads to three possible outcomes:
- no government dares to take action, and they keep to calling things “completely unacceptable” while accepting the actual situation. Things stay as they currently are, and the world keeps sliding into a more dangerous and harmful place.
- One government takes action against X, and the US regime retaliates so hard that no other government will dare to do meaningful enforcement against the massive harms created by X.
- One government takes enforcement action against X, creating a permission structure for other governments to also take actions.
All three options have a meaningful impact on the open social web. For the first two outcomes, it further cements X as the place of power for the neo-royalty, further cementing its dominance in the political sphere. This position of power is also what prevents the alternative open networks to become a place of political power in it’s own right. The third option, of mass enforcement, is what creates an opening for open social networks to not just be an alternative, but to be a source of political power as well.
I do not know what the outcome will be, and with how rapidly the world has been changing I do not know which option is likely either. It’s easy to be highly cynical, and that point of view has been extensively validated over the years, but I do choose to hold to hope that we can build a better, more ethical, social internet out of the toxic waste ground of the current state of the internet.
https://connectedplaces.online/reports/a-power-problem-not-a-platform-problem/
Gert V 🇵🇸
boosted
Got my 3rd rejection from #NLNet. 🎉
The first was around 2012 for an #openvideo distribution tool built on the (then) excellent Miro Player. The 2nd in 2023 was to improve #CiviCRM's user interface. We did it anyway :D
This one was to write a comprehensive free guide + research into funding content and platforms on the Fediverse, following on from our book on film funding (11,000+ copies sold).
It's kinda exhausting that they've such a long application form for stage 1, this one took a week.
Got my 3rd rejection from #NLNet. 🎉
The first was around 2012 for an #openvideo distribution tool built on the (then) excellent Miro Player. The 2nd in 2023 was to improve #CiviCRM's user interface. We did it anyway :D
This one was to write a comprehensive free guide + research into funding content and platforms on the Fediverse, following on from our book on film funding (11,000+ copies sold).
It's kinda exhausting that they've such a long application form for stage 1, this one took a week.
@kaffeeringe @hng @blog No it can't, yet. Because in federation recurrance is still a problem without a clear solution. But #NLnet currently supports filing a #FEP for that. However, for that FEP I guess some (user) research will be needed. You're welcome to join, or just wait, cause once that first step is made adding that feature to the event bridge will come next.
Tim Chambers
boosted
@tchambers Things that also come 2026 thanks to #NLnet:
#WordPress users will be able to create polls for the Fediverse.
WordPress in combination with #GatherPress will offer federated RSVP for events.
Interoperability of Events will improve, e.g. Mobilizon will be able to receive events from other applications.
#Lauti, an event management tool for medium sized communities, will join the Fediverse.
@tchambers Things that also come 2026 thanks to #NLnet:
#WordPress users will be able to create polls for the Fediverse.
WordPress in combination with #GatherPress will offer federated RSVP for events.
Interoperability of Events will improve, e.g. Mobilizon will be able to receive events from other applications.
#Lauti, an event management tool for medium sized communities, will join the Fediverse.
