Skip to main content

Today we released #Vernissage 1.37.0 🎉

You can now follow hashtags: photos marked with followed tags will appear in your private timeline. Status details also show common hashtags and clearly highlight tags you follow.

The iOS app with hashtag-following support has been submitted to Apple for review and should be available soon.

This release also includes notification improvements, minor fixes, documentation updates, and other refinements.

Thank you for your support! 😊

v2026.4.2
A patch release for a calm Wednesday morning.

  • Fixed quote authorization stripping post metadata
  • Fixed broken navigation when "automatically reload when losing connection to the server" is enabled
  • Hid voting options for logged out users
  • Build system has been rewritten
  • Dependencies have been updated

There Are a Million Fediverses. Some of Them Are Louder than Others.

In my work supporting a broad array of open social web community managers, moderators, administrators, and developers, I regularly hear sweeping generalisations like "the fediverse is anti-this", "the fediverse doesn’t approve of that", "the fediverse holds this very specific opinion".

I’ve always struggled with these statements. How can anyone possibly come to a definitive conclusion about what "the fediverse" believes? As I've said before, there isn't just one fediverse, there are a […]

In my work supporting a broad array of open social web community managers, moderators, administrators, and developers, I regularly hear sweeping generalisations like “the fediverse is anti-this”, “the fediverse doesn’t approve of that”, “the fediverse holds this very specific opinion”.

I’ve always struggled with these statements. How can anyone possibly come to a definitive conclusion about what “the fediverse” believes? As I’ve said before, there isn’t just one fediverse, there are a million fediverses. You’re only ever subjected to the specific fediverse you inhabit, the community you reside on and the people you choose to follow.

If you happen to follow high-signal accounts – people who post prolifically, loudly, and relentlessly – their sheer output completely shapes your reality. Their volume becomes your perception of what the entire network thinks.

For years, we have held up the chronological timeline as our great escape from this kind of distortion. In the open social web, we often point to our lack of an engagement-driven algorithm as a moral high ground. We don’t have a black box sorting our conversations for outrage and engagement, we just have time. First in, first out.

But I’ve long believed this is an incorrect oversimplification. Pure chronological timelines are incredibly easily dominated. They do not naturally create a balanced feed; rather, they inherently privilege whoever has the most time, the most grievance, and the loudest voice.

I’ve struggled to communicate this clearly, but after reading Tobias Rose-Stockwell’s brilliant breakdown at The Noisy Room I’m happy to defer to someone way smarter than me. He outlines a simple metaphor for how commercial social media distorts our reality.

A depiction of a crowd of people, three are highlighted in red and are louder than the others.

 

Imagine walking into a pub with a hundred people inside. Ninety-seven of them are having perfectly normal, nuanced conversations. Three of them, however, are screaming at the top of their lungs about politics, about each other, about whatever gets a reaction.

Now, imagine the pub employs a bouncer who gets paid by the minute you spend staring at the spectacle. To keep your attention, the bouncer wires those three screaming people into the pub’s PA system and turns it up to eleven. You walk in, you hear the deafening roar of the three-voiced extremes, and you conclude – logically, based on what you are hearing – that the room is entirely full of unhinged trolls.

In the commercial, centralised web, that bouncer is the algorithm. It amplifies the 3% of users who post severely toxic content because toxicity drives engagement and sells ads.

We look at that and say “thank goodness we fired that bouncer, he’s useless”. We boast that our decentralised, chronological feeds don’t have algorithms manipulating our conversations. But it turns out we don’t need a bouncer to amplify the toxicity via the PA when our timeline does it automatically for us.

The Chronological Illusion

When you build a network on a purely chronological feed you replace algorithmic amplification with sheer volume. If those same 3% of vocal, toxic, aggrieved accounts are posting twenty times a day while the 97% of us post once or twice, who dominates your timeline? They do. They don’t need a viral algorithm to amplify them, they just need access to the firehose.

First in, first out simply means the most frequent posters are the most frequently seen/heard.

This creates the exact same distortion that The Noisy Room identifies. We perform an environment scan of our fediverse feeds and conclude that a select few dominant voices reflect the zeitgeist. Maybe we see acrimony, rapid-fire hot takes, relentless indignation, and we believe – falsely – that we are in the minority.

This leads to the same tragic outcomes we see on commercial networks. The quiet majority goes silent. People self-censor. They step away from the keyboard, or they leave the platform entirely, ceding the space to the most extreme voices. The loudest users start to believe they are the majority.

Everyone gets each other wrong.

We designed a protocol to save us from algorithms, but we forgot or failed to design for human perception.

We need to design for people, not protocols

If we’re going to build better social media, we have to acknowledge that a pure, unfiltered chronological timeline is not a neutral arbiter. It is a megaphone for the relentless. So, how do we enhance the fediverse to fix this?

1. We need client-side enhancements that recognise when a single account is dominating a timeline. If one account posts say ten times in an hour, those posts could collapse into a single stack on my timeline. Let me choose to expand them. Give me back my chronological view of my community, not one person’s stream of consciousness.

2. We need to stop treating curation as a dirty word. An algorithm designed by a corporation to maximise ad revenue is harmful, yes. An algorithm, which is just a set of robust filtering tools, designed by you to protect your peace and balance your feed is empowering. We need apps and clients that allow members to easily dial down the volume on highly active accounts without having to fully block or unfollow them, softening the edges without severing relationships.

3. The Noisy Room suggests a “Community Check“, a representative layer of polling shown below contentious issues to show what the silent majority actually thinks. While this would be complex to implement across a decentralised network, we can build tools that gauge consensus without relying on the loudest voices. We need to find ways to measure and display community sentiment that isn’t just counting the number of angry, rapid-fire replies. Community managers should be able to add community notes to posts, or reply with them in a fashion that pins them to the OP. Emelia and the W3 SWCG Trust and Safety taskforce has some thoughts on this.

You’re not in the minority

Most people want their own space, shaped by their needs and their values. They want to connect, share, and learn without being shouted at. The social web is full of these people. I’m one of them. We are the 97% having a normal conversation in the pub while the 3% scream near the bar.

Our goal shouldn’t just be preserving a chronological feed at all costs. Our goal should be building better relationships. We need to ensure that when a new member joins any one of our million fediverses, they see the whole room, not just the people shouting the loudest.

Let’s find our common ground, let’s build tools that reflect the reality of our communities, and let’s give the quiet majority their voice.

PieFed v1.6.23 is released

  • Only notify the banned person about their ban if they previously posted in the community they were banned from. This stops abusive mods from using bans as a form of harassment.
  • Modlog: Reason for the action is only shown from trusted instances, so abusive mods won’t have an audience. Admins can still see the reason though. Which instances are trusted is set in the admin UI.
  • Instance silencing similar to Mastodon. A silenced instance is not defederated from but their posts do not show in the Popular or All feeds and their communities are not shown in Starter packs aka Topics. Their communities can still be found in the communities list and joined in the normal way. Once joined, posts in there show up in the subscribed feed as usual.
  • Per-user number of posts per page. Previously this was an admin setting but now individuals can override this to be a lesser number if they wish.
  • Spinning “loading” animations on slow forms.

Week in Fediverse 2026-05-01

Servers

- Hollo v0.8.0
- ActivityPub for WordPress v8.2.0
- NodeBB v4.11.0
- PieFed v1.6.20
- TinyAP v0.1.8

Clients

- Pachli v3.6.0
- tooi v0.24.0
- Tusker v2026.1
- Nicolium v0.2.0
- Jerboa v0.0.87
- PeerTube Mobile v2.1.2
- Mitra Mini v0.3.0
- Holos v1.5.2

For developers

- Fedify v2.2.0

Articles

- The Seven Deadly Fediverse UX Sins: A Redemption Report Card
- The Architecture of Autonomy and Freedom
- What would a LinkedIn alternative look like on the open web?

-----

#WeekInFediverse #Fediverse #ActivityPub

Previous edition: https://mitra.social/objects/019dc0ea-e585-7be3-8174-327c55ab28d5

No replies yet

Be the first to share your thoughts.