Blogs are one of the historical features of Movim, the upcoming 0.32 release will redefine how…

What is Tyr?

We're taught that email must go through servers. Why? Because the Internet was built around centralized infrastructure. Every email you send travels through multiple servers - your provider's server, maybe a few relay servers, and finally your recipient's provider's server. Each hop is a potential point of surveillance, censorship, or failure.

Even "encrypted" email solutions still rely on these centralized servers. They encrypt the message content but the metadata - who you're talking to, when, how often - is visible to anyone watching the servers.

But there is a network, called Yggdrasil, that gives everyone a free IPv6 and doesn't need a blessing from your ISP. We finally have this possibility to use true P2P email. And moreover, this network has strong encryption to protect all data that flows from one IP to another.

Tyr brings true peer-to-peer email to your Android device using these unusual conditions. Unlike traditional email clients, Tyr doesn't need:

❌ Centralized mail servers (the connections are straight P2P)
❌ Message encryption layers (the network takes care of that)
❌ Port forwarding or STUN/TURN servers (Yggdrasil handles NAT traversal)

When Cloudflare went dark, half of the internet staggered with it. People kept calling it an “outage.” Let’s be honest: it was a structural failure baked into how today’s web works.

A centralized DNS stack is incredibly efficient right up to the moment it collapses. And when that single point of failure snaps, nothing downstream matters. Millions of websites freeze because one company sneezes.

We’ve normalized this fragility for way too long.

If yesterday proved anything, it’s this: the modern internet still depends on chokepoints that have no business existing in 2025.

Centralization wasn’t a mistake. It was a shortcut. And shortcuts always invoice us later.

The alternative isn’t theoretical. Decentralized naming systems are finally maturing, and they don’t break just because a single provider does. Not because they’re magical or perfect, but because mathematically they can’t collapse the same way centralized DNS does.

Several experimental architectures have been exploring this direction for years, including ledger-based distributed name systems that remove the root-layer bottleneck entirely. The point is: the path forward exists — we just haven’t committed to it as an internet community.

Yesterday wasn’t a warning. It was a preview.

The next outage won’t be a wake-up call. It’ll be a consequence.

It’s time to rethink the root layer of the internet, not patch it.

Resilient systems aren’t optional anymore. They’re overdue.

Eugen Rochko: "After nearly 10 years, I am stepping down as the CEO of Mastodon and transferring my ownership of the trademark and other assets to the Mastodon non-profit."

How could a robust decentralised file system be useful?

Would you use one if one was available?

If so, to what use (storing, sharing, building apps on top of it, ...)?

If not, are there some specific reasons like difficulty to set up, legal, you already use one, or other?

I'm making one and it is fully functional but adoption is not here yet so I'm trying to figure out why.

Cheers

Edit: I'm referring to a decentralised online storage, accessible from anywhere.

This page provides a measurement of Bluesky PBC's control over various components of the AT Protocol social network infrastructure. It tracks the distribution of power across key protocol elements, helping to assess the current state of decentralization and identify areas where centralized control may need to be reduced to achieve the protocol's long-term vision of a truly distributed social network.

This post is a blogpost version of a recent talk that Daniel Holmgren gave at AtmosphereConf (March 2025).

AT Protocol (or atproto) is a protocol for creating decentralized social applications.

It's not the first protocol with that aim to exist. In the history of decentralized social media protocols, atproto takes a unique approach which is still deeply influenced by technologies and movements that came before it.

The phrase “atproto ethos” often comes up during our protocol design discussions. It's a fuzzy term, but we use it to refer to the philosophical and aesthetic principles that underlie the design of the network.

In this post, we'll distill that ethos. First, we look at the movements in technology that have most directly influenced atproto.Then, we pull out the core innovations that atproto brought to the table. Finally, we highlight some opinionated ways of thinking that influenced the design.

• Added a Mastodon importer to move your Mastodon posts to your WordPress site.
• User setting to enable/disable Likes and Reblogs
• + a ton of fixes

Mar 13, 2025

We have a temporary glitch with GitHub—probably some contributor was geolocated in a sanctioned region (no details yet). All required documents to unlock the account have been uploaded. Don't blame Microsoft/GitHub - it is just U.S. law. Please be patient. It should be unblocked soon.

Mar 27, 2025

GitHub has gone - long live Forgejo (@forgejo).

Fully migrated out of Microsoft’s walled garden after they blocked us:

• 54k commits
• 9.5k issues
• 4.3k pull requests
• 100k comments

Everything moved. Nothing left behind.

https://git.omaps.dev/organicmaps/organicmaps

What cons of decentralised social media do you see? I read that it's mostly about moderation challenges and regulatory compliance. What do you think?

I'm working on a project called WebMirror that allows visitors to access websites in a decentralised fashion by fetching their content from mirrors (like BitTorrent but) just using a browser (without having to install any additional software including browser extensions).

I've a technology demo for it already (https://webmirror-demo.netlify.app/) but it's mostly interesting from a technical standpoint. I'm trying to find a kind of website and/or content on the Web that would benefit from being mirrored. For example, OpenFreeMap is a great example of that (free map tiles for everyone), but it's just that the dataset is much bigger than my project could support (28.33 million files).

―Bora

More discussion on Mastodon: https://mastodon.online/@mastodonmigration/113819523685288481