Horizontal and vertical feeds

January 25, 2024

Fediverse client Phanpy is great, and one of the reasons I’ve given in a shoutout a few times already is that it combines the regular vertical feed with horizontal feeds, to make distinctions between types of content.

What Phanpy does, is to take all the boosted posts in your feed, and relegate them to a ‘carousel’. The carousel is a horizontally scrolling feed that intersperses the regularly scrolling horizontal feed. This creates a distinction between posts by the accounts I follow and posts that are boosted into my timeline. Posts by the accounts that I follow stand out more, and I read them all when I scroll through my timeline. Boosted posts are a bit less important to me, and I sometimes scroll past and skip the carousel if I have a little less time and only want to read the more important posts.

I think it is valuable to create distinction between different types of posts:

Posts by my mutuals are the most important, and I don’t want to miss those.
Posts by accounts I follow is something I try to keep up with.
Boosted posts are interesting, I want to see most of them but I’m okay if I don’t see all of them.
Followed hashtag posts are not a priority, and I don’t mind if I don’t check those for a few days.

By relegating boosts to a separate horizontal feed, it becomes easier for me to prioritise, and make sure that I see the posts by the accounts I follow. As a result I sometimes do not see all boosted posts, but thats fine with me. I would like have the same with posts that I only see because of the hashtag that I follow, those posts are not a priority either, and I think it would be nice if they are also placed in the carousel.

Bluesky has 3 million accounts, invites going away soon

Some news from Bluesky: yesterday the network reached the milestone of 3 million accounts. This milestone comes 2 months after hitting the 2 million mark. While there are no official numbers for Monthly Active Users (MAU), Kuba Suder estimates around 500k MAU, based on the daily and weekly active users. For comparision, Mastodon has between 1 million and 1.5 million MAU, depending on the source.

Bluesky’s growth is far from over, as Bluesky said that that ‘Invite codes are going away soon’. This is different from opening up the network for federation, which is also planned for early 2024.

Bluesky is explicitly positioning itself as a place that is welcoming and suitable for news organisations. Bluesky team member Emily published an explainer how newsrooms can use Bluesky for the upcoming election season. Not everyone is convinced the network is ready yet though, as ændra explains some features that the network needs to be fit for purpose.

Dutch public broadcaster quits X

The Dutch broadcasting company KRO-NCRV has announced it is immediately ceasing the use of X. The news comes after a week of racism and hatred on the trending hashtag #Akwasi. Akwasi is a Dutch rapper who was a participant on the TV Show ‘The Smartest Person’. Other Dutch public broadcasters, notably including the main public broadcasting organisation NOS, is also considering leaving X over the amount of hatred and vitriol on the platform.

What strikes me about this announcement, and the consideration of other organisations like the NOS to do the same, is that it indicates my thinking of how platform migrations happen has been off. In the beginning of the Twitter migration wave I assumed that the process would happen as follows:

Someone sees that platform A is bad and becoming worse. They search for alternatives, and if they find alternative platform B, and deem it good enough of a replacement, they migrate to platform B. My thinking has mainly been focused on platform B: is B providing a good enough alternative to A?

Instead, the current situation with KRO-NRCV indicates that alternative platforms play a minor role in the consideration to leave. The decision is more straightforward: at some point a platform becomes bad enough that an organisation leaves. Instead of pointing to direct microblogging alternative platform where they will become active, they point towards their entire social media stack instead.

Threads and Tumblr on fediverse connections

Both Threads and Tumblr have publicly stated that they are planning on joining the fediverse, and from both companies we received some more information this week.

Threads held a meeting, titled “Meta’s Threads Interoperating in the Fediverse Data Dialogue” with various people who are active in the fediverse in some ways. Johannes Ernst shared extensive notes of the meeting, which are worth reading. Some of the items that stand out to me, based on these notes:

Meta’s decision to roll out federation in a step-by-step approach makes a lot of sense from a technical perspective, but does have a major cultural challenge. Many people already distrust Meta, and an implementation that starts with only partial federation of one or two features will likely increase, not decrease trust in Meta. Combating this will require significantly more open communication to the entire community; as Meta again risks increasing distrust in the community by only partially communicating to only some people within the fediverse community.

According to Johannes Ernst, Meta could not provide a clear answer to the main question as to why the fediverse integration is happening, and what has changed at Meta that the company is now touting openness as a strategy. A lot of the distrust around Meta revolves around a lack of clarity of the motives of the company, and it seems so far that it is difficult for the company to produce a clear and concise answer.

Tumblr is still working on it

Tumblr also confirmed that the fediverse is still being worked on. Over a year ago, Automattic CEO Matt Mullenweg made a statement that Tumblr would “soon” add ActivityPub support. During the year, not much news came about, and over the summer an engineer for Tumblr claimed that the project got cancelled quickly after the announcement. Now, Sarah Perez reports that Tumblr did move over someone to work on the project, but that Mullenweg also cautioned that they have not seen outsized user demand yet for federation.

I agree with Sarah’s take, that “reading between the lines, it seems the company isn’t ready to place a full bet on ActivityPub”.

Garbage day – the internet is different now

Ryan Broderick, author of the excellent internet-trendwatching blog Garbage Day, has posted a new video, about why the internet feels different now, and feels bad for many people. There are some fascinating ideas in there that I think are worth highlighting:

First is the observation that how TikTok is operating is a good representation for the current state of the entire internet; in the sense that it has become impossible to form a coherent narrative on what is happening on the internet. We used to (pretend!) be able to form a coherent idea of what the internet was thinking, mainly by just taking some viral tweets and sticking them in a news article.

This period is now truly over, and helpfully demonstrated by this article by The Verge, ‘TikTok’s biggest hits are videos you’ve probably never seen’. It showcases that the internet has fragmented, and that even posts with half a billion views can stay within their own corner of the internet, barely making an impact outside of their own space.

Ryan also says that he is a big believer in the idea that once you have a name for something that is happening online, it is pretty much already over. Based on this, he assumes that whatever will be the new status quo has already arrived, and that the rut that we feel might already be over, with a small pocket of the web already having figured out the new way of using the internet.

Personally I agree with the idea, but I also think that it is just a bit too early to say so. From my observations it seems like the next status quo way of using the internet is on the cusp of arriving, without fully being there yet. As Jeremiah Lee says here, ‘the next big social network is just the Web’, and that is something I strongly agree with. More specifically though, I think that the next big social network will be build on top of all the current social networks. I think the winner will not be Mastodon, Bluesky or Threads, but instead the next generation of products that can seamlessly interact will all of them. From there it can expand to incorporate the entire Web. With more and more work being done both from the protocol level (bridges) as well as the client level, that vision is now starting to appear.

Innovation in decentralised social networks

Flipboard’s Mike McCue recently released the first episode of the new podcast Dot Social with Mike Masnick, where they discuss protocols, platforms, and the decentralised internet, and it’s worth listening to.

Johannes Ernst gives a thread with a summary and responses here, which is worth reading. One thing I’d like to comment on is Mike Masnick’s comment that he expects innovation more to happen on Bluesky’s ATProto than on fediverse’s ActivityPub.

I agree that innovation in the decentralised network space is happening to a signficant extend outside of the fediverse sphere, but I disagree with the idea that this will happen on Bluesky and ATProto. Instead, I think that Nostr is a more likely candidate:

Innovation in a decentralised network is currently largely dependent on individual hobbyist developers that are experimenting. For an individual developer the accesibility and difficulty of working with the protocol is an important consideration. From my understanding talking to developers is Nostr the easiest to work with. ActivityPub differs a lot, but can certainly be difficult, especially regarding actual interoperability. I have been told that ATProto is the hardest of the three to develop for, plus that it is simply not even put into practice yet.

Culture of the network is even more important though in driving innovation. The fediverse has cultures and etiquette that say that some innovations in the network are unwelcome, especially regarding search and consent. One of the things that interest me about the fediverse is that the social impact of technology is taken into account. We’re building these networks for people. But making features off-limits in a network does limit innovation as well, there is a cost to it.

Bluesky is threading a difficult middle ground here with the culture. The developers seem to have more of a technologist mindset to protocol design, and concerns about how federation will interact with content moderation are not given much care. At the same time, a core group of Bluesky users is not particularly interested in federation, and wants a simple Twitter replacement. That puts the team in a pretty difficult spot with regards to future innovation. They made great strides with custom algorithms, but they do experience significant pushback from the community on features that they themselves want to work on, especially relating to opening the network.

Nostr has an explicit culture of adverse interoperability, and a libertarian community who seems to be quite inspired by crypto’s mantra of ‘if we can build it we should build it’. This is not really grounds for a network that is safe for many people. It does provide a fertile ground for rapid experimentation and innovation. The network is by far the smallest of the three, but it has also created quite some innovations that the other two network haven’t, in the recent months. Multiple long-form article publishing sites, a torrent archive, an integrated payment system for subscriptions with crypto, and more. There are good reasons to be have some issues with some of these innovations, but it is hard to deny that they are developing at a rapid speed.

Overall I think that innovation often happens at the fringes where there is reasons for experimentation. But also, cultural reasons that inhibit innovation speed can actually be pretty good from the human perspective.

Understanding Nostr

In the coming period I’ll be posting more regular about Nostr, as there are some interesting developments happening with implications for the fediverse. This blog is meant as a reference and explainer of what Nostr actually is, to provide more context of the news and developments that are happening.

Nostr (Notes and Other Stuff Through Relays) is an open-source protocol, that enables a decentralised social network. It draws inspiration from crypto with its focus on keys instead of usernames and passwords, and has a significant crypto community as well.

There are four parts to the Nostr network to understand: Clients, Relays, Public keys and Private keys:

Your client is an application that you log into. This can be a website or a mobile app, and anyone can create clients for the network. The client is where you post and read messages. It does not store any of your data.

The first time you visit a Nostr client, you create a public key and a private key. Together these effectively form your account.

Your Public key is your username, and it is a long hexadecimal string of random letters and numbers. You find other people on the network by searching for their public key.

Your Private key is your password. There is no authority that controls this, meaning that if you loose your private key, or someone else accesses your private key, your account is permanently compromised.

Relays are the back-end of the Nostr network, and are effectively a group of servers that store all the data of the network. Your account is not hosted on a specific relay, nor is your data. Instead when you create a post, you select which relays (plural) you send your information to. Sending your post to multiple relays creates redundancy, and is the core part of the censorship-resistance that people on the protocol want. Similarly, when you open a client, you get the data from most relays, in order to make sure you get all the posts of the people you follow.

Finally, some personal thoughts regarding Nostr. A social network that prioritises censorship resistance with little to no content moderation tools available for users is simply not safe to use for a large intersectionality of different groups of people. I write and care about decentralised social networks because I value that everyone can have a place on the internet that is safe for them to use, especially marginalised people. As such, I don’t feel that Nostr’s ideology aligns with mine, nor do I think that they set themselves up to escape the ‘nostr bitcoin bubble’, as Jack Dorsey calls it.

Jack Dorsey on Nostr

Jack Dorsey wrote a post about Nostr the other day that I think is worth reading, and take stock how he thinks about social networks. Fediverse-accesible link here (side note: Nostr can connect to the fediverse via a bridge, so I follow Jack directly from my Mastodon account).

First to note is that Jack is fully focused on Nostr, not on Bluesky. He deleted his Bluesky account a while ago, even though he is still on the board of the Bluesky organisation. Reading this post makes it clear where his attention is though, and its not at Bluesky.

Reframing Twitter as an information network over a social network is interesting, and I can see why. It also indicates the impossible dilemma Mastodon is in: Eugen Rochko explicitly frames it as a Twitter competitor, but by making search opt-in, it can never come close to the information network that Twitter was.

Nostr is positioned as working towards a multi-app/use-case ecosystem. Nostr struggles with the same problem that the fediverse has here; both protocols allow for a large variety of different networks and products, but they are currently heavily dominated by microblogging. The challenge for both networks is how to grow beyond just microblogging, and providing other use-cases.

Jack also mentions the ‘completely open and wild API’ as a benefit for Nostr. It is here that the fediverse is ahead of both Bluesky and Nostr. An open API is great for a whole lot of purposes, but at the same time it violates people’s consent to actually use their data. Nostr positions itself as the wild west. This comes with advantages, such as development speed, but makes the community unsafer. In turn, this makes it all the harder for the Nostr community to expand beyond the bitcoin bubble.

Overall, the post illustrates what makes me excited about decentralised social networks, and that we finally have choices. I personally disagree with part of Jack’s vision how a social network should look like. I can contribute to the network that aligns with my values, while there is still space for people to have their own network with their own values.

Biden and the White House join Threads

The White House has joined Threads with a variety of new accounts. The government announced accounts for President Biden, first lady Jill Biden, Vice President Kamala Harris and second gentleman Doug Emhoff, as well as accounts for the White House and a Spanish version, La Casa Blanca.

The move comes after a week in which Elon Musk continued his descent into white nationalism and antisemitism, as well as major advertisers fleeing the platform after having their ads placed next to nazi posts.

The White House has condemned Musks posts as “unacceptable”. The Hill reports that the move towards Threads has been in the work for several weeks. While the latest controversies might be a good confirmation for the White House, it does not seem like it might be the direct reason.

Overall this should boost Threads uptake significantly, with it being endorsed by the most powerful political figures. One thing to watch for his how politicians from both sides of the political spectrum will approach this move. For democratic politicians there is now a significant incentive to follow Biden to Threads. Republicans have been aligning themselves more closely with Elon Musk however, and might be less incentivised to join another platform.

A Mammoth Task, a Zeit interview with Eugen Rochko

A Mammoth Task, is the headline of the interview by German newspaper Zeit with Eugen Rochko. It details the history and background of how Eugen Rochko came to build Mastodon, and his vision for social networks. It introduces him as a private person, who rarely gives interviews, does not go to conferences, and has not met most of Mastodon’s employees in person (a shame, as I very much enjoyed meeting Mastodon CTO Renaud Chaput at Next Generation Internet conference last week)

The article does not shy away from pointing out some uncomfortable facts either, noting his title as a Benevolent Dictator For Life. They write: ‘“I feel quite comfortable in this role,” he [Eugen Rochko] says, “an individual with a vision always builds a better product than a group.” He almost sounds like the Mark Zuckerbergs and Elon Musks of this world. Is Rochko really interested in building a democratic network? Or does he just want to be at the top himself?’.

The entire interview is worth reading (non-paywalled here, auto translation works well enough), as it shows that some press is moving beyond an easy narrative of Mastodon as an replacement for Twitter, and is willing to look deeper into how decisions get made on this new generation of decentralised social networks.

Laurens Hof

Blogs by Author: Laurens Hof