Social networks: what if we had the solution to escape US Big tech?

When you look at topics around privacy or sovereignty, one problem regularly stands out: social networks.

And this subject of networks is particularly sensitive. Because while they are often seen as simple leisure tools where you post cat pictures, they are actually real information highways. Media, politics, sports, culture, social and family ties, romantic relationships, everything goes through networks.

But as you should know, information is power.

What happens when a handful of people control these networks? When they can collect and analyze a considerable amount of information about us, decide what we see or don't see, based on their own interests?

In short, we would be wrong not to take them seriously, and it's an issue, particularly in Europe at a time when we regularly talk about sovereignty.

But it's difficult to create social networks. It's expensive in the first place. Few investors in Europe are willing to finance the next LinkedIn, the next Twitter, or the next Youtube. And most importantly, to hope for success, you need to convince you, but also convince your friends, your family, your colleagues because otherwise, it has no interest. This is what we call the network effect: the value of the product is proportional to the number of users using it, which makes the arrival of new players almost impossible.

And yet, there is today another way.

What if, instead of trying to build the next Twitter, we completely changed the way social networks work?

What if we returned, in a sense, to the origins of the Web: a more open and decentralized Web, where you could choose your provider without losing access to the rest of the network?

In short, let me introduce you to a protocol that could help us redefine our social networks and create opportunities, particularly in Europe: ATProto.

Bluesky is not ATProto

You may know Bluesky?

Bluesky is a new Twitter with around 43 million users. If its success remains modest compared to its predecessor, its growth has been rather dynamic over 4 years and varies depending on regular waves of departures from X. Thanks Elon…

But you might say, Bluesky is just another US company, so another monopoly in the making. Yes, certainly, but you need to look under the hood.

Bluesky is based on a protocol: AT Protocol. This protocol defines the data present on the network, how we store it, how we view it, how we can moderate it (label it), in short, how the network works.

More precisely, it means that Bluesky is not required to own your data to display it. On Twitter, Substack or Medium, the data is owned by these companies and can be filtered, modified, or monetized without your consent.

Here it's different. Your data is hosted on a PDS (Personal Data Server), and this PDS doesn't necessarily belong to Bluesky. For example, my data is hosted by Eurosky, a European PDS.

And I'd like us to pause for a moment on the Eurosky homepage:

Eurosky is also behind mu.social, an alternative Bluesky client, and mu is the first prefiguring thousands of other social apps.

Because yes, the ATProto protocol is so open that each piece of the puzzle can be replaced or extended. And that's where we can start having fun.

Everyone can build on top of it. We can define our data, so we can build alternatives to Instagram (flashes, grain), to Pinterest (current.is), to Substack and Medium (Leaflet, Writizzy), to Tiktok (Spark, Skylight), to InoReader (Gleen), or even to Github (Tangled), all based on the same architecture, but especially, already benefiting from a user base of 45 million people.

The network effect is already there.

Bluesky is not ATProto, it's one piece of a much larger ecosystem: the atmosphere.

The lexicons

By default, the data present on ATProto is made to display short messages. It's the famous Bluesky message you already know: a message, an author, a number of likes etc…

But, at first, you can't post long messages and besides it's not really designed for reading blog articles, videos, or source code.

But we can extend all of this via lexicons. A lexicon is a common language, a data schema, it's roughly the set of vocabulary on a network record, and that certain applications will use.

For example, the lexicon of a Bluesky post is defined by app.bsky.feed.post and we'll find fields like text, createdAt, langs, embed (when there's an image) etc…

But anyone can define a new lexicon. Obviously it won't be read by all applications. Bluesky won't read your lexicon but your application can.

And precisely not long ago I came across a lexicon that particularly interested me for blogging platforms: standard.site.

The initiative comes from three platforms that partnered together: Leaflet, pckt.blog and Offprint and they proposed this new format for long-form publications.

This format pleased so many that it is now available via a plugin in the WordPress ecosystem, for static blog generators via Sequoia, and now it's also becoming central in Writizzy, the product I'm building and which runs this blog (and probably soon also Bloggrify which I also maintain).

This integration therefore allows you to have native understanding of your Writizzy publication in Bluesky/muSocial. Note the "view publication" link which doesn't normally appear for a simple base link.

But most importantly, this allows all posts published on Writizzy to be visible in all readers that scan the atmosphere for blog articles: standard-reader, Heron, potatonet, docs.surf, leaflet etc… In short, in one step, Writizzy becomes a member of the atmosphere and can expose a new article to millions of users.

And this example shows us that we can now create new apps, with their own vocabulary, on existing PDSs and benefiting from an already established network.

The return of an open and decentralized web

Beyond this specific example, ATProto offers many opportunities to return to a more decentralized web under your control. You could, for example, have your own PDS (personal data server) that participates in the AT Proto network.

And you can go even further. The entire protocol provides that each feature is composable and that includes moderation or feeds. It's not Bluesky that imposes them on you. Again, Bluesky provides a service, but it's not ATProto.

If you want to use another labeler (a tool that allows moderation), or if you want to create your own feed with your topics, you can. I could create in the future a feed of all Writizzy tech blogs for example.

In short, you can regain control over your data, the algorithms that push content to you, and the associated moderation.

But more than that, it's huge opportunities to rebuild all the social networks we're missing in Europe: LinkedIn, Youtube, Tiktok, Instagram to name a few.

Hoping that indeed, mu.social is only the first among thousands.

This open and composable web won't happen by itself, and it won't come from Silicon Valley. Today, the infrastructure is ready, the user network exists, and the foundational building blocks are in place. It's up to us, European developers, creators, and entrepreneurs, to seize ATProto to build tomorrow's platforms, on our own terms.

PS: oh yes, and if you comment below the post on Bluesky/mu.social, the thread should come back as a comment on the blog