With the repeated crazy statements from Musk, many people want to leave Twitter/X, but where to go? This question has become a real battlefield for Mastodon and Bluesky enthusiasts, two decentralized alternative to Twitter/X.

I must admit that I am quite uncomfortable with the current debate. On the one hand, we have OSS engineers advocating for Mastodon while ignoring all its issues making it unsuitable for a broad audience. On the other hand, we have mainstream media praising Bluesky as the perfect alternative to Twitter/X, while not questioning whether it could become toxic like Twitter/X.

My point is simple: none of them are the future of healthy social media (yet). On the one hand, Mastodon offers a real decentralization, but requires clear improvements to attract a broad audience. On the other hand, Bluesky is able to attract a broad audience but its organization does not guarantee a healthy social media.

Recently, I’ve often been asked my POV on this debate. It is always hard to answer with structured and clear arguments. This post is then my answer to this tricky question. More than just answering it, I also highlight what is needed to build a healthy Social Web.

1. Mastodon has issues, I know

To start this discussion, a criticism of Mastodon is a nice starting point. I am some sort of Mastodon enthusiast, but I am aware of its current flaws.

Onbooardbing issues

The most obvious issue with Mastodon is its onboarding process (i.e., how do you join Mastodon?). Any non-tech user would be lost during its first steps in Mastodon.

Mastodon advocates rightfully remind that its architecture is similar to the email system. While it is a fair point, we must remember that all other digital systems used by average Internet users are highly centralized. Even the email system is partially centralized: most people use GMail, Outlook, or Apple.

Thus, the decentralized architecture is rather unsettling for most users. We must provide better UI and communication to make the decentralization accessible or even seamless.

As a “cryptographer”, I am quite inspired by Signal success: while cryptography was only for Tech experts 20 years ago, Signal enabled anyone to use cryptography… without even knowing what it means. Tech advocates should be aware that average users do not have the time (or the knowledge) to understand the protocols. Hence, open protocols are essential, but they are not a selling point for them. We must make these nice technical properties accessible.

An ecosystem under development

This accessibility point leads me to a key element: Mastodon (and the rest of ActivityPub-based social media) is still under development. The community is also figuring out content moderation and many other things. A lot of amazing efforts are being pushed by organizations such as the Social Web Foundation or IFTAS.

We should promote what the Social Web could become, not what it currently is. It is flawed and it is all fine. Considering the negligible resources invested into ActivityPub-based social media, I am already amazed of the quality. This should be a selling point for public funders: look what these engineers are capable of with no money, imagine what they could do with even a small investment!

My biggest fear: a premature migration

All these problems in the Fediverse created a concern within me: a premature migration. I believe the Fediverse (i.e., the ActivityPub-based social media) is a great framework for the future of social media. However, it is not ready to greet billions of Internet users. If a migration to the Fediverse happens too early, this place will explode and this ideal will have no second chance.

My dream would be to call for a migration when the Fediverse is ready. Such perspectives should help building the Fediverse with a clear goal. It should also force developers/administrators/users to see the current problems as a roadmap instead of a shame.

2. But Bluesky isn’t a real solution either

Now, let’s talk about Bluesky. It is a clearly nice place, much nicer than Twitter/X. However, I have major concerns about it. While the founders claim it cannot become like Twitter/X, I have a few elements creating doubts.

Decentralization washing

First, there is obviously the decentralization. You can find many Mastodon advocates dissecting its pseudo-decentraliztion, so I won’t do it. In short, Bluesky protocol decentralizes some features, but not all of them (e.g., the identity management). Bluesky is a decentralized social media with a single server, while the Fediverse currently has 30K servers; this is real decentralization.

Bluesky publishes most of its code in open-source to make it easy to redeploy in case of “Musk-like” problem. Open-source is important, but it is no longer sufficient in the Tech regulation. Big Tech companies have become the main contributors to open-source, and they are still toxic. Big Tech manages to keep their audience captive, because massive user migration is hard to organize (see Twitter/X). Thus, being able to redeploy a social media does not prevent this social media from keeping millions of people in its toxic environment.

What is missing then? An open governence. I am not satisfied by the semi-centralized architecture of Bluesky, but it also facilitates some operations. Thus, it could be acceptable. For example, Wikipedia is centralized and nobody can contest its success. However, to make such a system a success, you need an open governance. I cannot trust a company based in the US to handle properly my digital space. Thus, if your technology is semi-centralized, decentralize at least the governance.

What about the “Nazi problem”?

To judge the governance of a decentralized system, the “Nazi problem” is a great use case. The “Nazi problem” could be formulated as follows: you live in a peacefully digital place with like-minded people, and suddenly, a Nazi arrives. What happens? Do you ban the Nazi? Do you isolate them? Do you let them talk?

The Fediverse has some clear and decentralized answers (e.g., the isolation of far-right servers such as Gab). I really wonder how it would happen in Bluesky: how do they handle harmful content? Considering the semi-centralized architecture of Bluesky, I feel like they have to handle (at least partially) the moderation centrally. For example, they need to remove illegal content such as CSAM.

This semi-centralized governance creates an uncertainty: who decides what is illegal? What if the governance in 5 years decides that it is unacceptable to talk about politics? I am not a Bluesky expert, but I am sure that they provide some safeguards. However, their semi-centralized architecture must enable some centralized censorship if needed. The sole existence of such things is toxic for social media used all across the globe.

Like what we have seen in Big Tech social media, they could impose US social norms. I ask for a social media where every one can have a digital space respecting its personal values. This is the main selling point of Mastodon: each server has its own culture, rules, and moderation policies. There is no standard imposed to the whole world.

A problematic absence of business model

All this Bluesky phenomenon looks like a “decentralization” washing to me: it sounds nice, even if they do not really do it. I believe the founders have good intentions, but they are supported by venture capitalists whose intentions are clear: profits. Bluesky could be a novel manner to make surveillance capitalism acceptable. As long as they do not find a healthy business model, I will not believe that they can be the future of healthy social media.

They have no real source of income. One day or another, the VCs will force the profitable changes. The fact that Bluesky is not fully decentralized looks like a way to keep control on the network… and to make profit from it.

3. Social media is MUCH MORE than microblogging

Now, I can tell you the most important of my message: the debate Mastodon vs. Bluesky does not matter. Microblogging is only a small part of social media (https://sproutsocial.com/insights/new-social-media-demographics/). I know that many people hate Musk (as I do), but Twitter/X should not be our only priority.

Instagram, YouTube and TikTok are more popular, and have massive impact on their mental health. We need to look beyond this worthless Mastodon/Bluesky debate. In this context, the Fediverse has a clear head start compared to Bluesky. Indeed, the Fediverse is also Peertube (alternative to Youtube), PixelFed (alternative to Instagram), Lemmy (alternative to Reddit), Loops (alternative to TikTok), etc.

Even if Musk is crazy, mainstream social media were problematic even before him. Surveillance capitalism is poisoning our digital world and our real lives. We must reinvent the Social Web and Mastodon/Bluesky could at best be a small part of it.

4. The path to a healthy social web

Until now, my post could have been quite depressing. However, I want to end on a positive note. All the previous paragraphs aimed to put everything into perspective. The Internet can be rebuild in a healthy manner but we must stop listening to VC-backed narratives. To build a Social Web, I would have three main suggestions.

Ask for an open and decentralized governance

As mentioned for Bluesky, an open and decentralized governance is absolutely necessary. The Internet is more than a tool; it is a digital world we live in. Thus, I want more than just “regulating” the Internet, I want the people to “govern” the Internet. Citizens and politicians should reconsider their relationship to the digital world. It deserves more democracy. We are much more than users: we are Internet citizens.

Each Internet citizen has its own culture and values. The Social Web should enable all values and cultures to exist. Up until now, mainstream social media has also been a soft-power tool that pushed US social norms. Decentralization is the occasion to give a real digital home to all Internet citizens.

Be realistic in our expectations

We cannot expect 3 engineers programming during their weekends to provide a system replacing billion-dollar companies. It is unfair and it insults their work. We should acknowledge their amazing results with so small resources. We should consider their current product as a reason to finally fund their work.

Rebuilding the Internet is an ambitious project that cannot succeed in one or two years thanks to three part-time volunteer engineers. Rebuilding the Internet will require investment, and we must be aware of it. Big Tech made us believe that it could be for free… it was not. The cost has been our personal lives, our mental health, and our democratic stability.

Find a business model for the Social Web

While I am vocal against surveillance capitalism, we have to propose a viable and healthy business model. Fediverse (and Bluesky) advocates are quite silent on the matter.

Thus, I have a suggestion: public subventions. Thanks to the subvention model, the social web could be for free or cheap AND not rely on surveillance capitalism. I would take inspiration from the French newspapers: each newspaper receives a state subvention proportional to its number of readers. We could do the same for decentralized social media: each server would receive a subvention based on its number of users. This subvention is not meant to cover all operational costs. Each French newspaper has to complement its revenues with ads, subscriptions, or anything else. Servers in the Social Web could do the same. This business model would inject money into alternative social media and structure the public investment in our digital space.

However, this model is more suitable with the Fediverse than with Bluesky. The semi-centralization of Bluesky would complicate public subventions because many countries would be reluctant to fund a US-based company. The real decentralization of Mastodon/Fediverse allows countries to only local servers. Thanks to the ActivityPub protocol, all servers would still be interconnected, we would have a global social media, but each country would can fund and regulate their share of the digital world.