Bluesky Rant

Watching Twitter implode has been a really interesting experience. Watching it go from a dumpster fire of a service to something even worse has been quite the experience. And its only now that we’re seeing a real mass exodus of people moving away from Twitter. And while this is undeniably a good thing, I seriously have to question why seemingly everyone thinks Bluesky is the most viable (or only viable) alternative.

Now, I should make it clear that I am very much biased against social media as a concept. Content recommendation algorithms on most social media sites spread misinformation, cause mental health issues, and rely on gross violations of user privacy. That being said, I’m still aware people use them despite these clear negatives. Recently, a friend compared it to junk food. Sure, its bad for you, but you still eat it anyway. I can’t even claim moral superiority because even I’m partaking in it. I think we just need to keep in mind how much we’re consuming and if healthier alternatives exist (Spoiler Alert: they do).

So if you aren’t in the know, the main reason people are leaving in mass is because of a recent change to the terms of service that allows AI to train on user’s posts. Now, of course, this is after a long string of events that resulted in things like allowing more hate speech, banning the word “cis”, and completely invalidated the point of verification. All this stemming from one person coming in and buying out the platform in order to have complete control. So the question then becomes: why switch from Twitter to Bluesky when it has all of the same issues?

I think its first very important to understand the difference in how the two work. Twitter is simple, all your data lives with them. Your data is stored on Twitter’s servers, things are accessed via Twitter’s servers, and all the algorithms are run on Twitter’s servers. Bluesky changes things, quite a bit actually. They introduce the AT Protocol, which is quite a bit more complex given its federated nature. You can read their run down here, but I’ll try to put it in less technical terms. The AT Protocol functions like Twitter does, but in a completely distributed nature. This meaning anyone can run their own server that holds data. Anyone can run their own server that handles the algorithms. Heck, anyone can run a service that takes the data from the AT Protocol and uses it in a transformative way (see WhiteWind, Smoke Signal, and Frontpage). There are some more technical details that involve how data is stored and aggregated, but the core idea you need to know is that the data and algorithm are two separate things that can be run by different people.

So what is wrong with this? It sounds like you get all the benefits of something like Twitter while remaining in control of your data, and I agree, thats what it sounds like, but that isn’t whats really happening. The key issue with Twitter was its centralization, one person was in control and it lead to them changing things in a direction that ended up being worse. Now, I have to ask, who is in control with Bluesky? If you signed up for an account, you’re using their algorithms and your data is stored on their server. Even if you went through the effort to set up your own personal data server, all your data still needs to be sent to them for you to use their app. In effect, they have all the control. And sure, in theory any one could start up their own version of a social networking site that uses the AT Protocol, but Bluesky will always have the advantage because they already have a massive user base. They will pretty much always be able to provide a better service given they have the most users and the most data to make the best algorithms. The AT Protocol gives the illusion of freedom with its talk of federation, but in practice it just ends up recentralizating everything, bringing back the same issue.

The issues don’t end there. Part of the AT Protocol is separating the “speech” layer and the “reach” layer. This means accounts are highly portable and there is nothing really preventing bad actors from posting harmful content in the first place. All the power is now placed into the “reach” layer, as there is no practical way to moderate the “speech” layer given its design. And who controls that “reach” layer? Bluesky, because the vast majority of the AT Protocol ecosystem is concentrated there. Again, this means it can fall victim to the same issues Twitter did because there isn’t even the slightest hint of real decentralization. Sure, the protocol allows for easy account portability, but whats the point when there is no where to go?

Now I know people will ignore all that for the simple fact that someone with ill intentions isn’t running Bluesky and all their friends are on there. And while I can’t deny that the network effect is the strongest factor of choice in any social service, I can put forth a good argument as to why Bluesky might not remain safe. I originally was going to write how Jack Dorsey being on the board of directors was concerning given how he is a massive fan of Twitter’s current owner, but it turns out he left the board of directors. No doubt he has some lingering influence, either directly or indirectly, from his tenure there. It especially makes me wary given the recent news that Bluesky raised $15 million in funding, primarily by a crypto company. Having investors means keeping said investors happy, and that generally involves making a stock price go up. That could be in the form of selling user data, aggressive ads, integrating crypto into the app, or various other user hostile things. By being a for profit company, Bluesky is compelled to make decisions that benefit the shareholders, even if it comes at the cost of the user experience. Twitter has definitely made decisions that degrade the user experience in the name of profit (Twitter Premium). Given these are the people and organizations Bluesky is affiliating with, it makes me doubt their long term ability to avoid enshittification.

Now should be the point where I shill the Fediverse because it solves all the problems that Bluesky introduces or something like that, but in reality is is just as flawed. According to FediDB, about 11 million users are on the Fediverse, with about 2.2 million of those being on one main server. Thats 20% of the network in one place. Over 40% of the network is concentrated in the top 10 most populous servers. Its clear to see that a change to any of these would have a large impact on the network as a whole. To be fair, this is a much less pressing issue than Bluesky given that the Fediverse still has actual decentralization and other options for social networking software other than just Mastodon, but its still an issue.

The other glaring issue is that you still don’t have control over your data. All your data is stored on a server you don’t own (unless you host your own instance), and regardless, that data has to get shared with other instances. I do have to acknowledge, that the whole point of social media is to have this data be publicly available. If these servers didn’t share data, a federated system could not work. This would be a good spot to go into a discussion on what data you should and shouldn’t be posting online, but that is for another time.

I think regardless of the issues the Fediverse has, it still is the better option when compared with Bluesky. Why? Because it has actual diversity, actual decentralization. Sure, it may be imperfect, but it has so many less red flags than Bluesky does. There is actual decentralization, there is an actual ecosystem of apps surrounding the Activity Pub protocol, and there are no for profit companies running it. As a social network, the Fediverse can actually avoid the pit falls that caused Twitter to go down the drain.