There has been a ton of discussion about Meta's impending entrance into the Fediverse. Lots of folks are concerned, and for a variety of good reasons. The tech bros want you to know that it doesn't matter though; Meta is so big that they're going to scrape the Fediverse and consume all of your data, whether you Federate or not. There's nothing you can do, so why care?
If you’re just catching up, the current discourse on the Fediverse looks like this:
Person 1: I really don't like the idea of Meta joining the Fediverse, they're not a good company and I don't want to participate in a social network that they are a part of.
Person 2: Why are you concerned about Meta? They're probably already scraping all of your data already, and they’re not going to stop. And they already have all your data, so it doesn’t matter.
First and foremost, this argument that "Facebook can scrape your data right now without federating so it doesn't matter" is not an acceptable answer to someone who is concerned with how Meta uses their data. If someone tells you they don't want something to happen, responding with "well there is nothing you can do about it" doesn't change the fact that they'd like for that thing to not happen. Just because something can happen, or because something is happening, does not mean that it is acceptable for it to happen.
It is, in fact, acceptable for individuals on the internet to desire, and even demand, a space on the internet that doesn't get scraped, harvested, and used to further the surveillance capitalism apparatus built by mega corporations like Meta. The world would be a better place if more people in the world fully understood how surveillance capitalism works, because then they’d be better able to demand that it stop. I want to live in a world where everything I put out there isn't scraped to oblivion and used to help Meta, Microsoft, Google, OpenAI, or other companies hoard more wealth.
Pointing out that something can be scraped, and is probably being scraped, does nothing to help the situation. What you're attempting to do is simply dismiss someone's valid complaint, because you simply do not care about Surveillance Capitalism. You look like that jerk from that meme comic:
And while I have you here: you’re right, they totally can scrape some of the Fediverse… but you don't actually have any evidence they are doing that. Sorry bud, you really don't. Also, admins can enable
AUTHORIZED_FETCH on their servers which can help mitigate people from scraping Mastodon. Finally, the Fediverse isn't just Mastodon. There are other ActivityPub applications, like Friendica, which have privacy features built in, that limit who can see your posts.
But again, none of this matters! The point isn't whether they can, it's whether they should.
At this point I should just call this blog post done, because I've already invalidated the point these folks are trying to make. But, unfortunately for all of us, this argument isn't just flawed in one way, it's flawed in several ways!
I have seen it argued that "defederation is zero defense" against Meta extracting data, and that it's no different if they federate or scrape. This argument woefully underestimates the value Meta gains by joining the Fediverse.
Let's pretend for a minute that these folks are correct, and Meta is actively scraping the Fediverse in real time. There's no evidence this is actually happening, so we really do need to pretend here. What does Meta get when they scrape the Fediverse? Well, they get all of the public parts of the Fediverse: a bunch of things we've said or posted. They'll know what the underlying social graph looks like. This information is in theory valuable: if you have a bunch of content from someone, you could analyze it to figure out as much as possible about them, then use that information to power your advertising network to target them around the greater parts of the internet. But that’s the basic “value of data” in 2023, right? Nothing new there.
What does Meta get when they look at their own apps? Well, all the same stuff that the Fediverse would give them, but also so much more. They have their websites and apps highly instrumented, so they can retain very specific details about how you consume the media on their platforms. They know you like certain types of content, not just because you hit the like button, or follow someone, but because of how many times you've watched the video, looked at the photo, how many times you viewed someone's profile, how quickly you stopped watching that one video, how many minutes a day you spend looking at your ex wife's profile, that the third time you viewed that ad for those shoes you bought a pair, and you paid full price for them. In short, they get a shitload more information from their apps, and don't forget, they have this information about billions of people.
Hey, quick question, how many times have you clicked on a link in Mastodon, then been navigated to the website for a totally different Fediverse app like Pleroma, or Akkoma? Every day right? Yeah, me too, this place is wild. I wonder what would happen if you clicked a link that takes you to the new Meta app? I bet they'd know how long you looked at the page, whose profile you viewed next, how many times you watched that video, and yeah, they know you paid full price for those shoes bud, we all know you did.
Ok, let's pretend you're really good at not clicking on links to other Fediverse Apps. (We get it, you're a real pro). What else could Meta possibly get by connecting one of their apps to the Fediverse? Quite a lot actually. By making a federated app, and federating with as many Fediverse servers as possible, Meta will be merging their social graph into the social graph of the Fediverse. All of the data that they can ingest (and supposedly already are ingesting) is much more valuable when considered in relation to the data that Meta already has on its existing users.
Meta can only get so much information out of the existing public data on the Fediverse, but they cannot get the in-depth metrics they get on their own users via their apps. What they can start to do is compare you to the users they do know everything about. If they know you're like these other 3 million users they know everything about, then they now know quite a lot about you. Not only that, but they're probably going to bring a lot of users to the Fediverse, which gives them many more opportunities to gather more data on you, as you interact with those accounts. They'll know in real time that you were up at 3 am, and that you looked at every photo on your ex’s profile. Which, you know, you should really stop doing.
Meta will then use all of this information to help target you everywhere you go on the internet. Even if you think you've done a good job keeping your account totally anonymous, and you've never once clicked on a single link to their app, it is very possible that you'll leak enough information over time to reveal who you are. It could be as simple as them looking at which accounts you're following, who is following you, and how everyone interacts. (This is actually a common technique used to detect so-called sock puppet accounts.)
Meta may not be able to directly push an ad to you on the platform of your choice, but you know what they can do? They can target all of the people on their platform with ads, and when the people you follow interact with those ads do you know what you get to see? An ad in the Fediverse. It's that simple. Brands know this, which is why they've all evolved to have exceptionally annoying personalities online, so people will interact with them, and push their content to you without ever having to pay for an ad. Just take a look at Wendy's on Twitter.
Think of the situation like this:
Scraping the Fediverse is like watching a hate5six concert video on YouTube. You can see the band perfectly, the audio is top notch, and you see how crazy the crowd is going, it's almost as if you're at the show. Except, of course, you’re not. You can close that window at any time.
Building an app that actually federates with the greater Fediverse is like actually being at the concert. Not only do you see the band, you can feel the music reverberating through your body, you can smell the sweat of the mosh pit, and you can catch people stage diving. Oh, and since you are literally the venue you can put giant banner ads all around the stage, and sell the data you gathered from ticket sales to data brokers, so they can sell everyone a new pair of shoes. You are there, and they’ve got you.
Scraping data has some value, but it becomes much, much more valuable if it's collected as part of a larger social graph that extends into Meta's platforms.
If you see a bunch of strangers standing in a crowd, and you focus on a single person, ignoring everyone around them, you're not going to learn too much about that person. If you zoom out a bit and view them within the context of the group they're hanging out in you can learn a lot more. "Oh look, that person is hanging out in a group of people, a lot with binoculars looking at birds, they must be birders". If, however you're in that crowd with some of your friends that you know really well, you'll also start to learn a lot more about the strangers in the crowd, based purely on how they're interacting with them; "Oh wow, all of these birders seem really cool, within this group there are few clear leaders who hold a lot of social status, and then there's Krang; Krang has the worst shoes, and I bet he paid full price for them, too".
Ok, I hope I’ve sufficiently made my case, and we can all stop saying “there’s nothing we can do, so why try?” Let’s all stop pretending that the fatalistic argument is a valid one, and lets acknowledge that there is real value in federation for Meta. So, on that note:
And finally, if you liked my article please consider giving me a follow, I'm @fancysandwiches over on Mastodon. Feel free to reach out and tell me what you liked. If you're willing to operate in good faith you can even tell me what you didn't like.
Tremendous thanks to my partner Adrienne for editing this for style and clarity.