What are your thoughts on Meta's plans to federate through ActivityPub?
Meta's new app, Instagram Threads, is planning to be compatible with the ActivityPub protocol. This means it will be possible to federate with services such as Mastodon.
I became aware of this as my Mastodon admin talked about being approached by Meta to discuss federation. Many other large instances have been approached too. There is a general apprehension about federating with Threads, with many instances (including mine) saying outright that they will not federate with any Meta product.
Personally, I think this is an advantage of federation. I can continue to use Mastodon and choose whether I want an instance that interacts with Meta or not.
I definitely understand the apprehension with providing all of your data/information to a company not exactly known for their good handling of data/information, but I'm also not as against it as some people seem to be. If they are going to create a service like this, I'd rather it was federated than a walled garden.
How do you feel?
I'm extremely suspicious of Meta's intentions. I highly doubt that Meta is supporting ActivityPub and joining the Fediverse out of the kindness of their hearts. More likely, they're doing it to exploit the Fediverse for users, engagement, and data.
I can see Meta going two ways.
1. Meta are supporting ActivityPub purely to capitalise on the Fediverse's current popularity. In this case, they'll probably ditch it once it's no longer financially convenient. That will create a rift in the Fediverse, where non-Meta users are heavily pressured to create a Threads account to continue talking to their established friends.
2. Meta is seeking to gain influence over ActivityPub and the Fediverse to shape them for their own purposes. I expect that Threads will rapidly grow in popularity, and eventually dwarf the userbase of other Fediverse apps. This will put increasing pressure on Fediverse app developers to support Meta's features and goals, granting them more influence. Eventually, Meta may as well be the sole company that defines the Fediverse.
Either way, the outcome isn't good for the Fediverse at large. At best, we benefit from a sudden surge in users that we wouldn't otherwise get. Then they could either be yanked away again, or leveraged to influence the Fediverse. I don't think that's worth it.
In my view, the best course of action is for instances to defederate from Meta and effectively lock then out of the Fediverse. If we can prevent Meta from meaningfully participating in the Fediverse, we can prevent them from gaining control and influence over it.
This isn't to say that all businesses should be excluded from the Fediverse by default - but Meta is a huge company seeking to maintain its position as one of the dominant (maybe the dominant) social network companies, and they have done little to gain our trust in the past. I highly doubt they want to support the Fediverse for the sake of the Fediverse.
The fediverse's "current popularity" is nothing of the sort. Mastodon/etc conspicuously failed to replace twitter despite being handed the opportunity on a silver platter; Lemmy/etc are firmly on course to fail to replace reddit. Sure, a small group of people have jumped ship and from the fediverse's point of view that's a lot of people - but the vast majority of the userbases haven't. Nobody in the large-scale game cares about those services apart from them recently having proven their inadequacy, which frankly was a pretty safe bet from day one.
So on point (1), no. Meta don't care about the tiny minority of nerds (the entire fediverse is less than 0.1% of instagram alone) who currently do fediverse stuff. They're a rounding error in the sort of userbase Threads will likely have by the end of this month. Meta need to compete with the current federated services in the same way that elephants need to compete with bacteria.
So why are they using ActivityPub? They have zero need to. It could be because there are a lot of engineers at Meta who actually believe in open, collaborative standards. As a company they certain have a very aggressive profit-seeking aspect but their software engineers are just the same type of nerd as hang out here.. (some of them are my friends). Even Zuck, before he went off on his stupid metaverse trip, was talking very enthusiastically about that sort of thing - although back then it was in the context of messaging, but microblogging is basically the same thing. Meta desperately wants to be a platform provider rather than a couple of websites, especially given FB is on a slow but clear death trajectory and IG is looking a bit peaky. Meta want to sell shovels, not dig for gold. For a while they - well, Zuck at least - thought it would be via the metaverse but again, anyone with a clue could have predicted that was never going to work. But microblogging? Could actually work for them, especially given the growing chasm that the lack of a decent Twitter competitor is leaving.
Will they monetise their services? Obviously they will. But one of the reasons the fediverse will otherwise fail to get mainstream is because any large scale internet service has to be monetised. The internet costs money, especially when you're talking about hundreds of millions, let alone billions, of users. You have to make money or your thing will never get big. Like tildes. I love tildes. I've donated to tildes and will continue to do so. But it's never going to scale, it's never going to replace reddit (I know it's not intended to do either). Donation based systems don't scale because the minority of power users who do pay, don't make up for the vast majority who don't.
So I think you're probably broadly right on point (2). Do I trust Meta? Obviously I do not. But they will make the fediverse work in a way that without their support would never otherwise happen. Will it be the perfect system? No. But it might just be better than the various walled gardens we've had so far.
Regarding point 1, I think I've overestimated the popularity of the Fediverse and underestimated the size of Instagram. Either way, I think it drives home the risk of what's likely to be an extremely large ecosystem joining onto an extremely small but established one; it could fundamentally change the Fediverse forever.
I like your thoughts on Tildes, and that's exactly how I feel about the Fediverse. A lot of people probably won't agree with me, but I don't think it needs to or necessarily should get big. Like Tildes, I'd rather see it grow organically and be enjoyed for what it is rather than be influenced and changed by a big tech company.
I mean, Meta owns Insta too. I suspect them supporting ActivityPub is more about them being able to federate their own services into each other than having any real connection with the outside.
That's an interesting idea.
However, if that were Meta's primary goal, surely it would be easier to implement their own bespoke data sharing solution? I get the impression that choosing ActivityPub would add a lot of rigidity and baggage that they wouldn't necessarily need.
I reckon federation externally has to be a primary goal, or I don't see why they'd invest in the effort over a bespoke solution.
See my other reply which talks more about EEE.
The Fediverse is not a threat now. But it has major potential. Meta rightly saw Insta and Whatsapp as threats, and gobbled them up. Embracing an open protocol also gives them the advantage of being able to quickly onboard devs who used that protocol.
I'm not very familiar with Mastodon and ActivityPub but does Threads using ActivityPub mean there's suddenly a load of 3rd party clients that people like using available to be used with Threads..?
I don't think so. ActivityPub is just used to federate with other instances. The third-party clients use the platform's own API.
There are a few Fediverse platforms besides Mastodon that all essentially use the same API so that they can be compatible with Mastodon apps. However, I doubt Meta will go that route because they'll want to funnel users into their own apps and hoover up their data.
Amusingly, the ActivityPub standard does include a client to server API, but no one actually uses it (Pleroma has it implemented, but I cannot recall there being any clients that make use of it).
They don't need to use ActivityPub to do that. If they want to federate Threads and IG (and FB and possibly even Whatsapp and Workplaces), they can easily make their own standards for their own software to communicate with. They could even just use a closed implementation of ActivityPub internally and not tell anyone.
That they are choosing to federate externally is a choice they've made for reasons. It's interesting to speculate but who really knows what those reasons are..
Where it's said they are gonna federate with external sources/entities?
Um. In the post these comments are on?
"I became aware of this as my Mastodon admin talked about being approached by Meta to discuss federation. "
I'm not sure if there's an official announcement yet, but reporting that Threads would support ActivityPub began months ago, and Meta has been seeking private meetings with administrators of major Mastodon instances. This would seem to all but confirm that they are seeking to federate externally.
Fediverse users currently estimated around 2.5 million. Instagram users are 2.5 BILLION. Facebook I think is approaching 3 billion these days. Active users is a different count but still. Meta are vast.
I like the small places like tildes and a couple of other forums I'm active on, but I would also like a social network which is big enough to have a decent proportion of my RL friends and family on it. I don't really care very much who provides that, although I'd rather it wasn't Elon Musk (and I think a lot of Twitter users are increasingly looking for somewhere else). I don't have any ideological stake in the fediverse, it's currently not fit for my purposes and if Meta makes it such then great.
My understanding is that the current non-muggle/non-meta fediverse can continue just fine regardless of what Meta do, and probably will.
It'll be interesting to see where things go.
I hear your points. But why do you think Meta made overtures to meet with Mastodon instance admin? It sounds like they were gaging interest in Mastodon federating with Threads. Why would Meta go out of their way to set up meetings with Mastodon (with signed NDAs from the Mastodon admins required) ?
Also, why do you believe the Fediverse wants to scale to the same degree as Reddit? That kind of talk seems to be coming from people who have recently left Reddit for Lemmy and kbin, not really from people who have been there for a while.
I suspect Meta does want to be part of an open, federated community. That's sort of what I was getting at. Their motives are not entirely profit driven. They do want to be part of the greater fediverse, but they don't need to be. If they wanted to effectively squash the fediverse as it stands that would be relatively easy, albeit fairly pointless from their perspective. They don't seem to be acting like they consider the existing services viable competition.
I don't pretend to know what "the Fediverse" wants. I don't really use anything there right now. I personally do want it to scale to insta/reddit size because that benefits me. Lemmy and Kbin and beehaw and so on just aren't doing it for me, I want half a billion people there because that's exactly why reddit was good. I don't actually want everywhere to be like here (and here is great!). I like the breadth of communities that only come from large userbases. I liked that reddit was so big that every little niche had it's corner to exist in, and the fediverse doesn't supply that right now. So, selfishly speaking, I want it to get big.
I'm not sure it matters what the current Fediverse wants, if it even has a clear collective desire - because the current userbase will be probably lucky to be 1% of the fediverse by the end of the month. And anyway, I believe I am right in saying that all the little places that currently exist can continue existing without Threads being involved. So nothing is getting lost, just added to.
Re your last sentence, I hope you're right. I have FB, IG, and WhatsApp. I know what Meta is about and use them accordingly. I could see myself creating a Threads account out of curiosity but I never was a big Twitter user (I created a Mastodon account and peek in occasionally). But I'm on kbin.social and follow their and Lemmy's communities - I want to believe that what I post there only lives there and no data is being collected (naive?). I'm hoping it becomes Reddit circa 2013. I am loving Tildes, too, and find I am commenting and posting here way more than I ever did Reddit (I have basically left Reddit after the API hubbub, but I've kept my account just in case). I guess the answer is to assume nothing posted ever goes away and could be mined for profit. (I'm even circumspect when texting.)
You know that's true here too, right? Anything on the public internet is the same. I don't think Deimos will ever sell our "data" (whatever that even means) straight out of the database but your comments here are going to be scraped and archived and probably end up training LLMs and being used for large scale analyses and so on. If you care about that stuff, don't post things on the internet.
From the short amount of time I've been poking around on the Fediverse it seems like a lot of the users there don't really recognize this. It's simply not possible to have the number of users that, for example, reddit did and not monetize. Donations work on the small scale, but they're unreliable and don't scale fast enough. Unfortunately the best way to do this is with advertisements, and while I do personally still hold that there is a spot where ads could be that is both 1. enough for the company and 2. not too many for the user, no company really wants to hit that sweet spot.
One cool thing I actually liked that reddit had that they seem to have gotten rid of was there was a counter that showed how many days of server uptime your account had provided. So if you bought gold or someone gifted you gold it would show something like "this account has paid for 4 days of server uptime" and it was cool to be able to see that.
I see people say this kind of thing pretty often but I don't actually think its a truism here—that thinking is too wrapped up in centralisation. Donations and volunteering scale perfectly fine as long as ratio of instances to users remains fairly low.
I guess you could think of this as horizontal vs vertical scaling; if we have enough people willing and able to host instances then I don't believe we will ever hit the point where it needs to be monetised.
I think tending towards suspicion is a good course of action when it comes to companies like Meta. I also agree that there doesn't seem to be any real motiviation for them to participate in the "spirit" of the Fediverse.
There is a pipe dream future where this endorsement of federation and open protocols encourages the idea as a whole, and leads to a more open and decentralised web. However, this seems very unlikely given the history of Meta, as well as case studies like Google Talk and IRC.
There might also be a #3, arguably more benign (in that it doesn't try to extinguish the Fediverse). Meta could simply try to federate with as many instances as possible and collect all of the user data it can. I haven’t heard yet what will happen to non-Threads data that lives in Threads instances.
User data is something I'm personally not concerned about in this scenario.
The majority of user data on the Fediverse is public, and Meta could obtain it without needing to federate.
What's not public?
The real threat to user data, I think, is using an official Meta app or platform like Threads. They have a track record for hoovering up all kinds of unnecessary user data. If you can avoid that, I don't think there's too much to worry about.
A little while ago this post on Tildes was made that links to an article called "How to Kill a Decentralised Network (such as the Fediverse).
It specifically talks about Meta's plans and why your train of thought
Might actually be less benign as you think it is. It does this by talking about what happened to XMPP, specifically when google joined XMPP federation through talk. It is well worth a read if you are involved with all of this.
Here is another mastadon thread about it as well: https://social.coop/@loshmi/110594900719666868
It's the old embrace, extend, and extinguish play. It is the best legal way to kill an open source project. Zuck's terrified of legit competition, which is why he buys anything that's remotely a threat to his empire. Since he can't buy the Fediverse, he'll adopt the fediverse and extinguish it. The word 'fediverse' will die and be replaced with 'metaverse.' Zuck will see that the people on the fediverse now are outnumbered ten thousand to one at least, so their opinions and desires will be washed out by the 'just want something that works' crowd.
Google is doing this to Android right now. Won't be long before the entire platform and everything that runs on it is proprietary. Frankly, GPLv3 has failed as a license. It wasn't strong enough to force the code and all derivatives open in perpetuity. Corporations know they can use it for a jump start and then subvert it later. That's why newer copyleft licenses like AGPLv3 (which Tildes uses) exist. In theory the newer licenses close this loophole.
Thanks for the links.
It does seem that the more I consider this the less attractive it looks. On the surface, it seems as simple as my instance not federating with Threads. However, Meta doesn't actually need every instance to federate in order to exert some sort of control over things. Adding ~500 million users to the Fediverse will vastly alter the ecosystem, and the social pressure and FOMO will be enough to move a lot of people onto servers that federate with Threads, or Threads itself.
I think perhaps I feel less strongly about it as I am a more casual user of Mastodon, as I was with Twitter and Reddit before it. However, I can certainly appreciate that this move could pose an existential threat to many, many small communities which thought they had found a safe haven in federation.
That link was my first thought on this. "Embrace and extinguish" might be what Meta is going for. Easier to kill something while it's small rather than wait for it to become a larger threat later.
Or maybe Meta just wants to use the tech for themselves. I remain skeptical.
I mean, I get their reasoning. They're worried about the whole Embrace,Extend,Extinguish-thing. Since we can assume Threads to rapidly get very large comparing other AP-instances, they would control so much of the space that if they announce defederation in the future, the comparatively tiny amount of non-Threads users would be under social pressure to also create a Threads-account and move over there.
It's exactly what happened with Google Talk -> Hangouts, after all.
But, I also don't quite understand that argument if I'm honest. Meta, being as big as they are, can also that very thing without ever supporting federation. They are still huge. They can just wrangle people onto Threads, using their existing market penetration from Facebook, Messenger and Instagram to get enough criticial mass of people to use Threads to enact social pressure for the rest to also make an account. They're one of the few companies easily big enough to do this considering the size of the existing champion, Twitter.
So at least IMO, it's still a win they support federation, and beyond worries about the load this might put on existing instances I cannot quite get the criticism. Yes, it might be temporary and they might go proprietary if they get big enough, but the alternative is them simply skipping Embrace and Extend and going straight to Extinguish, which they can still do. The negative results will be the same, only without the benefit at the start.
I don't think this is actually the same thing as Embrace, Extend, Extinguish.
If Threads can grow larger than the Fediverse and thrive without federating with it, then good for them - they'll have a very popular and lucrative social network service with zero control over the Fediverse itself.
Sure, people will be incentivised to join Meta for its popularity, and maybe that will mean fewer people have Fediverse accounts. But the Fediverse will still remain for those who want to use it.
In an Embrace, Extend, Extinguish scenario, a company doesn't simply seek to outcompete its competitor on the basis of popularity. Rather, it intends to gain influence over the competitor to subvert it. If Meta were to apply EEE to the Fediverse, they would seek to reshape it for their own ends. It would eventually no longer be the Fediverse as we know it today; just an extension of the Meta platform.
That would be decidedly worse than if Threads succeeded independently and Meta left the Fediverse alone.
Their goal isn't to outcompete federation. Their goal is to extend it, in other words warp it and slowly destroy it so that they no longer have to compete with it at all.
Correct. Here is the full path for almost every open service a big player joins:
Right now, Microsoft, Apple, and Google could probably kill email if they wanted. They have so much influence, if the collaborated, they'd be able to in the name of security. Large companies love that buzzword and are already using Office, and would migrate in a heartbeat.
Meta rightly sees the Fediverse as a threat. Because most people (the billions of users) don't care about interacting with the entire world. That's always been a tiny minority. The majority are interacting locally with people they know. If some nerd manages to get their family onboarded to a fediverse node, most of their family will have little need to 'reach out'.
A world where you can have your local instance, and just tangentially connect to other instances is a dangerous one. It destroys the powerful social network affect that keeps users tied to a platform.
Its similiar to how Matrix is dangerous to Discord and Slack. If a Matrix client is as easy to use as Discord, the chances of having a 'normie' hop ship to a Matrix client if their other friends use it. Which reduces the chance somebody pays for Nitro.
Threads in itself is Embrace and Extend (I'm assuming the Extend part will be that it's easier to create a Threads account and have an algorithm based on your Facebook, IG or WhatsApp follows/friends set you up automatically with people and groups to follow). They can Extinguish without doing much more than existing and getting all 1 billion of their IG users to create Threads accounts.
I don't know anything about Threads. I would need to know more about how Threads works and how other Fediverse servers normally work to tell what kind of problems we'll see. I guess we'll find out? Here are some thoughts about possibilities. Perhaps people who know more can help narrow it down some.
Consider sending an email. When you send an email to someone who has a gmail address, your email will be stored on Google's servers. You don't expect to be able to delete an email you've already sent them. Too late! This is well-known and accepted.
When you post something on Facebook and then you change your mind and delete it, you expect it to be gone. Do you have a guarantee that Facebook didn't squirrel it away somewhere? Maybe not? Maybe you expect them to abide by the GPDR? This is a little fuzzy, but in day-to-day practice, deleting it and having your friends still being able to read the post would be an unusual, severe violation of our expectations. It would be considered broken and worth reporting as news. (A friend could still have taken a screenshot, though.)
When you write a post in Mastodon and a follower is on a different server, and you delete it, what do you expect to happen? What usually happens, in practice? What guarantee is there that it actually happens?
The Fediverse seems low on guarantees? I don't know how accurate it is, but here's a post about The Delete Activity And Its Misconceptions.
It might mostly work in practice. Maybe sometimes it's broken? How good will Threads be at following established customs? How well-established are those customs, anyway? Are the ways deleted posts get handled by other websites to be considered rules, conventions, or mistakes?
One good way to find out more would be to actually try it. Someone should do some experiments and report back.
Also, what has Facebook written about their plans? Did they say anything about wanting to be a "good citizen?" What do they think that means?
Deletes are federated out to instances that your instance knows has the post (those of your followers, and those that have liked/replied/etc). Some implementations send the object out to every known instance (also reasonable, but has some scaling issues).
Usually those other instances delete the post, as expected. Some of them either intentionally do not, or do not due to a technical issue (e.g. they might be temporarily down, and not get the delete request). This isn't really that different from Facebook or Twitter. Deleting something from Twitter doesn't mean it's gone from archive.org or anywhere else that might be archiving it.
You are entirely right that the Fediverse is low on guarantees. There's plenty of implementations that do things very differently.
All I know is that I want no part in anything coming from Meta.
Seconded. I know it’s preferred to deliver detailed and thoughtful responses on Tildes, but only one thing really needs to be said here;
Fuck the Zuck.
As an EU citizen I'm not that worried. Upcoming regulations on tech gatekeepers will force Meta and others to open up their walled gardens which is probably why they're going straight for Activity Pub based solution. Even if they go for EEE it'll be harder to execute with regulators breathing down their necks and possibly introducing even more regulations if they don't play nice.
I think it’s mostly just a throwaway gesture. People talk about EEE but let’s be honest, the current fediverse isn’t worth extending, embracing, or extinguishing to meta on a strategic level. It’s a tiny ant next to an elephant. They don’t care.
They mention it like how companies mention their emission or charity initiatives. It’s some positive PR.
Honestly I think any “normal” instance that federated with threads will immediately die because their server load would increase by such an absurd magnitude it would ddos them until they stop federating anyway.
I don't think they'll succeed at all. The Fediverse is largely driven by FOSS users, and FOSS advocates are pretty universally anti-Meta.
I'm not a fan of the products that Facebook (the company) ships.
But I am a fan of a lot of what they do in open source. They build a lot of interesting stuff (e.g. sapling a new version control system a la Git), cinder a fork of CPython that runs some of Instagram's infrastructure). Some of those projects have grown to become pretty important in their ecosystems (jest, react, yarn).
I absolutely think it is a bad idea to let them. Google did the same to Jabber/XMPP with GTalk, the network haven't recovered from that since.
The answer is obvious, they're turning the Fediverse into the Metaverse
I hate everything Meta is and what it stands for.