Thoughts regarding #Facebook and the #Fediverse. A lot of this grows out of my previous thread.
First, methodology. There are some bad ones floating around. Some seem to be based on a supposed metaphysical moral taint to Facebook or capital generally. Others draw allusions to gentrification or big chains or whatnot, but break down under closer analysis- the dynamics are completely different. We can set those aside.
The key questions are "what does Facebook gain by federating that they can't gain without", conversely, "what loss do they prevent by federating that they can't prevent otherwise?"
This differentiates the options, and highlights what their goals would be.
The second most important thing is to understand the dynamics governing this space. I'll get to that in closer analysis.
So let's run down a couple of possibilities floating around...
@emc2 the most credible reason I have read for why Meta is seeking federation is so that they can claim to be open or not a monopoly and thus have some defense against antitrust lawsuits. IANAL but this seems logical because Meta's size and dominance means that they can just launch a proprietary Twitter alternative and completely ignore AP.
@seav I could be wrong, but I don't see any real risk of an anti-trust suit (unfortunately, neither major party is interested in pursuing that).
The political pressure seems to predominantly being right-wing politicians claiming that platforms are censoring them by banning far-right content, but this falls apart under close scrutiny, so I doubt this will support any actual legal action.
This one falls apart almost immediately. They can already do that, and federating gets them nothing.
Facebook's ability to surveil users depends on them using their website/app (both do a lot of spying). ActivityPub will not give them that sort of insight into other instances. Trying to rework ActivityPub to snoop on other instances won't work, for a number of reasons.
1- Content is already there, when they launch the app, it won't be empty for the first wave of users that sign up, it'll feel like a social media that's been there for a while
2- We've seen dozens of "new" social media platforms this year, most of which are direct Twitter competitors. Facebook is very late to the party, and they know people are done looking for a Twitter alternative. So what better way to advertise it than piggybacking off Mastodon's free promo it got earlier this year.
"Same familiar experience but with Facebook's trust"
> "1: they want to monetize our data. 2: "they want to spam ads and disinfo"
In neither of these point have you really acknowledged the true nature of the problem here.3. The embrace, extend, extinguish (EEE) strategy.
Think about what happened when most people moved to Gmail as their email provider. This made it very difficult for smaller email providers to survive, especially when Gmail started treating all other email providers that are not Gmail, Yahoo, or Outlook, as potential sources of spam.
Now suppose the Mastodon users are on Facebook, outnumbering people in more organic communities by (lets say, for example) 19 to 1, or even 9 to 1, this is essentially Facebook buying up clout. Will these other users ever care to change over to another Mastodon instance, or will they see the rest of the fediverse as just a worse version of Facebook?
Will mastodon introduce breaking changes to the ActivityPub API (the second E "extend") to allow more spying? Other Mastodon instances might be pressured to adopt those changes or else risk becoming incompatible and defederating from the majority of fediverse users.
Suppose Facebook starts censoring Mastodon accounts that do not spread their propaganda? Maybe some instances (e.g porn-friendly ones) won't care, but other instances that think of themselves as wholesome may find themselves policing their users out of fear of being defederated from Facebook.
The end game is to destroy communities (the third E: "extinguish"), which is definitely possible for them.
So it is not really a matter of merely spamming ads and disinformation, it is the fact that they will have such a huge population of people using their service, and this gives them a lot more control over communities than you seem to think.
Check out this post from @paco for an affirmation of what I am saying.
I am not sure I buy your analogy between the fediverse and free/open source software.
Firstly, because there is a distinction between free/libre software and open source software (OSS). OSS is friendly toward industry at the expense of free software, it allows industry to exploit free labor of software developers working on problems in their free time, and draws people away from free software, which is more strictly oriented toward community building, and discouraging of than exploitation. Using your analogy, Fediblock is more like "free software" distinguishing itself from "open source software."
But secondly, the analogy breaks down because there is a difference between running an server and developing the server software itself. Once software is written, it exists and anyone can use it. But a service requires constant upkeep to exist. It is much more labor-intensive, and difficult to compete with a massive corporation like Meta in running servers. It is not really the same as maintaining software, which can exist without the constant upkeep. The economics between servers and software is really what makes your analogy not work.
So I don't find your argument all that convincing. Preemptively defederating from Facebook generally sounds like precisely the kind of community defense that Mastodon needs to do to keep the communities they have built from being affected by the sheer number of victims who are being weaponized by Meta against them.
This reads like a metaphysical taint argument. We may just have an irreconcilable difference. I view the ability of OSS to capture funding, resources, and products from the for-profit sector as one of its most significant achievements, not as a failure.
Also, a long-term success condition of the FediVerse is precisely to eliminate or reduce the need for a huge SRE organization. One part of that is we don't do all the data-mining that Facebook does.
We may just have an irreconcilable difference. I view the ability of OSS to capture funding, resources, and products from the for-profit sector as one of its most significant achievements, not as a failure.
Yes, we might indeed have an irreconcilable difference. I do not see the for-profit sector exploiting the free labor of millions of developers as a "significant achievement," I see it as a loss to all but the corporations who use those profits to further their economic and political power.You do seem to have some regard for Cory Doctorow'ts concept of "enshittification," though. The for-profit sector may seem, for a time, to be using the results of all of that free labor (open source software) to provide a net benefit to society — that is until they feel like they are entitled to more profit, and use their power to further exploit the people who provide free content for their services by selling their private information to advertisers.
I am not saying this means Facebook will somehow be able to directly sell private information of fediverse users to advertisers. My point here is just that the widespread use of OSS by corporations is not necessarily the "achievement" you say it is.
Corporations do what they do because they're short-sighted and profit driven, not because they temporarily want to benefit society. OSS has played the long game, rewritten the rulebook so that they're driven to act in our interest.
This is achieved by understanding the dynamics, which comes from outcomes-focused analysis and strategy, not appeal to vague notions of metaphysical taint and purity.
> " not appeal to vague notions of metaphysical taint and purity."
We are at an impasse as to the benefit of OSS, but I just want to clarify one last thing: exploitation of free labor is not "metaphysical taint," it is deeply unethical. It is unethical to profit off of the work people do and not pay them. This is not some "metaphysical" notion. Work is physical, profits and payment have real physical consequences. There is no metaphysics in anything I have said.
> "And what is your methodology for addressing that?"
My methodology would be for Facebook to pay Mastodon instances for federating with them. Facebook will, after all, be presenting content created by the fediverse, which they do not own, to users of their platform as if they own it, and with ads.If Meta offers payment to Mastodon servers for federating with them, then I might reevaluate my position.
(I am not naive, I know this will never happen. Therefore block them.)
ActivityPub is an open standard. There is no one to pay to use it. But if you mean pay for the right to not be blocked?
That breaks down immediately. And Fediblocking is a security tool not a "walled garden to enforce payment."
That would be like paying every email provider to not be blocked.
Who would they pay for the right to federate to all Wordpress ActivtyPub plugin users? Should every activitypub dev need separate deals with 22,000+ servers?
Plus, we have all the tools we need to defend against ads posted into the fedi without fediblocking. And every fediverse server does this now with what they federate in:
"presenting content created by the fediverse, which they do not own, to users of their platform as if they own it."
> "There is no one to pay to use it. But if you mean pay for the right to not be blocked?"
> "Who would they pay for the right to federate to all Wordpress ActivtyPub plugin users?"
I mean the people who run the Mastodon instances with whom Facebook would federate should be paid. Right now there are hundreds of Mastodon instances being run, that bandwidth and those servers are not free. But they would be providing free content to Facebook which Facebook could use to profit from with advertisements. On top of that, Facebook would be using up the bandwidth of each instance only to provide an incentive for the users of those smaller instance to switch over to Facebook. (This is the "extinguish" part of the EEE strategy.) So in what world is this fair?Going back to what you said about fighting them on the "extend" part of the EEE strategy: Facebook could also pay the developers of various ActivityPub applications, like Mastodon, Kbin, Pixelfed, PeerTube, and so on. They might just do that. They could also have their own paid engineers contribute code to those various projects. This is fine with me as long as the process of deciding features is fair and democratic, i.e. Meta gets one vote just like every other stakeholder.
But that is not going to pay for the upkeep of the Mastodon communities that have grown organically, and it will not take the pressure off of the users of those communities to abandon their respective instances for Facebook+ActivityPub. Blocking Facebook unless they pay solves this problem.
There are actually 22K+ servers on the Fedi, not counting the unknown number of super tiny ones. Not sure you have thought thru the feasibility of what you are suggesting.
Plus: Many of us would view turning "fediblock" from a security system used as a last resort, into a pay for play paywall, would be a profound step backward. "Don't become a monster to fight a monster" and all that.
There are much smarter ways to defend against EEE.
Plus: Many of us would view turning "fediblock" from a security system used as a last resort, into a pay for play paywall, would be a profound step backward. "Don't become a monster to fight a monster" and all that.
This is not true, at least, not by my understanding of FediPact.So let's leave aside discussing blocking of bad actors (e.g. terrorists, human traffickers) who we both agree should always be blocked. If all Mastodon instances block only Meta in exchange for pay, and freely associate with all other smaller servers, what is the problem with that? The way I see it, it is the state of affairs we have now, Facebook and Mastodon being two separate services, so no big deal.
But suppose we try your strategy of allowing Meta (or any other megacorp who wants to federate) free access to any content on the fediverse without pay. This puts undue burden on the people running the fediverse instances. This could cause them to shut their instances down. This could encourage their users to abandon their instance for Facebook instead. And they are still not being paid, so these smaller instances all shut down. "Extinguish" is a success.
In order to remain viable, the fediverse must develop the means to repel the likes of Russian GRU, kiwifarms, and gab. All of these are able to employ methods and/or resources unavailable to Facebook.
If we can't deal with Facebook, we were never long-term viable.
> "I addressed this already. In order to remain viable, the fediverse must develop the means to repel the likes of Russian GRU, kiwifarms, and gab."
No, I specifically said, let's leave that issue aside because we all already agree on this.What I would like you to address is this:
If all Mastodon instances block only Meta in exchange for pay, and freely associate with all other smaller servers, what is the problem with that? The way I see it, it is the state of affairs we have now, Facebook and Mastodon being two separate services, so no big deal.
But suppose we try your strategy of allowing Meta (or any other megacorp who wants to federate) free access to any content on the fediverse without pay. This puts undue burden on the people running the fediverse instances. This could cause them to shut their instances down. This could encourage their users to abandon their instance for Facebook instead. And they are still not being paid, so these smaller instances all shut down. "Extinguish" is a success.
> "No, you don't get to say 'let's leave aside a critical flaw in my position'
My position is (1) that terrorists and human traffickers should be blocked because they are illegal.My position is also that (2) Facebook should be blocked because it pressures community-run ActivityPub instances to shut down.
These are two separate arguments. I thought we both agreed on (1) and so there is no need to discuss it further. So now you are saying (1) is not separate from (2) and is also a critical flaw in argument (2)? How so?
> "Illegal or not, the fediverse will have to deal with it. If you think otherwise, you are severely misinformed."
I do not think otherwise, on the contrary (like I have repeatedly said) you and I already agree on that point.But I would really like to know why you think this has anything to do with my second argument (about which we very clearly disagree) that allowing Facebook free access to Mastodon content is not going to extinguish Mastodon over the long term. And I would also like to know why you think Mastodon contending with illegal content has anything to do with my argument about Facebook's attempt to extinguish the existing fediverse to the greatest extent that they can?
As far as I can see, these are two completely separate issues. But you think one is a flaw in my argument about the other?
Oh, so a cyber-threat being illegal means we don't have to worry about it?! Well shit! Why didn't someone say so?! Well, infosec folks, close it up! Looks like we're all out of work!
On a more serious note, if you're that poorly informed, you really ought to learn more before you start theorizing.
> "Oh, so a cyber-threat being illegal means we don't have to worry about it?!"
I never said any such thing, I never suggested any such thing. Again, all I said was that I agree with you, cyber-threats (illegal activity) are a problem that need to be dealt with.But I still do not see how this amounts to a flaw in any of my arguments, anywhere, ever. You are the one who claims that me saying "illegal activity is bad," is somehow a flaw in my argument, so you can either defend that assertion, or you can just agree that you misunderstood me.
> "So you go from 'OSS getting funding and resources from for-profits is ontologically bad' to 'the fediverse should charge for access'?
OK, I think I understand your confusion: I think you are conflating the developers of the OSS and the corporations who profit from it.Let me put it this way: who is doing the work? Whoever does the work should be paid. If people are developing OSS and a for-profit company is benefiting from it, the developers should be paid. The point you have made (as I understand it) is that if an OSS project gets big enough, the software becomes essential to the corporations who use it and they begin paying the developers for maintaining it. What I am saying is, this is an exception (such as with Linux), not the rule. Usually OSS is labor exploited for free which is not ethical, at least, not until the OSS projects gets really, really big, which most don't. So the ethics depends on who is profiting and who getting paid, if at all.
You realize that would destroy the commons and re-create something akin to the AOL/compuserve internet, right?
So I think you are making an analogy between ActivityPub and the Internet Protocol now — interesting, let me run with it for a bit. So it depends on how well regulated the market is. In parts of the US, people can buy Internet access from only one provider (e.g. Comcast). This gives Comcast a monopoly in their region, which is horrible for the consumers. Yes, I can see this same thing happening with ActivityPub, and Meta would be the one monopolizing the common resource.
@ramin_hal9001@moralpanic@tchambers@paco Comcast's monopoly exists because the costs of building the infrastructure naturally creates monopolies. The same does not apply to ActivityPub.
That aside, your position seems to be a kind of separatism, where for-profits have top pay to use OSS. You realize this is the same as the OSS business models you seem to be against, right? This is literally what sleepycat did with BerkeleyDB.
> "That aside, your position seems to be a kind of separatism, where for-profits have top pay to use OSS. You realize this is the same as the OSS business models you seem to be against, right? This is literally what sleepycat did with BerkeleyDB."So I think this is another example of where your analogy between online services and open source software breaks down.If Facebook chooses to use an open source software product, the project grows and it has the potential to become profitable (though until it does, the work done by the developers of that software is exploited, and likely never to be profitable).
On the other hand, if everyone uses on Facebook Mastodon, this increases the bandwidth cost of running the Mastodon instance, presuring the instance to shut down. Unlike source code, when an instance shuts down, all the people using it loose it forever.
Also if Meta uses open source software, the users of that software get to benefit from the developments Meta makes to that code base. Unlike source code, if Mastodon users can use Facebook, this incentivizes them to leave the smaller instance for Facebook, which eventually could lead to the instance shutting down.
The whole goal of Meta "embracing" ActivityPub is to "extinguish" these smaller servers.
So I guess maybe we could say that OSS is not a zero-sum game, but running a social network service is, and this is where your analogy fails.
Your argument implies that the fediverse itself isn't viable. The same issues with bandwidth apply to everyone else. Same as your argument that we can't deal with blocking Facebook's traffic (which implies we can't deal with the most serious threats)
If that's your position, fine. I don't agree, but I'm going to have to say "time will tell".
(I will also note that you've completely abdicated the "OSS getting money from capital is bad" position)
> "Your argument implies that the fediverse itself isn't viable. The same issues with bandwidth apply to everyone else."
No, I don't think it does imply that. I mean, it is definitely a challenge keeping Mastodon instances running due to costs, especially for servers with lots of users. But everyone paying what they can afford through LiberaPay or what have you can help with that problem.But what benefit is there to a server for letting Facebook have access to all the content and people who use your server for free?
I say there is no benefit at all, it only increases the total cost of operations while incentivizing people to leave your service for Facebook, and therefore Meta's strategy of extinguishing a competing service succeeds.
I think this is the most likely outcome unless Facebook (or any other megacorporate ActivityPub service) is universally blocked.
> "It sounds to me like you don't actually think fediverse is viable, unless it remains small, obscure, and highly exclusive."
I never said the Fediverse should remain small, obscure, and exclusive. I am saying it should compete with and become bigger and better than Facebook. You cannot compete if you are giving away your most valuable resource to your competitor for free.
So we're back to instances charging for access, then...
Look, you're running in circles, because your positions are incoherent, because your methodology is non-falsifiable and therefore can't analyze systems and develop strategic plans.
I do have to work, so I can't go on running in endless circles with you. I'm content with the arguments I've already made.
> "No, you don't get to say "let's leave aside a critical flaw in my position". My argument stands."
So, this is a little arrogant of you, but I remain open to the possibility that I misunderstand you. Although let me be honest here, it seems to me like you do not even understand your own argument anymore.> > "Look, you're running in circles, because your positions are incoherent, because your methodology is non-falsifiable"
These are not circles, I am trying to explain to you the argument I am trying to make here. I keep explaining the same concept again and again because you clearly aren't understanding me. Every time I come back to this very simple, and completely coherent idea of "if you work you should be paid," you veer the conversation in some other direction with lots of false analogies.It's fine, we do not need to keep discussing this. But you really do not understand my position at all, and I am not sure how else I can explain it to you.
Please, do not let me keep you from work. Myself, I have to sleep on my side of the Earth. Cheers.
You don't understand my arguments because you don't understand the protocols, the technology, or the cybersecurity concerns well enough to realize that the solutions you suggest aren't viable, and the problems Facebook poses are eclipsed by the threats I outlined. This much is obvious.
Your only position is "well it should be this way", with no clue on how to actually that let alone get there from here. "Should" is neither analysis nor strategy.
> "you don't understand the protocols, the technology, or the cybersecurity concerns well enough to realize that the solutions you suggest aren't viable"
Oh, I understand the technology well enough. And I know the technology is (brace yourself...) not relevant here at all... **because this is a social problem, not a technical one.**The question is, what should the majority of Mastodon instance sysadmins do in response to Facebook "embracing" the ActivityPub protocol. As a social problem, the people only have to make choice between two options:
Allow Facebook free access to their servers in the form of federation.
Block Facebook from federating with their servers.
Allow or Block. That's it. Simple.
> "Your only position is 'well it should be this way', with no clue on how to actually that let alone get there from here. 'Should' is neither analysis nor strategy."
The strategy is also simple: convince people choose to block Facebook federation. It is so simple, in fact, there is little that I have to do, it is already beginning. And the reasoning for blocking them is very simple (although you do not seem to be able to comprehend it), the fediverse ought not create free content to strengthen a competing service (Facebook). Federating means Facebook appropriates content without paying. There is no benefit to this, instance operators already understand that.
You can go ahead and try to obfuscate the issue with false analogies to open source software, if that is your prerogative.
Will this prevent Facebook from scraping data off of the fediverse servers? No. But preventing them from stealing data is not the goal. The goal is to prevent them from "extinguishing" the communities that have grown organically here on the fediverse.
@ramin_hal9001@moralpanic@emc2@paco Defederation does zero to defend against a EEE attack….specifically zero to defend against the danger of the toxic attempt to proprietary “Extend” variants. But other things OSS and open standards efforts have succeeded in doing for years do work to defend against that.
> "Defederation does zero to defend against a EEE attack….specifically zero to defend against the danger of the toxic attempt to proprietary 'Extend' variants."
I think defederating Facebook from Mastodon does more to defend against the "extinguish" phase. Treating Facebook as an isolated, separate service (like it always has been) is to prevent Facebook from appropriating the freely produced content of the fediverse in order to augment their own content and increase the value of their service.End users are forced to choose, they cannot just join Facebook and expect the best of both worlds. Nor can get angry at the fediverse servers when they inevitably fail to keep up with the breaking changes that Facebook will introduce to their ActivityPub service.
My objection to that claim Ramin, is that hours after launch Meta will have more users than than the entire active Fediverse user base for the month, and then go from there. And by definition you can't stop someone from "embracing" an open standard like ActivityPub. So that embrace is coming fully, even if the existing fedierse fully preemptively blocked them, which I think wisely, they won't.
But you can stop Extend. That is where the fight should be.
This is subsumed by a problem FediVerse has to/will have to deal with anyway. We already deal with spam and bots and such, and by actors who can behave far, far worse than Facebook. Facebook has resources, yes, but they're not going to be able to pull the same stunts as the likes of Gab, KiwiFarms, or the Russian troll farms.
FediVerse has and is refining methods for dealing with this.
This is actually the first really serious threat, and something to be defended against. More broadly, it's something the #OSS movement needs to develop defenses against.
If Facebook can gain dominance over the ActivityPub protocols, they can shape it to favor big instances, recreate their kinds of walled gardens, and break interoperability.
Add comment