Aral Balkan

Dear Fediverse,

Soon, some venture capitalist will put $5 million into a startup that offers free fediverse accounts with unlimited storage, etc.

Unless every other instance instantly blocks such a service from Day 1, you can kiss your fediverse goodbye and say hello to a new Gmail. Once network effects kick in, they will make the rules.

It _can_ happen here but you have an opportunity to kill it before it does.

@aral In that context, would you block the Birdsite, if it became ActivityPub compatible?

@aral Makes me think a lot about hashbase for the same concern.

@aral And $5 million is a small investment by venture capital standards. It would be peanuts for Google.

@bob @aral But it would unfortunately be enough money to start one mega instance and turn it into Twitter.

@bob @aral Please, let's not suggest Gruegle invest in the fediverse. They already had a moderately respected and loved network called Orkut. Remember that one? They f*cked it up and made Google+ out of it's bones.

Here’s an idea: form a voluntary “Fediverse Code” that limits the size of instances to a small number. Say 100 or, at most 1,000. Any instances larger than that size are blocked.

The mega-instances we have today, like .social, voluntarily agree to cull their numbers over a period of time; aiding the migration to other instances.

The fediverse is already hugely centralised and unless we do something about it, it will be ruled by a handful of feudal lords/Mini-Jacks.

CC @Gargron

@aral While I completely agree with what you want to achieve, maybe another (and way more smart approach) to that would be: Figure out *why* people (user, ...) are heading for bigger instances, why we *see* these bigger instances pop up rather than many small ones? Why regulate things technically instead of figuring out and eventually meeting the users needs? 😉

@z428 Looks like simple network effects at play, here. People learn of instances proportional to the people on those instances. That's why fighting Zipf's law requires an active effort.

@kensanata Yes. But I think it's also a matter of making this decentralized, instance-based approach accessible to users who aren't just little-trained but actually not trained in technical things at all. So even understanding what an "instance" is might be difficult especially for people who never used e-mail. Not even talking about making a smart choice. And not even talking, too, about people "buying products" and identifying "things" with "products" - "is" mastodon?

@z428 I understand your concerns. All I'm saying is you said in the first toot that we should figure out "why" people are heading for the bigger instances and I feel I provided an explanation for it. You then asked "why we see these bigger instances pop up" and I think my explanation covers that, too. Therefore I conclude that we do need to "regulate things" because we figured it out: this is not about "users needs" but about how information spreads.

@kensanata Maybe, yes. Maybe it's, however, also about technology - getting back to that "getting started" approach again. If we want to have a fediverse with extremely small instances (ideally single-user-instances), the technology must be hidden from the user, same as the internals of my car are hidden from me (and I don't want to know at all). Our challenge is to cater single-user instances to users whose skills are little advanced beyond installing apps on smartphones. 😉

You know, there is that seems very official (it is) where under the signup section suggests a bunch of instances to register in. Then with v2.5.0 we have relays, which in some ways is a good reaction to information-spreading-with-bigger-instance things by allowing small instance to reach others faster.

@z428 @aral

@dragnucs Yeah, I try to point people to But even if half of us do this, and so half the newbies are well distributed: the other half will get on the same instances that their friends are on and the effect is slower, but still there. And it will always be there, if at least some people join instances directly. It'd be something different if bigger instances weren't listed on, for example: then we'd introduce a negative bias to fight this.
@z428 @aral

@kensanata Still think we need to reconsider our threat model here. What if that $5m budget is not spent on building a "single large" instance but rather on buing a hosting service such as masto<dot>host to allow for cheap and "cost efficient" creation of small instances by arbitrary users? This way, we still had a nice clean #fediverse without big instances, and yet all ideas of decentralization would be completely lost.

@dragnucs @aral

@z428 @aral @dragnucs @kensanata
Good that noone is using one of the big hosters or dns-providers here.
People in generel don't take the effort to even know that there are different email-providers or ISPs. They take what they heard of, which is always the big ones. It'a not different here and that's why they first join .social. Abd that is exactly why spawning cheap instances is a terrible business-model.
And a 5Mio$ monster-instance would just compete with twitter and FB.

The threat might be .social becoming too big. A solution could be nodging users after a month or so to move their account to another instance and directly recommend one based on the user's social graph. Then having a migration service that also relays incoming mentions etc. to the new account for a while and notifying all followers to unfollow-follow with one click.

@z428 @kensanata @dragnucs How Venture Capital works: A VC invests $5 million in the _exit_ of a business. That is, they invest in the _sale_ of the business in several years time, ideally in the $500M+ range (because they’ve already invested 9 other $5M lots in other startups and know they will lose that $45M – one of the $5M bets has to be a unicorn for them to get a return). Now ask: if this VC-backed initiative is successful who will buy it for $500M+ and what will their business model be?

@z428 @Gargron @aral I expect that the reason for that is something very simple such as when articles appear about Mastodon they link to .social rather than any other instances. Or it could be that .social is the highest ranked in Google search.
@z428 @Gargron @aral If I do a Google search for "Twitter alternative" then is one of the top links with no other instances listed.

@bob Yes. Plus, on the other side: "Big" instances also inevitably will come to life in cases in which educated users do trust a certain admin and its management and moderation crowd to provide a good service. From that point of view, forcing them to other instances rather than being with people or choosing depending upon whom they trust seems totally counter-intuitive in terms of "the #fediverse approach". 😉
@gargron @aral

@z428 @aral @Gargron

Man, if Masto-as-a-whole somehow started regulating instances to that level, “you can only have xxx people,” I would totally be out of here.

but otoh, since my home instance was hovering around 100 and is probably still under that in re. active users, I’m not the target for this discussion. Just *shudder* “prune users.”

@aldersprig Same here. 🙂 Actually, I am still trying to figure out what's the exact rationale for keeping / demanding instances to stay "small". I'd rather have a network similar to e-mail here, with a load of "small" players, a bunch of "big players" (Google, Apple, Microsoft, some Telcos), a bunch of public structures (universities, ...), and maybe some others - but only "user count" seems too simple an approach to build that structure. 😉
@aral @gargron

@z428 @Gargron @aral @aldersprig
My concern about .social is more that what if it becomes too big for a small team of voluntiers to manage.

@vi Yes. I definitely see that. Plus, well: I *do* work to earn a living, too. I want to use FLOSS and open services. But I do not *expect* people provide me with these things for "free". I still firmly believe that we only will see a real change if we come to sustainable, robust business models allowing people to earn a living by developing FLOSS software or providing services such as mastodon, without being volunteers depending upon donations.

@gargron @aral @aldersprig

@z428 @aldersprig @Gargron In which case, we can all go home now because we already have this exact same system today. It’s called surveillance capitalism and it’s centralised. It’s a system where the big players make the rules and violate your rights.

The alternative is a system in which every node is equal.

@aral @Gargron @aldersprig @z428
No, that's called buzzword-bingo. Politely speaking.

@vi @aral @Gargron @z428 If we’re back to yelling at the big companies, could we please take this out of my mentions? Thanks! :-)

@aral Yes. The alternative would be a system where every node is equal. Such as e-mail where my small self-hosted domains MX is equal to GMail or T-Online mail accounts. I don't even disagree with your points, no offense. My criticism is: We don't seem to do decentralization "right". How "decentralized" would this eventually be given a majority of "decentralized" mastodon instances runs on Amazon AWS or Google Cloud?

@aldersprig @gargron

@aral Likewise, I don't really buy the term "surveilance capitalism". I don't want "big players" to be able to dictate the rules. But spending effort on trying to demand instances to remain small won't change anything about that. The earlier we accept that companies such as WhatsApp, Google, ... got big by providing an overwhelmingly large amount of "untrained" average users access to a technology they never used before, the better. WhatsApp is the "best" (worst?) ...

@aldersprig @gargron

@z428 @aldersprig @Gargron That’s cool, it’s still, sadly, the system we live under today, whether you buy it or not.

@aral Maybe, yes. But are we really willing to change it? We're lost in a load of mostly technical discussions - matrix, xmpp, mastodon, the fediverse and the federation, and feel like we're moving something. Are we, really? I bet the majority of people using WhatsApp, Google, ... is completely ignoring all this - not out of stupidity or ignorance but simply because they don't even understand the actual problem. And even worse: The very moment we give bad names ...

@aldersprig @gargron

@z428 @aral @aldersprig

We had a barcamp at the FrOSCon where we discussed the topic.

I summarized the barcamp here (in German, sorry)

We concluded that we will never have the means to outperform the big companies. So while a brilliant UX is paramount it boils down to an ethical decision. We have to make it cristal clear: a capitalist company can claim "Don't be evil" but it will never be good. We need a #coopweb.


Fully agree. We need a #coopweb and ethical hosting. My idea on that, a while ago, was #libresaas generally for Software-as-a-Service offerings in situations in which a pure FLOSS licensing approach doesn't help because it seems the wrong level of abstraction... :

@aral @aldersprig

@z428 @aral @aldersprig

Cool. Your Libre Saas approach is exactly what we discussed at the FrOSCon. The cooperative webhoster @hostsharing complies with your five rules and we – the coop members – already discussed to offer the services you mentioned. But this project needs a lot of ressources. Our membership is growing, so we might be able to realize these ideas someday.

Perhaps you would like join us in Essen? We meet regularly and guests are always welcome.

@juh I'd really love to be somehow involved with this, because I think it's the right idea, because I think I might be of help and also because, given my current professional engagement, I have time and energy left to be spent on side projects. My only limit however is that I am virtually unable to do travelling anytime soon, given Essen is more than just an afternoon trip. I'll see what I can do.
@aral @aldersprig @hostsharing

@aral ... to the tools they're able to use with their skill sets (in example all the Google stack), chances are they will simply ignore us. We need to convince people if we want them to make a different choice. Using wording such as "surveillance capitalism" or "Fuck off Google" is not very likely to help here in my opinion - it just might drive away people, even those who eventually would be willing to support because they share similar mindsets.

@aldersprig @gargron

@z428 @aral

The partly different positions may also be related to what we expect from Mastodon. Should it be a global replacement for Twitter? Or a playground for nerds? For me personally it's a tool to create a kind of think tank or lab. A nucleus that helps to make society freer and more democratic. And I can run my instance in this way while others pursue other goals. That's great.

@aldersprig @Gargron

@christian Fully agree. That's why I am even more uneasy with that focus on trying to "block" "larger" instances. In such an environment, I would *merrily* have a load of small instances live right next to some city or company instance or a "commercial" (paid?) service, maybe similar to posteo or tutanota (for e-mail) operated by a commercial entity and paid by its users? The only thing that needs to be ensured is that all these entities actually *remain* equal...
@aral @aldersprig @gargron

Have you ever wondered how they are able to provide such good services for free?

Of course not. But I don't need to be convinced. 😉 On the other side, I've been into FLOSS for more than two decades now. I have learnt to have operating systems, desktop environments, browsers, ... available "for free" (-as-in-free-beer). And, too, this seems a difficult excuse: Even if I was willing to pay, I don't see any paid "open" service on par with Google Apps for Enterprise (except maybe Microsofts Office Online stuff which seems little better a choice).

@hypolite @aral

I’m not trying to convine to switch to anything, just to show you what surveillance capitalism is. You already know why FLOSS is free, and surveillance capitalism is simply why Google/Facebook/Twitter services are free.

@hypolite Ok, agreed. Where, in your definition, is a difference between surveillance capitalism and advertisement-based services?

If the ad is displayed in a web page, then there's no difference. For example podcast sponsorship isn't surveillance capitalism. As long as your browser is making a request to one (or, usually, many) third-party domains on any given website, then it is most likely part of surveillance capitalism.

@hypolite I'm uneasy with this, a bit. It seems a weird and very special technical aspect and just one side of the problem. But more than that: Why are we spending so much time fighting this technically and in communication? Wouldn't it be better, for media outlets, app developers, software creators, ..., to instead have business models that allow them to earn a living without having to rely upon advertisement and (subsequently) tracking? There doesn't seem much in this field... 😟

Because most people have absolutely no idea about all the data they unknowingly provide to Google, Facebook and lesser-known third-parties when they visit their favorite websites. And even if they knew, they may not even care.

Consequently, most ad tracking-free subscription-based business model tries (and there are plenty) never get very popular, because their market can only be comprised of privacy-conscious consumers, currently a minority. It definitely is shifting after the recent Facebook scandal, but there's still a long ways to go to really challenge surveillance capitalism as a whole.

@hypolite While I mostly agree, I am unsure whether that's all there is, to it. From where I stand, I see online publishing *rather* difficult: People don't want ads and trackers because, well, it's a fairly blatant invasion of privacy. Agreed. Most people I know, also, strongly disagree with paywalls and login-to-read models (because it effectively destroys the idea to easily link to articles once read). Even more and possibly worse: Some people would even be willing to pay for ...


@hypolite ... high-quality journalism but either fail to do so because the payment itself is done in a not very privacy-friendly manner (*cough* PayPal) or it is way too expensive as a all-or-nothing approach or it is too difficult to handle. From that point of view I really think it would dearly need, in example, *trivial* and anonymous micropayments, maybe all along the lines of what #brave did with the Basic Attention Token. Something you can use to pay for content with a ...


@hypolite ... single "click". Much like pressing a "like" button but actually by paying a small amount of money. Another difficult (maybe different) example, from where I stand, is #fdroid. I really like this very much, but as far as I see, right now there is no chance to build FLOSS Android apps for a living if you don't want to rely upon donations. There's no business model for that, especially for games and considering that some even see "in-app purchases" as an anti-feature. ...


@hypolite I've been into Software Libre and GNU ever since reading the GNU manifesto in the mid-1990s. I always believed in "Libre" being more important than "free-as-in-free-beer". Today, however, I see that, apparently, there is only "free-beer" based business models on the internet, and either people refund using shady approaches such as tracking and ads, or they use none of these (and possibly don't have an income if no one donates once in a while). Not sure how to resolve this. 😐


I hear you and I agree, I'm just saying that this conundrum is why surveillance capitalism as I described it is so successful today.

@aral ... example, in my opinion. All they did would have been possible "just" with stock XMPP before, as well - but no one really cared to actually package this into something Jane & John Doe could use. WhatsApp did. Same about e-mail. Love it or hate it - GMail provided a dead-simple access to this service virtually *everyone* could use. Why don't we just stop bashing these services and get back to work to provide services same as easy-to-use and same as reliable?


@z428 @Gargron @aldersprig @aral

It's a mistake to think that people adopt those systems because they are easy. Things like this are usually adopted because they're installed by default or have corresponding antipatterns.

For example, it's hard for the average user to set up an Android phone after having just bought it without having or creating a Gmail account. The install process assumes that and strongly guides the user to making a Gmail account.

@bob Yes. But these are two different things in my opinion. Setting up an Android device *with* a GMail account is dead-simple. Getting WhatsApp installed and configured to talk and text to virtually everyone in her family is something my aunt (who will turn 70 next year) did completely on her own, without us even knowing she actually bought a smartphone. Looking closer ...

@gargron @aldersprig @aral

@z428 @aral @aldersprig @Gargron

Yes, and that's not a happy accident. Google does make it as easy as possible to set up Android with Gmail. With anything else it's an annoyance. It's part of the lock-in of Google.

A while ago I was trying to set up a vanilla Android image for testing purposes. It kept bugging me for a Gmail address and I was like "Mr G, I don't have a Gmail address and I don't *want* a Gmail address. I have my own email server which you won't accept".

@bob Fully agree. I'm exactly on the same side here. And yet, my question would be: Getting back to my 70-year aunt. With "stock" Android she managed to do so. Maybe with iOS as well, which seems no better. Can we give her a FLOSS based smartphone she can get started with just as easily? I don't see even that, right now. And *this*, in my opinion, is the problem we need to tackle, not trying to blame users for making a choice where there aren't "better" alternatives.

@aral @aldersprig @gargron

@z428 @Gargron @aldersprig @aral

Likewise I don't think users can be blamed for this. As Jono Bacon might once have said, "we don't really have a good FOSS story on mobile".

Instead we have Android. Android is FOSS in principle, but actually existing Android has problems on multiple levels and usually comes with a lot of pre-installed closed apps which you're almost strong-armed into using.

We need better mobile systems. There are things like LineageOS, PostmarketOS and whatever UbuntuTouch is now called, but all of these are definitely non-ideal and mired in closed source legacy to a greater or lesser degree. There is no fully FOSS phone that I can recommend where all the hardware will just work. Maybe Purism will be it, but that's way too expensive.


I have nothing more to add here. That's exactly my point and, for me, the most crucial issue. We can't blame the user for making choices. We only can try to work to give people a "better" option to choose, "better" both judged by *our* and by *their* requirements. Maybe Purism wil be there, at some point. At least I hope so.

@gargron @aldersprig @aral

@z428 @aral @aldersprig @Gargron

Critically, I think we have to control the platform and that will mean open hardware designs for mobile. Unless we have control of the default user experience there's always going to be the Microsoft effect of "well, I'm just using what it came with".

@bob As far as I see, these things are even worse: We're faced with a large set of users that doesn't want to control the platform but actually just *use* it - and doesn't make a difference between "hardware" and "software" either. There are these tablets, smartphones, ... and software is as an integral, inseparable part of the device. I am not sure how to change this.

@aral @aldersprig @gargron

Sign in to participate in the conversation
Aral’s Mastodon

This is my personal Mastodon.