Beyond overt bans how resilient is the Fediverse (and Lemmy specifically) to the new forms of digital control and censorship?

Hey fellow inhabitants of the Fediverse, particularly those lurking on Lemmy,

I've been thinking a lot lately about the nature of information, discourse, and where genuine human connection can still thrive online. It leads me back to platforms like this one.

We often talk about censorship in terms of direct bans or content removal, which is obviously a critical concern. But what about the more insidious forms of control? I'm talking about the subtle fiddling of algorithms, the deliberate hiding of certain content without outright deletion, the 'shadowbanning' that makes you feel like you're shouting into a void. How resistant is the decentralized nature of Lemmy, and the wider fediverse, to those kinds of pressures? It feels like the very architecture here might offer a unique defense, but I'm curious about the community's thoughts.

I know we're not exactly bursting at the seams with users, and frankly, if you're not already clued into how something like Lemmy works, you're probably never finding it through a casual search – SEO seems like a foreign concept here, battling potential duplicate content issues across instances. Is this quiet corner its strength, or its eventual downfall if the 'outside' world becomes too noisy?

Speaking of noise, it feels like nearly 90% of the content generated on the broader internet these days is starting to feel like it's churned out by LLMs. Autogenerated articles, comments, even entire 'conversations' that ring hollow. Is the Fediverse, specifically, a safe haven from that rising tide of artificial content? Does the human-centric, community-driven nature of these instances inherently push back against such automation?

I've looked into ActivityPub and other federation tools in the past, and my observation has often been that they've been adopted primarily by marginalized groups in society, seeking refuge from mainstream platforms. While that's incredibly valuable and a testament to their utility, what could truly happen to extend this concept, to genuinely get more people involved without compromising the very principles that make it appealing – decentralization, human curation, and resilience against algorithmic manipulation?

Just throwing it out there. Would appreciate any insights or theories.

20 comments

I've kinda wished there were more tools to create my own algorithms and move them between accounts or even share them with others. Like grouping cross-instance comms of the same or highly similar content into one feed. Would be nice to just lump all the asks together to scroll through when I want to interact or something. Or having my own search terms that find me more positive content. Would be cool to have more ways of finding stuff I'm interested in it would just help a lot if there was some transparency and self determinism in the process.
- Or a way when you search to hide all the communities with zero posts (or below a threshold)
- Well.. this could be your project/task if you know coding :) it's FOSS
  
  The most code I know is turtle LOL (and to be clear I'm talking about the one for like, kindergartners where you draw geometric shapes by lifting and putting down the pen with the commands)
  also if I did know something else the first thing I would do would be something like powerdeletesuite for reddit. I know it doesn't get rid of stuff posted to other instances but it would likely auto-delete a lot of them unless the instance specifically wanted to block me from doing so, and a big part of why I do it regularly is that it keeps me fairly obscure. Every time I comment on this people are like "what you put on the internet is forever" and like, yeah, but I don't have to make it easy to find.

Speaking of noise, it feels like nearly 90% of the content generated on the broader internet these days is starting to feel like it’s churned out by LLMs. Autogenerated articles, comments, even entire ‘conversations’ that ring hollow. Is the Fediverse, specifically, a safe haven from that rising tide of artificial content? Does the human-centric, community-driven nature of these instances inherently push back against such automation?
It may be, I don't know for the very simple reason that I don't read this 'broader' Web that is supposedly AI-infested or made of low-effort content.
I read personal websites whose authors I care about that I don't think will make me waste my (precious) time. For the same reason I quit using all social (beside the Fediverse, mostly here on Lemmy). It's a waste of time (time that is so fucking precious) and their content is of little value, if any.
So, to me, the real question would be: how many of us are still consuming that crap instead of focusing on more... I was about to write 'quality content' but I think it would be safer to say 'focusing on human-made content'?
As long as we're willing to eat their crap, and are willing to pay (one way or the other) to get fed that very crap, why would those websites and corps stop feeding us crap? It's making them money. A lot of it. And as far as I can tell all that matters in our so wonderful world is how much money one makes.
edit: typos.

I don't know much about it either, but one thing for sure is that Lemmy.world is too big. If the admins would go rogue, a big chunk of the user base will be affected.
I know I'm kinda part of the problem, but I'm too lazy to switch
- Yeah, that's one reason I researched other instances, but I get it, new people like to join platforms where they see many... What would you loose if you switch? Isn't there an import/export?
  
  Sometimes it doesn't work too great, I've jumped instances every few months and had to manually copy pasta saved posts and comments plus resubscribe to communities

I run multiple instances and work at several projects involving privacy and social resilience.
It is pretty astonishing how much one can be free of disinformation if they keep a disciplined attitude towards content.
Lemmy for example works well if you're aware that you need to exclude and or report harmful content, semi regularly search new communities and take the leap to leave those who are not your cup of tea.
The fediverse is a great teacher of anarcho syndicalism since it is essentially free of censorship in a totalitarian sense but allows people, communities, servers and even groups (fediseer) to exclude what they deem harmful. It isnt the case in any other place, just here.
It would of course be easier if there were simple explanations of the concepts but then again, progressive ideas arent in the broader society's interest and therefore need to be taught slowly.
I think having the option to make your phone or instance guess content for further radicalization or just information based on your configuration could be beneficial. Then again, its not without dangers.
- Hi thanks.
  Regarding running your own instance, is there any issues/security vulnerabilities you've noticed? I remember reading the mastodon docs and even things like pleroma tend to use a lot of RAM/CPU and disk space... I know lemmy is said to be lightweight, but what complications does it bring down the road? I even remember matrix being a HOG.
  
  Of course it depends highly on your configuration. Mastodon, lemmy, matrix and peertube themselves have not ever been a security problem in the years I run them. Of course, you can run into problems with all surrounding stuff like firefall, updates, installation methods, admin accounts, ssh, etc. I can recommend hosting a fedi instance for people who are experienced in hosting public services. Otherwise i suggest asking a friend who is experienced to host the server or buying the managed server as a service and just doing the fedi service alone. You can of course learn if from scratch but that involves many days of reading, trial and error and vigilance.
  In terms of space and resources, I run many services on two threads and 8 gigs of ram, both are not even half utilized. Diskspace is a little more precarious. You should plan for at least 10 gigs per service on lemmy and mastodon and also have auto delete on for a private instance. A public instance goes bonkers because of the pictures and videos that are uploaded. You can store these on object storage i guess. Peertube is fine as long as you dont upload and only show curated federated content. But videos, especially long ones will likely need many gigabytes of space.
  And last points: make backups and dont host public services on private networks.

Someone please let me know if I'm wrong, but I don't see how Lemmy could withstand any amount or coordinated cyber attack or operation.
If dedicated security teams from fortune 500 companies can't beat them, I don't see how someone's server in their garage is going to do it.
- well Lemmy isn't just one thing, its decentralized, and historically decentralized stuff is a lot harder to hit.
  
  It makes some things hard and some things easier. For example, you can more easily defend against DoS attacks because there's just more targets.
  But decentralized makes it easier for bot manipulation because you can hide your actions across multiple users on different instances and those instances can't easily identify bot signatures like IP addresses to ban many accounts.
  
  I don't really understand how it would be, since things get duplicated across any federated instances. If anything it amplifies the spam since it auto-dupes it everywhere.
  You might say everyone could just defederate, but at that point the damage is done, no? Not to mention if they're using bot armies they can just send them at any new instances that spin up.
- The best defence is having nothing of value
- There are several types of threats, some that Lemmy is quite well protected against, some that Lemmy is very exposed to.
  DoS/DDoS - An attack that overwhelms a system, preventing legitimate traffic to/from the server. I would rate this as a medium risk currently, Lemmy in it's concept is actually quite resilient against attacks like that due to it's distributed nature, I actually don't believe it is technically possible to take down Lemmy completely through a denial of service attack, as you would need to attack all instances and hope no one has a backup. The main reason I still rate the risk medium is the consolidation of communities and users on a few servers like lemmy.world, lemmy.ml and lemmy.ee. This makes lemmy vulnerable to a fairly small attack, communities should be spread around the instances to reduce the impact of an attack.
  Government censoring - Low/medium, the government of some countries can order that an instance owner in their country takes down a post/comment, but they don't have the power to order a foreign instance owner to do the same, so lemmy as a whole is not super exposed to this threat.
  Influencing attack/Psyop/Harrassment - Very high, Lemmy lacks an efficient global tool to prevent users from escaping a permanent ban over the whole network. I would not be surprised to learn that Lemmy had a few seemingly normal instances that are being run by national security agencies all over the world, and when they need to they can just set up new users.
  
  This makes sense, thanks for the answer!
  Consolidation is inevitable I think. The most important feature of a social network is who's on it, and federation is a technical enough concept that it's going to be very hard to convince people not to go to the biggest instance.
  I'm not a dev, but I believe the only way to avoid consolidation is to build some kind of neutral signup site that automatically assigns people to an instance based on load/location. This would also be good for adoption cause people could sign up without having to figure out what instances are.

20 comments