We tried to run a social media site and it was awful
ft.com
ft.com
I should be able to follow @ClarkKent@mastodon.daily-planet.info - that'll tell me that the account is who it says it is.
But mastodon.daily-planet.info absolutely shouldn't have a public registration.
Running a social network isn't for the faint-hearted. But running a publishing platform should be as simple as running a website.
Why is this consider stupidly easy? What kind of website are talking about, because the websites I run and have run in the past are not so simple
If you're running a basic WordPress / CMS - then running Mastodon should be as simple as that (i.e. pay someone to host and run).
If you're running a global website with a load of subdomains, custom features, CDNs, etc - then running Mastodon should be as simple as your existing infrastructure.
Except it's not just spam. Most of the wider fediverse will block your instance if you federate with instances that host nazis, child porn, etc. Which means the first thing you need to do when setting up your instance is gather a list of all those types and block them. Sure, you can say you're committed to "free speech" or whatever, but if all the large servers slam the door in your face your users will have nobody to talk to.
Also, with the way Mastodon works, if your instance federates with another, your instance hosts a cached copy of all of their stuff that appears on your users' timelines. Which means you might run afoul of your host's TOS or even local laws. If you're in Germany and some American nazi's post saying "kill the Jews" gets federated onto your instance, you might be looking at serious legal trouble. Or even if you're not in one of those countries, you might get booted off AWS or Azure for hosting content against their TOS.
So sure, it's as easy enough as any other webapp to launch into your infra, but maintaining a functioning Mastodon instance that doesn't get blocked by the wider fediverse or brings down the wrath of your hosting company or law enforcement on you is hard.
What about the tech of Mastadon makes other nodes refuse to "federate with" me if I "federate with" something they don't like?
Nothing. The blocking situation on the fediverse is arcane at best.Go to any large instance, hit up their about section and look at their list of banned servers: a good chunk of them have ban reasons like "'free speech' server that federates with gab and other nazis", etc.
Basically if you block both left and right wing hatred and CP you're alone in the Fedi lmao
The Fedi is like 4chan but worse, and normies think it's going to be better than Twitter even though it's worse in literally every aspect that they're upset about Twitter: petty censorship by server admins, horrible content, and identity verification are all -worse- on federated networks
Whether that's good or bad is a different discussion, but the fact is that it's resulted in some fairly aggressive federation policies. In general, people on mastodon are pretty okay with this because they value their personal safety and comfort over "free speech" ideals.
The thinking is that your instance has the right to moderate and federate however you want. Having posts from your server federated into mine is a privilege and totally at my discretion. If your server federates with hate speech servers and you don't intend to change that, I'm going to block you to prevent anything from the hate speech server from reaching mine.
The overall goal is to protect users first and allow free speech second. Mastodon users want a safe and friendly place to have fun and hang out. It's meant to be a place to escape from free speech hellholes like Twitter.
So servers defederate quickly and aggressively. Whether or not defederating from your server causes you problems is very strictly a "you" problem. The safety and comfort of users is paramount.
In practice this basically means that while the fediverse is decentralized and all that cool stuff, it's made up of two main groups of instances ('free speech' ones and not 'free-speech') that don't wanna have anything to do with each other. Self hosting an instance for yourself without blocking anything may work, but the second you open it to the public, questionable people are gonna join and retweet questionable stuff from other instances. If you don't take action you're gonna get blocked, and if you take action you're gonna have to eventually de-federate with these problematic instances yourself.
That is not what we are talking about with this sub-thread; the top-level comment carefully argued that people shouldn't do that at all--that you should only host your own accounts and that it undermines the concept to host random people at your domain (which I agree with 100%)--and the comment I am directly responding to buys into the premise and talks about servers my server federates with, not accounts my server itself hosts.
> Self hosting an instance for yourself without blocking anything may work...
Well, does it? The comment I was responding to claims that, if I host my own Mastadon server--which of course isn't going to let randos get accounts... you can't sign up for an @saurik.com email address either (nor should you be able to)--and then have the gall to merely federate with everyone everywhere, I will get blocked by common big servers.
I don't see why or even how this would be the case, but maybe Mastadon has some weird technical thing allowing this kind of weird transitive block which I don't understand. I am thereby trying to get some kind of explanation for how this works and then why people would go out of their way to do it once they figured out how, as it seems ridiculous to me, and I am using the same analogy used for hosting one's own email server to show why.
For avoidance of any doubt, the claim I am responding to--the thesis of the "second paragraph" that I say needed to have more explanation--was that "most of the wider fediverse will block your instance if you federate with instances that host nazis, child porn, etc.". Do you, I guess, disagree with this (and thereby agree with me and my complaint/confusion)? Again: this has nothing to do with letting people sign up for accounts on my server.
If you host your own Mastadon server, you can federate with everyone. Other servers will only block you if you interact with the nazis, child porn, et cetera. If you keep the problematic content read-only then it's gonna be fine. You can hide the list of federated/blocked servers, so no one will know (unless of course, again, you re-tweet, like, or comment something from a specific server).
There is no weird thing allowing transitive blocks, but mind that there are block lists so if your server is inserted in one you will be blocked by a lot of servers at once.
If everyone was as risk-averse as the FT comes off in this article, we would not have any creations of note.
I don't get why newspapers don't use Mastodon as a standard for their comment sections. Social networks are all about credibility. By moderating their users and requiring a fee, newspapers could become the service for credible accounts and reclaim the social network space from Facebook.
Again, the newspaper assuredly did all the homework around building policies around moderating social media decades ago. They just aren’t putting two and two together.
Quote: "It is therefore with relief and regret that we announce the shutdown of Alphaville.club, this blog’s completely unofficial home on the Fediverse. Our reasons are listed below in full but, to summarise, Mastodon has proved more hassle than its worth."
I also think Pleroma or one of its forks (Akkoma, etc.) might be far less hassle to run while still being compatible with the same metaverse protocol, because Elixir (with erlang/OTP underneath) is far more resource-efficient than Ruby is (we're talking orders of magnitude more efficient in both CPU and memory efficiency, although content storage demands would likely remain necessary to keep a lasso on)
At least with Nostr (compared to Mastodon) this won't be a problem anymore, at least some decent implementation of the declaration of independence of Cyberspace.
"the problem is that most major states are pressuring corporations to limit speech."
Corporations want family friendly sites with only happy thoughts.
What you are saying has absolutely nothing to do with legitimate medical doctors being deplatformed because they have an off script opinion of how society should respond to covid.
It has nothing to do with kicking trump off Twitter.
What you are saying has nothing to do with any of the 'moderation' that anyone is actually complaining about.
that made me laugh out loud
Though we all love archive.ph, FT Alphaville isn't behind a paywall. Email registration -- on the far right of the barrier page -- gives unlimited access.
I set up the server during the peak exodus from Twitter. We use Twitter a lot, both to get stories out and to listen to readers, so the risk of it dying was very real. We made the decision to get a minimal-viable-product up asap, to give our Twitter refugee readers somewhere to go.
Urgency to act meant avoiding committees and working the difficult stuff out later. But Twitter survived, 'Don user engagement was meh, and the difficult stuff became not worth it.
The 160gb thing isn't just a Johnny Mnemonic reference. I'm a lazy coder who relies on cludges, so my 'Don deployment was a mess of duplication and surplus. About three weeks after launch the server hit 100% storage and crashed. I had the option of fixing the code or buying more space. Because I'm lazy I took the second option. A week later it hit 100% and crashed again.
Though it was obvious the setup was broken, the long-term strategy of repair always lost out to the short-term fix of buying more space.
We talked about keeping the server going but restricting it to staff only. Ultimately though, it still involved jumping through many hoops to make everything compliant, and while we all like the Fediverse to varying degrees it hasn't yet become essential to our work. We just couldn't justify the time required.
Maybe the FT can be convinced into doing something official eventually. I'm certainly not against it, so long as I'm not involved in the IT or legal side.
was it the sheer amount of content, or just that it was outside the sphere of support (both legal and technical) to be useful?
They really do have some of the ingredients, but they are missing key ones. I'd say the biggest limitation to building a social platform from even as rarefied a captive demographic as their esteemed readership is that when your livelihood is based on relationships and narrative over concrete physical skills, the stakes on being controversial are too high.
If you are a financial advisor and you are controversial in public and lose, among peers whose stock in trade is being aligned to the highest powers available, you lose your credibility and place in your pecking order. Whereas, if someone decides I'm on the wrong side of a narrative, I can still write code, build something, or make it work.
Opinions are what we have when we're not actually doing the thing we have the opinion about, so it's my indulgence to be provocative about how well the world is being run - because I'm not the one doing it. All the world is indeed a stage, and getting all the best lines is almost fair compensation for having to put up with its managers. The issue of our time is that the people running the world think people like me have a bit too much freedom to mock their degenerate incompetence, and they think controlling public forums is going to be easier than doing a more credible job.
There is an iron law about the trade off between autonomy and power, where power is to act through others, and autonomy is to act without others. Quality discourse (social media content), and news stories, require some kind of friction or conflict to make them interesting and compelling. This disqualifies powerful people from participating as themselves because the risk of alienating the people through whom they act is too high. (Musk is the exception that makes the rule.)
The "social" in social media means that it's for kids and plebes like us who can afford to have drunk pictures of themselves on the internet because we aren't engaged in the all-against-all political power struggle that defines elite competition - the world most FT/Economist readers inhabit. My opinions have nothing to do with my ability to fix computers, where for the typical FT reader, their opinions signal their alignment, status, and reputation. They can't risk their reputations on making the kind of piquant online comment that makes this all so good.
I'm glad they learned you can't just "start your own" social media platform, but that wasn't the real obstacle for them. It's that, they'll never (shit)post like common people.
This got me wondering, I created a Mastodon user @ defcon.social and my first couple of posts were images, just to test things out and get started with my chosen client - Ivory by Tapbots (the creators of Tweetdeck, RIP) after tptacek spoke glowingly of it here. Then I realised that by posting those I'm basically taking up space on a community service without paying for it, and I wasn't sure exactly what the etiquette was. On a commercial, monetised service like Twitter it's different as there are advertisers (or at least there were...) but afaik "defcon.social" is just a non-profit and I'd really like not to be a big cost sink.
Also I wondered if there's any overhead involved in following users across many different instances - if my @defcon.social user follows a bunch of users on various other instances does that place additional burden on their mastodon instance or just on my local client?
Might be time for me to read the defcon.social fine print and T&Cs (I only glanced over it) or even about Mastodon/ActivityPub generally...
So if 10,000 people on server A subscribe to someone on server B then server B will send 10,000 duplicate messages to server A.
I hope I’m wrong but that’s what I’ve heard.
Anyone with expertise here who can speak to the scaling of ActivityPub?
On each individual server they are ideally saved based on their unique identifier (which are RFC3987 IRIs) and I doubt that there is any software that will do it individually for each inbox instead of storing one copy to which each inbox links to.
To minimize even this, each server can store just the IRI(which is mandated to be de-referenceable) and load the payload when prompted by UI. I doubt that anyone does this also.
> This program posts news to thousands of machines throughout the entire civilized world. You message will cost the net hundreds if not thousands of dollars to send everywhere. Please be sure you know what you are doing.
From the rn usenet client, back in the day.
This is why we can't have nice things: https://en.wikipedia.org/wiki/Laurence_Canter_and_Martha_Sie...
[0] https://www.w3.org/TR/activitypub/#x7-1-3-shared-inbox-deliv...
[1] https://github.com/mastodon/mastodon/blob/a5a00d7f7adff5e0af...
It caches all media that you view, and I assume it might prefetch stuff to. It has a tool that prunes out old cached images and such. However that script isn't setup to run automatically (or if it is, its far too loosey-goosey)
However the biggest killer is that it doesn't clear cached user page header images. Here is the output of my cache. Headers are not cleaned, and need to be deleted manually.
Attachments: 1.43 GB (8.19 MB local)
Custom emoji: 61.9 MB (0 Bytes local)
Preview cards: 344 MB
Avatars: 3.01 GB (71.7 KB local)
Headers: 6.69 GB (172 KB local)
Backups: 0 Bytes
Imports: 0 Bytes
Settings: 36.2 KBMy personal instance has been up two months with 12 GB space used, but only 88 MB of that is my stuff; the rest is just the cache of others' posts. That can be fairly aggressively cleared out, and I think Mastodon lets you configure an automatic purging of this data.
I don’t allow people who I don’t know to make accounts so I haven’t had any spam or abuse trouble.
I haven’t had to deal with any downtime or technical issues since I got it running, which is a bit of a disappointment tbh
The usage can be contained with a few cron jobs to purge old cached content aggressively. Running
`tootctl statuses remove && tootctl media remove` once a week is enough to make it sane in my case.
Every rule (if you already have them) will be broken, and you will soon find yourself surrounded by a shitton of moral and legal dilemmas. Not to mention the insane amount of work and attention that has to be dedicated to it. It's a thankless job that never ends. Also, brace for people who have a different cultural, moral or legal understanding who want to argue against your (imperfect) decisions. It's a lot of responsibility and potential liability.
gotosocial: du -h | tail -1
3.8G .Bullshit! I could do that, and I'm broke!
Thats the entire point of ftalphaville. Its a great read, and free, barring email signup
I mean it's a common thing to not have end-to-end encryption. Gmail, Instagram, Discord, all can read DMs of logged in users.
Something being open source doesn't mean it's end to end encrypted.
Only some messengers have end to end encryption, like WhatsApp, Threema, Signal, etc. There are plans for mastodon to get E2EE as well, there is some backend work done but it seems to not be used in the mobile apps yet.
https://www.bbc.co.uk/news/world-us-canada-62688532
https://en.wikipedia.org/wiki/Twitter_Files
Democracy is a bit of a prisoner's dilemma.
There will always be people who want to take away your freedom of speech; especially other countries. Any exception to freedom of speech basically can never be weaponized. "Hate Speech" is readily and easily weaponized against your political opponents.
Everyone gets the right to speak or nobody does. Which is what we have right now. I realize where this post shall go.
A democracy where you cannot speak to your political opponents is a democracy doomed to collapse.
I've never seen anyone say this before. Could you elaborate how? Like speaking over someone so they may not speak?
That's one of the reasons the whole hate speech topic became an issue.
How?
edit: you've just picked another (tautological) form of speech for which you're not explaining how it suppresses the freedom of others to speak. The question-begging word "bullying" here is the hint. The "bully" would refer to it as commentary, not bullying.
Can Donald Trump be bullied? Can a racist mass shooter? "Bullying" is a value judgement.
I think people who kill themselves are pretty much finally suppressed in their freedom of speech.
If everyone is free to speak their minds, how can it be used to silence others?
Even more improbable in the context of cyberbullying. Tyler the Creator's cyberbullying tweet appears to be more relevant than this take.
There aren't many people for whom this would have no effect on the nature and quantity of their comments.
For that, speech is usually more than enough, especially if performed by a large ingroup against a small outgroup.
In terms of paradox of tolerance. Even in the USA, there are limits on free speech.
https://en.wikipedia.org/wiki/United_States_free_speech_exce...
You cant advocate for violence in the USA. Not only is this not protected by free speech, it's a crime.
The assertion of "tolerance without limit" doesnt apply. There's your limits right there.
Elon's twitter never changed this rule. You may not advocate for violence on twitter. This is not in contradiction with Elon's free speech absolutism.
What's important to understand is say Reddit who has eliminated free speech. What do you earn? You earn echo chambers, you divide your populous. You never solve your problems. Each side sees only the worst of the other side. Nobody talks to each other again. Political polarization is 100% because of no free speech on social media.
>You're posting too fast. Please slow down. Thanks.
I'm shadow banned and flagged on HN so I can't respond to you until it's lifted. Guess it was more than a 1 hour ban. Imagine attempting to have a conversation on HN and not being able to reply
Good ole -10, goto jail for me. Thanks prisoners dilemma.
How ironic, but not surprising. What a waste of time. So long and thanks for all the fish.