Egregoros

@strypey @p @icedquinn @fish @sampler @silverpill @light Eventually that will be the case when they decide to do something dumb. A good example of what is to come is the quote "consent" they decided to do. Nobody but Mastodon cared to implement it fully yet and those that implemented it, implemented it partially so quotes actually work and plan to use the "consent" revocation differently than what Mastodon wants.

If they make a change that causes a full on network split that is hard to deal with, few developers would care enough to fix it I think. Same with ActivityPub 2.0 being backwards incompatible. Before this thread I genuinely don't remember the last time I had a pleasant interaction with someone using Mastodon that isn't from this Fediverse part. Loosing Mastodon means nothing to me.

@strypey @icedquinn @fish @phnt @sampler @silverpill @light

> Ignoring what the implementation used by the vast majority of people in the network is not how protocol standardisation works.

Neither is allowing them to dictate the protocol by letting them ignore the process and then forcing other implementers to play catch-up. At some point, you just let them do whatever and give them "best effort as long as it's not too much of a pain in the ass".

> But you do you.

I can't be stopped.

@icedquinn @p @silverpill
>mastodon't is inherently a void screaming platform.
Mostly because Twitter is also that and Mastodon wants to be Twitter. You can't have a meaningful discussion in 500 chars, because it limits your expression into sentences with keywords more than a proper way of typing out your thoughts. Or you are forced to do 🧵1/♾ like has been the case in this thread. People made virtual thread composers that let you write a post and then convert it to 🧵1/♾ once you are finished. And people made the inverse for people that are used to writing 🧵1/♾.

Doesn't help that 🧵1/♾ people should at some point just write a blogpost, because nobody is reading a twenty reply thread.

@silverpill @icedquinn @fish @sampler @p @strypey @light

https://git.pleroma.social/pleroma/pleroma/pulls/7875

This does not implement anything meaningful in the spec and only makes quotes work on Mastodon (because they even fucked up that which worked for years before their attempt).

Trust me when I say that if the proper workflow as envisioned by Mastodon gets implemented in Pleroma, it gets patched out by most larger Pleroma instances almost immediately. Myself included, so if Nicole doesn't make it a toggle, I will. Nobody but Mastodon wants this, and I don't want the Twitter copy of hidden replies either.

@silverpill @fish @icedquinn @light @p @sampler @strypey
>Myself included, so if Nicole doesn't make it a toggle, I will. Nobody but Mastodon wants this, and I don't want the Twitter copy of hidden replies either.

That is not to say that I hate the "feature" completely. To me Pleroma is about choice, so if the Mastodon way™ is implemented (which it probably will), I don't have a problem with it as long as it is an optional feature, because clearly someone wants it. Same way signed fetch is a toggle. Doesn't really matter which way. Disabling it in the backend and returning revoked/non-approved quotes as normal instead of some "hidden" field in MastoAPI, or having a toggle in the frontend that makes "hidden" quotes the normal ones.

@silverpill @icedquinn @fish @sampler @p @strypey @light The Fediverse part I care about does not include Mastodon as I've said above. I loose almost nothing if they stop federating tomorrow. So if they want to split from me into some normie mainstream network, their loss, I didn't care anyway. This Fediverse will continue to exist and arguably be better without them.

But to answer your question properly in the perspective of someone helping with server development and not a user, I've been vocal about the issues with new proposals as you've probably seen many times already and I will continue to do so. But it's not like someone at Mastodon actually cares about criticism, they are the GNOME of this network. W3C is a different beast and so far has been focused on things nobody cares about (C2S and OAuth2 discovery for C2S). If they decide to do something stupid, like pushing more JSON-LD, I've been also vocal about that, but not like I can do much than that when Evan blocked me over the grant disagreement and he's mostly the one running the show. @a is the only one sane there even with the JSON-LD enthusiasm. So far they only introduced E2EE which is still half-baked and not a proper spec anyway (key distribution is missing completely), while being something very few will care about I think. If you want E2EE messaging, just use XMPP or one of the other 5+ apps for it.

If incompatibilities arise, like the quotes one, they will get fixed eventually like I've done previously with other servers. The quote issue has been very down on my list of things to do and I didn't have a good idea of how to fix the UX, so that's why I didn't touch it at all. But if the incompatibilities are too big, at some point you have to stop making excuses for the bully and just cut them off. They can't act like the sole arbitrator of the network/protocol and everyone keep bowing to them. In that way, the hidden quotes is a similar hijack to what they did with the subject/description field. Taking what they made and altering it to something they don't want.

At worst, you can always take ap-next, add some FEPs, strip JSON-LD and call it LitePub 1.0.

@silverpill @icedquinn @fish @phnt @p @strypey @light
The nice thing about quote retweets is they let people repost without necessarily putting it in on their timeline. If your account is followers only and all your posts are followers only but you retweet public posts there is still data leaking.


Pleroma slept on the implementing quote retweet, so I even though it took Mastodon like 6 years to add them, it's our punishment that we need to stick with their way of doing them because we dropped the ball.

@sampler @silverpill @fish @icedquinn @light @p @phnt @strypey

> If your account is followers only and all your posts are followers only but you retweet public posts there is still data leaking.

I'd be cool to give people more fine-grained control of scopes at the user-level, and I don't think anyone can force you to federate an RT, but there's no foolproof way to solve "data leaking" when something has to leave its origin. Superficial ways of making things not show up might be enough to satisfy most people: personally I'd like it if I didn't have to look at one-sided conversations on my timeline because it's private and I don't follow everyone involved. But people are bound to complain when they find out something doesn't *really* work the way they expected it to. And that's what's going to happen with Mastodon's proposal.

> it's our punishment that we need to stick with their way of doing them because we dropped the ball.

I don't like the idea of being punished over a feature I never liked anyway. I think QRTs are stupid and if anything I'd gut them out, but I get this is mostly a culture-clash between people who treat fedi to varying degrees more like Twitter and people who don't. But at least QRTs as they are now work exactly the way you'd expect them to, and Mastodon wants to do something else that won't.

@sicp @icedquinn @fish @sampler @p @strypey @silverpill @light Just the fact that people complain about leaking data on this network is proof that trying to prevent it is a never ending battle. Truth is, you cannot enforce what is effectively DRM on your posts in this network, this network is purely operated on implied trust everybody puts on all instance administrators without even thinking about it.

https://evilmaid.net/blog/trusting-trust-fediverse/index.html

Also related to more scopes, Pleroma develop now support scoping of repeats ported from Akkoma. IMHO a dumb feature that shouldn't exist, but to each their own.

>But people are bound to complain when they find out something doesn't *really* work the way they expected it to.

Once again, "We must not claim to prevent what we cannot prevent." Warnings that posts may leak when trying to compose a lockpost or a DM are the only solution to that. Users will freak out, but that's the reality. You cannot lie yourself out of reality.

@phnt @icedquinn @fish @sampler @p @strypey @silverpill @light

> https://evilmaid.net/blog/trusting-trust-fediverse/index.html

Good writeup.

> Also related to more scopes, Pleroma develop now support scoping of repeats ported from Akkoma. IMHO a dumb feature that shouldn't exist, but to each their own.

Neat. I also think it's dumb but I figured it's the kind of thing people'd want anyway. People can do dumb things as long as I don't have to join them (QRTs, private posts)

> You cannot lie yourself out of reality.

Again as long as I don't have to, if they want to screw everyone else out for choosing not to drink the kool-aid then I say let them: I share the sentiment if there was a big netsplit I can't imagine shedding any tears; nobody can make me upgrade anyway and I think its funny there's an overtone on their side of a netsplit being a worst-case scenario whereas over here most people are indifferent to the idea.

@silverpill @phnt @icedquinn @fish @sampler @strypey @light

> Like, what's the plan? Let Mastodon/W3C do their thing until incompatibilities accumulate and the network splits?

They are not meaningfully a part of the fediverse at this point. They have made a Twitter. It is no longer a platform that you use to communicate with friends: it is a broadcast-only circle-jerk and when it fails to be a broadcast-only circle-jerk people that want to "control the experience" and that can't distinguish between startup dipshittery and tech built by people to communicate with each other. Look at the early posts by the hachyderm dipshit, that guy's like "Oh, this isn't going to work, we have to eject these people and we have to make it advertiser-friendly because if we don't then fedi will never replace Twitter!" I don't want to replace Twitter with a shitty clone of Twitter: I want to replace Twitter with something good.

I intend to have a platform to communicate with friends and strangers. I don't really care if Mastodon decides to create a desert and call it peace. They have fundamentally different goals. Nobody complained when Gab fucked off (and they were, in many cases, fucked off before that: https://pastebin.com/E0k5fcd6 ), but somehow Mastodon is important? Go look at what they're putting onto the network: https://mastodon.social/explore .
gab_fears_the_fse2.png
gablins-fear-fse.jpg
trannydemonhackers.png

@Mammal @fish @icedquinn @light @phnt @sampler @silverpill @strypey Nothing of value was lost.

There was a lot of the same hand-wringing: "they're the most active instance!" and it's like...do you want to see that activity? For at least a month, the *majority* of their activity was a single account that posted three times per second. I saw the bot, killed the bot, could not tell Gab about the bot because Gab blocked FSE. (Gleason was working at Gab at the time, I told Gleason and offered to help Gab quell the bot problem and told him to pass along the offer to Colbert; Gleason agreed but Colbert says Gleason never relayed the message. At the time, it was in my own interest to stop the flood at the source because Gab was essentially DoSing FSE by constantly flooding us with bot shit. I got to do k-means and come up with interesting ways to automatically detect bots, I was having fun.)

And now it's handwringing, what if we lose Mastodon? What *if*? They're as bad as the Gablins. There are people using fedi to communicate and there are people using fedi to broadcast and try to push their agenda and, you know, fuck the latter category.
FSE vs GAB.jpg
blockfse.png
blockrecommendation.png
fse--subversive_neonazis_and_communists.jpg
gold.png
high_iq.png
make_fediverse_great_again.png
whole_mentality.png

@Mammal @fish @icedquinn @light @phnt @sampler @silverpill @strypey The main issue with federation was that Gab was cramming shit in that no one wanted and they had the clunkiest UI (which Gleason turned into Soapbox, which is still the worst fedi UI) so they were bleeding users into fedi. They had to cut it off to stop people from going "I can still talk to my friends but I don't have to use Gab?" and hopping to a sane instance. Regular shit on fedi was a "pro" feature on Gab, they were suffering massive downtime, they botched federation (by putting Gab on as many relays as would take them, so they were spending all their bandwidth delivering bot posts to the rest of fedi) and they covered it in ads for whatever low-rent hustle Torba was doing. They expected that they could use fedi to get people to join Gab, based on the way they were talking, and when it went the other direction, they had to kill federation.
torbasnitch.jpg

@p @icedquinn @fish @phnt @sampler @strypey @silverpill @light Gab was under intense pressure to be able to stay online after multiple hosts deplatformed them over allowing total free speech, aka criticism of zog, woke etc. Gab was a pilot program for the regime to eventually produce today's post-woke musk version of twitter. Free speech was winning the argument against the regime shills on gab so its federation was killed to recreate a scenario where gatekeeper guys could maintain authority.

@Mammal @fish @icedquinn @light @phnt @sampler @silverpill @strypey

> Gab was under intense pressure to be able to stay online after multiple hosts deplatformed them over allowing total free speech

"Complete freedom of speech unless it is anime or you ask questions about glowies or you notice how many bots there are or you are FSE, which is a hive of subversive neonazis and communists, or you point out a hole in Epik's VPN or you investigate Gab's finances or you point out that deleting your account silently fails if you have been verified."

They weren't kicked off for freedom of speech. They were kicked off because they antagonized literally everyone in order to prop up the persecution story. I met Utsav--having a PR guy that couldn't stop saying gamer words to journalists when they declined to go to his hotel room was a really stupid move to begin with--and half the sentences that come out of his mouth start with "We must fight the". They fork Mastodon and their announcement is them shitting on Mastodon. They fork Brave and they shit on Brave in the announcement. They were insular when they weren't antagonizing literally everyone. The only reason anyone heard of Gab was that there was a "left-wing bias in the tech press and Silicon Valley" thing going around in 2015 and so after Torba had Gab built, he started spamming an HN networking board (that is, a private board for people that had been HN-funded startup founders at some point) with "BUILD THE WALL I HOPE YOU ALL GET DEPORTED TRUMP MAGA 2016" and ignoring "Please stop that, this is a business board, not a politics board" until he got banned. That is, he committed suicide by janny in order to have a hook to call up journos and go "They banned me just for being a Republican! That is why I am making my own Twitter that doesn't have the left-wing censorship" and do a bunch of interviews. Literally no one would have heard of Gab if that PR push hadn't succeeded and that was the only move in his playbook: he shits on people so that he can complain about their reactions. It gets attention but it ensures that he has no friends, so when the social justice pitchforks come out (because he spends time antagonizing them too), he has no one willing to say a word in his defense: if you're in a position to do something about it, Torba had either burned you or burned one of your friends.
animeavatars.jpg
billstclair_v_pussyslayer2.png
fse--subversive_neonazis_and_communists.jpg
gold.png
gtfo.png

@p @icedquinn @fish @phnt @sampler @strypey @silverpill @light @Mammal I never took any of this stuff seriously 6 years ago. A lot of us were younger and thought this was all dumb drama about shitposting and blowing off steam when these people were trying to genuinely trying to subvert American culture.

But these days Torba seems to be exclusively running in these weird astroturfed "trad" circles whose only consistent position is shilling for random 3rd world countries. "Save the White race by siding against White countries no matter what the issue. In fact, don't even look into the issue. Being a misinformed vibes based gonzo idiot is BASED!"

It makes me wonder if all the deplatforming from those days that everyone said was "leftist" was actually targeted at people receiving foreign money.

You look at these types of people from back then and where they are now and they're all shilling for Orban when that election was totally irrelevant to everyone except Hungarians and Russians, they all have Russian wives for some reason, or they're promoting pro China, pro Iran, pro pan-Arab propaganda.

Basically anyone that has influence that isn't in the core American geopolitical coalition, the old Far Right influencers are carrying water for those people. I'm thinking more of these Torba types than we thought are just paid agitators of foreign governments. Even back then a lot of them were being promoted by RT.

@StarProphet @icedquinn @fish @phnt @sampler @strypey @silverpill @light @Mammal

> Gab used to have a bot account called “jsheistydiety” that posts hi-res screenshots of low res boomer memes at multiple times per minute.

So, that's something I saw; I suspected that the Gab Trends was gamed so I did some analysis but it was really difficult to get some sort of clean view of Gab Trends (clickwrapping plus downtime plus their shit was difficult for both humans and bots to read). But I found this group of accounts that would all post the same set of Pepes and then randomly they would post the same link as all the others did, and then that link would show up on Gab Trends.

There was some recipe bot that would shit recipes scraped from recipe sites and the guy was claiming to be a Vietnam War sniper (his account was something like "sniperss"), right, a member of a unit that didn't exist during the Vietnam War, and now he's a "hotel chef" and he purported to be posting his own recipes but he wasn't even using the same units consistently; his top posts were all spelled and punctuated properly but his replies were all weird shit definitely not written by an American and he'd end them with "thank you,,david", weird shit. So I noticed a lot of accounts had that pattern, their top-level posts were

The most reliable indicator was that the bots would engage in some really easy-to-analyze behavior: they'd post uniformly around the clock (humans sleep), the overwhelming majority of their activities would be posts/reposts. That is, very few likes and very few replies to people that didn't tag them, very few follow/unfollow/block; this is expected of spambots, right, they aren't going to put any effort into something that doesn't put something on your screen, but they would reply when you interacted with them because it's worth their time to evade detection. So their replies read like they were written by offshore clickfarm dudes, like someone was running a few hundred bot accounts and then they'd send the replies to some sort of centralized place, right, like so they could reply "no im not bot" or "thank you,,david". You could plausibly run a massive farm like that, like several hundred accounts per guy. (They didn't have an obvious commercial interest, so it's possible that people were right with their suspicion that Gab was using clickfarms to look more active than it was.)

But the behavior made it easy to detect them: I wrote a little script that would look at all of the activity an account did, look at the ratio of normal activity to broadcast-style activity and how flat their activity was over 24 hours, and then draw a sparkline graph so I could visually inspect. This was really helpful for identifying bots, no ML stuff required even. And then once that was easy to do, I could do k-means clustering to see who was following them, and I iced about a hundred accounts on FSE that had posted just gibberish or nothing at all but were following only Gab spambot accounts. (I hated the spam and it actually cost me money by forcing me to upgrade boxes on Frantech, but it was fun to come up with bot-detection stuff.)

> And it had a comped yellow circle during imagocaust 2024.

I don't know what that is, like I heard something had to do with making image-posting a "Pro" feature, but I also don't really have any interest in Gab except where they intersected with fedi.