Is knowing what is censored valuable information? Would this plan work?

[angertard]

Spoiler: blah blah blah funny tweet I saw like 5 years ago

So I was thinking about this funny tweet. It was something like

You are gay

And it had some characters at the end which just looked like an ascii imitation of fairydust, particles, bubbles or something.
And twitter slapped it with

Get the Facts about Covid-19!

Which was funny! The algorithm they were using to detect covid-19 disinfo did something basically just like checking boxes, and some of those were triggered just by all the ascii characters at the end. The filter was basically smacking a warning label down on anything that had words or synonyms for things related to the virus, respiratory symptoms, or 5g networks. I would argue this special attention is because the conspiracy theory most likely to be harmful to twitter's business (internet content distribution) was people blaming 5G networks for covid symptoms.

Spoiler: double blah blah, about internet peering

Companies like netflix highly prioritize peering with internet service providers who can provide high quality reliable transit, because even just a few artefacts or a moment of buffering ruin immersion. You can imagine a similar thing with twitter, right? If your internet buffers, your scrolling reflex is interrupted and you close the tab. Content hubs have huge benefits to reap from every new generation of internet technology.

So "oxygen" and "frequency" in a tweet could get a warning box, no covid mention required. One of the ascii characters in the tweet was a degree symbol, which I think was somehow labeled as related to the 5g networks, a lowercase letter o must have registered as being about oxygen.

Anyway, I was thinking: if google censors information, I could test what information they automatically censor and use that to either A. understand what type of thing they censor or even B. get guesses at the specific content of censored information confirmed.
Now I'm not sure if B is feasible at all. Like in the tweet example in the spoiler above, the conspiracy was probably chosen not because it was true but because it was most inconvenient. If I pitched a government agency on catching and hiding classified information whenever it appears on the internet... well, I'd imagine they'd give me a hard time when I tried to put the stuff on a server, even just by interacting with one of google's servers or something. If they did say, "Yeah, censor anything about the fitbit antarctica fiasco" I don't think they'd give me specific military base coordinates to censor, they'd just want me to censor all proposed coordinates, excepting maybe one decoy location of a legitimate and publicly known facility just to confuse and delegitimize where possible.

Then again, it isn't like this sort of thing (clever hacking based on a poorly considered release of information) is unprecedented. For example, I remember there was a twitter user who used natural language processing to data mine mentions of codenamed operations in publicly released government documents, clustering them by their context in documents where they were mentioned. I didn't dig deep into their results, but it shows that this kind of thing can work, when people don't think about all the implications while designing a process.

Is there value in type A censorship knowledge? Like, would it be interesting to know ALL the no-no terms that the chinese great firewall has configured in their cisco-provided censorship boxes? "Wait wait- I know about Tiananmen square, but what the hell is this stuff about a ______ _______?" Also wikipedia says Cuba, Iran, Vietnam, Zimbabwe and Belarus do similar things, it would be cool to see if any of them have slipped up and censored something currently obscure. Or like I mean, it would be a total bombshell if google was suppressing something like net neutrality (now that they've rolled out a bunch of fiber) or specific security/privacy concerns, right? Even if it gave us no clues about the validity of concerns/stances, it would at least look really bad.

On the other hand, boring "we knew that" censorship is just a widely known standard. Blackhatworld (SEO forum) shut down all chatter about the coronavirus on their site because google didn't like it. "Hey guys, don't mention the 'demic..." Not a big secret, they just couldn't handle the heat. Skip scanning these terms.

Spoiler: tech details

I think you could do china-style firewall term scanning pretty cheaply if I understand how it works correctly.
1. set up websites with a bunch of almost-plausibly censorable text on the external internet
2. rent a virtual private server inside censored country to be probed, e.g. China
3. try to reach some sites, if you can't get through then log what the site contents are using your non-china internet access

For step 1, I've seen weird websites show up in searches where the words are there for no plausible reason, I think they might be communication channels for botnet command and control? Anyway the point is you don't have to roll your own gibberish sites just to start trying this out. (I don't suggest going looking for these on your own computer, like since it's malware related?) Fresh sites are required every time, because the firewall remembers if you've been a bad boy- so you can't just trust that because a site is censored that something on it right now is a trvth nvke, it could be that they posted a trvke long ago or something. But presumably if a site wasn't blocked and then it became blocked, that would be a faster way to pinpoint "oh, they just tripped a censor tripwire"

Now with checking google for censorship, honestly I don't know. It isn't clear to me how hard it is to get a site put up high in a search. Difficulty can be broken up into two issues: consistency and expense. I can always pay a guy from new delphi to make a new website but how much does it cost to get backlinks or whatever so bing or googly will put it in a search for "Bill Gates Bolivia vampire double rape suicide" but if it really costs a lot that could suck. It also would be useless if backlink quality is just hard to pin down and I think that one of my links is blocked 1/10 times but on another trial it just turns out that I didn't have enough backlink clout or whatever. I guess if I think about it for a few minutes I can optimize this, it's probably optimal to over-boost your sites just so you minimize the chance that your trial was a failure for that reason.

Actually, you could just boost the site again after the first failure right? If it keeps out the search after the google bot has re-scanned then you know they aren't letting you in even with a bunch of clout. Or maybe the google bot would just not revisit sites that are suppressed for censorship purposes? If it does that, then that's the tell and that's actually a little more scalable than renting a residential proxy (so you can search google a bunch) because you would just need to check the logs to know that they're discriminating against your clickbait shit site.

Would google just manually check if you are a gibberish site and skip censoring you if you are an obvious nonsenseGOD? Running a small large small language model to generate more realistic content would be more slower, but they're the ones who have to deal with this stuff at internet scale so at least it is assymetric in my favor. Also they might have automated it with AI which would be easier to cheat, and checking if you are gibberish is a measure that only works if they think someone is doing this and they want to defend against it.

Okay post out

 

[MySpace Tom]

That type of knowledge is always valuable. People have uncovered some censorship subjects/tactics in the past, but given how often topics and censorship methods change, experiments would be interesting. However, you'd have to be careful given that shadowbanning tactics are pretty commonplace, so you'd have to make sure that your tests aren't still being made available to you while not available to anyone else, or even that your experiments aren't linked together in such a way that the inverse happens like in your tweet example

This reminds me of the censorship in Google Now (Google's search autocomplete service) that Project Veritas was shedding light on a few years ago (back before James O'Keefe's position was usurped) where Christian and conservative topics and sites were being censored as well as certain phrases linked with politicians. This was visible when trying to type in certain keywords or phrases. The autocomplete suggestions would hide results for some of these topics or would provide some unusual suggestions. This article on The Christian Post covers some of this censorship

Anyways, I hope your experiments go well. And good luck publishing your results when you're finished since the result could be blacklisted itself

 

[Ninetails]

(I'm going to try to reply with a sufficient answer, but my concentration has been broken multiple times due to my cat was pestering me. Therefore, this may come off as more redundant than I had initially planned it to be.)

So I was thinking about this funny tweet. It was something like
And it had some characters at the end which just looked like an ascii imitation of fairydust, particles, bubbles or something.
And twitter slapped it with
Which was funny!

I'm not sure if I'd consider Twitter's Fact Checker to be censorship. The information is still there and not removed, but you will be fact checked by an automated re-educator. If the post was removed for posting an emoji, then maybe that would be considered censorship.

For something like China, (or any foreign country, really) what language you use to search for things is a pretty huge factor. I only know a little bit of Chinese, and I've only tested the Baidu search engine using Chinese terms. The firewall doesn't block against as many English sites as they do with Chinese language sites, because most Chinese people don't know English. IIRC, about only 1% of the Chinese population know English.

Likewise, I've noticed that Google (and Duckduckgo) tend to punish you if you search anything related to Russia or Russian culture.... at least, in English. If you search any name into Google, like "Vladimir ____", and if it isn't an extremely well known figure, it will just redirect to articles about Vladimir Putin. (EDIT: I should mention it's been a few months since I tested it, and I don't actually remember which names I exactly used to test this. YMMV.) This doesn't happen with Yandex, even if you are typing the name with roman letters. Now, I haven't tested this with Cyrillic letters on Google or DDG, but if I had to guess, it wouldn't lead to articles about Putin because of the characters used. I wish these search engines didn't do this, but it has at least given me another reason and motivation towards learning a different language. I don't actually 100% know if this is intentional censorship, or just the fact that Russians won't be searching in English and/or roman letters in an American search engine, and thus will just redirect to English sites with more pings. Perhaps a little bit of A, and a little bit of B.

In summary, language barriers provide a sort of "soft censorship", and large organizations know this and use this to their advantage.

 

[Ian Moone]

You're doing something exceptionally interesting here. The censorship apparatus is extensive and fiddling with it will definitely warrant some attention from likely undesired parties. Yes, of course, you can play around with the filters, learn about their triggers and find ways to exploit what's being triggered. The most effective means of playing with these levers is always going to be via social vectors. The idea that you or anyone like you, outside of equally exceptional skill with network security, can do this digitally, by playing with the 1s and 0s at play, is a far cry. You are on to something with how you approached the sandbox, however.

It might be fun to find the right words to explain more, but I'll refrain for now. I may have about five months ago, though Blue is a slightly prettier color these days.

 

[K0WLOON]

One is unlikely, as a regular person, to stir the waters enough for the gogliathan to turn it's gaze towards thee... What is a likely thing is that one pisses into the ocean of piss and the thing goes unnoticed (because, even with the vastness of the Principality of Googleon Ground Forces, the ones are likely prioritizing certain things more than others and, all due respect, I have a sneaking suspish that you are not very high on the list). Whereas twitter is only scanning it's own internal dataspheres, Google scours the Web, not merely giggle.net or whatever. It is like reading one's personal library or one trying to read the entire country's libraries. And how would one differentiate between merely being too small to notice vs being censored??? It is a thing that Google has greatly decreased efforts at indexing small, non-commercialized websites as often as in previous decades (one has an HN post about this but it is lost in the digital junkdrawer somewhere)

 

[Some_porcupine]

my head hurts reading that (got brainrot i guess...)
you mean "this pic"?

 

[punisheddead]

rent a virtual private server inside censored country to be probed, e.g. China

I doubt that in most cases you'll need a VPS in a censored country as most of the firewalling setup is only one sided, to block outgoing connections rather then incoming ones. If a country is so heavily restricted that it restricts outside connections, it probably wouldn't offer a VPS you can buy to then access said censored information. If you want to probe the firewall, I'm pretty sure most corporate grade VPSes are exempt from the the firewall because they often deal with international clients.

The bigger problem would be, for example in this case of China, is how they limit access to many platforms to only those in the mainland. You would need a Chinese phone number at the least and their ID at the most to register accounts in many places. A VPS would not help in this case. There are ways around this, but they're risky and cost money. I doubt anyone would want to go that far. Some platforms even institute tests that you need to pass before being able to access most of the platform and engage with it, something that again can be bypassed but is difficult.

This all depends on how far you want to go. You can access a CN search engine anytime, same goes for a RU search engine, only thing you need to know is the language.

 

[Some_porcupine]

relevant threads

• 

  Thoughts on online-community operation?

  Let me start out by saying that I'm not here to pick a fight with any of the Agora Road administration or moderation. If you've got a bone to pick with them, do it somewhere else, this thread isn't the place. I've run and moderated my own fair share of discord servers, and similar online...

  forum.agoraroad.com
• 

  Operation Preinstall Minecraft: Creating Self-Replicating Decentralized Cells to Combat Gaslighting AI Propaganda & Oppressive Governments

  This post was originally created on 4chan, the TLDR version is posted below. Important to note, Ideology does not matter for this. :hspin:Agora Road is Apolitical. Posted here for archival purposes and reading entertainment. Please Enjoy.:hspin: Here is some discussion on the validity of...

  forum.agoraroad.com
  • (https://forum.agoraroad.com/index.p...op-or-a-legitimate-resistance-technique.3945/) (https://forum.agoraroad.com/index.p...no-proof-of-but-still-believe.4091/post-67297)(https://forum.agoraroad.com/index.p...g-them-down-financially-and-spiritually.7139/)
• https://forum.agoraroad.com/index.p...iency-privacy-anonymity-thread.7557/ [/LIST]
  • • needs more self-sufficiency or similar thread(s), i.e. if i forgot. something on type of ("mc"-one above and: )
      • https://forum.agoraroad.com/index.p...-the-dead-internet-and-fight-against-it.7613/
      • https://forum.agoraroad.com/index.p...-and-the-fall-of-labor-power.5971/post-122088
        • https://forum.agoraroad.com/index.php?threads/resisting-the-technocrats-general.6020/post-107532
          • and related, i.e. one (thread) about homesteading out here

---

antithesis

You are dumb because you want.

Think about it. Just for a second. In medieval times, learning was a privilege of nobles and wealthy families, and even they weren´t able to access to certain specific books kept in monasteries under the hands of jealous monks. Later on, with the invention of the printing press by Johannes...

forum.agoraroad.com

---

What we could have:

• WhiteSnake 46 minutes ago​

We are taught not to be connected to the world around us, when we come from a westernized society. It's cold and dead, layers of concrete. Lawns instead of natural ecosystems.

We are part of a web, interdependent and connected. But rarely few ever see this in any genuine way.

Why do the people who want to rebel against a corrupt system so bad, never consider the most practical things first? Before burning the building, try scattering local seeds in secret that will grow too fast and root too deep to be removed. Before bombing a car, consider first tearing apart a road and cluttering the space with beautiful and giant shrines that are easy to start and hard to remove.

For every destructive act, consider first what constructive act you can do that will have the same effect.

For the would-be member or cluster group of Operation Preinstall Minecraft, the Author recommends thoroughly studying Gene Sharp's 'from dictatorship to democracy'. Nonviolent warfare is impossible to contain, but becomes ruthlessly effective when spread decentralized everywhere.


With blessings,

White snake.

Attachments​

• 
  tis-the-season-v0-0bnya1wt99sd1.jpeg
  253.7 KB · Views: 1

BarnSwallow
Build a parallel society, not on lines of religion or ethnicity, but on the peaceful rejection of culture and institutions that harms us. Even if it's a society of one, build it strong so you may flourish.

Also, feds fear the parallel society, which is absolutely based.