TruthArchive.ai - Related Post Feed
@elonmusk - Elon Musk
Twitter recommendation source code now available to all on GitHub https://github.com/twitter/the-algorithm
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
Twitter revealed its algorithm to the world. But what does it mean for you? I spent the evening analyzing it. Here’s what you need to know:
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
1. Likes, then retweets, then replies Here’s the ranking parameters: • Each like gets a 30x boost • Each retweet a 20x • Each reply only 1x It’s much more impactful to earn likes and retweets than replies.
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
2. Images & videos help Both images and videos lead to a nice 2x boost.
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
3. Links hurt, unless you have enough engagement Generally external links get you marked as spam. Unless you have enough engagement.
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
4. Mutes & unfollows hurt All of the following hurt your engagement: • Mutes • Blocks • Unfollows • Spam reports • Abuse reports
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
5. Blue extends reach Paying the monthly fee gets you a healthy boost.
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
6. Misinformation is highly down-ranked Anything that is categorized as misinformation gets the rug pulled out from under it. Surprisingly, so are posts about Ukraine.
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
7. You are clustered into a group The algorithm puts you into a grouping of similar profiles. It uses that to extend tweet reach beyond your followers to similar people.
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
8. Posting outside your cluster hurts If you do “out of network” content, it’s not going to do as well. That’s why hammering home points about your niche works.
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
9. Making up words or misspelling hurts Words that are identified as “unknown language” are given 0.01, which is a huge penalty. Anything under 1 is bad. This is really bad.
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
10. Followers, engagement & user data are the three data points If you take away anything, remember this - the models take in 3 inputs: • Likes, retweets, replies: engagement data • Mutes, unfollows, spam reports: user data • Who follows you: the follower graph
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
Shoutout to all the people analyzing: @NFT_GOD @amasad @mxpoliakov @0xCygaar @xerocooleth
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
If you enjoyed this, 1. I write daily threads to help you grow. You may like to follow: @aakashg0 (But if you’re going to unfollow, go ahead and don’t!) 2. Consider RTing the first tweet so others can benefit:
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
As much as it's fun to analyze the Twitter algorithm, it's also fickle. Most of my content doesn't make it to your feed. Subscribe to the newsletter to get my best and deepest work: http://aakashgupta.substack.com
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
How to optimize for the algorithm: Likes, then retweets, then replies You are clustered - posting outside it hurts Links hurt. Mutes & unfollows hurt Misinformation is down-ranked Images & videos help Blue extends reach Making up words or misspelling hurts
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
New learning: There’s also something known as “Heavy Ranker” This heavily weights replies to replies and time spent on Tweet. https://t.co/WBKTMJC5Ze
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
Additional learning: Your follower to following ratio matters. Following way more than follow you hurts. Use lists.
@aakashg0 - Aakash Gupta 🚀 Product Growth Guy
The big open question is: what about bookmarks? The predominant opinion right now is favcountparams() 30x multiplier's formula is: Likes + Bookmarks = Favorites Count It doesn't look to be in the code right now. Part of the problem here is what's on GitHub is incomplete.
@tibo_maker - Tibo
I spent hours digging through every bit of the Twitter Algorithm. Read this to 10x your Twitter growth - in 5 minutes:
@tibo_maker - Tibo
1. Get Twitter Blue As a Twitter Blue subscriber, you get a whopping 4x algorithm to boost when you engage with tweets from those in your network. If the person you're engaging with isn't in your network, you still get a 2x boost. My advice: Get Twitter Blue.
@tibo_maker - Tibo
2. Maintain A High Follower-Following Ratio Don't follow too many people. Twitter calculates your ratio of followings to followers and reduces your rank if it's too high. Keep your following count low and hence your rank high.
@tibo_maker - Tibo
3. Keep your `Tweepcred` high Get your tweets featured by boosting your Tweepcred score. It's based on your interactions, account age, followers, and device usage. Aim for healthy engagement to keep it high and watch out for spam, blocks, and mutes that can bring it down.
@tibo_maker - Tibo
4. Niche down and become a trusted voice in one niche. Get noticed within your niche by engaging with the right accounts. Out-of-network replies can hurt your visibility, so focus on building connections with prominent accounts in your community first.
@tibo_maker - Tibo
5. Engage the most in your tweets in the first 6 hours. A Tweet's relevancy score will decrease by 50% every 6 hours. Older Tweets become less relevant over time. Engaging and replying to comments during this time will work in your favor.
@tibo_maker - Tibo
6. Add images & videos. In the current ranking model, tweets with images & videos get a 2x boost. So use images or videos more.
@tibo_maker - Tibo
7. Be less controversial Keep your Twitter reputation score high by avoiding negative feedback loops. Being blocked, muted, or reported for abuse or spam, and losing followers can tank your Tweepscore.
@tibo_maker - Tibo
8. Create a reason for people to engage Once you're ranked in a user's feed, the likelihood of being seen depends on factors like time spent on the tweet, profile visits, and replies. Create content that encourages users to take action and engage with your tweet.
@tibo_maker - Tibo
9. Reply to every reply on your account Boost your tweet's score by engaging with replies. If someone liking your tweet gives that tweet 1 point, your engaging with someone’s reply to your tweet gives you 150 points. So have more conversations.
@tibo_maker - Tibo
10. Don’t add links unless you know it’s going to get engagement Avoid being marked as spam by limiting external links in your tweets. Posting only URLs can drastically reduce your rank unless your tweet gets high engagement in the first few hours.
@tibo_maker - Tibo
If you like this, RT the top tweet ♻️ Would help a lot ❤️ And feel free to follow me at @tibo_maker for more like this
@tibo_maker - Tibo
And if you're looking for a more in-depth analysis, find it here: https://tweethunter.io/blog/twitter-algorithm-full-analysis (careful, it gets quite technical)
@The1Parzival - THE PARZIVAL
🧵 TWITTER SUPPRESSION ALGORITHM THREAD 🧵 If you have ever been censored on Twitter or other platforms you will want to read this. This Thread will Expose the ALGORITHM LIKELY USED BY TWITTER FOR MASS CENSORSHIP. Let me introduce you to the Tombstone Generator! 💪@elonmusk
@The1Parzival - THE PARZIVAL
Yes, there is really an ALGORITHM that is called "TombstoneGenerator", which is buried deep in the layers of the coding for Twitter. We can see some of the Coding regarding "VISIBILITY" here in the first lines of Code.
@The1Parzival - THE PARZIVAL
Here is some of the following Code expanding on the previous Code pictured above. It has coding for "visibilityParams" and "statsReceiver". Not to mention the Coding for "LocalLawsWithheldMedia".
@The1Parzival - THE PARZIVAL
The final parts of the ALGORITHM seem to be the most intrusive of all the Code. Hopefully this is something @elonmusk and @Twitter can fix with urgency!
@The1Parzival - THE PARZIVAL
It appears that this algorithm is giving Users the Metaphorical Tombstone by the Undertaker when it comes to visibility. If you want to fact check you can start here.
@The1Parzival - THE PARZIVAL
Follow the path to find the Code and check for yourself.
@The1Parzival - THE PARZIVAL
BOOM! 🎤⬇️
@The1Parzival - THE PARZIVAL
🎤⬆️ I decided to plug the TombstoneGenerator Code into ChatGPT and ask how it is used to censor users. Here is the output response I received.
@The1Parzival - THE PARZIVAL
Here is the direct link to the Algo: https://github.com/twitter/the-algorithm/blob/90d7ea370e4db804fb8f57fcb133a84af767dbfb/visibilitylib/src/main/scala/com/twitter/visibility/generators/TombstoneGenerator.scala
@The1Parzival - THE PARZIVAL
@The1Parzival - THE PARZIVAL
@The1Parzival - THE PARZIVAL
@The1Parzival - THE PARZIVAL
@The1Parzival - THE PARZIVAL
@The1Parzival - THE PARZIVAL
@The1Parzival - THE PARZIVAL
@The1Parzival - THE PARZIVAL
@The1Parzival - THE PARZIVAL
@The1Parzival - THE PARZIVAL
🧵 TWITTER RANKING & BOOST/DEBOOST ALGORITHM THREAD 🧵 📢 SHARE FAR AND WIDE 📢 https://t.co/lc8V9J3P2E
@The1Parzival - THE PARZIVAL
Twitter has been actively updating and modifying its Algorithm Coding since making it open source 2 months ago. We can see the updates on certain parts of the code and the when they happened here. https://t.co/wpdRacdZCg
@The1Parzival - THE PARZIVAL
We can also see that the updated this Ranking Algo last month to make some fixes regarding grammer, BUT MORE IMPORTANTLY TO ADD 1 LINE OF CODE FOR THE BOOST OF VERIFIED TWITTER BLUE USERS. https://t.co/7pfg7T047t
@The1Parzival - THE PARZIVAL
The Algo shows that there was a line of code added to give users an absolutely pitiful boost if they are a verified twitter blue account. THIS LINE OF CODE IS WHAT YOU ARE PAYING $8 FOR! https://t.co/Ugdd5G3hmD
@The1Parzival - THE PARZIVAL
It is evident that this Algorithm is ranking USERS based on large number of parameters, including REPUTATION, YOUR SOCIAL CIRCLE, AND YES, VERIFIED TWITTER BLUE ACCOUNTS. There is even an "OutOfNetworkReplyPenalty" THAT WEIGHS 10X MORE THAN THE BOOST TWITTER BLUE ASSIGNS! https://t.co/JFfCiWJODG
@The1Parzival - THE PARZIVAL
Basically the "OutOfNetworkReplyPenalty" makes you VIRTUALLY INVISIBLE even with your Blue Checkmark. It also keeps Retweets from getting any traction, even if Retweeted by large accounts of 100k+ Followers. There is also Coding for "EarlyBird" and "Blender" Parameters. https://t.co/Hw2guY129u
@The1Parzival - THE PARZIVAL
Looking into the EarlyBird Settings we come accross this information provided by Twiiter on what it is used for. We can see that information here. https://t.co/zlfu5upzQG
@The1Parzival - THE PARZIVAL
It should be clear by now that Censorship on Twitter is far from being eliminated! There are still yet even more Algos that contribute to this which I will continue to expose.
@The1Parzival - THE PARZIVAL
There is so much important information out there, BUT UNTIL WE COMPLETELY END CENSORSHIP THE TRUTH WILL REMAIN IN THE DARK.
@The1Parzival - THE PARZIVAL
🧵 TWEEPCRED AND REPUTATION TWITTER ALGORITHM THREAD 🧵 Ever wonder why once your Twitter Account gets BLACKLISTED it seems almost impossible to get off? This thread will explain why this likely happening. https://t.co/t0CsHxDSuW
@The1Parzival - THE PARZIVAL
There actually an Algorithm that seems to base your REPUTATION on a SCORE OF 15-95. It is theoretically possible to be outside these limits, since the range is coded to be from 0-100. We can see the ALGO and Code for RANKING here. https://t.co/A0nTGLgGbV
@The1Parzival - THE PARZIVAL
Within this Code there also seems to be RANKING functions based on your follow/following ratios. You can see the Thresholds for those in the Code here. https://t.co/KObqxy2GWH
@The1Parzival - THE PARZIVAL
There seems to be many Algorithms that are responsible for determining a users visibility, yet its the summation of all them that produce these results. There is unquestionably much to fix in the Twitter Algorithm, BUT IT IS DOABLE.
@The1Parzival - THE PARZIVAL
The question that remains is, now that we all know about it, WILL anything be done to fix it? Fortunately, with the Code being Open Source, it is publicly visible the effort and time frames these fixes are taking place. https://t.co/iquddx0rsV
@The1Parzival - THE PARZIVAL
Twitter completely silenced me, then took credit for my work. I even tagged both Elon and Twitter Engineering in dozens of posts to no avail. It must be a pride thing, since they don't want to admit some anonymous stranger on Twitter figured this out before Twitter Engineering.
@The1Parzival - THE PARZIVAL
🚨🚨🚨 BREAKING 🚨🚨🚨 Someone from Twitter 1.0 left a LITERAL BREADCRUMB, which led me to this Code. It appears the STORM AGGREGATE has Counters to track activity of RIGHT and LEFT Accounts. - RightDataRecordCounter - AuthorRightDataRecord Counter This is likely the reason many conservative accounts seem to get much less visibility. As you post Right leaning material "AuthorRight" and interact with others Right leaning material you get counters. These counters then become part of the users stats and are then subsequently used when importing the "StatsReceiver" for future tweets or retweets. Here is the direct link to the code on Github: https://github.com/twitter/the-algorithm/blob/fb54d8b54984f89f7dba90a18e7c3048421464c3/src/scala/com/twitter/timelines/prediction/common/aggregates/real_time/StormAggregateSourceUtils.scala#L240…
@The1Parzival - THE PARZIVAL
@The1Parzival - THE PARZIVAL
@USAB4L @elonmusk Bingo!
@The1Parzival - THE PARZIVAL
@WarrenFahy Thanks Warren!
@The1Parzival - THE PARZIVAL
@whykikiasks Nope
@The1Parzival - THE PARZIVAL
@mind_vince Thank you!
@The1Parzival - THE PARZIVAL
@Jeremibullfrog2 @TheNotoriousLMC For reals! 🤣
@The1Parzival - THE PARZIVAL
@adamlund @AwakenedOutlaw The imported input parameters are not available, nor is all the source code. I have asked for this information to no avail. I hope this answers your question.
@The1Parzival - THE PARZIVAL
@Pammywho @TheNotoriousLMC Right back at ya Fam!
@The1Parzival - THE PARZIVAL
@icu_luci Yep, see my other Threads too. They are in my pinned tweet.
@The1Parzival - THE PARZIVAL
@Cowlesdc @benshapiro @CommunityNotes They won't even respond to my first post. Responding means they have to acknowledge me first. 😏
@The1Parzival - THE PARZIVAL
@dnwiebe Nope
@sparklingruby - Kristen Ruby
This discovery is significant. “Google’s Jigsaw Perspective was used for sentiment analysis.” - The Ruby Files Is Twitter still using Jigsaw Perspective?
@The1Parzival - THE PARZIVAL
BREAKING ALGORITHM FOUND IN THE CODE THAT WAS USED BY TWITTER TO SUSPEND AND SILENCE @realDonaldTrump ON JANUARY 8TH, 2021 AFTER J6. IT HAS A CONSTANT FEATURE THAT KEEPS THE STATE OF HIS ACCOUNT AS SUSPENDED AND COMES FROM AN IMPORTED DATA SET.
@Emilio2763 - 𝕰𝖒𝕲
It’s Just The Algorithm…
@MarioNawfal - Mario Nawfal
Twitter 1.0’s Censorship Engine Exposed: IS YOUR ACCOUNT STILL BEING CENSORED? Wondering why your account has slowed down or why some of your posts don't get views? 𝕏 made their Algorithm open source, allowing us to analyze what impacts your account and your reach. When examining the algorithm, we found Tweepcred, a chunk of code across 32 files from Twitter 1.0, and it’s STILL IMPACTING YOUR ACCOUNT! HOW DOES TWEEPCRED IMPACT YOUR ACCOUNT: 1. Account Reputation Score & Post Deboosting (file name: Reputation.scala) This file contains a class of code called Reputation, which gives a reputation score to each account from 0 to 100. This incorporates many factors, such as the Following/Follower Ratio (discussed below), your account suspension status, verification, and more. One important factor we will focus on today is YOUR POST'S PERFORMANCE, which plays a very important role on your overall account score. To make sure your posts get a high score, make sure you apply the following: - Include media (video or image) - Add links to news pages (does not specify which) - Write in English if that’s your audience (English gets boosted more than other languages) - Do not include explicit/NSFW content - Avoid external links (except news pages) - Focus on trending topics - Do not use multiple hashtags or too many topics - Aim for high engagement (likes, reposts, replies, and quote reposts) If your account score is too low, a function BadUserRepFilter will filter your account out. This will result in people not seeing your content, aka your account will be ‘deboosted’. 2. Your Following/Follower Ratio Matters! - (file name: ExtractTweepcred.scala) This is the ratio of how many people an account follows versus how many people follow them. Accounts with more followers than people they follow have a much higher score. The Reputation.scala file has a function that checks if you're following over 2500 people. If not, your tweepcred score stays the same. If you are, keep your following-to-followers ratio below 60% to avoid score adjustments. If you are over 60%, it lowers your score. Unfortunately, to avoid your Tweepcred score from being lowered, you have to also check the ratio of those you INTERACT with! ELON IS REMOVING TWEEPCRED @ElonMusk will be removing this social credit score for accounts, however it's important to note that this will be an uphill battle for the development team as it is embedded into the algorithm and plays a key role on the platform. I don’t agree with everything Elon says or does, but the level of transparency he and the 𝕏 team are bringing is not only needed, but is working! 𝕏 CEO Linda Yaccarino just announced a few hours ago that 𝕏 video views are up 90% compared to last year. This follows multiple announcements in the past few weeks and months of 𝕏 breaking new records. This all reminds me of a post by @ehikian at 𝕏: “excited about continuing to improve transparency on X, mostly so I never have to hear the word shadowban ever again” mic drop
@t3dotgg - Theo - t3.gg
Okay, let's go through this point by point 1. Account Reputation Score & Post Deboosting I still think you guys are reading WAY too deep into this score. First, there's a really easy way to get a "perfect 100", just pay for Premium. Also re:suspension, obviously your tweets get reset to 0 if you are currently suspended? The score is applied at an ACCOUNT level, not a post level, so this whole section is outright incorrect So...what actually IS TweepCred? Very simple: tweepcred determines how likely it is that your content is spam "deboosting" I think it's important we distinguish between three different things: deboosted posts, deboosted accounts, and BadRep (Low TweepCred) accounts Deboosted posts contain content Twitter/X has decided it does not want to promote in recommended feeds. Usually this is because it contains keywords or links that fail some filter check. this is on a post level, not an account level Deboosted accounts are accounts that have been flagged as bad actors. This is the closest thing to a shadowban, and this is what y'all actually want to talk about. From what I know, these flags are almost always temporary, and will go away in a few days. BadRep accounts are accounts that Twitter can't confirm are real human posters. This is what the post is about, so it's what I'll be focused on going forward. The code is pretty straightforward. It checks a few things: - Are you CURRENTLY suspended? - Are you CURRENTLY verified? - How old is your account? - How many devices have you signed in on? - Follower/following ratio (IF FOLLOWING OVER 2500 PEOPLE) I highlighted the bit about account age and device count below, as those are the strongest tells that this is code to check for spam bots. Reminder that verified accounts skip this check entirely NONE of this takes your content into account. TweepCred is basically just a check for if someone should be put under the "see more" fold. 2. Your Following/Follower Ratio Matters! - (file name: ExtractTweepcred.scala) As you said here, this only matters if you follow over 2500 people. You don't need to follow 2500 people, especially if you're a new account. New accounts with tons of followers are usually spam, so they are hidden under the fold. "Social Credit Score for Accounts" You could make an argument that deboosted accounts are determined by some "social credit" system, but none of that is visible in the code discussed here. This post is trying to conflate TweepCred, suspensions, post ranking and account status. Make good content, and you'll be fine
@t3dotgg - Theo - t3.gg
@MarioNawfal @RyanRozbiani lmk if anything here is unclear 🙏
@Emilio2763 - 𝕰𝖒𝕲
It’s Just The Algorithm… Had ENOUGH Yet?…
@Emilio2763 - 𝕰𝖒𝕲
It’s Just The Algorithm …https://t.co/LITECEAp6W
@The1Parzival - THE PARZIVAL
🧵 TWITTER ALGO FINALE THREAD 🧵 This thread will serve the purpose of providing a mountain of evidence to @realannapaulina for her congressional hearing against Twitter 1.0 employees for their role in mass censoring Conservative Americans.
@The1Parzival - THE PARZIVAL
Let's first show the open sourced code on github that was provided by Elon Musk.
@elonmusk - Elon Musk
Twitter recommendation source code now available to all on GitHub https://github.com/twitter/the-algorithm
@The1Parzival - THE PARZIVAL
Using the code from this repository we can then begin to confirm the codes connection to the Twitter Files released by @bariweiss. From this post we can see that there was account labels placed on popular conservative accounts.
@The1Parzival - THE PARZIVAL
From the images we can see the labels that were placed on those accounts were: - Recent Abuse Strike - Trends Blacklist - Notifications Spike - Search Blacklist - Do Not Amplify Consequently all these labels show up in the open source code found here. https://github.com/twitter/the-algorithm/blob/72eda9a24f815f6d566818cbf8518138e29d83e9/visibilitylib/src/main/scala/com/twitter/visibility/models/UserLabel.scala#L11
@The1Parzival - THE PARZIVAL
Not only did Twitter apply labels to accounts, but they also applied labels to individual tweets. This served to further reduce the reach of conservative accounts like those shown above. Here is the link to the full list of Tweet Safety Labels. https://github.com/twitter/the-algorithm/blob/72eda9a24f815f6d566818cbf8518138e29d83e9/visibilitylib/src/main/scala/com/twitter/visibility/models/TweetSafetyLabel.scala#L12
@The1Parzival - THE PARZIVAL
These Tweet Safety Labels reduced visibility on some of the most important information for the public regarding Covid-19 and Elections. We can see these overreaching labels here for: - Misinfo Covid-19 - Misinfo Covid-19 Vaccine - Misinfo US Elections https://t.co/sXWz1aXz1B
@The1Parzival - THE PARZIVAL
There is also a label they can apply to tweets for simply not liking a specific user and is specifically worded as Persona Non Grata. For those that don't know what this means here is the definition. https://t.co/zMQwMdMWZy
@The1Parzival - THE PARZIVAL
We will now move away from labels and show yet other forms of censorship mechanisms in the code. There's actually a Social Credit Score built into the Algo and applies scores for the following: - User Mass Score - Reputation Score (Tweepcred) - Toxicity Score - Follow Score https://t.co/tbdK44DXep
@The1Parzival - THE PARZIVAL
Here is the full thread I did on the Reputation Score that goes into detail about how it works in the code. The Reputation Score was also confirmed by Elon Musk and his Engineering Team 3 weeks after I disclosed it. https://t.co/cHK6Gv67Ph https://t.co/AoDUvHcTpT
@The1Parzival - THE PARZIVAL
🧵 TWEEPCRED AND REPUTATION TWITTER ALGORITHM THREAD 🧵 Ever wonder why once your Twitter Account gets BLACKLISTED it seems almost impossible to get off? This thread will explain why this likely happening. https://t.co/t0CsHxDSuW
@The1Parzival - THE PARZIVAL
Additionally, Twitter 1.0 also went beyond just censoring individual users, but censored entire groups, mainly conservatives, by using a clustering method. By grouping these accounts in clusters they were able to silence entire groups and topics. It is shown in the code here. https://t.co/9wnaGZpGvy
@The1Parzival - THE PARZIVAL
Here is a deeper explanation on the clusters from a previous post I did. https://t.co/rZcaFykGyG
@The1Parzival - THE PARZIVAL
🚨🚨🚨 BREAKING 🚨🚨🚨 TWITTER ALGO HAS CODING TO PENALIZE CERTAIN GROUPS, KNOWN AS CLUSTERS, WHERE CLUSTERS GET PENALIZED BASED ON LOTS OF INTEREST AND POPULAR USERS! CONSERVATIVE CLUSTERS ARE BEING PENALIZED, WHICH LIMITS VISIBILITY FOR ALL THAT ARE IN IT!! https://t.co/O33Q5gJgXD
@The1Parzival - THE PARZIVAL
As if these censorship mechanisms are not enough their is actually coding for a Government Request to intervene on things they consider misinformation. https://t.co/zpvvC4cYbD
@The1Parzival - THE PARZIVAL
We can also see the code that was directly responsible for suspending and silencing the sitting President of the United States @realDonaldTrump. Here is the full thread on it I did previously. https://t.co/iRpoIfpaWl https://t.co/qBaFpniRSx
@The1Parzival - THE PARZIVAL
🧵 DONALD TRUMP TOMBSTONE THREAD 🧵 Now that we know the exact Code responsible for suspending and silencing a sitting PRESIDENT, let's see how the process unfolded on January 8th, 2021. https://t.co/ZsakECluCf
@The1Parzival - THE PARZIVAL
This is now a good time to bring up the Tombstone censorship mechanism, which was the first thing I found in the code. Here is the full thread on the Tombstone as well as an example of what the Tombstone looks like in use. https://t.co/eMUJFE5ld6 https://t.co/Yp8NT713ax
@The1Parzival - THE PARZIVAL
🧵 TWITTER SUPPRESSION ALGORITHM THREAD 🧵 If you have ever been censored on Twitter or other platforms you will want to read this. This Thread will Expose the ALGORITHM LIKELY USED BY TWITTER FOR MASS CENSORSHIP. Let me introduce you to the Tombstone Generator! 💪 @elonmusk https://t.co/VFapx85pOS
@The1Parzival - THE PARZIVAL
Now let's go a little deeper and talk about Models within the Algorithm. Here we can see several Trust and Safety Models that were released in the open source code and are as follows: - Abusive - NSFW - Toxicity https://t.co/zxoje0TogE
@The1Parzival - THE PARZIVAL
Looking into the Toxicity Model we find that there is a keyword list used in order to make computational decisions on whether a posts visibility is altered. The list includes words for Politics, Insults, and Race. These keywords however are not open sourced. https://t.co/umzyVdZCZv
@The1Parzival - THE PARZIVAL
Additionally, not all the models have been released and is even confirmed in the code. Here is what it says. https://t.co/GXrWKDcryy
@The1Parzival - THE PARZIVAL
It is also important to state that not once has any of my findings been disputed or Community Noted. In fact @CommunityNotes essentially confirmed my findings, but were too cowardly to actually apply the note. This is a post I did in response to this. https://t.co/xT2MRbMxlx
@The1Parzival - THE PARZIVAL
🚨🚨🚨 BREAKING 🚨🚨🚨 COMMUNITY NOTES AUDIT PROVES THAT THE GROUP IS COMPROMISED AND DOES NOT EVEN HONOR THEIR VERY OWN VOTING RESULTS. HERE ARE THE RESULTS OF THE DATA FROM THE COMMUNITY NOTE THAT CORRECTS THE FACT THAT I FOUND THE REPUTATION SCORE! https://t.co/t3Aj8njWyb
@The1Parzival - THE PARZIVAL
At this point it should be clear that Twitter Algo was designed for 1 purpose and 1 purpose alone, to silence conservatives and control the information narrative on social media. The depth that Twitter went to silence conservatives is nothing short of election interference.
@The1Parzival - THE PARZIVAL
This critical info should significantly help Rep Anna Paulina Luna and may also be of great interest to @Jim_Jordan as well in regards to his recent bombshell report on the Industrial Censorship Complex. https://t.co/y5mrdzcDRi
@The1Parzival - THE PARZIVAL
Let's hope that this information initiates some serious change on Social Media platforms to stop censoring free speech and limiting the reach of conservatives. With elections coming up in less than a year it is more important than ever that this be addressed immediately.
@The1Parzival - THE PARZIVAL
In conclusion these mechanisms are still in place on Social Media Platforms including X. Although Elon Musk claims a new Algorithm is about to roll out, until it does, it is safe to assume that most of the old Algorithm is still in place. The time to restore Free Speech is now!
@The1Parzival - THE PARZIVAL
Hey .@CommunityNotes will you finally do the right thing and apply the note that I found the REPUTATION SCORE to this post by Xdaily? Even Grok has confirmed it! Doing nothing proves bias and favoritism towards specific accounts. https://t.co/NgfaE1o95g
@xDaily - X News Daily
NEWS: The Twitter team continue to find shadowbans buried deep in the Twitter code. Just last week they found a measure that stopped accounts assigned a low 'Reputation score' from trending. This shadowban even applied to Elon's account and prevented his tweets from trending. https://t.co/6qjO3hXo5p
@The1Parzival - THE PARZIVAL
🚨 BREAKING: Special Coding from the Twitter algorithm called "Twadoop" has now been definitively linked to .@paraga, Ex CEO/CTO of Twitter. This function generates a TSV File for User Mass Scores which was used to censor conservative Americans and dates back as far as 2009. https://t.co/ersm6r7a3E
@The1Parzival - THE PARZIVAL
🧵 TWADOOP THERE IT IS 🎶 This thread will outline the creation of the function "TWADOOP", which is a special code used by the Algorithm to generate a list of "USER MASS SCORES". It will also discuss Parag Agrawals involvement as well as other Ex Twitter Employees. https://t.co/bQRwyIw98f
@The1Parzival - THE PARZIVAL
To begin let's first look at what a User Mass Score means and how it affects you. The User Mass is a Sub-Algorithm that assigns a Score to the account based on a variety of characteristics, many of which the values are hidden. The characteristics and code are shown here. https://t.co/pWL4aYnZqN
@The1Parzival - THE PARZIVAL
The User Mass is part of the "Tweepcred" group of Scoring Metrics used to determine an accounts visibility on the platform. The User Mass has a TSV File that is generated by the Twadoop function. A TSV File is simply a plain text data table that is native to many applications. https://t.co/Ve4DpHNGaM
@The1Parzival - THE PARZIVAL
This TSV File acted as a pseudo list of each accounts Social Credit Score on the platform. This Score could then be easily manipulated by changing values in the tables of the files manually or adding a weight multiplier to scores. https://t.co/QWTliLOcqT
@The1Parzival - THE PARZIVAL
Here is where we begin to connect previous employees that had a hand in the development of Twadoop, Tweepcred, and this Social Media Social Credit System. There is a correspondence regarding Twadoop Files between Parag and Kevin Weil in 2012, both very high level employees. https://t.co/TiG2zEM2fF
@The1Parzival - THE PARZIVAL
So who is Kevin Weil? He was Twitter's Product Head. He later went on in 2016 to join Facebook's Instagram. I'm very curious to know what exactly was in the files Weil was asking Parag about? This post from Parag just so happens to be deleted now. https://t.co/KvuSw8ZmnS
@The1Parzival - THE PARZIVAL
Now remember this interaction between them was in 2012, but Twadoop actually goes back even further, all the way to 2009. We can see references to Twadoop from another Ex Twitter Employee that "Helped Create the Data Platform Twitter". Notice he said Data not Social Media! https://t.co/ph5fWaI35d
@The1Parzival - THE PARZIVAL
It should be clear that Twadoop has been around for quite some time specifically since 2009. Now let's put some focus back on Parag who was there since 2011 through his reign as CEO to 2022 when Elon Musk took over and fired him. https://t.co/bhTbKBnJbn
@The1Parzival - THE PARZIVAL
Parag took over for Jack Dorsey as CEO and was previously the Chief Technology Officer. Before joining Twitter Parag got his Masters and PHD from Stanford in 2008. Stanford resides only 35 miles from Twitter Headquarters and has been accused in the proliferation of censorship. https://t.co/WiQYkmUtx2
@The1Parzival - THE PARZIVAL
Prior to his Studies at Stanford Parag lived abroad. He was born in India and gained his Bachelors Degree from the Indian Institute of Technology, Bombay. Within 6 years of arriving to the US in 2005 Parag obtained his PHD and began his work at Twitter CENSORING AMERICANS! https://t.co/2WDNom4JVS
@The1Parzival - THE PARZIVAL
Parag has also made his stance clear on Freedom of Speech. He said: "Our role is not to be bound by the First Amendment, but our role is to serve a healthy public conversation ... [and to] focus less on thinking about free speech, but thinking about how the times have changed." https://t.co/8R3XfcNyjl
@The1Parzival - THE PARZIVAL
It's very disturbing that someone can come from another country, obtain US Citizenship, reap the benefits of America, then go on to destroy the Constitution, the very thing that makes America great. My God given rights are not debatable and they will never take our freedom! https://t.co/SExvHGKe8o
@The1Parzival - THE PARZIVAL
🚨 BREAKING: Special coding found in the Twitter Algorithm called the "Tseng Takedown" which has a list of specific country codes and takedown reasons. Totally seems American! 🇺🇲 https://t.co/D0vkwXauCI
@imUrB00gieman - 𝐉𝐎𝐇𝐍 𝐖𝐈𝐂𝐊 𝕏ʰⁱᵗᵐᵃⁿ 🏴☠️
Impressions is controlled entirely by The 𝕏 #tweepcredalgo This has absolutely nothing to do with the quality of individual tweets. This is 𝕏 putting its thumb on the scale of social justice just like shady butchers of the past did. https://t.co/q1fZYFGZCA
@The1Parzival - THE PARZIVAL
🥇 🥈 🥉 - Author Specific Score Adjustments for relevance. - Adjustments for tweets posted by specific Authors. - This is favortism built directly into the code. 🌊 🌊 🌊 https://t.co/DE6Sk5xOsD
@The1Parzival - THE PARZIVAL
🧵 TWEEPCRED 2.0 - PART 1🧵 - It's time to take a deeper dive into TweepCred aka the Reputation Score used on this platform for users to establish their credibility. Is it really about credibility or is it being used as a social classification system like a social credit score? https://t.co/n2UTg22Iju
@The1Parzival - THE PARZIVAL
After researching this extensively I have found that TweepCred acts as social class system rather than a meter for a users credibility. The first clue comes from Twitter/X own words outlined right in the code. Here we see TweepCred is based on your Following to Follower Ratio. https://t.co/r1nC7ApKx7
@The1Parzival - THE PARZIVAL
Why is this so important to understand? Well, by using this ratio it would mean your credibility is based on some fictitious number of people following you and how many you follow in return. Additionally, it does not factor in the Bots which represent a large majority on X. https://t.co/PQdBwlKWkN
@The1Parzival - THE PARZIVAL
Clearly there is a significant bot presence still here on X with no sign of slowing down. Keeping this in mind let's now look at how your TweepCred is impacted by your following to follower ratio. This chart I created shows how this ratio is creating social scoring tiers. https://t.co/z1cVeoIk7p
@The1Parzival - THE PARZIVAL
There is a ton of information to process in this chart, so lets break it down piece by piece to better understand what were looking at. Let's first look at the legend to describe what we are seeing. Each follow ratio is based on the same number of following 1, 10, 100, 1k, etc. https://t.co/pbDnFGWQWy
@The1Parzival - THE PARZIVAL
Additionally, you can see where I have inserted a few points for notable accounts including mine to see where we all stack up in comparison. I showed my account in 3 different states for comparative purposes: - Actual - Max Ratio, following only 1 - Even, equal follow/follower https://t.co/eBWxJWaBMz
@The1Parzival - THE PARZIVAL
It's obvious that decreasing my follows by 1790 had a significantly greater impact to my Ratio than following an additional 22,600 accounts. This is because order of magnitude is king when doing this calculation. Notice there are 2 dead zones, the first being a coding limit. https://t.co/prwN3htBmj
@The1Parzival - THE PARZIVAL
Now let's look at the other dead zone which is defined by mathematical restrictions. Essentially what it shows is the maximum non-zero achievable ratio based on your total followers. The more followers you have the higher (Lower) you can get your ratio, which is 1/follower#. https://t.co/vDwwp9bmSZ
@The1Parzival - THE PARZIVAL
Therefore as a comparison your max ratio based on these follower numbers & only following 1 would be: - 1/1 = 1 - 1/10 = .1 - 1/100 = .01 - 1/1000 = .001 - ETC. As you can see your ratio jumps orders of magnitude directly with your number of followers, creating a ratio barrier. https://t.co/ETtxW4RU6k
@The1Parzival - THE PARZIVAL
This barrier prevents you from getting more visibility of accounts with more followers because you can never hit their same ratio until you have a near equal number of followers yourself. Therefore, we can see your follower number is creating a social classification system. https://t.co/9drm4Kp8DK
@The1Parzival - THE PARZIVAL
Let's discuss 1 more piece of this chart before we take a look at it as whole. This is the crazy zone aka the John Cena Zone where an account is basically a Follow Collector and follows a minimum of 500k accounts. Due to follow limits doing this will take years to achieve. https://t.co/OQtOuOqO8t
@The1Parzival - THE PARZIVAL
Hopefully you're tracking along so far, now let's get back to big picture of the chart. Not only are certain ratios unachievable until you hit a corresponding follower number, but there are bands based on how many you follow back. Following order of magnitude is important too. https://t.co/Snl9icGz0O
@The1Parzival - THE PARZIVAL
You should be grasping what this chart is showing and how your numbers do matter as far as the algorithm and math are concerned. If we look at the other prominent users in the chart we can see the maximum possible ratio they can achieve without hitting the mathematical barrier. https://t.co/44soID5RsJ
@The1Parzival - THE PARZIVAL
You may be asking yourself why you or I don't get the same visibility as these accounts even though you may be on a similar band or even a higher band? Well, the answer's simple, it's based on the Ratio order of magnitude which we cannot achieve as stated previously, classes. https://t.co/fsU9FPigoX
@The1Parzival - THE PARZIVAL
Now begs the question why doesn't myself or these other prominent accounts shed their folllowing to achieve a Maximized Ratio? For myself, it would destroy my account because people would be pissed. On the otherhand for the prominent accounts it's all about status & networks. https://t.co/SS070N13Jp
@The1Parzival - THE PARZIVAL
You can & should use this chart to see where your account stacks up, but I wouldn't recommend unfollowing a bunch of people to boost your ratio as that can have other effects. You should however shed accounts that you feel are bot like or of no value as they hurt your Tweepcred. https://t.co/IaKNOIAL60
@The1Parzival - THE PARZIVAL
Another important thing to keep in mind is that following the big account grifters, even if it's "just to watch them", is helping their ratio and hurting your ratio simultaneously, a double whammy. You are hurting your own visibility simply by "just watching them". https://t.co/VM4hC1f3OO
@The1Parzival - THE PARZIVAL
It should be very clear now how TweepCred alone is creating social classes here on X, but it is just 1 of many being used. You can also include Mass Score, Toxicity Score, Follow Score, Account Labels, and Clustering just to name others, but the list is extensive. https://t.co/2A8aRpfZeu
@The1Parzival - THE PARZIVAL
If we consider what percentage of these big accounts following are actually just bots that are contributing to the compounding of these social class discrepancies, the ramifications become overwhelming. Depending on that percentage, the bots could already be in control. https://t.co/khyZGGAEzg
@The1Parzival - THE PARZIVAL
I planned on including a lot more info, but appears I'm close to hitting to 25 post thread cap, so looks like I will be doing this thread in parts. I think you've had enough for now so I'll leave it off here. https://t.co/apameOTF2U
