The Twitter Files, Part 3: The Removal of Donald Trump
From October, 2020 through January 6th, the whole thread.
The following is a reproduction of the “Twitter Files” thread from this past Friday evening. Because the entire thread with illustrations far exceeds Google’s email limit, I’m putting the document here, and mailing out a link to the material:
1. THREAD: The Twitter Files, Part 3. THE REMOVAL OF DONALD TRUMP: Part One: October 2020-January 6th
2. The world knows much of the story of what happened between riots at the Capitol on January 6th, and the removal of President Donald Trump from Twitter on January 8th.
3. We’ll show you what hasn’t been revealed: the erosion of standards within the company in months leading up to J6, conscious decisions by high-ranking executives to violate their own policies, and more, all against the backdrop of ongoing, documented interaction with federal agencies.
4. This first installment tonight will cover the period before the election through the first day of the crisis, January 6th. Tomorrow, on Saturday, @Shellenbergermd will tell a story in screenshots of the chaos inside Twitter on January 7th. On Sunday, @BariWeiss will guide you through the heretofore secret internal communications at the firm on the key January 8th.
5. Whatever your opinion on the decision to remove Trump that day, the internal communications at Twitter between January 6th-January 8th have clear historical import, in addition to being controversial. Even Twitter’s employees – who at times seem overwhelmed by the weight of it all – understand it’s a landmark moment in speech enforcement.
6. In fact, as soon as they finish banning Trump on 1/8, Twitter executives start processing their new power. They begin preparing to ban the accounts of future presidents and future White Houses – perhaps even president-elect Joe Biden. The “new administration,” says one Policy VP, “will not be suspended by Twitter unless absolutely necessary.”
7. Twitter executives removed Trump on January 8th, 2021, in part because of what one executive called the “context surrounding”: actions by Trump and his supporters “over the course of the election and frankly last 4+ years.” By then they’ve fully redefined moderation as a holistic process, that requires looking at a broad picture. But broad pictures can cut both ways.
8. The bulk of the internal debate leading to Trump’s ban happened in those three January days. However, the intellectual framework was laid in the months preceding the Capitol riots.
9. Before J6, Twitter was already a strange mix of automated and/or rules-based enforcement by the rank-and-file on the one hand, and more subjective moderation by senior executives on the other. As @BariWeiss reported, the firm developed a vast toolbox of hidden levers for manipulating visibility, nearly all of which were thrown at Trump (and others) long before J6.
10. As the election approached, and then after, when tensions soared as Trump insisted he’d won, senior executives – perhaps under pressure from federal law enforcement and intelligence agencies, with whom they met more as time progressed – increasingly viewed rules as annoyances, and began to speak of “vios” as mere pretexts to do what they would likely have done anyway.
11. Within weeks after J6, internal slack chats showed Twitter executives reveling in newly heightened roles as public-facing extensions of the intelligence community. Here’s Trust and Safety head Yoel Roth, Twitter’s de facto chief censor, lamenting his lack of “generic enough” calendar descriptions for use in concealing the identities of his “very interesting” meeting partners. This is 1/26/2022:
12. These initial reports are based on searches for documents linked to the company’s most prominent executives, whose names were already public. They include Roth, former trust and policy chief Vijaya Gadde, then-Chief Technology Officer and future CEO Parag Agrawal, Policy Officer Lauren Culbertson, Director of Public Policy Nick Pickles, General Counsel Sean Edgett, and recently plank-walked Deputy General Counsel (and former FBI General Counsel) Jim Baker.
13. One particular slack channel offers a window into the evolving thinking of top officials in late 2020 and early 2021.
14. On October 8th, 2020, Twitter executives opened a slack channel called “us2020_xfn_enforcement.” Through J6 and beyond, this would be the home for discussions about election-related removal decisions, especially ones that required “scaled investigations” or involved “high-profile” accounts (often called “VITs” or “Very Important Tweeters”), and “edge cases” requiring senior input.
15. There was tension at Twitter between Safety Operations – a much larger department whose staffers used a more rules-based process for addressing issues like porn, scams, and threats – and a smaller, more powerful cadre of senior policy execs like Roth and Gadde.
16. The latter group were a high-speed Supreme Court of moderation unto themselves, issuing content rulings on the fly, often in seconds and based on guesses, gut calls, even quick Google searches, even in cases involving the President.
17. During this time, executives were clearly liaising with federal law enforcement and intelligence agencies law enforcement about moderation of election-related content. Still at the start of reviewing the #TwitterFiles, we’re finding out more about these interactions on a daily basis.
18. Here, Policy Director Pickles answers a query from Marketing. He’s asked if they should say Twitter detects “misinfo” through “[machine learning], human review, and **partnerships with outside experts?*” The employee adds, “I know that’s been a slippery process… not sure if you want our public explanation to hang on that.”
19. Pickles quickly asks if they could “just say “partnerships.” After a pause, he says, “e.g. not sure we’d describe the FBI/DHS as experts.”
20. Because we now have a much bigger document set, we’re learning more about incidents like the handling of the Hunter Biden laptop story.
The following post about the laptop FUBAR shows Roth not only met weekly with the FBI and DHS, but with the Office of the Director of National Intelligence (DNI). Roth’s report to FBI/DHS/DNI about blocking the Hunter Biden laptop story on “hacked materials” grounds is almost farcical in its self-flagellating tone:
“We blocked the NYP story, then unblocked it (but said the opposite)… comms is angry, reporters think we’re idiots… in short, FML” (fuck my life).
23. Some of Roth’s later slack posts indicate his weekly confabs with federal law enforcement involved – at least in one case – separate meetings. Here, he ghosts the FBI and DHS, respectively, to go first to an “Aspen Institute thing,” then take a call with Apple.
24. How did the government flagging system work? Here, the FBI sends reports about a pair of tweets, the second of which involves a former Tiipecanoe County, Indiana Councilor and Republican named @JohnBasham claiming “Between 2% and 25% of Ballots by Mail are Being Rejected for Errors.” Part of the reason this is slow work is that we have to manually load in things like the ID number of Basham’s tweet to figure out what they’re discussing.
25. The FBI report then got circulated in the enforcement slack. It appears Twitter cited Politifact to say the first story was “proven to be false,” then noted the second was already deemed “no vio on numerous occasions.”
26. The group then decides to apply a “Learn how voting is safe and secure” label because one commenter says, “it’s totally normal to have a 2% error rate.” Roth then gives the final go-ahead to a process initiated by the FBI:
27. Examining the entire election enforcement slack, we didn’t see one reference to moderation requests from the Trump campaign, the Trump White House, or Republicans generally. We looked. They may exist elsewhere. We were told they do. However, in this channel, they were absent.
The only thing we could find that might qualify as a pro-Trump content flag involved a request of unknown provenance that hashtags like #TrumpTerrorism and #TrumpTerrorist be taken off the “trends denylist,” a visibility filtering tool that would have prevented the terms from trending.
29. ”There doesn’t seem to be clear grounds for trends denylisting,” comes the reply, meaning the group doesn’t believe those terms should be “visibility filtered” by preventing them from trending.
30. Moreover, while we found references in other data sets of action against left-wing sites (below, see former CEO Jack Dorsey intervening on behalf of the World Socialist Web Site), in the 2020 election enforcement slack we mostly found a double-standard in blue-red moderation attitudes bordering on the absurd.
31. In one ridiculous case, former Arkansas governor Mike Huckabee joke-tweets about mailing in ballots for his “deceased parents and grandparents.”
32. This inspires a long slack by Twitter personnel that reads like Titania McGrath parody. “I agree it’s a joke,” concedes one employee, “but he’s also literally admitting in a tweet a crime.” Tweeps declare Huck’s an “edge case,” and though one notes, “we don’t make exceptions for jokes or satire,” they ultimately leave him be, because “we’ve poked enough bears.”
33. Roth suggests moderation even in the above absurd case will depend on whether or not the joke results in “confusion.” This sounds silly but is actually important. In the docs, Twitter OPS staff are seen working factory-style to flag content based on clear protocols, while policy execs constantly expand criteria to subjective issues like intent (yes, the video is authentic, but why was it shown?), orientation (was a screenshot of a banned tweet shown to condemn, or express support?), or reception (does a joke cause “confusion”?). This thought reflex will become key after J6.
34. In another example, Twitter employees prepared to slap a “mail-in voting is safe” warning label on a Trump tweet about a postal screwup in Ohio, before realizing “the events took place,” which meant the tweet was “factually accurate”:
35. “GOOD JOB ON SPEED” Trump was being “visibility filtered” as late as a week before the election. Here, senior execs didn’t appear to have a particular violation, but still worked fast to make sure this fairly anodyne Trump tweet couldn’t be “replied to, shared, or liked”:
36. A seemingly innocuous follow-up involved a tweet from actor @realJamesWoods, whose ubiquitous presence in controversial Twitter data sets is already a #TwitterFiles in-joke.
37. After Woods’s angry tweet about Trump’s warning label, Twitter staff – in a preview of what ended up happening after J6 – despaired of a reason for action, but resolved to “hit him hard on future vio.” Note that Woods and his tweet also deemed candidates for DNA or “Do Not Amplify” status.
38. Here a label is applied to Georgia Republican congresswoman Jody Hice for saying, “Say NO to big tech censorship!” and, “Mailed ballots are more prone to fraud than in-person balloting… It’s just common sense.”
39. Twitter teams go easy on Hice, only applying “soft intervenion,” with Roth worrying about a “wah wah censorship” optics backlash:
40. Meanwhile, there are multiple instances of pre-election tweets involving pro-Biden tweets warning Trump “may try to steal the election” that get surfaced, only to be approved by senior executives. This one, they decide, just “expresses concern that mailed ballots might not make it on time.”
41. “THAT’S UNDERSTANDABLE”: Even the hashtag #StealOurVotes – referencing a theory that a combo of Amy Coney Barrett and Trump will steal the election – is approved by Twitter brass, because it’s “understandable” and a “reference to… a US Supreme Court decision.”
42. In this exchange, again unintentionally humorous, a post by former Attorney General Eric Holder claiming the U.S. Postal Service was “deliberately crippled” was initially hit with a generic quickly had a warning label taken off, once Roth caught wind of it:
44. Later in November 2020, Roth asked if staff had a “debunk moment” on the “SCYTL/Smartmantic vote-counting” stories, which his DHS contacts told him were a combination of “about 47” conspiracy theories:
45. On December 10th, as Trump was in the middle of firing off 25 tweets saying things like, “A coup is taking place in front of our eyes,” Twitter executives decided they’d had enough, and announced a new “L3 deamplification” policy, seemingly designed specifically for this situation:
46. Some executives wanted to use the new deamplification tool to silently limit Trump’s reach more right away, beginning with the following tweet::
47. However, in the end, it’s decided that the team will use older, less aggressive labeling tools at least for that day, but that the “L3 entities” can be deployed the following morning:
The significance of these and other exchanges is that it shows that Twitter, in 2020 at least was deploying a vast range of visible and invisible tools to rein in Trump’s engagement, long before J6.
For instance, in the docs, Twitter execs frequently refer to “bots,” e.g. “let’s put a bot on that.” A bot is just any automated heuristic moderation rule. It can be anything: every time a person in Brazil uses “green” and “blob” in the same sentence, action is taken.
In this instance, it appears moderators added a bot for a Trump claim made on Breitbart.
51. The bot ends up becoming an automated moderation tool invisibly watching both Trump and, apparently, Breitbart (“will add media ID to bot”). As you’ll learn, Trump headed into J6 already covered in bots:
There is no way to follow the frenzied exchanges among Twitter personnel from between January 6th and 8th without knowing the basics of the company’s vast lexicon of acronyms and Orwellish unwords. To “bounce” an account, for instance, is to put it in timeout, usually for a 12-hour review/cool-off:
“Interstitial,” one of many nouns regularly used as a verb in Twitterspeak (“denylist” is another), means placing a physical label atop a tweet, so it can’t be seen. PII has multiple meanings, one being “Public Interest Interstitial,” i.e. a covering label applied for “public interest” reasons. The post below also references “proactive V,” i.e. proactive visibility filtering. Every one of these actions supposedly required a justification. As with the military or the FBI, you had to speak acronym to work here.
This is all necessary background to January 6th. Before the Capitol riots, the company was engaged in an inherently insane/impossible project, trying to create an ever-expanding, ostensibly rational set of rules to regulate every conceivable speech situation that might arise between human beings.
This project was preposterous, yet its leaders were unable to see this, having become infected with the firm’s groupthink, coming to believe – sincerely – that it was their responsibility to control, as much as possible, what people could talk about, how often, and with whom.
The firm’s executives on day 1 of the January 6th crisis at least tried to pay lip service to its dizzying array of rules. By day 3, a million rules were reduced to one: what we say, goes.
When panic first breaks out there’s a fair share of WTF-type posts, mixed in with frantic calls in Twitterese to start emptying the firm’s whole drawer of tools at the situation. “What is the right remediation? Do we interstitial the video?” asks one employee, in despair:
Early on January 6th Twitter execs show a ton of shock, but still seem superficially tied to rules. This “Freedom or Death” tweet from National #StopTheSteal gadfly Mike Coudrey elicits colorful reactions:
Roth groans about Coudrey: “THIS asshole,” but still seems determined to stick at least superficially to rules, itching to act “if” this “constitutes incitement.
Technically this was true, but just minutes later, Roth executed the historic decision to “bounce” Trump, i.e. put him in timeout. “I hope you… are appropriately CorpSec’d,” says a colleague.
This theme of Policy feeling less and less interested in answering queries from Communications executives – who themselves have to answer the public’s questions – continued for all three days. Two days later, you see chatter about pulling Comms out of the loop:
The first company-wide email from Gadde on January 6th announced that 3 Trump tweets had been bounced, but more importantly signaled a determination to use “violations” as a guide for any possible permanent suspension:
“WHAT THE ACTUAL FUCK?” Safe to say Trump’s “Go home with love & in peace” tweet didn’t go over well at Twitter HQ:
Later in the day Roth is commenting on the enormous number of “duplicate” bot enrollments on Trump. The number will get higher:
By the end of the next day, January 7th, the top execs will be operating more or less unilaterally, and soon will be less worried about managing images coming out of Washington than preparing for the sonic PR boom of a permanent Trump suspension. Watch @ShellenbergerMD this weekend for the play-by-play of how all that went down.
By January 8th, which @BariWeiss will describe Sunday, Twitter will be receiving plaudits from “our partners” in Washington. And the fig leaf will be all the way gone.
One last note:
People on the left, right, and everywhere in between want to know what else is in the #TwitterFiles, from suppression/shadow-banning of everyone from leftists to lab-leak theorists, and amplification of everything from military propaganda to popular conservative accounts. Everyone has questions. We understand that.
We’ve stumbled on tidbits here and there about topics ranging from COVID to foreign policy:
Still, the reality is the data sets are enormous and we’re still working through them. More is coming, stay tuned.
Earlier this week, Elon Musk confirmed that Twitter deputy general counsel and former FBI general counsel James Baker had been “exited” from the company. He’d reportedly been “vetting” the first batch of internal Twitter files before their release to independent journalists Matt Taibbi and Bari Weiss.
While he was at the FBI, Baker played a key role in the Russia collusion probe, including the FISA warrant to surveil then-Trump 2016 presidential campaign aide Carter Page.
“James Baker was the one running blocking-and-tackle campaigns at the FBI to prevent the disclosure of documents we lawfully subpoenaed from Congress … Fast forward … he was running the same operation that he ran as FBI General Counsel inside of Twitter HQ,” says Kash Patel.
It has also now been revealed that the FBI had regular meetings with Twitter (and other tech giants) in the lead-up to the 2020 elections. Were these authorized directly by then-Attorney General Bill Barr and FBI director Christopher Wray?
“I want to know every contract the FBI has with Twitter,” says Patel. “I believe that they do have working-level engagement agreements with Twitter. And those need to be made public. Are taxpayer dollars funding part of this censorship scheme?”
Musk should release all the Twitter censorship files without redactions, Patel says, and he should also fire Perkins Coie, which was still representing Twitter as late as Dec. 6, according to court documents.
Below is a rush transcript of this Kash’s Corner episode from Dec 9, 2022. This transcript may not be in its final form and may be updated.
Kash Patel:
Hey everybody and welcome back to Kash’s Corner. Jan and I have done a lot of prep work, which means today’s episode might be the best yet. What are we going to talk about, Jan?Jan Jekielek:
It seems to me, well, I guess it seems to both of us, that Elon Musk’s purchase of Twitter has much greater ramifications than just Twitter. And this is kind of coming out in spades. And most recently, you know, you were out on Truth Social, not on Twitter, calling to see someone named James Baker basically removed from his position as counsel for Twitter.Mr. Patel:
Yeah, so, look, I mean, we have to rewind the clock here. And, of course, I always tell people whenever we talk about Truth Social, I’m on the board of Directors of Truth Social. So, while I have a fiduciary duty, to Truth Social, I agree with President Donald Trump who owns Truth Social. He’s not cheering against Twitter. He’s repeatedly said he hopes Elon Musk does what he said he’s going to do, which is restore it to being an actual free speech platform that’s censorship-free from political bias. So, all in on that mission.https://www.theepochtimes.com/kashs-corner-james-baker-and-perkins-coie-russiagate-architects-are-back-as-the-twitter-files-saga-unfolds_4914557.html