Cole Allen’s political and religious profile: what 4,000 tweets reveal

After my analysis of the manifesto and his Bluesky posts, I now also went through the more than 4,000 tweets on X (with same methodology) to see whether this corpus would confirm, nuance, or contradict my hitherto findings. Here is what I found:

I. Summary of most important findings

After my analysis of the manifesto and the Bluesky posts (which you can read here), I went back through the 4,000-plus tweets on X — Allen’s first online political phase, ended by his migration to Bluesky on 16 November 2024 — with the same methodology. I half-expected the larger corpus to nuance some of what I had written. It mostly didn’t. It strengthens my first blog post at every important point, with one real qualification I will come to at the end.

On Allen’s political development:

• Politicisation begins with the Russian war on Ukraine, not with Trump. Through 2021 and most of 2022, the X account is essentially gaming-only. The first political post in the corpus is an April 2023 retweet about a Russian missile strike. Pro-Ukraine politicisation precedes anti-Trump politicisation; the in-house Christian critique arrives later still and rests on top of an already-formed political analysis.

• The November 2024 election outcome doubly anchors the “traitor” charge against Trump. Within hours of the election call on 6 November, Allen forms a settled belief that the election itself was won by fraud and amplifies that belief at substantial volume across the following week. The “traitor” lexicon throughout both corpora picks up Trump’s pre-election conduct and a stolen-election conviction that crystallised the day after the call.

• The manifesto’s anchoring “complicity through inaction” argument is in Allen’s own voice twenty months before the act. It first appears, addressed to President Biden, in August 2024 over U.S. forbearance toward Russian attacks on Ukrainian civilians. The structure is identical to the manifesto’s: visible-but-tolerated mass violence + claim of helplessness = complicity. What changed between August 2024 and the manifesto is the target of the charge, not the argumentative form.

On Allen’s biblical hermeneutic:

• The X corpus confirms the broadly progressive shape of Allen’s biblical voice. The dominant Christian framing he amplifies is in-house critique of MAGA evangelicalism, led by Rev. Benjamin Cremer (Wesleyan-Methodist). Conservative-Christian and right-leaning voices register at essentially zero in his amplification network across both corpora.

• None of the biblical texts that anchor the manifesto’s first-objection rebuttal appears anywhere in his digital corpus. Allen never engages, in either his own posts or in retweets, with Mt 5:38–42 (the Sermon-on-the-Mount text the manifesto rebuts), with Romans 13, or with the Pauline / Petrine submission texts. His native scriptural materials are Revelation 13–14, John 8:44, James 2:19, 1 Corinthians 5 / 8–10, and Matthew 18 — none of them pacifism or political-submission texts. He reaches for Mt 5:39 only in the manifesto itself, in what reads as a reply to a left-pacifist Christian objection he expects.

• The X corpus confirms Allen as a careful and creative reader of both scripture and fiction. The figural-but-not-prophetic mode I described in the original Bluesky reading is on display across both corpora and across both kinds of source material: scripture (John 8:44, 1 Corinthians 5, Revelation) and speculative fiction (Warhammer 40K, Asimov’s Foundation, Tolkien, Pratchett, anime). The hermeneutical discipline runs in parallel.

• The Revelation emphasis appears more striking now in light of the X period’s near-silence on biblical material. In the X phase the biblical-figural mode is John 8:44 (Christian discursive tradition) and pop-mythological (Warhammer-Nurgle, Skaven). Revelation enters Allen’s vocabulary only after the November 2024 election. I still do not think this makes Allen an apocalyptic fanatic, but Revelation may have done more affective work in shaping his sense of the situation’s gravity, in the Bluesky and manifesto period, than I had earlier credited.

Where this leaves things:

• Allen’s own posts trace a slope toward greater willingness to accept political violence — and this slope is not mirrored in his sources. Five or six items between February 2025 and April 2026, in his own voice, constitute structural endorsement of armed citizen response: a verbatim Declaration-of-Independence complaint against George III applied to Trump, paired with redirection toward “2nd amendment-related actions”; “more guns than people”; “best time to buy a gun”; revolt as a legitimate option in a four-item taxonomy; “Jan6 but better”. None is operational; none names a target. The trajectory is real, and it is Allen’s own work — not modelled by anyone in his amplification network.

• The framings on which Allen builds the diagnosis, by contrast, are in the sources. Trump as fascism-coded, on the Hitler-comparable tier, in a broader Weimar-to-Reich template, is the cumulative reading the sources themselves support. What makes the case difficult is that those same sources also explicitly reject political violence as the response. For Allen the diagnosis and the response-framework are not independent commitments: the Reformed-Protestant tradition he brings converts sufficiently severe tyranny into an obligation of tyrannicide undertaken against the actor’s own life.

• My first blog post’s conclusion stands and is sharpened. Unless specific individual psychological factors made Allen perceive these widely shared framings with a severity nobody else around him will replicate, the threat of political violence from mainstream Christian readers — readers who are not in any meaningful sense fringe, who hold both the diagnosis and a Reformed-Protestant ethic on which sufficient tyranny converts tyrannicide into duty — should be taken more seriously than has so far been recognised.

How the X corpus was recovered, and the comparison limits between the X and Bluesky archives, are described in the Appendix. Every individual claim above is anchored to verbatim posts with date and Wayback URL.

II. Allen’s politicisation

Having analysed the corpus of over 4,000 tweets*, it seems to me that three developments can be discerned in Allen’s political voice on X.

The Russian war on Ukraine

The X corpus lets one date Allen’s politicisation more precisely than was previously possible. Through 2021 and most of 2022, the @CForce3000 account contains 2,424 tweets, almost all of them about Super Smash Bros. Ultimate: frame data, character-matchup analysis, tournament play, the YouTube channel where Allen posted dozens of short videos as a Mii Brawler main. Across 2023 and the first four months of 2024 — sixteen months in total — the X corpus contains eleven tweets. This is a near-silence the corpus alone cannot explain. Of those eleven tweets, eight are still gaming-coded (frame-data RTs, smash crew-battle commentary, Gintama anime celebration); only three are political, and all three are Russia–Ukraine-coded. The earliest political post in the corpus is a 16 April 2023 retweet from @MargoGontar about a Russian missile strike on Slov’iansk; a 6 June 2023 retweet from @StarskyUA about Russian troops digging trenches in the Chornobyl exclusion zone is the second.

Allen’s online political activation, as far as the present data make it accessible, thus seems to have begun in response to the Russian war on Ukraine, fourteen months after that war’s full-scale escalation in February 2022. The flood of explicitly political material then begins in May 2024, peaks across June through November 2024, and continues into a brief Bluesky-overlap period before Allen migrates to Bluesky on 16 November 2024 — explicitly framing the move as a departure from X:

leaving for bluer skies as soon as I get my data archive / I don’t think there’s much reason to be on here anymore

(Wayback snapshot, 16 November 2024). The Bluesky corpus that my first blog post analysed begins, in other words, after Allen’s most active political-X phase has ended.

The sequence has implications for my first blog post’s argumentative architecture. The Christian-critique voice that becomes so prominent on Bluesky — speaking from within Christianity, policing the boundary between real Christians and merely so-called ones, the explicit Reformed self-identification — is, on the X record, layered on top of an already-formed pro-Ukraine, anti-Russia, anti-Trump position whose foundations in Allen’s own voice predate any visible biblical engagement. The first datable political RT (April 2023) is pro-Ukraine. The first own-voice “Trump as traitor” framing (May 2024) is political-legal. The first own-voice biblical engagement (May 2024, John 8:44 against Trump) comes a few weeks after his return to political posting. The Christian critique of MAGA Christianity layers on top of an already-formed political analysis. At least with respect to its appearance online, politics is anterior to the in-house Christian polemic. This seems to add at least some more weight to my earlier analysis, according to which there are motivations running deeper than his reading of the Bible alone could explain.

The November 2024 election

One specific moment from that political foundation merits separate attention: Allen’s reading of the November 2024 election outcome. My first blog post read the “traitor” charge against Trump primarily as a judgement about Trump’s pre-election conduct — the Putin contacts, the Russia-asset framing, the foreign-policy moves the academic-anti-authoritarian commentariat had been documenting. The X corpus shows that within hours of the election call on 6 November 2024, Allen had also formed a settled belief that the November 2024 election itself had been won by fraud. Across 6 to 14 November 2024 he amplified, at substantial volume, a sequence of left-coded election-fraud claims: that “20 or so million votes are unaccounted for” (RT @PamKeithFL, 6 November); that there had been a “14 million swing” between the popular and Senate votes (RT @birdog456); that the result was “likely fraudulent” (RT @PamKeithFL, 8 November); the “#TrumpCheated” and “#DoNotConcedeKamala” hashtags; the under-vote-anomaly accusations (RT @PamKeithFL, 10 November, on North Carolina); and the claim that Elon Musk’s Starlink terminals were used to transmit election results in swing states (RT @Mollyploofkins, RT @AndreaChalupa, RT @JudithLanigan, all on 10 November). The cluster culminates on 14 November with a retweet of the Bonifaz computer-security-experts open letter to Vice President Harris urging a forensic recount.

The Bluesky-era “rubberstamping the election of a traitor” framing of 6 January 2025 reads, in light of the X material, as a judgement that is illegitimate twice over in Allen’s own account: it is a judgement about Trump’s pre-election conduct and about the legitimacy of the November 2024 outcome itself. The manifesto’s first-objection rebuttal is, I think, untouched by this finding — the rebuttal turns on the complicity-through-inaction logic, not on the legitimacy of the election. But the “traitor” lexicon throughout both corpora is doubly anchored: it picks up Trump’s pre-election conduct and a stolen-election conviction that crystallised the day after the call.

One detail of this sequence is hard to ignore: the fraud-narrative cluster accumulates within hours of Allen’s retweet of Snyder’s pre-emptive warning against exactly this kind of narrative. On election day itself (5 November 2024), Allen retweets Snyder warning that “This platform will be set all day and in days to come to promote lies about election fraud. If you see that, remember that it is part of a plan by losers to ruin our democracy.” Within twenty-four hours Allen begins amplifying exactly the kind of fraud narrative Snyder warned about, only with the political polarity flipped. He does not appear to register the symmetry. The worry I want to raise here is narrow: it concerns Allen’s factual reasoning about the 2024 election, not his biblical hermeneutic.

The shift to complicity framing

Both in the manifesto and in the Bluesky posts, Allen justified the attempted assassination with the argument that not acting would itself amount to complicity in Trump’s actions, which Allen evidently regarded as egregious. The X corpus sheds further light on Allen’s line of thought here.

The most rhetorically developed Bluesky formulation of the complicity-through-inaction argument — “if you permit lynchings/genocide/extrajudicial murder and abduction in your societies and say you can’t do anything about it, you have become complicit in it through lack of resistance” (Wayback snapshot, 17 November 2025) — was, in essentially the form the manifesto would later use. Earlier Bluesky posts already gestured at the same logic: the 10 December 2024 mockery of Biden’s “thoughts and prayers”; a 2 February 2025 charge that the New York Times is “almost certainly complicit at some level” in Russian war propaganda; a 17 March 2025 generalisation to “quite a lot of the media is complicit”; and the 18 March 2025 hardening of the move into doctrinal form: “at what point is someone who abdicates leadership but does NOT abdicate their position just straight up complicit?” (Wayback snapshot, 18 March 2025). The Bluesky logic was visibly in motion well before the November 2025 anchor.

In the X corpus, the earliest such language comes from Timothy Snyder, the Yale historian and author of On Tyranny, whom I return to in §III. Allen retweeted his post on 28 May 2024 (Wayback snapshot, 28 May 2024):

The ‘just imagine’ tweets are a form of genocide denial. They are meant to instill a false premise. ‘Just imagine if Putin…’ tells your brain that he didn’t when he did. Everyone who tweets this material is complicit in genocide denial which means complicit in genocide.

The topic of US treatment of the Ukraine war features prominently in Allen’s own voice both before and after this retweet. The language remains at a more neutral framing of inexcusable inaction, without drawing out the complicity-implication explicitly. On 18 August 2024, Allen demands that “ukraine doesn’t simply ignore biden lol when it comes to missiles on russia / like, he ain’t gonna be in office next year either way / what’s he gonna do, trickle aid more?” (Wayback snapshot, 18 August 2024). On 27 August 2024, he addresses President Joe Biden directly (Wayback snapshot, 27 August 2024):

@POTUS If I may make so bold, Mr. President, why is our response reduced to thoughts, prayers, and repairs? This insistence on Ukrainians to merely endure attack after attack without public explanation of our seeming inaction is grating beyond measure.

On 25 September 2024, he attacks Vance’s diplomatic restraint as morally insufficient: “this comment by Ukrainian President Zelenskyy is WAYYYYYYYY more restrained than I’d be in his shoes and Vance deserves much worse to be said about him by a lot more leaders” (Wayback snapshot, 25 September 2024).

While in his X posts Allen thus remains more cautious about explicitly framing inaction as complicity, we see a further push in that direction in the Bluesky posts. Here Allen not only takes up the explicit language of Snyder’s original post, he also broadens its applicability: from leadership-abdication-while-holding-position (March 2025), through passive permitting of mass violence by anyone within range of preventing it (November 2025), to the manifesto’s bystander move, where presence at an event hosted by an administration one holds responsible suffices for complicity.

III. Allen’s social media bubble(s)

The X corpus, with its much larger record of who Allen amplified, lets one see his online intellectual environment more clearly than the Bluesky-only material allowed. Across roughly 1,900 amplifications — retweets and quote-tweets, the posts in which Allen passed someone else’s content along rather than originating his own — in those 4,370 tweets, the most-retweeted single source is Rev. Benjamin Cremer (@Brcremer, 32 retweets), a United Methodist pastor at the Cathedral of the Rockies in Boise. Cremer’s biography is itself relevant context: trained in the Wesleyan-Holiness tradition and ordained in the Church of the Nazarene in 2018, he was pushed out of that denomination in 2021 over his LGBTQ+-affirming positions before being received into the UMC, which preserved his credentials. By the time Allen begins retweeting him heavily in May 2024, Cremer is a high-profile online voice writing as a Christian against the deformations of MAGA-Christianity. Behind Cremer trail @BlackKnight10k (28 retweets, anti-Trump activist), @MuellerSheWrote (23, the podcaster Allison Gill), @whstancil (22, progressive policy commentator), and @JayinKyiv (20, pro-Ukraine commentator) — all by a few posts. Cremer’s lead is real but not dominant. What stands out is the kind of voice that leads: not a politician, not a journalist, not a public intellectual, but a pastor — one whose own trajectory had taken him from a conservative Holiness denomination, by way of credentialing pressure over progressive social positions, into the mainline Methodist fold.

The 32 distinct Cremer retweets I have catalogued add up to a striking prefiguration of the in-house Christian critique I traced in my first blog post. Some examples:

• 24 May 2024: “Beware of any Christian movement that demands the government be an instrument of God’s wrath but never a source of God’s mercy, generosity, or compassion.”

• 31 May 2024: “Dear Christian, One of the biggest things that destroys your credibility is when you take an unflinchingly ridged stance towards the ‘sins’ of your fellow citizens but then claim the ‘system is rigged’ when your preferred presidential candidate is held accountable for his sins.”

• 16 June 2024: “A Christianity that needs presidents and governments to do its bidding is a Christianity that stopped believing in the power of its God and the truth of its message.”

• 13 September 2024: “’Thou shalt not bear false witness against thy neighbor’ is the 9th commandment. When we Christians bear false witness against our neighbors, we not only invite harm onto them, but we destroy our reputation and any chance we have to be treated as people who ‘speak the truth.’”

• 21 September 2024: “Christian hypocrisy looks like constantly condemning the sins of ordinary people we don’t like in our culture while simultaneously shrugging off the sins of people in political power we like.”

• 8 November 2024: “When Christians can’t even acknowledge the sins of the person they support as president, let alone demand they repent from them, they reveal to the world that they have traded all moral and theological integrity for the sake of political power.”

The 13 September 2024 retweet deserves a separate flag. The Reformed numbering of the Decalogue that I had identified on Bluesky as a confessional fingerprint distinguishing Reformed from Catholic and Lutheran enumeration is, on the X record, already prefigured in what Cremer himself uses, and amplified by Allen ten months before his own “as a Protestant, what in the 2nd / 3rd commandment violation is this iconography” post (Wayback snapshot, 5 April 2026). Cremer is supplying not just the no-true-Scotsman vocabulary but the specific scriptural infrastructure in which Allen’s later confessional moves operate.

After Cremer, the Christian voices Allen amplifies most heavily are uniformly mainline-progressive — @johnpavlovitz, @JohnFugelsang, @PastorBenMarsh, and a handful of smaller progressive-pastor accounts (Marotta, Lambert, Chafee, Hale), totalling around 51 retweets across the cluster. None of the conservative-Christian or right-leaning voices one might have expected — given Allen’s confessionally-Reformed background — register meaningfully in his amplification network.

After Cremer, the picture splits into two larger groups and three smaller ones.

The Pro-Ukraine commentariat is the loudest of them, with roughly 288 amplifications. Members include @JayinKyiv, @SpaghettiKozak, @KareemRifai, @BohuslavskaKate, @fellaraktar, @IAPonomarenko, @United24media, @reshetz, @ZelenskyyUa, @DarthPutinKGB, @TetyanaUkrainka, @CanadianKobzar, @AnnaOdesitka, @KhaterDiana, @P_Kallioniemi, @Kasparov63, @Maks_NAFO_FELLA, and dozens of smaller accounts. Some of them are themselves academics or former officials (Phillips P. O’Brien, Mark Hertling, Alexander Vindman, Michael McFaul); others are Ukrainian voices on the ground. The vocabulary they supply — appeasement, no red lines, untie Ukraine, words don’t shoot down missiles — becomes the vocabulary Allen reaches for. Crucially, this is where he picks up the moral-historical analogies he will later deploy: Munich 1938, Czechoslovakia, the Molotov–Ribbentrop pact, the broader Weimar-to-Reich template. A 23 May 2024 retweet of @fellaraktar — “If we had stood behind Czechoslovakia in 1938, the world may have been saved 6 years of slaughter and 50+ million lives. Instead we stabbed them in the back to provide ourselves with the delusion of safety” — is the earliest in-corpus appearance of the appeasement-equals-complicity argument that he carries through into Bluesky.

The Pro-Democrat / Resistance commentariat is the largest cluster by raw volume — about 317 amplifications. It includes @MuellerSheWrote, @whstancil, @Angry_Staffer, @BlackKnight10k, @FPWellman, @atrupar, @TristanSnell, @tedlieu, @JoJoFromJerz, @PamKeithFL, @donwinslow, @SethAbramson, @RonFilipkowski, @KamalaHQ, and similar accounts. What this cluster contributes is the affective texture — the traitor-axiom, the post-election fraud narratives, the MAGA-villain catalogue — but it does not, on close reading, supply the shape of the argument, the historical-analytical framework, or the ethical-philosophical vocabulary that the manifesto turns on. Its main contribution to his later vocabulary is the pervasive “Trump is a traitor” / “Russian asset” lexicon, which @PamKeithFL in particular pushes hard from June 2024 onwards, and which Allen carries forward almost verbatim into the Bluesky period.

The academic anti-authoritarian cluster (~79 amplifications) is the smallest of the four substantial groups but disproportionately important. The largest single voice within it by retweet volume is Ruth Ben-Ghiat (15 retweets); after her come Garry Kasparov (13), Snyder himself (9), Phillips P. O’Brien (6), Mark Hertling (4), Alexander Vindman (4), and a long tail of single- and double-retweet appearances by other historians, legal scholars, and former officials (J. Michael Luttig, Anne Applebaum, David Rothkopf, Anders Åslund, Lawrence Tribe, Andrew Seidel, Asha Rangappa, Jon Meacham, Carole Cadwalladr, Michael McFaul, David Blight, Jed Shapiro).

Although the Snyder retweet of “Do not obey in advance” on 26 October 2024 — Lesson 1 of On Tyranny — is the only place in the corpus where Allen amplifies a Snyder formulation that is identifiably a Snyder formulation as such (rather than a daily-news take), it is worthwhile to analyse this specific influence in more detail because of striking parallels in terminology and concepts. Eighteen months after the retweet, Allen revisits the same maxim in his own voice on Bluesky: “i mean, it wasn’t ‘do not comply EVER’” (Wayback snapshot, 8 March 2026). The immediate context of this remark — what Allen was replying to — is not preserved in the corpus, but the post itself shows him defending the maxim against an over-reading: he has internalised it enough to argue against a strawman version. He does not attribute it to Snyder or to On Tyranny here or anywhere else; the maxim had become a widely-circulated resistance commonplace by 2025–2026.

On 7 November 2024, two days after the election call, Allen formulates the Hindenburg parallel in his own voice (Wayback snapshot, 7 November 2024) — invoking the German president’s role in legitimising Hitler’s chancellorship as an analytical template for the second Trump term. Snyder, working in the same Weimar-to-Reich analytical mode, publishes on the closely related August 1934 loyalty-oath analogue the very next day. The Hindenburg framing as such is not Snyder-specific — it circulates in the Pro-Ukraine commentariat too, as noted above — and the chronology of these two posts rules out direct Snyder-to-Allen lineage on this particular move. What the convergence does suggest is that Allen and the analytical mode Snyder represents landed on the same Weimar-collapse template within twenty-four hours of the election result, in essentially the same idiom.

Allen was still reading Snyder shortly before the attack. On 8 March 2026 — six weeks before the events of 26 April 2026 — Allen quote-posted a Snyder Bluesky thread about The Desire for Terror, Snyder’s substack analysis of how the Iran war could provide the Trump administration with a pretext to “federalise” or postpone congressional elections. In his accompanying comment (Wayback snapshot, 9 March 2026), Allen translates Snyder’s argument into a schoolyard metaphor and adds a cynical postscript:

A purpose of the bully beating the shit out of that hooligan might be so that when the hooligan hits back, the bully has an excuse to beat the shit out of everyone else too for not stopping the hooligan for hitting the bully. If trump wants martial law he’s just gonna do it like everything else.

The Never-Trump Republican cluster (~70 amplifications) — Adam Kinzinger (16 retweets), Bill Kristol (4), Rick Wilson (4), George Conway (4), Project Lincoln (3), Liz Cheney (2), Sarah Longwell, Jennifer Rubin, J. Michael Luttig (3, also in the academic cluster) — supplies the constitutional-republican vocabulary, the “no kings” / “no one above the law” line. The July 2024 Trump v. United States immunity decision seems to be the moment everything tips: he stops just amplifying rule-of-law content and starts articulating it himself. This is the cluster that, fused with Founding-revolutionary vocabulary, will produce his verbatim Declaration-of-Independence deployment of 18 February 2025.

One thing worth stating plainly: Allen’s online intellectual environment, on the X record, was overwhelmingly the milieu of progressive American Christianity in confrontation with MAGA, embedded inside a broader Pro-Ukraine and academic-anti-authoritarian commentariat. The conservative-Christian or right-leaning voices one might have expected, given his confessional Reformed background, register at essentially zero in his amplification network. Whatever influence Allen’s original church milieu had on his political formation, his online reading was not the channel for it.

IV. Biblical hermeneutics

The biblical material on X is sparser than on Bluesky — partly because Allen’s posting habits shifted between the platforms (more retweets on X, more originals and threads on Bluesky), partly because his biblical engagement intensified after the platform migration. Three things from X deserve separate attention, plus one closing observation that cuts across both corpora.

The first is the John 8:44 deployment of 7 May 2024 — chronologically the earliest Allen-original biblical citation in either corpus. Allen self-quote-tweets a now-deleted earlier post of his own to complete a verbatim citation against a hypothetical defender of Trump (Wayback snapshot, 7 May 2024):

In response to ‘Trump has done nothing wrong’, I think John 8:44 is appropriate here: ‘You belong to your father, the devil, and you want to carry out your father’s desire. He was a murderer from the beginning, for there is no truth in him. When he lies, he speaks his native language, for he is a liar and the father of lies.’

The translation is NIV 1984 (the singular “father’s desire,” contrasted with the plural “desires” of the NIV 2011 revision). Allen redeploys the same passage on 5 November 2024, election day, in reply to a Tiffany Trump post defending her father (Wayback snapshot, 5 November 2024):

John 8: 42-44 seems most appropriate for this: ‘You belong to your father, the devil, and you want to carry out your father’s desire. He was a murderer from the beginning, not holding to the truth, for there is no truth in him. When he lies, he speaks his native language,’-

We observe Allen here operating from within the Christian discursive tradition rather than against it. This is not an isolated rhetorical flourish: it is a settled move, used twice in essentially the same form across eighteen months. From all I could see, this Johannine devil-paternity language is Allen’s own scriptural reflex and is not directly borrowed from other influences he interacts with. The first deployment predates his first Cremer retweet by four days.

The second is the 1 Corinthians 5:12–13 deployment of 28 October 2024. In a small thread about Trump and Christian voting choices, a fellow user appealed to the well-worn maxim “Do not judge, lest you be judged” (Mt 7:1) as a reason for restraint. Allen’s reply (Wayback snapshot, 28 October 2024):

Perfectly good passage But we’re discussing so-called Christians, so something else applies: 1 Cor 5: 12-13 ‘What business is it of mine to judge those outside the church? Are you not to judge those inside? God will judge those outside. Expel the wicked man from among you.’

The exegetical move here is exactly what the Reformed church-discipline tradition does with Matthew 7. Allen does not say Mt 7:1 is wrong; he says it is the wrong text for the case at hand, because the case is internal-Christian discipline, for which the relevant Pauline text is 1 Corinthians 5. He then quotes the Pauline text in NIV 1984 phrasing (“expel the wicked man,” compared with NIV 2011’s “wicked person”) — and the verb of the quotation is not pastoral exhortation but excommunication. “Expel the wicked man from among you” is the formula Calvin reaches for in his own writings on church discipline; it is the formula encoded in the disciplinary chapters of the Three Forms of Unity that govern the URCNA congregation Allen attended. The exegetical move — distinguishing the out-group judgement Mt 7:1 forbids from the in-group judgement Paul both permits and commands — is so distinctively Reformed-confessional that, on its own and even in the absence of all the Bluesky markers, it would suggest a confessional rather than a generically evangelical formation. Read together with the Bluesky markers, it confirms the placement very strongly.

The third is Cremer himself, and the question of how the in-house Christian critique transfers across the platform migration. The Cremer retweet pattern on X is dense and sustained (32 retweets, May 2024 – November 2024), and adds up to a long rehearsal of the precise rhetorical move Allen later carries in his own voice on Bluesky. The Bluesky-side picture takes some untangling. Cremer is in fact very active on Bluesky — @brcremer.bsky.social, in the platform’s top tier by follower count, posting several times a week — but he essentially does not appear in the recovered corpus of Allen’s Bluesky activity that I can verify (no quote-posts, no captured thread context). Given the bare-repost recovery limitation discussed in the Appendix, the most plausible reading is that Allen continued to interact with Cremer’s posts on Bluesky, but did so through bare reposts of the kind the Wayback corpus cannot preserve. The April 2025 “actual Christians” boundary-policing post that my first blog post cited as a prime piece of confessional evidence is in fact a quote-post replying to former Republican Congressman Adam Kinzinger’s Bluesky observation that “Christians are not persecuted in the United States” (Wayback snapshot, 24 April 2025). Allen also quote-posts Kinzinger’s “Never expected authoritarianism would be ushered in by law firms” with the line “the actual ruler of America is the Constitution, so it makes sense that the people who interpret the Constitution would be those subverted to help destroy it” (April 2025), and Kinzinger’s “Project Blue Beam” conspiracy-theory post with an ironic “NASA teaming up with the antichrist and DEFINITELY not the head of spacex” (December 2024). On the Bluesky-side visible record, Adam Kinzinger is the most-frequent Allen interlocutor of any X-cluster member.

Across both corpora, one observation about Allen’s biblical vocabulary bears directly on how the manifesto reads. The manifesto’s opening biblical move is the citation of Mt 5:39 (“do not resist an evil person”) and the reasoned rebuttal to it. I searched for any prior Allen engagement with: the Sermon on the Mount; “turn the other cheek”; “do not resist”; Mt 5:38 / 5:39 / 5:40 / 5:41 / 5:42; Lk 6:29; or any of the Pauline / Petrine New Testament texts on submission to authorities (Romans 13, 1 Pet 2:13–17, Mt 22:15–22 / Mk 12:13–17 / Lk 20:20–26). Not one of these texts surfaces in either Allen’s own posts or in the cluster material he amplifies. His native scriptural materials are Revelation 13–14, John 8:44, James 2:19, 1 Corinthians 5 / 8–10, and Matthew 18 — none of which are pacifism or political-submission texts. He reaches for Mt 5:39 only in the manifesto. The simplest reading is that he is responding to an objection from a particular kind of reader — a pacifist Christian critic — whom he expects to invoke this text against him. The choice of interlocutor matters: the manifesto is not written to the MAGA right (who would invoke Romans 13) and not to the secular left (who would not invoke scripture at all). It is written to a left-pacifist Christian, in his vocabulary. This is consistent with my first blog post’s reading of the manifesto’s audience-selection.

Alongside the biblical material runs a parallel strand of secular figural identification that my initial blog post did not take up. It is visible across both corpora; what the X material adds is the earliest dated instances of it — the Warhammer 40K Nurgle identification of 12 May 2024 and the Skaven identification of 18 June 2024 — which sit chronologically before the Bluesky archive begins and let one see the full arc of the figural mode from before the November 2024 election through to the manifesto window.

Allen’s online voice routinely identifies political figures with archetypes from speculative fiction. Trump is mapped successively onto:

• Warhammer 40K’s Nurgle (the disease-and-decay god) — first in a 12 May 2024 Allen-original quote-tweet of a Trump anti-vaccine-mandate statement: “’Tell me you’re a Nurgle cultist without telling me you’re a Nurgle cultist.’” The Nurgle identification recurs on Bluesky in February 2025 (“Trump is clearly a Nurgle devotee”) and forms an ongoing parallel thread through to the manifesto window.

• Warhammer’s Skaven (cowardly, treacherous, sadistic rat-creatures) — 18 June 2024 Allen-original: “honestly calling them orcs is an insult to orcs / Skaven would be more accurate: Cowardly, treacherous, sadistic, narcissistic rat-creatures hellbent on enslaving the world by means of poorly equipped cannon fodder, suicidally flawed inventions, and widespread sabotage & subversion.”

• Asimov’s Foundation — the Mule — Bluesky, May 2025: “Trump is the Mule lolol.”

• Tolkien’s Wormtongue — Bluesky, April 2026: “This level of pathetic is normally reserved for things like Grima Wormtongue / ‘Won’t you purchase a tiny flag of Rohan to remind Mordor that even though I betrayed my country, all is lost, and a Nazgul rules in Edoras, I’m a good and faithful servant who is independent of the Eye?’” (applied to Vance).

• Pratchett’s Discworld — Reacher Gilt — Bluesky, August 2025: “Pratchett having Gilt operate out of ‘Tump Tower.’”

• Frieren-style anime demons — Bluesky, April 2025: “there’s a very specific kind of demonic that they’re coding as and it’s the Frieren kind / ‘Monsters capable of speech.’”

Two things follow from this. First, the move is structurally identical to the biblical figural one: typological identification without singular-prophetic identification. Trump is the Nurgle cultist, the Mule, Wormtongue kinda, as a structural matter — but he is not fulfilment of prophecy. Allen’s hermeneutical approach to prophecy in the Bible is not isolated; it is entangled with a more general strategy of making sense of the present through tropes and types of narrative characters.

Second, this gives a more textured account of what the apocalyptic shift between platforms actually consists of. On X, the dominant religious-figural idiom against Trump is Johannine (John 8:44) and mythological (Nurgle, Skaven, Wormtongue). Revelation enters as a distinct vocabulary only after the November 2024 election. By the Bluesky period — the one I examined in the previous post — Revelation has displaced the mythological frame as the dominant figural target, though pop-cultural frames continue to recur in parallel (the Nurgle identification, the Wormtongue identification, the Pratchett-Discworld identification, the Foundation-Mule identification) right through the manifesto window. The hermeneutical mode (figural, ironic-hedged, anti-accelerationist) is continuous; the biblical locus moves toward Revelation only after November 2024. Perhaps this adds some nuance to my finding from the initial blog post — that on the Bluesky evidence Allen does not look like an apocalyptic fanatic. I still think that holds. But Revelation may have done more to make him feel the gravity of the situation than I had previously realised.

V. Where this leaves things

The X corpus does not change the headline reading I gave in my first blog post. It strengthens it on every point that mattered. Allen’s biblical hermeneutic remains, when both corpora are read together, recognisably mainstream Reformed Protestantism — careful, figural, anti-dispensationalist, and uninvolved in the kind of apocalyptic activism that some of the early reporting suspected. The act remains grounded primarily in a political-ethical reading of current events, secondarily in a Christian tradition, and not deduced from his reading of Scripture. To call him or his manifesto “anti-Christian” remains, on this evidence, indefensible. To overemphasise the role of religion in his motivation, or to read the apocalyptic vocabulary as evidence of feverish prophetic certainty, also remains unwarranted.

What the X corpus adds is a more precise account of how Allen got there. The complicity-through-inaction argument that anchors the manifesto’s first-objection rebuttal has a long preparation, with specific dated stations, beginning in his own voice in August 2024 and traceable to Snyder. The intellectual environment he was assembling his analysis from is mappable: progressive American Christianity confronting MAGA, embedded inside a Pro-Ukraine commentariat and an academic anti-authoritarian commentariat, with no significant exposure to conservative-Christian or right-leaning voices. The biographical sequence is also legible: a pure-gaming online life through 2022, near-silence through 2023 broken only by a handful of pro-Ukraine retweets, an explosive politicisation phase from May 2024 through the November 2024 election, and a migration to Bluesky immediately after the election that re-reads the same political analysis through an increasingly Revelation-framed apocalyptic idiom.

That leaves a deeper question. Does the additional material now in view produce a clearer picture of how Allen’s general ethics — shaped not least by Christian belief — accommodated his decision to attempt to assassinate Trump?

There is no post in either corpus, by either Allen himself or someone he retweets, that directly calls for Trump’s death in any tone, ironic included. The Butler attempt of 13 July 2024 prompts no celebration: Allen’s posts that day are entirely about Smash Bros. Ultimate gameplay, and he retweets no one celebrating the shooter. The Routh attempt of 15 September 2024 prompts denialism — the attempt was staged, no one shot at Trump that day — which is, structurally, the opposite of celebration. If “death” is mentioned in conjunction with Trump, it is uniformly about Trump’s policies causing others to die (anti-vaccine, ICE, healthcare cuts), not about Trump dying. There is no operational planning, no methods discussion, no targeting, no location reconnaissance. Allen never expresses personal intent in connection with violence (no I will, I’m going to, I’m prepared to); he names none of the obvious figures as targets, and his verbs against them are vote against, impeach, sanction, prosecute, excommunicate, arrest, deport, fire, censure. The tyrannicide tradition is not invoked with any explicit application-claim — no Stauffenberg, no Bonhoeffer, no July 20, no Brutus-and-Caesar; the historical analogies Allen does deploy are the American Revolution, the Civil War, and Allied force against Nazi Germany, justification frameworks rather than tradition-invocations. There is no smoking gun.

What the corpus does contain in Allen’s own voice is something more like a slope. From early 2025 onwards there is a small but real set of posts — five or six items — that endorse armed citizen response in structural terms, without naming a target and without using a verb of violence. The most charged is a two-post unit dated 18 February 2025, in which Allen first quotes verbatim the Declaration of Independence’s complaint against George III about “transporting large Armies of foreign Mercenaries to compleat the works of death, desolation, and tyranny” and applies it to Trump (Wayback snapshot, 18 February 2025), then in the immediately following post redirects political attention away from electoral remedies — “not as much on the midterms or congress getting its act together as they should be on understanding 2nd amendment-related actions / Because ngl it doesn’t look like we’re getting to the midterms (or even july)” (Wayback snapshot, 18 February 2025). Read as a single rhetorical unit, the pair places the founding document of justified armed revolt against a tyrant alongside the constitutional provision the American gun-rights tradition treats as authorising armed citizen response. There is no verb of violence in either post; there is no named target; there is no operational claim. There is, nonetheless, an argument being built.

Other items extend the same pattern. A 27 February 2025 post in Allen’s voice observes that the U.S. citizenry has “more guns than people” and is “just waiting for the treasonous idiots to realise that this is very suicidal on their part” (Wayback snapshot, 27 February 2025). A 4 September 2025 quote-post boosts another user’s gun-buying counsel (“Best time to buy a gun was [your state’s waiting period] days ago / Second best time is today”) in the immediate context of expanded ICE actions (Wayback snapshot, 4 September 2025). A 31 January 2026 post lists the legitimate options for an American citizen who “can’t afford to risk arrest” as “a) leave / b) die / c) revolt / d) capitulate” (Wayback snapshot, 31 January 2026); the framing is taxonomy-of-options, not direct exhortation, but listing revolt alongside leave, die, capitulate as a legitimate option is itself the move. An 8 March 2026 quote-post, six weeks before the events of 26 April, sardonically warns Trump that his policies will produce “the armed, depressed, and lonely young men of this country with the choice between ‘go die in the middle east to distract from a pedophile’s incessant crimes’ and ‘do Jan6 but better’ / what could possibly go wrong for you here?” (Wayback snapshot, 8 March 2026) — the “Jan6 but better” reference, meaning more successful, with its violent objectives realised, is the most specific predictive item in the corpus. Two earlier items belong here. On 15 September 2024, hours after the Routh attempt, Allen retweeted @covie_93’s “Be careful trump, too many of these assassination attempts and people will become numb to them like school shootings” (Wayback snapshot, 15 September 2024) — a sardonic warning whose rhetorical effect is normalisation rather than celebration. On 8 April 2026 Allen quote-posted another user’s claim that Trump “is literally one of those villains that if you beat his ass hard enough, he’ll join your team” with the comment “not really actionable cause no way schumer just canes him into acting his age, but, like, it would probably literally work on him” (Wayback snapshot, 8 April 2026) — caning being a reference to Preston Brooks’s near-fatal 1856 caning of Senator Charles Sumner. Each item is sub-advocacy, ironic-warning, or option-listing rather than direct exhortation. None invokes the tyrannicide tradition; none names Trump as a target; none rises to operational severity. Taken together they trace an arc — rhetoric becoming gradually more tolerant of political violence across 2025 into early 2026, in the language of citizen-arming and revolutionary-tradition framing, without ever naming Trump as a target or planning anything. The framework is in the diet; what the manifesto adds is the targeting and the act.

The amplification network around these own-voice posts has a structure worth reading in two passes. First, the authors whose framings Allen consumed consistently and explicitly condemn political violence in democratic societies in their own published voices. This is not an inference; it is what they say. Timothy Snyder, the analyst whose argumentative shape Allen most directly took up, has written across the period in question that political violence in democracies is a category mistake. His Political Violence essay (Substack, July 2024), published the day after the Butler attempt on Trump’s life, opens with “We should all condemn political violence” and continues that “this next election will be settled by the number of votes, rather than by threats, coups, beatings, or murders.” In What a terrible thing it is (Substack, December 2025) he writes that “there cannot be a right to kill someone else.” And in The Desire for Terror (Substack, 8 March 2026) — the essay Allen quote-posted six weeks before the attack (Wayback snapshot, 9 March 2026) — he argues that politically motivated violence against the Trump administration would supply that administration with the operational pretext to consolidate authoritarian rule, and that “the self-terrorism chain is closed only if we play our assigned part” by refusing the pretext. On Tyranny is in the same key throughout: the model figures the book names — Rosa Parks, Solidarność, Václav Havel, Teresa Prekerowa rescuing Jews from the Warsaw ghetto — are nonviolent; Lesson 6 identifies paramilitaries as the antagonist of free society; Lesson 7 asks those who carry weapons in public service to refuse irregular orders; Lesson 18 closes with Arendt’s “I was no longer of the opinion that one can simply be a bystander” but is itself a warning not to react with violent escalation when a terror event arrives (“do not fall for it”). Beyond Snyder, the picture is the same. Ruth Ben-Ghiat identifies authoritarianism with the use of violence and locates resistance in institutional categories. Garry Kasparov, himself a target of political violence, advocates sanctions and isolation. The Pro-Ukraine commentariat licenses interstate armed defence by Ukraine against Russian targets but makes no transposition of that licence to domestic political violence. The Never-Trump cluster recommends impeachment, prosecution, and electoral defeat. Rev. Benjamin Cremer, whose voice leads Allen’s Christian-progressive amplifications, operates within the Wesleyan-Methodist tradition, which places a strong emphasis on social justice — codified in the United Methodist Church’s Social Principles. On pacifism specifically, the tradition is not a historic peace church (Mennonite, Quaker), but it engages with peace and nonviolence more seriously than much of American evangelicalism: war is condemned as incompatible with the teachings of Christ, and nonviolence is framed as a Christian moral ideal, without being made a binding denominational identity. None of the thirty-two retweeted Cremer items in Allen’s X feed contains a permission, even implicit, for armed response. What Allen was reading is, taken together, uniformly anti-violence; the most substantial of those voices explicitly forecloses, in essays Allen had read in the weeks before the attack, the act-type he committed.

Second — and this is what makes the case difficult — the same body of writing supplies a diagnosis. Trump is read as fascism-coded; the Hindenburg parallel, the August 1934 loyalty-oath analogue, and the broader Weimar-to-Reich vocabulary circulate continuously; Project 2025 is read as a blueprint for personalised authoritarian rule; the appeasement-equals-complicity argument carries the Munich-1938 reference; Trump v. United States is read as the constitutional break; Cremer marks the Christian-political settlement Trump enjoys as a betrayal of the gospel; the daily affective vocabulary of the Pro-Democrat resistance commentariat supplies the traitor and Russian asset lexicon. Read together, these sources place Trump on the Hitler-comparable tier — and this is not Allen’s distortion of what they say but the cumulative reading they themselves support. Allen took it up in his own voice: the 7 November 2024 Hindenburg formulation (Wayback snapshot, 7 November 2024); the 18 February 2025 application of the Declaration of Independence’s grievances against George III to Trump, with its immediate turn toward “2nd amendment-related actions”; the November 2025 complicity-through-inaction formulation that the manifesto’s first-objection rebuttal later carries forward almost verbatim.

I had set out in my first blog post the long Reformed reading of Mt 5:38–42, on which Christian love of neighbour permits — and, in extremity, on the same logic, may even require — violence in defence of the neighbour, including against political authority. The framework is real, not fringe, and remains live in confessional Reformed circles. Within it, where the threshold of diagnostic severity is sufficient, the question of tyrannicide is converted from a permission into an obligation, one the actor undertakes against his own interest and against his own life. That this logic is in operation in Allen’s case is visible in the manifesto’s own postscript:

if anyone is curious how doing something like this feels: it’s awful. I want to throw up; I want to cry for all the things I wanted to do and never will, for all the people whose trust this betrays.

The postscript is not the voice of someone who expects gain or even relief; it is the voice of someone who has performed a duty he understands as personally ruinous.

Two things put together — a diagnosis the sources cumulatively support, placing Trump on the Hitler-comparable tier on its own weight; and a Christian ethic on which sufficiently severe tyranny converts tyrannicide into an obligation of love of neighbour, undertaken against the actor’s own life — make the situation I had warned about earlier in general terms now visible in specific operation. The authors hold diagnosis and response as independent commitments. They can therefore say “Trump is fascism-coded; this is Hitler-comparable; do not respond with violence” without internal contradiction. For Allen the two commitments are not independent: the framework joins them. For a reader who takes the diagnosis at the seriousness the diagnosers themselves take it, and who holds the Reformed Christian doctrine on tyranny in which the love-of-neighbour obligation applies above a sufficient threshold, diagnosis-taken-seriously and framework-taken-seriously together push toward a conclusion the authors do not draw and explicitly reject.

What the additional tweets add to the earlier general observation is the specific shape of the present case. What Allen was reading was not a fringe set of materials; the framings he received were widely available across mainstream alarmed-Democratic and Pro-Ukraine commentary in 2024–2026; the tradition he brought from his church milieu is not fringe either, but a major strand of Protestant ethical reflection with continuous historical use against political extremity. Many millions of Americans were reading versions of the same material; some non-trivial number of them hold versions of the same framework; almost none did what Allen did.

We do not know whether individual psychological factors played a role in Allen’s perceiving the framings he read in a specifically intense manner — placing Trump not merely in the vicinity of a fascist but on the level of a tyrant like Hitler whose removal had to be effected at all costs, including the actor’s own life. I have no insight into that and do not want to substitute a textual analysis for the kind of empirical psychological assessment I am not in a position to make.

And yet, working through the new sources alongside the older ones, and even with that caveat in mind — which I am not trying to downplay — I am more concerned than I was before. How many of those millions of readers, holding diagnosis and framework together, have not acted as Allen acted but are willing to? How many quietly think the act might be a good thing, and are simply waiting? How many do not act because they lack the courage that, on the framework’s own terms, the act would call for?

Snyder’s Lesson 20 — “if none of us is prepared to die for freedom, then all of us will die under tyranny” — was written in a martyr-of-conscience key, with a clear argument against political violence elsewhere in the same body of work, as I have set out above. But the line, lifted from its setting and rendered as a maxim of generic courage in the face of tyranny, fits the substantive structure of the position Allen brought to it without difficulty. The harder reckoning with this case is therefore not with Allen alone, who acted, but with the question how many readers of the same materials, holding the same framework, are now in the same place he was, with perhaps the courage or its absence between them and the same conclusion. That is what I find more alarming, with everything in view, than I had been before.

*Appendix: Method note on the X corpus and the cross-platform comparison

The @CForce3000 account on X was taken offline at some point before the present incident. CNN itself reported having reviewed more than 4,000 posts preserved on the Internet Archive. Working from the Wayback CDX index, I was able to reconstruct 4,370 of 4,399 unique tweet IDs ever observed for the account — an effective coverage of 99.3%. The 29 IDs I could not recover are either URLs the Wayback Machine indexed but never actually fetched (26) or captures of post-suspension Twitter error pages that contain no content (3). The reconstruction itself was nontrivial because X changed its rendering during the period covered, and because some of Allen’s retweets were captured at the moment of an HTTP 302 redirect to the source-tweet URL; those technicalities I will spare the reader.

A specific point of methodological transparency that goes beyond what was possible for my first blog post. The X corpus contains explicit retweet metadata: when Allen retweeted someone, the source author and original text are preserved. This makes a quantitative map of his amplification network possible: 1,778 retweets, 138 quote-tweets, 2,454 originals across 984 distinct sources. The Bluesky corpus is structured differently. Bluesky’s takedown of the @coldforce.bsky.social account took down not only the public-web routes I described in my first blog post but also the routes through which one would normally inspect Allen’s bare-repost activity (i.e., reposts of someone else’s post without Allen’s own commentary). I attempted four independent recovery paths for that data and each returned a takedown-propagated null: the public AT-Protocol AppView at public.api.bsky.app returns a structured 400 error against the account’s DID; the AT-Protocol Personal Data Server returns RepoTakendown for listRecords against the app.bsky.feed.repost collection; the public Constellation reverse-index returns zero linking-records (its 10-day retention window for the takedown’s date range has long since lapsed); and the public ClearSky API returns 410 Gone for any DID-scoped query against this account. The Wayback Machine has captures of fourteen profile-root snapshots of the @coldforce.bsky.social URL between April 2025 and April 2026, but each is the bare 12-kilobyte SPA shell — the captured Bluesky pages were never rendered with the API responses Wayback would have needed to play back the feed in subresource form, and the AT-Protocol endpoints (getAuthorFeed, listRecords over the repost collection) do not appear in Wayback’s CDX index for this DID at all. Bluesky-side bare reposts are therefore genuinely not recoverable through any standard external research path that I have been able to test. The remaining theoretical avenue is private firehose archives held by AT-Protocol researchers or data-broker services, which would require direct outreach beyond the scope of this analysis.

A note on the comparison itself. Where I report below that an X-side amplification source “drops out” on Bluesky or that Allen interacts more with a particular set of accounts on Bluesky, those statements concern the visible record — Allen’s originals (734 posts), his quote-posts and replies (where the parent or quoted post was captured), and the thread-context I have for 498 of the 734 posts (of which 126 rendered fully). The visible record cannot rule out bare-repost activity that the Wayback captures did not preserve. I will say “on the visible Bluesky record” or similar where the limitation is material. No quantitative claim I make below depends on an assumption about whether Allen did or did not bare-repost any specific source on Bluesky.

The X corpus is in some respects thinner than the Bluesky corpus despite being numerically much larger. Allen’s posting habits shifted between the platforms — more retweets on X, more originals and threads on Bluesky. Where the X corpus is silent on something attested on Bluesky, that silence reflects his use of the two platforms rather than a contradiction. Where the X corpus genuinely contradicts something attested on Bluesky — which it almost never does — I will say so.