Hello and welcome to Regulator, a newsletter that takes Verge subscribers into the smoke-filled back rooms of Washington as they become the Zyn-filled back rooms of Washington. (Although Tucker Carlson’s ALP is more popular among a certain set.) Not a Verge subscriber yet? Sign up here today! I’m not exposing myself to all these carcinogens for nothing.
Do you know something that’s not in any of those rooms? Send all tips to tina.nguyen+tips@theverge.com, or to my Signal account @tina.nguyen19. At the very least, you’ll save me a dry cleaning bill.
There’s a story that I’ve been chasing for months, along with, apparently, every other political reporter in Washington covering the Trump administration: Who is responsible for the government’s racist tweets? Or more specifically: Who is the person within the Department of Homeland Security creating the memes with all the deep-cut white supremacist references?
It’s a legitimate question, given the way the DHS has operated over the past year. The job of a comms officer at any institution, government or private, is to shape the public’s understanding of what their organization does — its activities, goals, intentions, and so forth. In this case, ICE and DHS enforce immigration laws, and under the Trump administration, they’ve spent the last year aggressively targeting a broad range of minorities under dubious pretexts, with the intention of removing them from America. It is, therefore, notable when this agency publishes social media posts that contain references to that other historic, WWII-era regime that aggressively targeted a broad range of minorities under dubious pretexts (except this one was in Germany).
The problem is, everyone I’ve spoken to in MAGAworld content creation — comms staff, influencers, whomever — knows their identity. In off-the-record conversations, people will simply tell me their name as soon as I ask the question, as they’re familiar with the memelord’s style from interactions in the disappearing MAGA group chats. (I do have to pause for a second and acknowledge that the meme zoomers have better opsec than the senior officials who did Signalgate a year ago.) But the moment I ask if they’d be willing to go on background about it — surely they aren’t Nazis, and shouldn’t they call out behavior in their own ranks? — sources immediately clam up. If I prod, they’ll shrug it off as everyone’s just having fun.
Though I wouldn’t go so far as to call it an actual ~ conspiracy ~ of silence, it’s one of the more fascinating iterations of MAGA omerta that I’ve witnessed, especially because “leaked MAGA group chats” is a journalistic genre of its own at this point. We’ve seen it over and over: Operatives get too comfy with their peers over text, their comments get racist, someone screenshots them and sends them to a journalist outsider, chaos ensues. Two leaked group-chat scandals are roiling Floridian Republican politics at this very moment — one involving rampant racism in a young conservative activist group chat created by the chair of the Miami-Dade County GOP, the other involving white supremacist gubernatorial candidate James Fishback and his compounding financial problems. Why does MAGA seem (occasionally) willing to out private racists, but not this specific very, very public one?
To explain the DHS phenomenon, let’s think about the Machiavellian impulses that lead to MAGA group chat leaks: Someone has an incentive to ruin their rival’s reputation, ideally with no fingerprints attached, and they do that by letting the world know how secretly racist their rival is. (Or, in the case of Candace Owens publishing a leaked private TPUSA group chat, how secretly anti-Israel Charlie Kirk was before he died.) The first massive racist group-chat story, involving several senior members of the New York Young Republican Club, came about not because anyone had a crisis of conscience, but reportedly because of an internal feud over a stolen photo opp with Donald Trump. And a racist chat from October that took down Paul Ingrassia, Trump’s former nominee to the Office of Special Counsel, came about because the leaker told a Politico reporter that they wanted “the government to be staffed with experienced people who are taken seriously” — not necessarily because Ingrassia said he had a “Nazi streak” and used a cornucopia of racial slurs.
What’s instructive here is that Ingrassia didn’t really get taken down. Sure, he had his nomination to his Senate-confirmed position withdrawn after it became clear that he didn’t have the necessary votes, but within a month he’d been assigned as deputy counsel in the General Services Administration. And it’s not as if poor performance and horrific behavior is grounds to get anyone fired from the Trump administration anymore. Kash Patel still has a job, former Homeland Security Secretary Kristi Noem was shuffled off to a make-work position, and as The Atlantic editor-in-chief Jeffrey Goldberg recently noted, no one was ever fired for Signalgate.
The sad conclusion: The DHS white supremacist meme maker might be beloved, loathed, or simply just a person one rung higher on the career ladder than the next guy. (Honestly, in politics, that’s sometimes justification enough to take down a rival). But in Trump’s Washington, at least, there is no political upside for anyone to name and shame them — that is, unless you want to be labeled an MSM snitch.
The Pentagon’s legal ouroboros
We all knew that Anthropic was headed to litigation the moment that the Pentagon slapped it with the supply-chain risk designation a few weeks back, but this week, the filings started pouring in. On Monday, the AI company officially sued the Department of Defense, and hours later, a separate group of senior OpenAI and Google employees — including Jeff Dean, Google’s chief scientist and Gemini lead — filed an amicus brief in support of Anthropic.
The group is represented by the AI for Democracy Action Lab, and Ian Bassin, the cofounder of parent org Protect Democracy, told me on Tuesday that more groups were planning on submitting similar briefs. (There’s at least one high-profile one in the works: Dean Ball, formerly Donald Trump’s chief AI advisor, has publicly stated that he and the Foundation for American Innovation would be submitting a brief.) “I think it’s patently obvious, to everyone watching, that the administration is doing something that’s not only not grounded in the law, but that is operating to a degree in bad faith,” he told me, “and certainly in a manner that is not good for the public interest.”
Indeed, there’s an inherent paradox in the Pentagon’s actions. During contract negotiations over the acceptable-use policy, Secretary Pete Hegseth and Pentagon CTO Emil Michael were threatening to label Anthropic a supply-chain risk and thus a threat to national security. But they had also reportedly considered using the Defense Production Act — normally utilized in wartime scenarios — to force Anthropic to let them use their products.
“Those are self-contradictory,” Bassin told me. “It doesn’t make any sense that a product could be both so dangerous that it poses a supply chain risk and needs to be excised entirely from the system — and at the same time, so essential that it must be forced into the system.”
It’s so obvious that even Sen. Ted Cruz (R-TX), who’s generally anti-woke AI, told CNBC on Tuesday that he had “not seen a basis laid out for why the government would be prohibited from using Anthropic.” Alas, that hasn’t stopped the administration from trying to pull wacky anti-Anthropic shenanigans of dubious legality. On Monday night, Axios reported that the White House was preparing an executive order to “formalize” Trump’s Truth Social mandate to remove Anthropic products from the federal government. The normal caveats apply to any executive order that emerges from Trump’s desk — but was the point ever legality?
Last December, I wrote a profile about Pubkey, a new dive bar in DC with a not-so-secret aspiration to be a diplomatic cultural outpost for Bitcoin policy. Earlier this month, the owners turned the tables on me and asked if I would come hang with founder Thomas Pacchia for their podcast, to explain venal DC political culture to the crypto audience. I figured it was a fair trade — especially since it would be the first podcast they taped out of their DC studio. (And I got to go on the pod before Treasury Secretary Scott Bessent did, so there.)
You can watch it on YouTube here, or listen on Spotify here.