Listen on Spotify (transcripts available)

Listen on Apple (transcripts available)

Hagen Scotten, the lead prosecutor on the Eric Adams case, has joined at least six other attorneys in leaving DOJ rather than carry out Trump’s alleged quid pro quo with Eric Adams.

His letter (posted by NYT) is one for the ages:

I have received correspondence indicating that I refused your order to move to dismiss the indictment against Eric Adams without prejudice, subject to certain conditions, including the express possibility of reinstatement of the indictment. That is not exactly correct. The U.S. Attorney, Danielle R. Sassoon, never asked me to file such a motion, and I therefore never had an opportunity to refuse.

But I am entirely in agreement with her decision not to do so, for the reasons stated in her February 12, 2025 letter to the Attorney General. In short, the first justification for the motion—that Damian Williams’s role in the case somehow tainted a valid indictment supported by ample evidence, and pursued under four different U.S. attorneys—is so weak as to be transparently pretextual. The second justification is worse. No system of ordered liberty can allow the Government to use the carrot of dismissing charges, or the stick of threatening to bring them again, to induce an elected official to support its policy objectives.

There is a tradition in public service of resigning in a last-ditch effort to head off a serious mistake . Some will view the mistake you are committing here in the light of their generally negative views of the new Administration. I do not share those views. I can even understand how a Chief Executive whose background is in business and politics might see the contemplated dismissal-with-leverage as a good, if distasteful, deal. But any assistant U.S. attorney would know that our laws and traditions do not allow using the prosecutorial power to influence other citizens, much less elected officials, in this way. If no lawyer within earshot of the President is willing to give him that advice, then I expect you will eventually find someone who is enough of a fool, or enough of a coward, to file your motion. But it was never going to be me.

Please consider this my resignation. It has been an honor to serve as a prosecutor in the Southern District of New York

As NYT described, Scotten is a former Special Forces Officer and clerked for both Bret Kavanaugh and John Roberts.

Mr. Scotten served three combat tours in Iraq as a U.S. Army Special Forces Officer and earned two Bronze Stars. He graduated from Harvard Law School and clerked for Chief Justice John G. Roberts Jr. of the U.S. Supreme Court, and for Brett M. Kavanaugh before he, too, became an Supreme Court justice.

Earlier today, I perused the Murdoch press and there is nothing so far about this burgeoning scandal. I’ve seen no comment from Republican members of Congress, and frankly far too little from Democrats.

But this may already be teed up to go to SCOTUS. And when it does, some of the loudest voices will be those of conservative lawyers who refused to be party to an alleged quid pro quo.

Update: Brad Heath reports that DOJ finally found someone to sign the motion to dismiss.

The other day, Pete Hegseth capitulated to Vladimir Putin, dealing away Ukraine’s future and leverage, making Neville Chamberlain look not only stronger, but better dressed, by comparison.

He tried to walk back his capitulation the next day.

Everything is on the table in his conversations with Vladimir Putin and Zelenskyy. What he decides to allow or not allow is at the purview of the leader of the free world of President Trump. So I’m not going to stand at this podium and declare what President Trump will do or won’t do, what will be in or what will be out, what concessions will be made or what concessions are not made.

Remember, in response to questions from Tammy Duckworth, Hegseth confessed he had never been part of international negotiations. In his first day and second days learning on the job, he failed every rule of negotiation.

I may return to Pete Hegseth’s predictable failures.

For now, though, I want to note all the things put in place before Trump seemingly turned on a dime, effectively demoting his Ukraine negotiator Keith Kellogg in favor of Marco Rubio, John Ratcliffe, Mike Waltz, and Steve Witkoff (who has been liaising with people like Mohammed bin Salman and — reportedly, Kirill Dmitriev from Mueller Report fame) and taking a much more pro-Russian stance in this negotiation.

Between Trump, Attorney General Pam Bondi, and DHS Secretary Kristi Noem, they have protected precisely the kind of interference and corruption with which Russia kicked off Trump’s political career ten years ago. These moves have been covered already (see this post from Casey Michel and this from Cyberscoop). But I want to look at the kinds of DOJ and CISA actions against which Trump’s team may be reacting, not least because this pivot from Trump did not happen until they were all in place.

Non-prosecution of FCPA: Start with the decision to first limit (in Bondi’s adoption) and then pause (in Trump’s adoption, in a later Executive Order) prosecution of the Foreign Corrupt Practices Act, a law that prohibits businesses with a presence in the United States from engaging in bribery. Bondi actually put this provision in a memo otherwise eliminating approval requirements for investigations and prosecutions targeting trafficking, and with regards to FCPA, simply made using FCPA against traffickers the priority.

Foreign Corrupt Practices Act. The Criminal Division’s Foreign Corrupt Practices Act Unit shall prioritize investigations related to foreign bribery that facilitates the criminal operations of Cartels and TCOs, and shift focus away from investigations and cases that do not involve such a connection. Examples of such cases include bribery of foreign officials to facilitate human smuggling and the trafficking of narcotics and firearms.

Trump, on the other had, halted its use for six months and then maybe another six months.

Most coverage of this move noted its use, under Trump, to penalize Goldman Sachs for bribing Malaysia’s 1MDB sovereign wealth fund, an investigation the aftermath of which sucked in Trump associate Elliott Broidy before Trump pardoned him. But it might be better to consider how such bribery statutes limit transnational investment companies like Trump’s own and Jared Kushner’s. That is, Trump’s intervention in FCPA might be personal to Trump.

Elimination of KleptoCapture Task Force: In the same memo, buried under a shift of focus for Money Laundering cases to traffickers and away from Trump’s buddies, Bondi also included this language about the KleptoCapture program that has been a key prong of Joe Biden and Merrick Garland’s response to the Ukraine invasion.

Money Laundering and Asset Forfeiture. The Criminal Division’s Money Laundering and Asset Recovery Section shall prioritize investigations, prosecutions, and asset forfeiture actions that target activities of Cartels and TCOs.

Task Force KleptoCapture, the Department’s Kleptocracy Team, and the Kleptocracy Asset Recovery Initiative, shall be disbanded. Attorneys assigned to those initiatives shall return to their prior posts, and resources currently devoted to those efforts shall be committed to the total elimination of Cartels and TCOs.

It’s not yet clear whether this means DOJ will start giving yachts back to the sanctioned Russian oligarchs that Biden seized them from.

But what this does imply is that the sanctioned oligarchs who had invested in property and other facilities in the US — people like Oleg Deripaska and Andrii Derkach, both of whom were identified to have ties to Russian influence operations in election years — might be free to invest in the US again.

Shift away from FARA: Buried in Section IV of a different memo innocuously titled “General policy regarding charging, plea negotiations, and sentencing,” are two paragraphs describing changes in the National Security Division’s focus.

Shifting Resources in the National Security Division. To free resources to address more pressing priorities, and end risks of further weaponization and abuses of prosecutorial discretion, the Foreign Influence Task Force shall be disbanded. Recourse to criminal charges under the Foreign Agents Registration Act (FARA) and 18 U.S.C. § 951 shall be limited to instances of alleged conduct similar to more traditional espionage by foreign government actors. With respect to FARA and § 951, the Counterintelligence and Export Control Section, including the FARA Unit, shall focus on civil enforcement, regulatory initiatives, and public guidance.

The National Security Division’s Corporate Enforcement Unit is also disbanded. Personnel assigned to the Unit shall return to their previous posts.

Let’s take them in reverse order. The FARA statement basically says that only people akin to spies will be charged criminally with it; everyone else will be subject to the same civil sanctions DOJ used before the Paul Manafort case. That of course means Manafort’s ongoing work is in the clear (a point that Ken Vogel makes in a column hilariously titled, “Moves by Trump and Bondi Raise Hopes of Those Accused of Foreign Corruption“). It also makes things far easier for Pam Bondi’s former colleagues at Ballard Partners, the most powerful foreign influence peddlers under the first and undoubtedly the second Trump term. This will save Bondi’s friends a whole lot of money in compliance worries.

But here’s the problem with this move: Most of the people DOJ has charged with criminal FARA in recent years were being handled by foreign spies. FARA, as it was used under Mueller and since, was a way to neutralize people for being in the pay of foreign spies without having to prove — or having to declassify evidence to show — that they were themselves spies. It was a way to disable spying, even or especially if people receiving foreign money didn’t know they were being handled by spies.

But Bondi just said she won’t use that tool.

Elimination of FITF: I might have written this post weeks ago, except I keep staring at Bondi’s claim that the Foreign Influence Task Force (the website for which has been taken down) led to “abuses of prosecutorial discretion.” Now, Bondi often parrots the stupidest bullshit that Jim Jordan has floated (which includes a lot of false claims made by Matt Taibbi), and this may be an example — because FITF would not lead to prosecution of a US person, as I tried to lay out in this table (which first appeared in this post).

What the FITF did was to identify attempts by foreigners to clandestinely influence Americans (not just during elections). It played a key role in funneling intelligence to the private sector, especially social media companies. While the government has charged foreigners involved in such operations (such as the Iranians who hacked Trump’s campaign), Americans would almost always be victims.

Based on that assumption, I can only imagine Bondi’s reference to “abuses of prosecutorial discretion” pertains to one of three possible prosecutions:

• The prosecution of Douglass Mackey for duping Hillary Clinton voters into “texting” their vote rather than voting in person, a prosecution that in later years might have arisen out of election protection efforts (the second row in this table) put in place in the wake of 2016.
• A warning about the Andrii Derkach influence operation in 2020, which was managed by FITF, and which led the FBI to shut down some informants sharing information on Hunter Biden. Importantly, the entire right wing believes that a FITF staffer, Laura Dehmlow, should have breached the confidentiality of a non-public investigation in 2020 and told Facebook that the hard drive shared with New York Post derived from a Hunter Biden laptop in the FBI’s possession was “real” (notwithstanding that the FBI had not, and still has not, done the most basic things to test if it was packaged up). So it’s possible that Bondi believes, like Jim Jordan does, that the outcome of the Hunter Biden investigation would have been different if they could have relied more on the laptop.
• The Tenet operation, in which the RT funded right wing propagandists Dave Rubin, Tim Pool, and Benny Johnson. The operation was exposed with an indictment of foreigners shortly before the pre-election halt to such actions, but not even Canadian Lauren Chen has been charged, much less the right wing bros. That indictment, for money laundering and FARA, might not be viable under Bondi’s new restrictions on other prosecutorial focus.

But there are a whole bunch of things you throw out with that bathwater. If the FITF is disbanded, then social media companies might not have discovered that Iran was adopting the identities of the Proud Boys to suppress turnout among people of color. There’s the ongoing Doppelganger effort to create counterfeit versions of real US and European media outlets to spread disinformation — such as an attack on USAID that Elon Musk spread just days ago.

Or there’s the multiple influence operations that Jack Posobiec has been party to, starting with PizzaGate (the weaponization of the Podesta emails stolen by the GRU), the GRU MacronLeaks operation, as well as a more recent FSB campaign. Posobiec’s centrality to all this — as well as his involvement in other kinds of rat-fucking — is particularly pertinent because Pete Hegseth at least invited Jack Posobiec to travel with him to the Munich Security Conference where he sold Ukraine out.

Trump administration officials at the Pentagon invited a far-right activist, Jack Posobiec, to participate in Defense Secretary Pete Hegseth’s first trip overseas, according to a planning document obtained by The Washington Post and people familiar with the decision, triggering alarm among U.S. defense officials worried about the military being dragged into partisan warfare.

Posobiec was in Ukraine yesterday — it’s not yet clear whether he traveled to Europe with the Defense Secretary.

The most charitable explanation for Bondi’s decision to shut down FITF is that she’s suffering from delusions that Jim Jordan passed on. But if she really understands what this program did, then she has deliberately chosen to make it easier for hostile countries, especially Russia, China, and Iran, to affect US elections.

Administrative Leave of CISA Election Security Staff: Which brings me to the most recent effort to help foreign adversaries, something done by Kristi Noem, not Pam Bondi. On Monday, 17 of the people who were involved in keeping the 2024 election secure were put on leave, citing a focus on election disinformation.

In recent days, 17 employees of the U.S. Cybersecurity and Infrastructure Security Agency who have worked with election officials to provide assessments and trainings dealing with a range of threats — from cyber and ransomware attacks to physical security of election workers — have been placed on leave pending a review, according to a person familiar with the situation who was not authorized to speak publicly.

Ten of those employees are regional election security specialists hired as part of an effort to expand field staff and election security expertise ahead of the 2024 election. The regional staffers were told the internal review would examine efforts to combat attempts by foreign governments to influence U.S. elections, duties that were assigned to other agency staff, according to the person.

All were former state or local election officials who were brought in to build relationships across all 50 states and the nation’s more than 8,000 local election jurisdictions. They spent the past year meeting with election officials, attending conferences and trainings, and ensuring officials were aware of the agency’s various cybersecurity and physical security services.

[snip]

The other staffers placed on leave are current or former members of the agency’s Election Security and Resilience team, who were told the review was looking into agency efforts to combat misinformation and disinformation campaigns, according to the person familiar with the situation. The 10 election security specialists who worked with state and local election officials reported to a different team at CISA, the field operations division.

Now, the rationale offered for this decision is a review of CISA’s involvement in warnings about mis- and disinformation. As noted above, that’s not what CISA does. To the extent it shares information with social media companies, it is to provide correct information to make it easier for people to get quality information on voting.

But consider something that these 17 people might have been involved in: the effort, in real time, to respond to bomb threats called into electoral precincts in Democratic areas, many of which were sourced to Russian email domains. (Remember that Ohio Governor Mike DeWine attributed the bomb threats in Springfield — threats ginned up with the significant involvement of Jack Posobiec — to overseas actors.)

We still don’t know whether the bomb threats targeting Springfield and voting locations actually were Russian operations or whether they were funneled through Russia by American actors to obscure their origin. We still don’t have a report from the FBI explaining what happened.

And with the decision to shut down both the FITF and to pause CISA’s election protection work, we may never get it now. We may never learn whether Democratic precincts had to shut down due to Russian involvement or that of people laundering their work through Russia.

In the wake of Trump’s victory, key Putin advisor Nikolai Patrushev claimed that, to win, Trump “relied on certain forces to which he has corresponding obligations.”

In his future policies, including those on the Russian track US President-elect Donald Trump will rely on the commitments to the forces that brought him to power, rather than on election pledges, Russian presidential aide Nikolay Patrushev told the daily Kommersant in an interview.

“The election campaign is over,” Patrushev noted. “To achieve success in the election, Donald Trump relied on certain forces to which he has corresponding obligations. As a responsible person, he will be obliged to fulfill them.”

He agreed that Trump, when he was still a candidate, “made many statements critical of the destructive foreign and domestic policies pursued by the current administration.”

“But very often election pledges in the United States can iverge from subsequent actions,” he recalled.

When he gave that ominous warning, I concluded that Trump would soon sell out Ukraine and the rest of Europe. But that didn’t happen right away. Rather, for months, Trump feigned a hardline stance against Russia, all while teasing the number of calls he was having with Putin.

Until this week.

Trump didn’t move to “fulfill” the “corresponding obligations” he made to get help in the election, if indeed he did get help, until Pam Bondi instructed DOJ not to look for such things.

A Normal Person’s Explainer on What Generative AI is and Does

Part 1 – In the Beginning was the Chatbot

“Are you comfortably seated? Yes, well, let’s begin.” *Clears throat theatrically*

“Our experience, in natural theology, can never furnish a true and demonstrated science, because, like the discipline of practical reason, it can not take account of problematic principles. I assert that, so far as regards pure logic, the transcendental unity of apperception is what first gives rise to the never-ending regress in the series of empirical conditions. In this case it remains a mystery why the employment of the architectonic of human reason is just as necessary as the intelligible objects in space and time, as is proven in the ontological manuals. By means of analysis, it must not be supposed that the transcendental unity of apperception stands in need of our sense perceptions. Metaphysics, for example, occupies part of the sphere of the transcendental aesthetic concerning the existence of the phenomena in general…”

It was 1995, and several of us who worked in my community college’s Macintosh lab were hunting around the net for weird software to try out, back when weird software felt fun, not dangerous. Someone found a program on the nacent web that would almost instantly generate pages of thick and unlovely prose that wasn’t actually Kant, but looked like it. It was, to our definitionally untrained eyes, nearly indistinguishable from the Immanuel Kant used to torture undergrad college students.

The logo of the Kant Generator Pro

We’d found the Kant Generator Pro, a program from a somewhat legendary 90s programmer known for building programming tools. And being cheeky. It was great. (recent remake here) We read Faux Kant to each other for a while, breaking down in giggles while trying to get our mouths around Kant’s daunting vocabulary. The Kant Generator Pro was cheeky, but it was also doing something technically interesting.

The generator was based on a Markov chain: a mathematical way of picking some next thing, in this case, a word. This generator chose each next word using a random walk through all Kantian vocabulary. But in order to make coherent text rather than just random Kant words, it had to be weighted: unrandomized to some extent. The words had to be weighted enough to make it form human-readable Kantian sentences.

A text generator finds those weights using whatever text you tell the computer to train itself on. This one looked at Kant’s writing and built an index of how often words and symbol appeared together. Introducing this “unfairness” in the random word picking gives a higher chance for some words coming next based on the word that came before it. For instance, there is a high likelihood of starting a sentence with “The,” or “I,” or “Metaphysics,” rather than “Wizard” or “Oz.” Hence, in the Kant Generator Pro “The” could likely be followed by “categorical,” and when it is the next word will almost certainly be “imperative,” since Kant went on about that so damn much.

The Kant Generator Pro was a simple ancestor of ChatGPT, like the small and fuzzy ancestors of humans that spent so much time hiding from dinosaurs. All it knew, for whatever the value of “knowing” is in a case like this, was the the words that occurred in the works of Kant.

Systems like ChatGPT, Microsoft Copilot, and even the upstart Deepseek use all the information they can find on the net to relate not just one word to the next, like Kant Generator Pro did. They look back many words, and how likely they are to appear together over the span of full sentences. Sometimes a large language model takes a chunk as is, and appears to “memorize” text and feed it back to you, like a plagiarizing high schooler.

But it’s not clear when regurgitating a text verbatim is a machine copying and pasting, versus recording a statistical map of that given text and just running away with the math. It’s still copying, but not copying in a normal human way. Given the odds, it’s closer to winning a few rounds of Bingo in a row.

These chatbots index and preserve the statistical relationships words and phrases have to each other in any given language. They start by ingesting all the digital material their creators can find for them, words, and their relationships. This is the training people talk about, and it’s a massive amount of data. Not good or bad data, not meaningful or meaningless, just everything, everywhere people have built sentences and left them where bots could find them. This is why after cheeky Reddit users mentioned that you could keep toppings on pizza by using glue, and that ended up becoming a chatbot suggestion.

Because people kept talking about using glue on pizza, especially after the story of that hilarious AI mistake broke, AI kept suggesting it. Not because it thought it was a good idea. AI doesn’t think in a way familiar to people, but because the words kept occurring together where the training part of the AI could see them together.  The AI isn’t right here, we all know that, but it’s also not wrong. Because the task of the AI isn’t to make pizza, the task is to find a next likely word. And then the next, and next after that.

Despite no real knowing or memorizing happening, this vast preponderance of data lets these large language models usually predict what is likely to come next in any given sentence or conversation with a user. This is based on the prompt a user gives it, and how the user continues to interact with it. The AI looks back on the millions of linguistic things it has seen and built statistical models for. It is generally very good at picking a likely next word. Chatbots even to feel like a human talking most of the time, because they trained on humans talking to each other.

So, a modern chatbot, in contrast to the Kant Generator Pro, has most of the published conversations in modern history to look back on to pick a next good word. I put leash on the, blimp? Highly unlikely, the weighting will be very low. Véranda? Still statistically unlikely, though perhaps higher. British politician? Probably higher than you’d want to think, but still low. Table? That could be quite likely. But how about dog? That’s probably the most common word. Without a mention of blimps or parliamentarians or tables in the recent text, the statistics of all the words it knows means the chatbot will probably go with dog. A chatbot doesn’t know what a dog is, but it will “know” dog is associated with leash. How associated depends on the words that have come before the words “dog,” or “leash.”

It’s very expensive and difficult to build this data, but not very hard to run once you have built it. This is why chatbots seem so quick and smart, despite at their cores being neither. Not that they are slow and dumb — they are doing something wholly different than I am when I write this, or you as you read it.

Ultimately, we must remember that chatbots are next-word-predictors based on a great deal of statistics and vector math. Image generators use a different architecture, but still not a more human one. The text prompt part is still an AI chatbot, but one that replies with an image.

AI isn’t really a new thing in our lives. Text suggestions on our phones exists somewhere between the Kant Generator Pro and ChatGPT, and customize themselves to our particular habits over time. Your suggestions can even become a kind of statistical fingerprint for your writing, given enough time writing on a phone or either any other next word predictor.

We make a couple bad mistakes when we interact with these giant piles of vector math and statistics, running on servers all over the world. The first is assuming that they think like us, when they have no human-like thought, no internal world, just mapping between words and/or pixels.

The other is assuming that because they put out such human-like output, we must be like them. But we are not. We are terribly far from understanding our own minds completely. But we do know enough to know biological minds are shimmering and busy things faster and more robust than anything technologists have ever yet built. Still, it is tempting, especially for technologists, to have some affinity for this thing that seems so close to, but not exactly, us. It feels like it’s our first time getting to talk to an alien, without realizing it’s more like to talking to a database.

Humans are different. Despite some borrowing of nomenclature from biology, neural nets used in training AI have no human-style neurons. The difference shows. We learn to talk and read and write with a minuscule dataset, and that process involves mimicry, emotion, cognition, and love. It might also have statistical weighting, but if it does, we’ve never really found that mechanism in our minds or brains. It seems unlikely that it would be there in a similar form, since these AIs have to use so much information and processing power to do what a college freshman can with a bit of motivation. Motivation is our problem, but it’s never a problem for AIs. They just go until their instructions reach an end point, and then they cease. AIs are unliving at the start, unliving in the process, and unliving at the end.

We are different. So different we can’t help tripping ourselves up when we look at AI, and accidentally see ourselves, because we want to see ourselves. Because we are full of emotions and curiosity about the universe and wanting to understand our place in it. AI does not want.

It executes commands, and exits.

The Acting US Attorney for SDNY, Danielle Sassoon, who was hand-picked by Trump’s people, resigned today rather than do the dirty bidding of Trump’s defense attorney (and disgruntled former SDNY AUSA) Emil Bove, by dismissing the case against Eric Adams.

After she resigned, two attorneys in DOJ’s Public Integrity Division, Kevin Driscoll and John Keller, joined her rather than dismiss the case.

A letter, yesterday, from Sassoon to Pam Bondi and another, today, from Bove to Sassoon document much of what happened.

Sassoon documents that Bove likened the dismissal of charges against Adams to the Viktor Bout prisoner exchange (something that was in his original letter).

Mr. Bove proposes dismissing the charges against Adams in return for his assistance in enforcing the federal immigration laws, analogizing to the prisoner exchange in which the United States freed notorious Russian arms dealer Victor Bout in return for an American prisoner in Russia. Such an exchange with Adams violates commonsense beliefs in the equal administration of justice, the Justice Manual, and the Rules of Professional Conduct. The “commitment to the rule of law is nowhere more profoundly manifest” than in criminal justice. Cheney v. United States Dist. Ct., 542 U.S. 367, 384 (2004) (alterations and citation omitted). Impartial enforcement of the law is the bedrock of federal prosecutions. See Robert H. Jackson, The Federal Prosecutor, 24 J. Am. Jud. Soc’y 18 (1940). As the Justice Manual has long recognized, “the rule of law depends upon the evenhanded administration of justice. The legal judgments of the Department of Justice must be impartial and insulated from political influence.” JM § 1-8.100. But Adams has argued in substance—and Mr. Bove appears prepared to concede—that Adams should receive leniency for federal crimes solely because he occupies an important public position and can use that position to assist in the Administration’s policy priorities.

[snip]

Adams’s advocacy should be called out for what it is: an improper offer of immigration enforcement assistance in exchange for a dismissal of his case. Although Mr. Bove disclaimed any intention to exchange leniency in this case for Adams’s assistance in enforcing federal law,1 that is the nature of the bargain laid bare in Mr. Bove’s memo. That is especially so given Mr. Bove’s comparison to the Bout prisoner exchange, which was quite expressly a quid pro quo, but one carried out by the White House, and not the prosecutors in charge of Bout’s case.

The comparison to the Bout exchange is particularly alarming. That prisoner swap was an exchange of official acts between separate sovereigns (the United States and Russia), neither of which had any claim that the other should obey its laws. By contrast, Adams is an American citizen, and a local elected official, who is seeking a personal benefit—immunity from federal laws to which he is undoubtedly subject—in exchange for an act—enforcement of federal law—he has no right to refuse. Moreover, the Bout exchange was a widely criticized sacrifice of a valid American interest (the punishment of an infamous arms dealer) which Russia was able to extract only through a patently selective prosecution of a famous American athlete.2 It is difficult to imagine that the Department wishes to emulate that episode by granting Adams leverage over it akin to Russia’s influence in international affairs. It is a breathtaking and dangerous precedent to reward Adams’s opportunistic and shifting commitments on immigration and other policy matters with dismissal of a criminal indictment. Nor will a court likely find that such an improper exchange is consistent with the public interest. See United States v. N.V. Nederlandsche Combinatie Voor Chemische Industrie (“Nederlandsche Combinatie”), 428 F. Supp. 114, 116-17 (S.D.N.Y. 1977) (denying Government’s motion to dismiss where Government had agreed to dismiss charges against certain defendants in exchange for guilty pleas by others); cf. In re United States, 345 F.3d 450, 453 (7th Cir. 2003) (describing a prosecutor’s acceptance of a bribe as a clear example of a dismissal that should not be granted as contrary to the public interest).

[snip]

In particular, the rationale given by Mr. Bove—an exchange between a criminal defendant and the Department of Justice akin to the Bout exchange with Russia—is, as explained above, a bargain that a prosecutor should not make. Moreover, dismissing without prejudice and with the express option of again indicting Adams in the future creates obvious ethical problems, by implicitly threatening future prosecution if Adams’s cooperation with enforcing the immigration laws proves unsatisfactory to the Department. See In re Christoff, 690 N.E.2d 1135 (Ind. 1997) (disciplining prosecutor for threatening to renew a dormant criminal investigation against a potential candidate for public office in order to dissuade the candidate from running); Bruce A. Green & Rebecca Roiphe, Who Should Police Politicization of the DOJ?, 35 Notre Dame J.L. Ethics & Pub. Pol’y 671, 681 (2021) (noting that the Arizona Supreme Court disbarred the elected chief prosecutor of Maricopa County, Arizona, and his deputy, in part, for misusing their power to advance the chief prosecutor’s partisan political interests). Finally, given the highly generalized accusations of weaponization, weighed against the strength of the evidence against Adams, a court will likely question whether that basis is pretextual. See, e.g., United States v. Greater Blouse, Skirt & Neckwear Contractors, 228 F. Supp. 483, 487 (S.D.N.Y. 1964) (courts “should be satisfied that the reasons advanced for the proposed dismissal are substantial and the real grounds upon which the application is based”).

1 I attended a meeting on January 31, 2025, with Mr. Bove, Adams’s counsel, and members of my office. Adams’s attorneys repeatedly urged what amounted to a quid pro quo, indicating that Adams would be in a position to assist with the Department’s enforcement priorities only if the indictment were dismissed. Mr. Bove admonished a member of my team who took notes during that meeting and directed the collection of those notes at the meeting’s conclusion.

2 See, e.g., https://thehill.com/homenews/3767785-trump-pans-prisoner-swap-brittney-grinerhates-our-country/.

In response, Bove suggested that DOJ was adopting an unreviewable judgement of “weaponization” in disciplining lawyers.

The weaponization finding in my February 10, 2025 memorandum was made pursuant to a policy set forth by President Trump, who is the only elected official in the ExecutiveBranch, in connection with a decision that was authorized by the Senate-confirmed Attorney General ofthe United States, and entirely consistent with guidance issued by the Attorney General shortly after that confirmation. Your Office has no authority to contest the weaponization finding, or the second independent basis requiring dismissal set forth in my memorandum. The Justice Department will not tolerate the insubordination and apparent misconduct reflected in the approach that you and your office have taken in this matter.

You are well aware of the Department’s weaponization concerns regarding the handling of the investigation and prosecution of Mayor Adams. Those concerns include behavior that supports, at minimum, unacceptable appearances of impropriety and the politicization of your office. The investigation was accelerated after Mayor Adams publicly criticized President Biden’s failed immigration policies, and led by a former U.S. Attorney with deep connections to the former Attorney General who oversaw the weaponization ofthe Justice Department. Based on my review and our meetings, the charging decision was rushed as the 2024 Presidential election approached, and asthe former U.S. Attorney appears to have been pursuing potential political appointments in the event Kamala Harris won that election.

I’ve been wondering for some time when Bove would wildly overstep with his aggressiveness. He’s now facing documentation that supports a quid pro quo seeking political favors. And in response, he suggested his recourse is to adopt a label — weaponization — with no due process.

Trump may yet get his quid pro quo (though Judge Dale Ho now has abundant reason to refuse to dismiss this case).

But he may lose DOJ as a result.

Update: Note that the same day Sassoon sent the letter to Bondi, Bondi sued Tish James. And as this was going on, Trump rescinded FEMA funding for NY.

Those likely are not unrelated.

Update: NYT has published the original letter instructing Sassoon to dismiss the case.

There have and will be a slew of lawsuits in response to Trump’s attack on government. But this lawsuit, from eight of the Inspectors General that Trump fired on January 24, has been much anticipated. [docket]

That’s partly because Congress just strengthened the laws protecting Inspectors Generals, in response to Trump’s firing of some in his first term, as the suit lays out.

63. Congress responded in 2022 by further amending the IG Act. The Securing Inspector General Independence Act of 2022, see supra ¶6, enacted by overwhelming margins in both houses of Congress, procedural protections before an IG can be removed or placed on nonduty status, designated that a “first assistant” would automatically replace an IG in the event of a vacancy, and required the President to communicate reasons for not making a formal nomination to fill an IG vacancy after a certain period of time.

64. The 2022 amendments also strengthened the procedural safeguards on removing an IG. Prior to the amendments, the IG Act had required the President to provide 30 days’ notice to both houses of Congress and “reasons for any such removal.” The 2022 amendments require the President to provide 30 days’ notice to both houses of Congress, including appropriate congressional committees, and to “communicate in writing the substantive rationale, including detailed and case-specific reasons, for any such removal.” 5 U.S.C. §403(b). With the 2022 amendments included, the relevant provisions now reads as follows:

An Inspector General may be removed from office by the President. If an Inspector General is removed from office or is transferred to another position or location within an establishment, the President shall communicate in writing the substantive rationale, including detailed and case-specific reasons for any such removal or transfer to both Houses of Congress (including to the appropriate congressional committees), not later than 30 days before the removal or transfer. Nothing in this subsection shall prohibit a personnel action otherwise authorized by law, other than transfer or removal.

65. These procedural provisions ensure that Congress or members of Congress can, if it or they deem it appropriate, seek to persuade the President not to go forward with a noticed removal. Indeed, the legislative history of the Inspector General Reform Act indicates that Congress added the notice requirement to “allow for an appropriate dialogue with Congress in the event that the planned transfer or removal is viewed as an inappropriate or politically motivated attempt to terminate an effective Inspector General.” See S. Rep. No. 110-262, at 4 (2008)

If Congress has any power to limit how the President fires someone, then this suit will uphold that power (a large team from Wilmer Cutler, led by former Solicitor General Seth Waxman, are representing the plaintiffs).

But it’s also because the plaintiffs in this suit embody everything Trump claims he wants to do with DOGE. Elon Musk claims he’s hunting for waste, fraud, and corruption in government agencies he’s wildly unfamiliar with. These civil servants have been doing this, some of them, for four decades.

Indeed, one thing the suit lists, for each of the plaintiffs, is how much material impact they have had in their role (with one exception, exclusively in the IG position from which they were fired, which the report explains is:

“Monetary impact” describes the estimated financial savings or losses that could result from implementing recommendations made in an IG’s audits, inspections, or evaluations, essentially quantifying the potential cost-benefit of addressing issues like waste, fraud, and abuse in a government agency or program. See CIGIE, Toolkit for Identifying and Reporting Monetary Impact, at 1 (June 18, 2024), https://www.ignet.gov/sites/default/files/files/Toolkit%20for%20Identifying%20and%20Reporting%20Monetary%20Impact.pdf.

Some monetary-impact estimates reported herein also consider monetary benefits associated with IG investigations.

And while there’s some inconsistency in the reporting (for example, Sandra Bruce included stuff from when she was Acting IG during Trump’s first term whereas some of the others left out susbstantial terms in other IG roles, Larry Turner’s number — for Department of Labor — seems quite high, and Mike Ware does not include $30 billion seized or returned pursuant to investigations he oversaw), the Inspectors General describe identifying $183.5 billion in material impact.

As noted in this post, that includes substantial work cleaning up after COVID relief rolled out by Trump, particularly from Mike Ware, work which lead DOGE Treasury Official Thomas Krause relied on to suggest that DOGE could be effective. In Ware testimony to Congress that Krause cited, Ware described up to $200 billion in fraud just in Small Business related relief alone.

Using OIG’s investigative casework, prior OIG reporting, advanced data analytics, and additional review procedures, we estimate SBA disbursed more than $200 billion in potentially fraudulent COVID-19 EIDLs and PPP loans. This estimate represents approximately 17 percent of disbursed COVID-19 EIDLs and PPP funds — specifically, more than $136 billion COVID-19 EIDLs and $64 billion in PPP funds. Since SBA did not have an established strong internal control environment for approving and disbursing program funds, there was an insufficient barrier against fraudsters accessing funds that should have been available for eligible business owners adversely affected by the pandemic.

That’s what Trump did by firing Ware and the others: halt proven efforts to do what DOGE is incapable of — and only pretending — to do.

Which is another reason to keep an eye on this lawsuit.

In the last few days, Trump has started doing a better job of messaging with his responses to lawsuits. I’ll attempt to explain that going forward. But one instance is the Thomas Krause declaration filed in the Attorneys General challenge to the DOGE access to Treasury systems (which I also wrote about in this post). Krause — still serving as the hatchet man CEO of Citrix (which probably creates a serious conflict) — uses his declaration to claim that he is attempting to “improve the accuracy of financial reporting.”

I am responsible, among other duties, for reducing and eliminating improper and fraudulent payments; waste, fraud, and abuse; and improving the accuracy of financial reporting. To that end, I am focused on improving the controls, processes, and systems that facilitate payments and enable consolidated financial reporting.

Later in the declaration, he provides a notably different explanation for his job.

My role on the Treasury DOGE Team is to find ways to use technology to make the Treasury Department more effective, more efficient, and more responsive to the policy goals of this Administration.

To justify the focus of DOGE, Krause cites several Biden-era GAO reports.

7. As illustrated by several reports released by the Government Accountability Office (GAO), we have our work cut out for us. On January 16, 2025, GAO released a report entitled “Financial Audit: FY2024 and FY2023 Consolidated Financial Statements of the U.S. Government.” In the report, GAO summarizes that they were not able to determine if the Financial Report of the U.S. Government is fairly presented. Among other reasons, GAO highlighted “problems in accounting for transactions between federal agencies.” GAO found many material weaknesses including “the federal government’s inability to determine the full extent to which improper payments, including fraud, occur and reasonably assure that appropriate actions are taken to reduce them.” GAO also reported that Treasury and Office of Management and Budget (OMB) officials expressed their continuing commitment to addressing the problems this report outlines. In short, the GAO report identifies the Federal government’s inability to account for all of the improper payments including waste, fraud and abuse across federal agencies.

8. On September 10, 2024, the GAO released a report entitled “Payment Integrity: Significant Improvements are Needed to Address Improper Payments and Fraud.” The report found that since 2003, cumulative improper payments1 by executive branch agencies have totaled about $2.7 trillion dollars. Some of GAO’s top concerns [1] included fraudulent or improper Earned Income Tax Credit refunds, Social Security payments, unemployment and Medicare and Medicaid payments. In fiscal year 2023 alone, federal agencies estimated $236 billion in improper payments across more than 70 federal programs. In addition, GAO estimated that the total annual financial losses across the government from fraud are between $233 and $521 billion. These numbers are truly staggering—billions and billions in hardearned American taxpayer dollars are being misspent every year. GAO highlighted a number of steps that Congress and federal agencies could take to help reduce fraud and improper payments, including that “[a]gencies should improve oversight to ensure that funds aren’t paid to ineligible recipients” [2] and that “[a]gencies should improve their collection and use of data for preventing and detecting fraud.” [3]

9. Similarly, GAO has identified areas for improvement in BFS’s systems related to identifying and tracing transactions to determine whether they were complete and properly recorded in the correct general ledger accounts and line items within the Schedules of the General Fund. See GAO Report, “Financial Statement Audit: Bureau of the Fiscal Service’s FY22 Schedules of the General Fund” (March 30, 2023). Specifically, GAO has found inconsistent reporting, lack of traceability, and need for improved controls with the Treasury’s Central Accounting and Reporting System (CARS), which federal agencies use to track their spending for budgetary and accounting purposes. These kinds of improvements and others can enhance BFS’s ability to ensure accountability in the spending of taxpayer dollars.

1 Improper payments and fraudulent payments are related but distinct concepts. An improper payment is a payment that should not have been made, or that was made with an incorrect amount; fraudulent payments occur due to willful misrepresentation. All fraudulent payments are improper, but not all improper payments are fraudulent. [emphasis and links added]

Elon Musk parroted a lot of this language at his presser at the White House yesterday (which is one reason I say they’re beginning to coordinate this better).

If you don’t look too closely, the declaration almost makes DOGE look smart. Except I decided to look at one of the reports — the second one — more closely.

And once I did, I realized that Thomas Krause is, in part, using Trump’s management failures during COVID as an excuse to start shutting down government. Start with the fact that the first agency Krause focused on after arriving at Treasury was USAID — pursuing his goal of making Treasury, “more responsive to the policy goals of this Administration.” But that’s not one of the high risk agencies, all of which have to do with direct payments.

Since 2003, which is when they first tracked the data, the amount of improper payments has steadily increased. But it has declined in recent years, under Biden.

There’s a reason for that. Look more closely at the estimated improper payments, their sources, and their timing. 

For longstanding programs — Medicare and Social Security, the ones Krause mentions in his declaration — the number of improper payments in recent years is about what it was under Trump. What has spiked in recent years (and then receded) are programs that expanded under COVID: Expanded Medicaid and unemployment access, and the PPP program rolled out under Trump, something Krause neglects to mention at [1]. A key thing this report measures is COVID mispayments — that is, improper payments made under programs set up under President Donald Trump, 1.0.

The quotes at [2] and [3] are not actually from the report. They’re from this website (which links to this report).

Many of the recommendations and data used in this report pertain to COVID or lessons learned from it. For example, the report recommends making the payment tracking center set up in response to COVID permanent.

Establish a permanent analytics center of excellence to aid the oversight community in identifying improper payments and fraud.28 This could be achieved by building upon and expanding PACE and making it permanent.

And it recommended building in such collection in case of any future emergency response — in part, to avoid the two to three year delay in finding these payments reflected in the table above.

Require OMB to (1) provide guidance for agencies to proactively develop internal control plans that would be ready for use in, or adaptation for, future emergencies or crises and (2) require agencies to report these plans to OMB and Congress.

Amend PIIA. Quickly reporting improper payment estimates for emergency relief programs is critical for agency accountability and transparency over whether appropriated funds were spent for their intended purposes. In addition, estimating improper payments and identifying root causes help ensure that agencies develop and implement corrective actions to reduce them.

In November 2020, we recommended that Congress consider, in any future legislation appropriating COVID-19 relief funds, designating all executive agency programs and activities that made more than $100 million in payments from COVID-19 relief funds as “susceptible to significant improper payments.31 Such a designation would require, among other things, agencies to report improper payment estimates for such a program and develop corrective actions to reduce improper payments. In March 2022, we recommended that Congress amend PIIA to apply this criterion to all new federal programs for their initial years of operation.32 The current approach resulted in 2-to-3 year delays in reporting improper payment estimates for short-term and emergency spending COVID relief programs.

Much of the fraud, too, pertains to COVID relief.

When it is discovered, the Department of Justice (DOJ) can bring charges of fraud against the alleged fraudsters. For example, DOJ has prosecuted over 2,000 COVID-19 fraud-related cases, and hundreds of additional cases are pending. We analyzed the department’s public statements and court documentation and found that, from March 2020 through March 2024, at least 1,998 individuals or entities facing fraud-related charges were found guilty or liable.16 This includes charges in cases involving SBA’s loan programs, DOL’s Unemployment Insurance (UI) programs, and Treasury’s economic impact payments. Of the individuals found guilty, at least 1,596 had been sentenced as of March 31, 2024, and many have also been ordered to pay restitution and fines. There were also federal fraud-related charges pending against at least 632 other individuals or entities involving federal COVID-19 relief programs, as of March 31, 2024.17 We expect the number to continue to increase as investigations take time to develop and given the significant number of investigative leads. For instance, SBA’s IG office reported that its actionable leads represent more than 100 years of investigative case work.18 The government has 10 years to prosecute individuals who committed fraud related to the Paycheck Protection Program (PPP) and the COVID-19 Economic Injury Disaster Loan (EIDL) program.19 DOL’s IG has requested Congress similarly extend the statute of limitations for the pandemic relief UI programs as well.20 Additionally, in a June 2024 press release, the Internal Revenue Service requested to Congress that the statute of limitations for fraud be extended for the Employee Retention Credit.21 We support their requests.

Now, to be fair, there was likely to be overpayments and fraud regardless of who was in charge when COVID hit (or when avian flu and measles become pandemics in months ahead).

This is not all attributable to Trump’s COVID failures.

But one other thing about this report deserves mention: It is full of discussion of the role of Inspectors General in finding this fraud, including a bunch of the people Trump fired four days into his term — six of whom just filed suit today (which I’ll turn to shortly). Indeed, one of them — Mike Ware — is cited in the report Krause invoked.

18 Stolen Taxpayer Funds: Reviewing the SBA and OIG Reports of Fraud in Pandemic Lending Programs Hearing Before the House Committee on Small Business, 118th Cong. 45 (2023) (statement of Hannibal “Mike” Ware, Inspector General of U.S. Small Business Administration).

Thomas Krause says we need to fix the errors created by Trump’s poor management of COVID. But one of the first things Trump in his second term did was to fire the people who’ve done the most to do so.

Update: Corrected Mike Ware’s last name.

Update: I’ve linked Ware’s testimony, above. Among other things, he estimated that SBC IG identified up to $200 billion in fraudulent COVID relief.

Using OIG’s investigative casework, prior OIG reporting, advanced data analytics, and additional review procedures, we estimate SBA disbursed more than $200 billion in potentially fraudulent COVID-19 EIDLs and PPP loans. This estimate represents approximately 17 percent of disbursed COVID-19 EIDLs and PPP funds — specifically, more than $136 billion COVID-19 EIDLs and $64 billion in PPP funds. Since SBA did not have an established strong internal control environment for approving and disbursing program funds, there was an insufficient barrier against fraudsters accessing funds that should have been available for eligible business owners adversely affected by the pandemic.

In other words, a huge chunk of the fraud Krause says he is looking for was IDed by one of the guys Trump fired on day four.

According to the currently operative story, Marko Elez — the DOGE [sic] boy who had source code for Treasury’s payments system — resigned in response to a query from WSJ reporter Katherine Long about his social media posts in support of

A key DOGE staff member who gained access to the Treasury Department’s central-payments system resigned Thursday after he was linked to a deleted social-media account that advocated racism and eugenics.

Marko Elez, a 25-year-old who is part of a cadre of Elon Musk lieutenants deployed by the Department of Government Efficiency to scrutinize federal spending, resigned after The Wall Street Journal asked the White House about his connection to the account.

“Just for the record, I was racist before it was cool,” the account posted in July, according to the Journal’s review of archived posts.

“You could not pay me to marry outside of my ethnicity,” the account wrote on X in September. “Normalize Indian hate,” the account wrote the same month, in reference to a post noting the prevalence of people from India in Silicon Valley.

After the Journal inquired about the account, White House spokesperson Karoline Leavitt said that Elez had resigned from his role.

But that belief is only based on correlation, not any proof of causation. Long asked about posts that are in no way exceptional for the far right boys Elon has infiltrated into the government. And Elez resigned that same day.

Sure, Elon implied that Elez quit because the boy’s far right ideology was exposed — he led a campaign for his reinstatement. That campaign — and JD Vance’s support for it — similarly led a lot of people to believe that Elez had been reinstalled at Treasury. But multiple court filings claim that Elez resigned and never came back, at least not to Treasury.

In fact, there are two things that might provide better explanations than the discovery that like Elon himself, Elez is a racist.

As WSJ itself notes, Elez resigned the same day that Colleen Kollar-Kotelly ordered that Elez, then still identified as a Special Government Employee, be granted only read-only access to Treasury’s networks. Once Elez no longer worked for the defendants in that case — starting with Scott Bessent — then any access he had would be exempted from the order.

More importantly, as a court filing submitted yesterday reveals, Elez’ resignation happened the same day that Treasury discovered Elez’s Bureau laptop, “had mistakenly been configured with read/write permissions instead of read-only.” The filing is a declaration from Joseph Gioeli, who has been employed as the “Deputy Commissioner for Transformation and Modernization in the Bureau of the Fiscal Service” since 2023 and is a civil servant first hired in the first year of Trump’s first term.

His declaration describes how the 4-6 week “payment process engagement plan” initiated (per Thomas Krause) on January 26 required giving Elez risky access to payment systems. Gioeli describes how they tried to mitigate those risks.

11. The scope of work as envisioned in the engagement plan required access to Fiscal Service source code, applications, and databases across all these Fiscal Service payment and accounting systems and their hosting environments. This broad access presented risks, which included potential operational disruptions to Fiscal Service’s payment systems, access to sensitive data elements, insider threat risk, and other risks that are inherent to any user access to sensitive IT systems. In light of these risks, BFS and Treasury Departmental Office employees developed mitigation strategies that sought to reduce these risks.

12. These measures included the requirement that Mr. Elez be provided with a BFS laptop, which would be his only method of connecting to the Treasury payments systems, both in connecting with the source code repository and for his read-only access of the systems. He had previously been provided a Treasury laptop from the Department shortly after he onboarded, but due to Bureau security policy, that device was restricted from accessing the BFS systems and services he had requested. BFS used several cybersecurity tools to monitor Mr. Elez’s usage of his BFS laptop at all times and continuously log his activity. Additionally, the Bureau enabled enhanced monitoring on his laptop, which included the ability to monitor and block website access, block the use of external peripherals (such as USB drives or mass storage devices), monitor any scripts or commands executed on the device, and block access to cloud-based storage services. Additionally, the device contained data exfiltration detection, which alerts the Bureau to attempts to transmit sensitive data types. The laptop is also encrypted in accordance with Bureau policy, which, if the laptop were stolen or lost, would prevent unauthorized users from accessing data contained within the laptop.

13. Additional mitigation measures that were adopted included that Mr. Elez would receive “read-only” access to the systems, and that any reviews conducted using the “read-only” access would occur during low-utilization time periods, to minimize the possibility of operational disruptions. While providing a single individual with access to multiple systems and data records accessed here was broader in scope than what has occurred in the past, this read-only approach is similar to the kind of limited access the Bureau has provided to auditors for other Treasury non-payment systems, though even in those scenarios the availability of production data was significantly limited. [my emphasis]

Gioeli goes on to describe how, starting on February January 28, the Bureau gave Elez source code in a sandbox environment.

16. On January 28, 2025, the Bureau provided Mr. Elez with the Bureau laptop and with copies of the source code for PAM, SPS, and ASAP in a separate, secure coding environment known as a “secure code repository” or “sandbox.” Mr. Elez could review and make changes locally to copies of the source code in the cordoned-off code repository; however, he did not have the authority or capability to publish any code changes to the production system or underlying test environments. This repository was separate from Fiscal Service’s typical code development environment, and unlike the usual code development environment, this new repository was segmented, to ensure that no changes to the operative source code could be made. [my emphasis]

Then, six days after giving him that sandbox access, using the same laptop, they gave him read-only access to first two and then one more systems.

17. On February 3, 2025, consistent with the engagement plan and mitigation measures developed, Mr. Elez was provided with read-only access, through his Bureau laptop, to the certain BFS systems. The read-only access that Mr. Elez was provided gives the user the ability to view and query information and data but does not allow for any changes to that information and data within its source system. While this reduces risk, it does not fully eliminate the risks identified in the assessment (for example, the risk of overburdening the system with a complex read-only query). Specifically, Mr. Elez was provided read-only access to the Payment Automation Manager (PAM) Database, Payment Automation Manager (PAM) File System, and, subsequently on February 5, the Secure Payment System (SPS) Database.

After he got that access, per a review of the logs, Elez copied some files from the active database onto his Bureau laptop, on which he had the source code.

18. ISS configured his network access and assisted him in setting up the necessary tools to connect to the PAM database on February 3. His access was closely monitored by multiple BFS administrators throughout the process on February 3. That same day, he received a “walk-through” demonstration of two BFS payment systems, the PAM database and the PAM file system (the system that controls the payment file “landing zone” discussed above), to see how the systems worked. He logged in with his read-only access to these systems on February 3 during this “walk-through” demonstration. The Bureau is in the process of reviewing the logs of Mr. Elez’s activity on his Bureau laptop, and this review remains ongoing. Based on the preliminary log reviews conducted to date, it appears that on February 3, Mr. Elez copied two USAID files directly from the PAM database to his BFS laptop; on February 4 and 5, Mr. Elez accessed the PAM file system; and on February 5, Mr. Elez accessed the PAM payment processing database. These activities are consistent with the read-only access that Mr. Elez was provided and did not change or alter any BFS payment system or record within their source systems. As noted, reviews of Mr. Elez’s work are still actively occurring; I do not have any more detail to provide at this time about his activities with respect to PAM. [my emphasis]

Then, on February 5, Elez got access to the payment system itself — again, with the same laptop on which he had source code.

19. Due to scheduling constraints, Mr. Elez was unable to meet with Bureau personnel to set up his access to the SPS database until February 5. On that date, lSS held a virtual walk-through session to help him to connect to the SPS database. He accessed this database exclusively under the supervision of Bureau database administrators in a virtual walkthrough session. According to the preliminary review of logs the Bureau has conducted to date, it appears Mr. Elez accessed the SPS database only once during that walk-through demonstration on February 5. It does not appear that he accessed the database again. As part of the ongoing review, additional log reviews are currently underway to confirm this. Mr. Elez never logged into ASAP, CARS, or ITS.gov, as technical access to those systems was never established for him. [my emphasis]

The next day, “it was discovered” (Gioeli does not say by whom, which means we’re not seeing a declaration from that person) that Elez actually “had mistakenly been configured” with read/write access, rather than “read only.”

20. On the morning of February 6, it was discovered that Mr. Elez’s database access to SPS on February 5 had mistakenly been configured with read/write permissions instead of read-only. A forensic investigation was immediately initiated by database administrators to review all activities performed on that server and database. The initial investigation confirmed that all of Mr. Elez’s interactions with the SPS system occurred within the supervised, walk-through session and that no unauthorized actions had taken place. His access was promptly corrected to read-only, and he did not log into the system again after his initial virtual over-the-shoulder session on February 5. To the best of our knowledge, Mr. Elez never knew of the fact that he briefly had read/write permissions for the SPS database, and never took any action to exercise the “write” privileges in order to modify anything within the SPS database—indeed, he never logged in during the time that he had read/write privileges, other than during the virtual walk-through – and forensic analysis is currently underway to confirm this. [my emphasis]

Throughout his declaration, Gioeli stresses that they’re still investigating whether Elez did something unauthorized with his access.

21. As noted above, the Bureau used several cybersecurity tools to monitor Mr. Elez’s usage and continuously log his activity. While forensic analysis is still ongoing, Bureau personnel have conducted preliminary reviews of logs of his activity both on his laptop and within the systems and at this time have found no indication of any unauthorized use, of any use outside the scope that was directed by Treasury leadership, or that Mr. Elez used his BFS laptop to share any BFS payment systems data outside the U.S. Government.

And his declaration stops short of saying that some mitigation techniques were used. For example, Gioeli says that the enhanced monitoring on Elez’ Bureau laptop had the ability to “block the use of external peripherals (such as USB drives or mass storage devices)” and access to the cloud, but it doesn’t say whether those functionalities was used.

Plus, Gioeli’s declaration makes it clear that they have’t involved Elez in any review of his access. No one has asked Elez, for example, whether he knew he had write access to the payment system.

Similarly, in an earlier declaration submitted in the parallel DC case, Thomas Krause gave a very couched answer about whether Elez had has any ongoing access.

I currently have no reason to believe Mr. Elez retains access to any BFS payment data, source code, or systems.”

Did anyone think to ask the guy? Does anyone know where that guy is? Are you going to interview him? Or is someone deliberately trying to keep him from being questioned further?

Worse still, Thomas Krause declaration submitted in the NY case doesn’t even say that Elez has left Treasury — only that he has resigned from the role of, “working closely with engineers at the Bureau of the Fiscal Service (BFS) on information technology (IT) matters in service of BFS’s mission to promote financial integrity and operational efficiency of the federal government through accounting, financing, collection, payment, and other relevant BFS services.”

On February 6, 2025, Mr. Elez submitted his resignation from this role. On that same day, he turned in his Treasury laptop, BFS laptop, access card, and other government devices; his BFS systems access was terminated; and he has not conducted any work related to the BFS payment systems since that date.

Elez was made a Treasury employee — contrary to early reports, he was not a SGE. That may make it easier to shuffle him off somewhere else.

What Gioeli describes is the panic that ensues when a guy who had high level access quits unexpectedly. And to date, we’ve never been given a formal explanation of why he quit — or whether he was asked to do so. We certainly can’t reconcile the claims that he has been reinstated with claims that he’s not doing what he was doing at Treasury.

Everyone has always assumed that Elez quit because his racism was discovered. But given the timeline, we can’t rule out that he quit because of the access concerns (and ongoing investigation) at Treasury.

Timeline

January 21: Elez hired.

January 23: Krause hired.

January 26: Treasury focuses on USAD. Treasury also adopts a 4-6 week engagement plan.

January 28: Bureau provides Elez with Bureau laptop copies of the source code for PAM, SPS, and ASAP in sandbox.

January 31: Treasury focuses on TAS codes; Elez assists in “automating” manual review of payments. “A high-ranking career official at Treasury also raised the issue of risks from DOGE access in a memo to Treasury Secretary Scott Bessent.”

February 3: Treasury gives Elez access to PAM. Booz threat contractor delivers report warning of grave insider threat.

February 5: Treasury gives Elez access to SPS, the payment system.

February 6 (afternoon): Elez resignation.

February 7: Treasury flags but then approves four payments. WaPo publishes story about Booz report and Booz contractor is fired.

February 8: Paul Engelmeyer limits Krause’s access.

February 10: Millenium Challenge Corporation submits, but then requests not to process, a payment.

Documents

Opposition to Stay

Thomas Krause Declaration: Describing the plan to use technology to provide more oversight over payments (citing three Biden-era GAO reports, not anything DOGE has discovered).

Vona Robinson Declaration: Describing that the only payment that has been intercepted at Treasury was a payment to the Millenium Challenge Corporation.

Michael Wenzler Declaration: Describing the hiring, employment status, revisions thereof, of Thomas Krause and Marko Elez, and also confirming Elez’ resignation from Treasury.

Joseph Gioeli Declaration: Describing the circumstances of Elez’ access and the investigation into what he did with it.

At a weird appearance in the Oval Office rife with awkward projections that Elon Musk believes he is more powerful than Trump (here’s the full CSPAN video), a journalist asked Elon how — given the egregious error he made about condoms and Gaza — we should believe anything he said.

The exchange is bad enough: Elon basically confessed, in front of Trump, that a hoax Elon started that traveled first to Trump propagandist Karoline Leavitt and from there, through Jesse Watters’ exaggerations on Fox News, into several repetitions of the false claim by Trump was wrong.

The entire point of this presser was to substantiate Trump’s false (and undocumented) claim that DOGE [sic] had found billions of dollars of waste, fraud, and abuse and use that to, first, pressure judges who are putting brakes on DOGE and, then, justify giving DOGE [sic] authority to fire a bunch of people via Executive Order.

When Trump asked Elon to substantiate such claims, Elon instead vaguely pointed to people who were wealthy even though they had meager salaries — not something that should be under his review. He listed other things that are known — and were known, during Trump’s first term — which are archaic but not fraud.

And in that appearance, a journalist called Elon out for inventing something about Gaza that led Trump to lie publicly.

That should have led to stories about how, in Trump’s presence, Elon admitted he makes shit up and Trump repeats them.

For the most part, it didn’t happen:

• NYT noted that Elon offered no proof of fraud, but did not mention the proof that Elon got caught in a lie.
• WaPo focused on the EO, but later explained that neither Trump nor Musk offered proof — but didn’t mention he got caught in a lie.
• Politico focused on the EO, but later noted that Elon said he would police his own conflicts.
• In an analytical piece, CNN claimed that Elon offered examples of fraud (which is false), but didn’t mention Gaza.

After airing Elon about scrutiny he claimed he was getting, WSJ did mention the Gaza question.

Asked about the Trump administration’s false assertion that the federal government sent $50 million worth of condoms to Gaza, the billionaire acknowledged that he might at times promote erroneous information. “Some of the things that I say will be incorrect and should be corrected,” Musk said. “Nobody is going to bat 1,000.”

But WSJ didn’t pursue the implication of it: that Elon got caught in a false claim.

Indeed, the only specific example that Trump mentioned was funding FEMA spent in NYC to house migrants — something approved by Congress — for which the staffers have been fired (as I’ll return to, Trump’s DOJ is already misrepresenting this in courts), was also based on an Elon lie.

The Trump administration said on Tuesday that it had fired four employees from the Federal Emergency Management Agency, including the agency’s chief financial officer, over their roles in disbursing federal funds to house migrants in New York City hotels.

The firings capped a startling chain of events that began on Monday with an early-morning social media post by Elon Musk who claimed, misleadingly, that FEMA had recently sent $59 million meant for disaster relief to New York City to pay for “high end hotels” for migrants, and who called the expenditure unlawful.

New York City officials raced to clarify that the federal money had been properly allocated by FEMA under President Joseph R. Biden Jr. last year, adding that it was not a disaster relief grant and had not been spent on luxury hotels.

Nonetheless, just two hours after Mr. Musk’s post, FEMA’s acting director, Cameron Hamilton, announced that the payments in question “have all been suspended” — even though most of the money had already been disbursed — and that “personnel will be held accountable.”

By Tuesday morning, roughly 24 hours after Mr. Musk’s post, the Trump administration had followed through on one part of its pledge.

Elon also made a false claim that they had turned on AIDS prevention — in one of the state lawsuits, Washington State presented a case where funds for AIDS prevention programs was being withheld.

This press conference consisted of Elon (and Trump) making false claim after false claim.

It also consisted of Trump lying over and over, without proof, about how one only needed to look for fraud to find it. No one asked why he hadn’t looked in his first term. Indeed, several times he blamed Biden for problems that have existed for decades.

And yet, at best, journalists instead claimed only that Elon and Trump simply presented no proof.

Yesterday, I argued that Trump would not yet defy courts because he wants to invite the Supreme Court to sanction his dictatorial powers, and so wants a clear appellate record.

Boy howdy was that a short-lived theory. Trump says he is appealing two orders that are not yet ripe for appeal in two lawsuits involving Democratic Attorneys General — RI Judge John McConnell’s order and follow-up order that the government pay grants to the states [appeal] and Paul Engelmeyer’s order ordering Treasury to stay out of the payment system [request for stay pending appeal] — as well as in Special Counsel Hampton Dellinger’s challenge to his dismissal.

So by the time Republicans figure out how they’re going to use reconciliation to pass Trump’s policies, SCOTUS may have already agreed to gut Congress’ power of the purse.

But the record in the spending cases is anything but clean.

In one of the two cases challenging DOGE’s [sic] access to Treasury systems — the DC case before Colleen Kollar-Kotelly — DOJ decided after the fact that Marko Elez, the DOGE [sic] boy who had been granted a copy of Treasury systems to sandbox, was actually a Treasury employee.

With the benefit of more time to investigate the facts over the weekend, Defendants came to understand that Marko Elez, who, at the time of the hearing was employed by the Department of the Treasury, had not, in fact, been designated by the Treasury Department as a Special Government Employee (SGE), as counsel stated at the February 5 hearing. Mr. Elez, was, however, a Treasury Department employee. Treasury hired Mr. Elez as Special Advisor for Information Technology and Modernization, Departmental Offices, Office of the Chief of Staff, under Treasury’s authority to establish temporary transitional Schedule C positions. See 5 C.F.R. § 213.3302. Although Mr. Elez could have been designated as an SGE because he was slated to perform temporary duties either on a full-time or intermittent basis for not more than 130 days, the Treasury department Ethics office did not designate Mr. Elez as a Special Government Employee, meaning that he in fact had to comply with additional ethics requirements that are not required for SGE positions.

[snip]

Defendants also wish to notify the Court that, as stated in the Declaration of Thomas Krause, Jr., filed yesterday, in State of New York v. U.S. Department of the Treasury, Case No. 25 Civ. 01144 (JAV) (S.D.N.Y.), Mr. Elez resigned from Treasury on February 6, 2025, and he returned all Treasury and BFS equipment and credentials the same day. See Exhibit 1, ¶ 11. Moreover, in that case, on February 8, the Court entered a temporary restraining order restricting who may access Treasury systems. See Ex. 2. Those restrictions are in addition to those imposed by this Court’s Order entered February 6.

This filing included Thomas Krause’ declaration (submitted in the Treasury suit filed by states, which Trump is appealing) describing that Elez had resigned (but not addressing whether he has been reinstated; in retrospect, it seems the declaration was written specifically to avoid calling Elez a DGE). But it didn’t include the underlying filing in the case, which in a footnote confesses that Elez had a full copy of the BFS system in a sandbox, falsely claiming that Krause addressed this in his declaration.

2 Since January 20, 2025, one other Treasury employee—Marco Elez—had “read only” access to or copies of certain data in BFS payment systems, subject to restrictions, and access to a copy of certain BFS payments systems’ source code in a “sandbox” environment. Krause Decl. ¶ 11. Mr. Elez resigned on February 6, 2025 and returned all Treasury and BFS equipment and credentials the same day. Id

This means that this correction doesn’t correct another false claim DOJ made to Kollar-Kotelly: that Elez’ access had been “read only.” And DOJ hasn’t told Judge Jeanette Vargas (to whom the New York case was assigned after Engelmeyer issued the TRO) that Elez is a full Treasury employee and so, if he has been reinstated, potentially excluded from Engelmeyer’s order.

In the USAID case, where Trump might believe he can coax a favorable ruling from his own first term appointee, Carl Nichols, Peter Marocco submitted a long, obnoxious declaration claiming they had to shut down USAID because of widespread insubordination among USAID employees. (I’d quote from it but the declaration breaks local rules requiring OCR filings.)

But after Marocco submitted that filing, the career AUSAs on the case submitted a declaration that included this correction.

Additionally, although Secretary Rubio’s January 24, 2025 directive only froze future contract obligations, id. ¶ 3, payments on existing contracts were paused as well as part of efforts by agency leadership to regain control of the organization’s spending and conduct a comprehensive review of its programs. See id. ¶¶ 5–10. Counsel for Defendants was unaware of this development prior to the hearing. [my emphasis]

Marocco confesses that existing contracts “were paused” by him this way:

Furthermore, many of USAID’s pre-existing programs were in conflict with the directives and priorities of the President and Secretary, and therefore were inconsistent with the public interest and foreign policy judgments of the Executive Branch. Given the scale of these programs, an ad hoc review of these conflicting programs would unduly burden the execution of the President’s other foreign policy priorities. A blanket pause with a waive-in process was the more efficient and effective path.

He describes this notice Marco Rubio sent to Congress, which makes no mention of pausing ongoing work. Then he continues to describe how existing programs “were paused” by him.

The first step of this review, in essence, involved the majority of USAID pausing a substantial portion of its ongoing work — going “pencils down” — so the Secretary and USAID leadership could gain control of the organization that included some employees who had refused to comply with lawful directives by the President and Secretary, directives designed to identify wasteful or fraudulent programs or those contrary to the foreign policy interests of the United States. The pause of ongoing work and use of paid administrative leave have enabled Agency leadership to begin a thorough review of USAID’s operations and align its functions to the President’s and Secretary’s priorities, without continued noncompliance by former Agency leadership and management undermining those priorities. Pausing a majority of USAID’s work was, and remains, necessary to continue this thorough review into the noncompliance issues first identified, as well as to continue to examine USAID’s processes and the manner in which USAID funds its programs.

In other words, the people that Marocco calls noncompliant are noncompliant because they’re following the law, a law uncontroverted by Trump’s order or even Rubio’s notice to Congress.

As Nichols said when he issued the TRO ordering USAID to reinstate employees, whether or not this involved existing or only prospective contracts was an issue of some contention in the hearing.

Plaintiffs finally seek a TRO as to Secretary Rubio’s January 24, 2025 order freezing funding to USAID’s contractors. As a threshold matter, the Court notes that there are significant factual questions about what the practical effect of that order is. The government argued at the hearing that the order only prevents USAID from entering “new obligations of funding”—leaving it free to pay out contracts that it entered into prior to January 24, 2025—and indeed, the text of the order does seem to permit that result. Dep’t of State, Memo. 25 STATE 6828. Yet, plaintiffs maintained at the TRO hearing that payments on existing USAID grants have been frozen, preventing certain “contracting officers” employed by USAID from using agency funds to fulfill monetary commitments that the agency had already made.

This factual dispute is relevant to plaintiffs’ TRO arguments, but ultimately is not dispositive of them. Plaintiffs allege that, by some legal mechanism, USAID contracting officers can be held personally liable for existing contractual expenses that USAID is supposed to, but does not, pay. Plaintiffs thus argue that those officers face irreparable harm as a result of the funding freeze because they will be left “holding the bag” when USAID imminently fails to disburse funds. Separately, plaintiffs argue that the general population of USAID employees will be emotionally harmed by the agency’s inability to pay its contractors because they will be stuck “watching a slow speed train wreck” as the agency reneges on its humanitarian commitments.

Even assuming the funding freeze indeed prevents payments on existing grants in the way plaintiffs claim (instead of merely preventing USAID from entering new obligations, as the government suggested during the hearing), the Court concludes that plaintiffs have not demonstrated resulting irreparable harm.

But because this suit involves employees, rather than states or other recipients of funds from Treasury (as is the case in the two suits where DOJ has said it will appeal), these plaintiffs themselves are not being injured because they’re still being paid.

DOJ is hiding behind career AUSAs making claims they likely do not know are false so as to shut down appropriations that have already been approved.

And they are appealing each instance in which a plaintiff has genuinely been injured (the states and Hampton Dellinger’s firing) in hopes — or maybe expectation? — after the Circuits deny appeals that are not yet ripe, SCOTUS will step in and render Congress impotent.

Update: USAID Inspector General somehow managed to put together a report on the damage the chaos is having. Among other things, it finds that the cuts have incapacitated any means of vetting disbursements to keep them out of the hands of terrorists.

USAID describes partner vetting as a risk-mitigation tool to “ensure that American taxpayer funds do not benefit terrorists and their supporters.” Currently, partner vetting is required for programming in Afghanistan, Iraq, Lebanon, Pakistan, Syria, West Bank/Gaza, and Yemen where designated terrorist organizations such as Hamas, Hezbollah, ISIS, and Ansar Allah (also known as the Houthis) operate. Before the Agency awards a contract, grant, or cooperative agreement in these locations, the proposed awardee must submit to USAID data needed to vet the organization and its key personnel. The same vetting must be undertaken before an aid organization issues a subaward. While USAID OIG has previously identified gaps in the scope of partner vetting, 10 USAID staff have reported that the counter-terrorism vetting unit supporting humanitarian assistance programming has in recent days been told not to report to work (because staff have been furloughed or placed on administrative leave) and thus cannot conduct any partner vetting. This gap leaves USAID susceptible to inadvertently funding entities or salaries of individuals associated with U.S.-designated terrorist organizations.