Read more

Read more

The Make America Read Again (MARA) Book Club will be meeting this evening at 6:00 pm Eastern time to discuss James T. Campbell’s fascinating book, “Middle Passages: African American Journeys to Africa, 1787-2005.” In lieu of a fiction wine pairing, the author has designated the following poems by Langston Hughes as the wine pairing for this book—for reasons that will become obvious once you begin the introduction:

• Afro-American Fragment

• The Negro Speaks of Rivers

• Aunt Sue’s Stories

• Negro

• The Weary Blues

• Misery

• Theme for English B

• I, Too

• Cross

• The Negro Mother

MARA Book Club events are for paid subscribers only. If you are not a paid subscriber and want to join, I’ll give you a hint about how you can fix that:

Subscribe now

Information about how to attend is below the paywall.

Friday on #DogShirtTV, the estimable proved the revolutionary nature of the show by talking without an agenda or chosen topic for an hour:

Recently On Lawfare

Compiled by the estimable Sarah Willrich

When Reporting an AI Security Incident Is Not Mandatory

Mackenzie Arnold and Stephan Llerena highlight the surprising point that OpenAI was likely not required to disclose the recent hacking of Hugging Face conducted by rogue OpenAI artificial intelligence (AI) agents. Arnold and Llerena explain why existing state laws do not compel disclosure and suggest that, rather than sharing partial reports after attacks, companies and state agencies collaborate to create a baseline of information that enables a more rapid response.

So long as reporting remains almost entirely voluntary, detailed answers to these questions will be hard to come by, and the companies that offer information voluntarily will be subjected to greater scrutiny than those that are less cooperative. Well-crafted rulemaking authority will be key both to ensure that information is adequate and that companies are well informed about their obligations. In many circumstances, authorities will not know all the details they require until after a serious event occurs. In those cases, they will need investigative powers to obtain the needed information.

Legal-ism Mission Creep

Matthew Forst reviews Doyle K. Hodges’s book, “The Rise of US Military Legalism: A New Theory of Military Professionalism.” Forst praises the book’s account of how military decision-making has shifted from a premise of individual judgment and ethics norms to careful parsings of legal texts to justify military action.

Borrowing Samuel Huntington’s characterization of commanders as “managers of violence,” Hodges describes military professionals as those with expertise in weapons systems and battlefield engagement. He contends that they are entitled to autonomy within this sphere and are obligated to stay out of everything else, including the political and policy matters reserved for civilian principals. As Hodges explains, the civil-military dynamic between policymakers and military officers is a principal-agent relationship in which the former sets the political agenda and the latter executes the corresponding military mission. The two remain interconnected, but both halves of the relationship have their domains, and military ethics requires warfighters to remain committed to obeying civilian control. Yet as the number of constraints on the use of force from policymakers mounts, military autonomy diminishes. Military legalism is thus the path of least resistance for a profession that values both autonomy and obedience to civilian control.

Knives Are Out for Open-Weight AI Models

In the latest edition of the Seriously Risky Business cybersecurity newsletter, Tom Uren discusses signals from the Trump administration and China indicating their intention to regulate open-weight AI models, the recent hacking of Hugging Face, sentencing in the U.K. of the Scattered Spider group hackers, and more.

Despite that, it seems that internal debates within the Trump administration regarding open-weight Chinese models have been settled: The Chinese are stealing American intellectual property, and there needs to be a reckoning.

That’s a shame, because these open-weight models are currently filling a gap not served by America’s frontier labs.

Podcasts

On Lawfare Daily, Maggie Haberman and Jonathan Swan speak with Roger Parloff about their recent book, “Regime Change: Inside the Imperial Presidency of Donald Trump.” They discussed how close Trump has come to suspending habeas corpus, how the Supreme Court’s immunity decision has shaped his conduct, his vindictive prosecutions, and his firings of national security professionals at the behest of Laura Loomer.

On Scaling Laws, Alex Bores joins Kevin Frazier and Alan Z. Rozenshtein to discuss his congressional campaign platform centered on AI policy. They talk about what he heard from voters about AI, how the millions of dollars coming into the race changed its dynamics, and the proper allocation of AI regulatory authority between states and the federal government.

Videos

On Lawfare Live, Anna Bower, Eric Columbus, and Parloff discuss this week’s hearing on the subpoenas sent to New York Times reporters, the recently ousted court-appointed U.S. attorney in the Western District of Washington’s challenge of his termination, and the first convening of the Alien Terrorist Removal Court.

Today’s #BeastOfTheDay is the horse, seen here exercising truly extraordinary patience by not stepping on a dog:

Video Source

Today’s #BeastOfTheDay is not the dog, which should not be rewarded for its idiocy.

Today’s #BeastOfTheDay is not the Grout White Shark, which is ineligible because it is not a real Beast:

Read more

And we did it! Phase I of #OperationTaya is now complete. I will go to the bank and wire the first payment of her tuition to the College of Europe on Monday. Thanks to the dozens of you who contributed. I think we have proof of concept.

Subscribe now

Yes, it can. And no, it doesn’t need to.

No.

Getting older. Becoming middle aged protesters.

No. Everything is bad for you. Even not drinking any alcohol is bad for you, because you’ll be drinking something else that will be bad for you. There’s no winning. It’s all bad for you for in the end.

No. It’s turned you into something even more annoying: which is a person who ruminates about whether he’s a sports guy. Watch sports. Don’t watch sports. Enjoy sports. Think watching other people you don’t know play sports is dumb. Whatever. Nobody needs your psychodrama about it.

Lift heavy things over and over again. Do wall squats. Do pilates.

No, I want to bet on your marriage, which won’t last.

Wednesday on #DogShirtTV, the estimable had various questions, to which the estimable and I attempted to provide answers:

Yesterday on #DogShirtTV, the estimable Mike Feinberg and I talked AI and quantum computers:

Recently On Lawfare

Compiled by the estimable Sarah Willrich

When Commercial Satellites Become Wartime Intelligence

Joshua Villanueva explains the difficulty of categorizing commercial satellites under international law given the overlap between civilian and military uses. Villanueva argues that military utility alone is insufficient to deem satellites a valid wartime target, proposing instead a six-factor framework in which operational integration into military activity is the most significant element.

One state may view commercial remote sensing as part of the standard information infrastructure, while an adversary may consider it an intelligence asset supporting enemy operations. Russian officials have warned that Western commercial satellites could become targets if used to aid Ukraine, and the United States has stated that attacks on U.S. infrastructure would trigger a response. The legal status of commercial space infrastructure can become contested in real time, with serious implications for crisis stability.

The Next Mythos Moments

Anton Leicht and Alasdair Phillips-Robins warn that U.S. policy is unprepared to respond to future novel threats posed by leaps in artificial intelligence (AI) development. They argue that without preemptive action—primarily by the executive branch, with Congress eventually following through with legislation—the U.S. response will struggle to balance the risks of governmental overreach against the danger of moving too slowly to address vulnerabilities.

Spur-of-the-moment decisions are also more likely to be based on thin evidence and political happenstance. The risk is governance by streetlight: Whatever the administration becomes aware of receives attention, and whatever the developers successfully conceal remains unregulated. Given that precedent, if an AI developer were to spot dangerous biological capabilities in its next model and no mechanism existed to compel the company to inform the government, executives might think twice about approaching the administration, and policymakers would miss out on an important window to prepare defenses.

The Monroe Doctrine Was Never a Blank Check for Presidential War

Katherine Yon Ebright argues that those who encourage the return of the Monroe Doctrine, and its detractors, misunderstand its original intentions. Yon Ebright explains that the Monroe Doctrine was originally interpreted as supporting anti-imperialism and constitutional limitations on the president’s war powers, and that this conception only began changing later, in the 20th century.

Those who seek to aggrandize presidential power or to critique the Monroe Doctrine often note that the message does not foreclose U.S. intervention, as opposed to European intervention, in Latin America. But this glosses over the message’s disavowal of U.S. adventurism and its recognition of Latin American independence based “on just principles.” It also overlooks the well-established position of Monroe’s secretary of state and the architect of the Monroe Doctrine, John Quincy Adams. Two years earlier, Adams had famously proclaimed that the United States “goes not abroad in search of monsters to destroy.” He lauded the U.S. policy of nonintervention, crediting it with keeping the nation out of wars that would “usurp the standard of freedom,” “insensibly change [U.S. policy] from liberty to force,” and transform America into “the dictatress of the world.”

Why U.S. Technology Controls Keep Faltering, and How to Save Them

Doni Bloomfield and Jeff Gordon consider the inconsistencies of the Trump administration’s export controls policy, arguing that the oscillation between stringent limitations on sales and permissiveness harms U.S. companies’ ability to generate revenue from sales to China, in turn limiting their capacity for the long-term research and development that maintains the U.S.’s technological edge. To rectify this, Bloomfield and Gordon propose tightening export controls while providing a tax credit to offset the revenue loss.

The administration’s seesawing risks becoming the worst of both worlds: giving China intermittent legal access to advanced U.S. technology while undercutting U.S. companies’ stable incentive to invest in R&D. The tax on U.S. companies is sizable. Researchers at the Federal Reserve have found that export controls on chips led U.S. companies’ market cap to fall by more than $150 billion, and that Chinese firms turn to non-U.S. suppliers, including those in China, to fill in gaps. And a mountain of economic evidence shows that how much companies invest in R&D depends in large part on the size of the market they can sell to. When the U.S. restricts sales to China—and all the more when it restricts sales entirely—it burdens the very firms it relies on to maintain its technological lead.

Podcasts

Lawfare Daily: Inside Zelensky’s Disastrous Decision to Fire Ukraine’s Minister of Defense: Anastasiia Lapatina sits down with Francis Farrell to discuss the Ukrainian president’s decision to replace the country’s minister of defense, the mass protests that happened as a result of his firing, and how the decision will impact the course of the war.

Lawfare Daily: Can State Laws Protect Against Misuse of the National Guard?: Natalie Orpett sits down with Elizabeth Goitein, Joseph Nunn, and Loren Voss to discuss the state laws governing the National Guard and what they mean for domestic deployments in the United States.

Rational Security: The “(Deep State) Friends We Make Along the Way” Edition: Scott Anderson sits down with Eric Columbus, Julia Curlee, and Molly Roberts to talk through the week’s big national security news stories, including President Trump’s election speech and document release, the Justice Department’s issuance of subpoenas to reporters from The New York Times, and the confirmation hearings of Jay Clayton and Todd Blanche.

Today’s #BeastOfTheDay is the likweli, a newly discovered monkey with a face that looks like a plastic surgeon’s error:

Congratulations to the likweli on being officially documented as a species! In honor of today’s Beast, check out the paper announcing the discovery. Alternatively, you can honor today’s Beast by wearing way too much orange lipstick.

Subscribe now

Read more

Thanks to everyone who has leapt to contribute to the #GreekChorusFoundation’s first project, #OperationTaya, which is our scheme to send a young staffer working at the Ukrainian Ministry of Defense to graduate school so she can become a master of the universe. In case you missed Monday’s announcement of the foundation and its initial undertaking, here it is:

Taya’s first tuition payment of ten thousand euros is due at month’s end, so there is some urgency to the matter. If you haven’t yet contributed, please do so. And please share this project with others you think may be interested.

Support #OperationTaya

And of course:

Subscribe now

Monday on #DogShirtTV, we announced the launch of the Greek Chorus Foundation:

Yesterday on #DogShirtTV, the estimable came on the show with two agenda items: xenophobic violence in South Africa—and Canoe Guy:

Recently On Lawfare

Compiled by the estimable Sarah Willrich

Everything Trump’s Election Speech Did Not Prove

Anna Bower, Julia Curlee, Renée DiResta, Michael Feinberg, Molly Roberts, and Loren Voss unpack President Trump’s recent speech on the integrity of the 2020 elections and the declassified document releases that accompanied it. They break down each of Trump’s allegations, the administration’s playbook for politicizing intelligence, and the consequences such rhetoric could have ahead of the 2026 midterm elections.

The audience is invited to treat these separate allegations as mutually reinforcing evidence of a single stolen-election plot. But they do not corroborate one another. Evidence that China acquired voter data is not evidence that it altered registrations. Evidence that intelligence analysts disagreed about the goals of Chinese influence is not evidence that any intelligence was concealed, or that analysts believe the influence effort had meaningful impact. A vulnerability is not an intrusion; a fraudulent registration application is not a fraudulent vote.

The White House’s case depends on erasing those lines.

The Short-War Illusion: What Iran Teaches Beijing About Taiwan

Curlee considers what lessons China may be taking from the U.S and Israeli war with Iran for its own potential invasion of Taiwan. She highlights three key takeaways: decapitation strikes have limited efficacy alone, chokepoints are cheap to deny but costly to reopen, and together, these factors mean the illusion of a short conflict can quickly turn into a drawn-out war of attrition—one Taiwan may not be prepared to endure.

Beijing’s likely takeaway is not that decapitation is useless, but that it cannot be the failure point for the entire plan. We should expect the PLA to pair leadership strikes with moves to deepen the island’s isolation—severing the two dozen undersea cables that carry nearly all of Taiwan’s connectivity, a vulnerability Chinese-linked vessels have already rehearsed cutting. Beijing is likely to leverage cyberattacks to cripple domestic communications so that a surviving government cannot communicate with the outside world, coordinate its military, or reassure its public, and expand the on-island subversion and disinformation capabilities it has spent years building to ensure Taiwan cannot—as Iran did—quickly name a successor.

Governance Fragmentation and Collapsing Health in Wartime Yemen

In the latest edition of Lawfare’s Foreign Policy Essay series, Safa Al-Saeedi and Christina Bouri analyze how the ongoing conflict in Yemen has driven the collapse of the country’s healthcare system. They explain that, in addition to the U.S. cutting funding last year, Yemen’s split between the Houthi authorities and the internationally recognized government has left aid organizations navigating rival health systems, identification documents, and supply chains—resulting in a patchwork system susceptible to politicization.

Rival factions have also treated humanitarian assistance and health facilities as patronage resources and instruments of control, further disrupting the delivery of health services. They have diverted essential commodities—including food, medicine, and fuel—to loyalists and cut off populations that are not supportive enough. For example, in 2018, Houthi authorities diverted an estimated 15,000 food baskets per month from the education ministry; these had been intended for vulnerable families but were instead sold on the black market or redirected to frontline fighters. This strategy has created deep inefficiencies in aid delivery, pervasive shortages, and the breakdown of health services, thereby accelerating outbreaks of disease and unnecessary loss of life.

Within Bounds: State Authority to Regulate Federal Contractors

David S. Rubenstein analyzes how courts have allowed and disallowed states from regulating federal contractors operating in their jurisdictions. He explains that states are largely allowed to enforce generally applicable laws against contractors, but cannot use regulation to single them out. Rubenstein also explores how states can leverage their market influence to indirectly regulate federal contractors.

So long as states do not directly regulate the federal government or single out federal contractors for discriminatory treatment, the intergovernmental immunity doctrine leaves room for neutral state health and safety rules. Under these principles, states concerned about an influx of new immigration detention facilities, for instance, can require health department inspections, fire suppression, and adequate egress for private facilities housing large numbers of people.

The Legal Risks That Chill Good-Faith Security Research

Belen Pisaniello, Sunoo Park, and Daniel R. Thomas discuss the chilling effects that contract and anti-hacking law can have on researchers who expose vulnerabilities in good faith.

One particularly alarming manifestation of these chilling effects is “stockpiling,” in which researchers discover security vulnerabilities but choose not to disclose them out of fear of legal consequences. Instead, they hold onto this information, making it a prime target for hacking, bribery, and coercion. One researcher reported being approached by a mysterious individual offering a large sum of money to buy their stockpile and their silence.

Podcasts

Lawfare Daily: The Trials of the Trump Administration, July 17: I sit down with Eric Columbus, Bower, Roberts, and Roger Parloff to discuss the Senate confirmation hearing for acting Attorney General Todd Blanche, subpoenas sent to New York Times reporters, a judicial opinion voiding President Trump’s settlement with the IRS, and more.

Lawfare No Bull: Confirmation Hearing for Director of National Intelligence Nominee Jay Clayton: Sarah Willrich shares edited audio of the July 15 confirmation hearing before the Senate Intelligence Committee for Jay Clayton, President Trump’s nominee for director of national intelligence.

Lawfare Daily: How To Make a Patriot, with Fabian Hoffman: Anastasiia Lapatina talks to Fabian Hoffman about what it would take for Ukraine to produce Patriot interceptor missiles, if the United States gave it the license to do so.

Scaling Laws: Courtney Cox on AI in Adjudication: Courtney Cox joins Kevin Frazier to discuss her forthcoming article, “Hardwiring Hercules?,” in which Cox argues that the debate over a “right to a human decision” has missed a central problem: normative uncertainty.

Today’s #BeastOfTheDay is the raven, which new research finds does not follow wolves in order to feed on their kills. Instead, the ravens predict where wolves are likely to kill things and show up in advance. A March 2026 study finds that:

Ravens benefited from their associations with wolves but demonstrated a more dynamic strategy than simply following individual wolves or packs. Instead of consistently remaining near carnivores, ravens periodically returned to areas of high carnivore activity or past kill sites based on anticipated food availability, while primarily foraging elsewhere…

Ravens’ movements toward wolf kills were often highly directional, covering distances of up to 155 km in a single day. Because wolf kills are clustered in particular areas of northern Yellowstone and not randomly distributed, we propose that ravens find kills by relying on spatial memory, possibly represented as a cognitive map of long-term average wolf presence and/or wolf kill abundance…[R]avens may learn and remember areas where wolf kills are abundant and where they are not.

In honor of today’s Beast, read the study. They put GPS trackers on all the Beasts and tracked them for years. It’s neat.

Subscribe now

Read more

I’m delighted to announce that the #GreekChorusFoundation is now live and accepting donations for #OperationTaya, our first project.

What is the #GreekChorusFoundation? It’s the institutionalization of #ProjectBattery, my effort earlier this year with the estimable to raise money to buy batteries for Ukrainian families who had no electricity in a frigid winter. With the help fo the audience of #DogShirtTV and the readership of #DogShirtDaily, we raised about $130,000 and spent it all very quickly helping real people do real things. In response, the #GreekChorus—our name for the audience of the show and the readership of the site—decided to create a foundation so that donations for future projects would be tax deductible and wouldn’t be running through my personal accounts.

And thus the #GreekChorusFoundation was born.

Support the #GreekChorusFoundation

The Foundation will operate as a communal project of #DogShirtTV and #DogShirtDaily. Major decisions about projects will be discussed on the show and voted upon in polls on #DogShirtDaily. The decision to constitute the Foundation itself was approved by the readership in almost Stalin-level numbers:

And the decision to take on our first fundraising project was adopted with an overwhelming vote as well:

And that brings me to Taya, a young staffer who works with the Ukrainian Ministry of Defense. We are not releasing Taya’s full name publicly right now, because of the circumstances of her employment. But she has been accepted into a two-year joint Master of Arts in Transatlantic Affairs at the College of Europe and the Fletcher School, the cost of which is prohibitively expensive. We figure we can do something about that.

Taya writes:

During my lifetime, I never had the luxury of experiencing a stable, safe Ukraine. Instead, it was full of life-changing events: two revolutions, waves of economic and political instability, and now a war. These experiences taught me lessons no classroom ever could. I learned early that ignorance is not an option here; it is a privilege we simply cannot afford. And, perhaps most importantly, that impact requires more than good intentions: it requires knowing your own capabilities, finding others who share your purpose, and building something together.

These were not abstract lessons. They became the logic for every decision I made about my education and my career. From a young age, I understood that the value of my work would be measured by its social impact, by whether it moved something forward for the people around me.

My career has taken me through regional administration, the Parliament, the Ukrainian Red Cross, and now the Ministry of Defence. In each of these roles, one thing has remained constant: the extraordinary commitment of the people around me, colleagues determined to reform our institutions, deepen ties with our allies, and build a country worth the sacrifices being made for it.

But this experience has also taught me something less obvious. We are very good at articulating what Ukraine needs. We are far less practiced at understanding how our partners think: how their institutions weigh decisions, build consensus, and shape policy. As Ukraine’s path toward the EU and NATO advances, I believe closing that gap will matter as much as any technical expertise we bring to the table. What is needed now are people fluent in both worlds, grounded in Ukraine’s realities, yet equally capable of reading and navigating the logic of our international partners.

That realization is one of the main reasons I applied for Political and Governance Studies in the Master of Arts in Transatlantic Affairs programme. I don’t simply want to gain another degree, I want to challenge the way I think, learn directly from people who shape transatlantic policy, and bring that perspective back to Ukraine. I believe that would allow me to contribute to my country on a much greater level than I can today.

Taya’s tuition will cost €30,000 for each of the two years of the program. We mean to cover it all.

One matter of urgency related to #OperationTaya. Taya’s first tuition payment of €10,000 is due on July 30.

So let’s get on it:

Support #OperationTaya

Wednesday on #DogShirtTV, the estimable Mike Feinberg and I welcomed the estimable Julia Curlee, fired NSC director, to talk about Trump’s destruction of the NSC:

Thursday on #DogShirtTV, the estimable Molly Roberts came on the show to speculate wildly about Trump’s speech:

And Friday on #DogShirtTV, the speech had occurred, and Molly came back, along with the estimable , to discuss it:

Recently On Lawfare

Compiled by the estimable Sarah Willrich

Takeaways for Iran a Year After the Syria Sanctions Rollback

Rachel Alpert, Claire O’Neill McCleskey, and Naomi Maxwell consider what lessons from the end of U.S. sanctions on Syria could be applied to Iran. The authors explain that the rollback of sanctions in Syria required continuous effort from the executive branch, Congress, and its international partners and argue that the process would likely be even more complicated in the case of Iran.

Syria’s SST designation proved to be sticky even after a decisive regime change. In Iran, unlike Syria, the conduct underlying its SST designation is ongoing, and the country is now led by what some observers have described as a new generation of “technocratic” IRGC leaders. As long as Iran remains a designated SST and the IRGC remains designated as a FTO, U.S. companies will be exposed to significant legal risk for dealing with anyone linked to the IRGC. Because the IRGC is deeply embedded across Iran’s economy—from energy and construction to telecommunications—that exposure is difficult to avoid in practice, and it will continue to deter investment even if broader sanctions are lifted.

Driven by AI, the Future of Cybersecurity Will Resemble the Past

Jason Healey reviews how cyberattacks historically transitioned from taking advantage of vulnerabilities to relying on deception. He argues that the development of artificial intelligence (AI) will reset cybersecurity to a world where there are abundant vulnerabilities that defensive technologies, even those powered by AI, are not yet equipped to address.

Mythos-class AI technologies are not only more adept at zero-days, but they can do more even with known vulnerabilities (called n-days). Anthropic found that within just a few hours, using known vulnerabilities, “Mythos Preview … not only produced a full chain exploit, but produced eight distinct exploits” for well under $20,000. And AI-discovered vulnerabilities are worrying; even before Mythos, the number of flaws that were both highly exploitable and high severity jumped 36 percent in one year.

The impact on defense will be profound. Former White House cyber coordinators Rob Joyce and Michael Daniel have complementary assessments of the dangers as, respectively, a great cleansing forest fire of old software and “our legacy technical debt” coming due.

Implementing Universal Jurisdiction in Ukraine

Anna Sayko and Jeanne Sulzer analyze the practical implications of implementing universal jurisdiction in Ukraine, which allows for the prosecution of foreign nationals for core international crimes, regardless of when and where the crime initially took place. Sayko and Sulzer explain that while the law makes prosecution theoretically possible, clearer guidance is needed on procedure and how prosecution interacts with international law, especially in the context of Ukraine’s ongoing war with Russia.

In cases where an alleged perpetrator is apprehended in Ukraine but no request for extradition from a foreign state is made, the legislation does not clarify whether the absence of such a request is sufficient to satisfy the conditions for exercising universal jurisdiction. Given that Article 8(2) of the CCU establishes a framework for exercising UJ in accordance with Ukraine’s international obligations, in such cases, the principle of aut dedere aut judicare (extradite or prosecute) should apply. More broadly, under the Geneva Conventions and the UN Convention against Torture and Other Cruel, Inhuman, or Degrading Treatment of Punishment, a state party is obligated to carry out criminal proceedings against suspects on its territory if it chooses not to extradite them to the state in which the crime was committed.

Responding to AI Distillation Without Panic

Bahrad A. Sokhansanj argues that policymakers rushing to crack down on artificial intelligence (AI) “distillation attacks”—training a model on a rival’s outputs at scale—are diagnosing the problem incorrectly. Sokhansanj proposes that, rather than treating all distillation attacks as theft, which they are often not, policymakers should identify the specific harm such attacks cause, such as bypassing model safeguards or threatening cybersecurity, and tailor policy accordingly.

Mass distillation merits a policy response, even if it isn’t theft. But policymakers first need to identify the problem they are trying to solve. If the concern is illegitimate access to U.S. frontier models by foreign competitors and state actors, then policy should help labs secure access, share threat information, and identify fraudulent accounts and proxy networks used to disguise who is accessing the model. If the concern is cybersecurity, then the problem is account abuse and getting around access controls. If the concern is the diffusion of dangerous model capabilities, then the first step is to determine whether distillation meaningfully improves those capabilities or helps remove safeguards.

How the U.S. Military Learned to Embrace AI Warfare

Jakub Kraus reviews Katrina Manson’s book, “Project Maven: A Marine Colonel, His Team, and the Dawn of AI Warfare,” an account of the U.S. military’s collaboration with Palantir to develop artificial intelligence (AI) for warfare. Kraus praises Manson’s documentation of such a secretive project and highlights how she tied Project Maven to broader questions about ethics in AI warfare.

Given that leaders of military projects generally seek to control information about their efforts, readers might wonder whether Manson sometimes received a curated view of Maven. Nonetheless, her tone is consistently neutral and journalistic; she often spells out conflicting perspectives, allowing readers to draw their own conclusions. This fact-finding, even-handed storytelling makes her work stand out. It is rare to find a book that unpacks a U.S. military project with such depth and neutrality, particularly when the project involves fundamentally shifting America’s approach to war.

Vermilion Tide

Dominic Brennan and Charlie Prior consider the implications of China’s plan to deploy floating nuclear power plants (FNPPs) to the South China Sea. FNPP deployment, Brennan and Prior argue, could both improve China’s ability to sustain operations in the region and increase the risk of nuclear escalation in the highly contested waters.

U.S. strategic objectives may nevertheless require military action against Chinese bases in the region, compelling U.S. forces to confront FNPPs. Allowing a radioactive tripwire to deter any form of engagement would set a dangerous precedent. Accepting this logic would grant de facto immunity to any military facility housing a nuclear reactor, a position that becomes increasingly problematic as countries, including the U.S., explore the deployment of nuclear reactors on military bases. To maintain credibility that the mere presence of FNPPs will not deter the United States, the U.S. military should thoroughly prepare for the potential consequences of operations on or near FNPPs. These preparations take two routes: kinetic strikes and on-site operations.

Ransomware Uses AI to Amp Up Negotiations

In the latest edition of the Seriously Risky Business cybersecurity newsletter, Tom Uren discusses how AI is enabling more effective ransomware extortion, how AI-supported bug patches may expose small companies that cannot quickly implement patches, the White House’s announcement of a new cybersecurity “clearinghouse,” and more.

The INC/Lynx strategy here is to use AI to hack more often to make up for less frequent paydays. FulcrumSec is taking the opposite approach and trying to maximize profits by amplifying their leverage in negotiations.

Our take-home message here is that nerdy, technically focused ransomware actors use AI to hack, but sophisticated ransomware actors? They use it to negotiate.

Podcasts

Lawfare Daily: AI Targeting Systems Are Coming—But Not as Fast as You Think: Kate Klonick speaks with Steve Feldstein about his recent Bulletin of the Atomic Scientists essay on AI targeting systems. Feldstein discussed what the conventional wisdom about AI warfare gets wrong, what the Iran War’s staggering Maven numbers do and don’t prove, how Israel became the case study in what it actually takes to build an AI kill chain, and whether export controls can realistically slow any of it down.

Lawfare Daily: What Do the Russians Actually Think About the War?: Anastasiia Lapatina speaks to Maria Snegovaya about how the Ukrainian attacks on Russian energy sites are shifting attitudes towards the war inside Russia, Snegovaya’s latest report called “Russian Attitudes Are Shifting as the War’s Effects Come Home”, and more.

Rational Security: The “Hip to Be a Square State” Edition: Scott Anderson sits down with Daniel Byman, Julia Curlee, and Ariane Tabatabai to talk through the week’s big national security news stories, including the collapse of the ceasefire with Iran, announcements from last week’s NATO summit, and growing tensions in the U.S.-Israel relationship.

Lawfare Daily: Consent in the Age of AI: Renée DiResta speaks with Kate Klonick and Elissa Redmiles about the difficulties of preventing AI systems from generating exploitative content. They consider technical approaches to prevention and discuss the cultural and legal concerns that the different approaches raise regarding privacy, consent, and responsibility.

Scaling Laws: Nils Tracy of Authorship: Nils Tracy joins Kevin Frazier to talk about the challenges and solutions with deploying AI in highly sensitive contexts. The two discuss how Tracy’s company, Authorship, enables users to map their AI policies to myriad use cases—from on-device AI tools to those connected to the cloud.

Videos

On Lawfare Live, I sit down with Michael Feinberg, Molly Roberts, and DiResta to unpack President Trump’s primetime address on July 16 about election security.

Also on Lawfare Live, I join Anna Bower, Roberts, Eric Columbus, and Roger Parloff to discuss the Justice Department sending grand jury subpoenas to New York Times reporters, the White House firing the court-appointed U.S. attorney in Washington state, and more.

Today’s #BeastOfTheDay is the goose, seen here undeterred:

In honor of today’s Beast, be undeterred and undeterrable.

Tell Me Something Interesting

Several days ago, I——was contacted by an acquaintance of mine. She was researching the question of what exactly constituted formal table manners in Persia circa 1810. Why? Why not. In the process, she discovered a 1775 work entitled Travels in Asia Minor: or, an Account of a Tour made at the expense of the Society of Dilettanti, and, because she is a kind and responsible person, she immediately brought the matter to my attention.

Now, being informed of the existence of this document presented me with a rare phenomenon: a historical question for which all possible answers are funny. Consider the potential explanations for a Society of Dilettanti.

1. Some people decided to get together and form a Society dedicated to dilettantism. This society then sponsored at least one serious research expedition.

2. Some people formed a Society of Dilettanti before the word “dilettante” gained its modern meaning. At some point thereafter, the word “dilettante” came to mean “superficial and affected dabbler,” leaving the poor Society of Dilettanti in the linguistic lurch.

3. Some people formed a Society of Dilettanti before the word “dilettante” gained its modern meaning. The word then entered the English vocabulary with reference specifically to the members of this Society.

Because sometimes history really is here for our amusement, all three possibilities are at least partially correct.

Read more

Subscribe now

Monday on #DogShirtTV, the estimable and I welcomed the estimable for a celebration of the life of Lindsey Graham:

On the topic of obituaries that really do justice to the genre, estimable Greek Chorus member directs our attention to Hugh Massingberd, former obituaries editor of the Daily Telegraph. The New York Times obituary for Massingberd includes some examples of his work:

[D]uring Mr. Massingberd’s tenure, observers widely agreed, every obit in the paper bore his droll, distinctive stamp. Naturally, he covered the presidents, kings and captains of industry who are the grist of obit pages everywhere. But Mr. Massingberd also sought out eccentrics; having the good fortune to live in Britain, he found them.

One Daily Telegraph obituary, from 1991, opened this way: “The Third Lord Moynihan, who has died in Manila, aged 55, provided through his character and career ample ammunition for critics of the hereditary principle. His chief occupations were bongo drummer, confidence trickster, brothel-keeper, drug-smuggler and police informer.”

Another, from 1988, memorialized Peter Langan, a London restaurateur: “Often he would pass out amid the cutlery before doing any damage, but occasionally he would cruise menacingly beneath the tables, biting unwary customers’ ankles.”

The very best of Massingberd’s obituaries are apparently collected in this book. I may have to acquire a copy.

Today on #DogShirtTV, the estimable Mike Feinberg and I discussed Hannah Arendt, and the estimable discussed the distinctive political category of white evangelicals:

Recently On Lawfare

Compiled by the estimable Sarah Willrich

The Purge: How the White House Broke the NSC and How to Fix It

Julia Curlee explains why firings at the National Security Council (NSC) are critical to understanding the foreign policy of the second Trump administration. Curlee describes how the administration’s purge of career staff dismantled the NSC’s ability to coordinate decision-making between different agencies, clearing the way for a foreign policy that moves faster, considers fewer options, and relies on political loyalty rather than expertise.

NSC staff require a specific kind of knowledge and rigor that political loyalty cannot supply. A director managing the Iran portfolio needs to know how the CIA structures a covert action finding, how the Pentagon builds a target package, and how the Treasury Department writes sanctions designations to survive judicial review. That working knowledge exists in no manual or organization chart, takes 12 to 18 months to acquire on the job, and rests on competencies built over careers. This expertise determines whether a president’s policy direction turns into executable options or aspirational statements agencies cannot implement.

These competencies exist in a person—and when the person is walked off the compound, they walk out with them.

The U.S.-Iran MOU and the Future of Maritime Shipping Through the Strait of Hormuz

Liron Libman analyzes the U.S.’s memorandum of understanding with Iran and its implications for the Strait of Hormuz. The memorandum, Libman argues, offers little clarity about what Iran (or the U.S.) can and cannot do, perpetuating an ambiguity that endangers traffic through the Strait of Hormuz and freedom of navigation more broadly.

To be sure, the memorandum appears to impose one limitation: Any future arrangement must be “in line with applicable international law,” but given the well-known disagreement between Iran and most of the world over what international law requires in the Strait of Hormuz, this formulation does little more than kick the can down the road.

U.S. v. Ullah and the Problem With the ‘Lone Wolf’

Peter Beck, Troy Edwards, Alexandra Hughes, and Barry Jonas analyze the U.S. Court of Appeals for the Second Circuit’s decision in U.S. v. Ullah, which held that Akayed Ullah, who attempted a suicide bombing inspired by ISIS in 2017, could not be charged with providing material support to a terrorist organization. The authors posit that the decision substantially undermines a critical instrument the Justice Department needs to prosecute lone wolf actors, which is particularly concerning as terrorist groups increasingly turn to digital channels to inspire recruits.

The decision has significant implications for future material support investigations and prosecutions. Many terrorists do not directly communicate with FTOs. Far from being “lone wolves,” though, they facilitate or, worse, directly carry out violent acts following instructions FTOs provide, spreading FTO symbols and messages, and doing it all in the FTOs’ names. After Ullah, absent another federal offense, investigators and prosecutors may have their hands tied until these kinds of subjects either attempt to, or succeed in, making direct contact with a terrorist group, commit another violation, or commit the attack itself. This may result in longer, riskier investigations and will require law enforcement to have sufficient investigative tripwires in place to know when subjects’ conduct establishes an arrestable offense. This article discusses the Ullah prosecution, the material support statute designed to stop terrorists earlier, and the potential ramifications of this decision for future counterterrorism cases.

GAO Goes After State (Again) for Treaty Practice

Duncan B. Hollis unpacks the Government Accountability Office’s recent report finding that the State Department significantly delayed reporting U.S. international agreements to Congress and the public. Hollis explains the State Department’s failure to meet its Case Act reporting requirements is due to systemic issues, such as an out-of-date website and high administrative burdens, but emphasizes that despite these problems, such reporting is critical to maintaining transparency in U.S. foreign policy.

I’m not sure the State Department has earned the benefit of the doubt given the decades of difficulties it has had in meeting the Case-Zablocki Act’s requirements. If the need to get agencies to confirm their own compliance is novel, history suggests such novelty is warranted. Short of funding delays, there are insufficient incentives to report treaties. The reality is that careers get made in negotiating and concluding agreements, but reporting them to Congress is a thankless, entirely administrative task.

Podcasts

Lawfare Daily: The Trials of the Trump Administration, Jul 10: I sit down with Eric Columbus, Anna Bower, Molly Roberts, and Roger Parloff to discuss the Justice Department settling a second suit with Michael Flynn, developments in the E. Jean Carroll litigation, the D.C. Circuit denying a stay pending appeal of the order to take Trump’s name off the Kennedy Center, and more.

Lawfare Daily: Taking Stock of the Ukraine-Russia Talks: Mykhailo Soldatenko sits down with Eric Ciaramella and Samuel Charap to take stock of the U.S.-led negotiations between Ukraine and Russia. They discuss the improvements in Ukraine’s position, the structure of the negotiations, territorial questions, and security commitments to Ukraine.

Scaling Laws: New Paths to Social Mobility with Joe Seddon: Joe Seddon joins Kevin Frazier to discuss artificial intelligence (AI), social mobility, higher education, and the future of opportunity. They explore whether AI can scale the kinds of guidance and tutoring that have historically been available only to students with strong networks, the role of universities in preparing students for an AI-transformed labor market, and whether AI will disrupt entry-level white-collar work.

Today’s #BeastOfTheDay is the mouse, seen here nibbling on the body of Christ in c. 1185:

I am not going to tell you how to honor today’s Beast. Today’s Beast deserves no honor. Today’s Beast has taken up residence in my kitchen, and it can honor me by departing before I kill it.

Subscribe now

Read more