Iran After Khamenei & Pentagon vs. Anthropic (with Karim Sadjadpour)

Summary

THE LONG GAME WITH JAKE SULLIVAN & JOHN FINER
Episode: Iran After Khamenei & Pentagon vs. Anthropic (with Karim Sadjadpour)
Date: March 5, 2026
Guest: Karim Sadjadpour (Senior Fellow, Carnegie Endowment)

---

OVERVIEW

This episode tackles two enormous subjects:

1. The aftermath and implications of the death of Iran’s Supreme Leader, Ayatollah Khamenei, focusing on power transitions, regime survival, and the prospects for democracy or further authoritarianism within Iran.
2. The Pentagon’s contentious standoff with the AI company Anthropic, paralleled by OpenAI’s entry into U.S. defense contracting, raising urgent questions about public-private technology partnerships, AI in warfare, and the boundaries companies can set for their government customers.

Throughout, Jake Sullivan and Jon Finer leverage guest expertise from Karim Sadjadpour, one of the most respected Western analysts of Iran, to disentangle the fraught, dynamic situation in Tehran, connect it to U.S. and Israeli policymaking, and reflect on broader dilemmas facing American power in the age of both revolutionary change and revolutionary technology.

---

SECTION 1: THE IRANIAN POWER VACUUM POST-KHAMENEI

Karim Sadjadpour’s Personal Perspective

• Opening Reflection: Karim separates his hopes as an Iranian-American from empirical reality, citing data that the vast majority of authoritarian regime transitions yield a successor authoritarian regime—especially when violence sparks the change.
• Quote:
  “What I've learned over the years on Iran…is that around four out of five authoritarian transitions lead to another form of authoritarian government. And when that transition is triggered by either internal or external violence, the likelihood of a democratic outcome is much, much lower.” (03:18)

Khamenei’s Worldview—Resistance Over Compromise

• Khamenei typified the “hedgehog” (one big idea) approach (06:10). His regime was predicated on “resistance” to U.S. and Israeli pressure, preferring intransigence, even under threat, to any hint of concession.
• Khamenei wanted enmity with the U.S. for internal legitimacy—a strategy openly admitted by former President Khatami:
  “We need enmity with the United States, the revolution needs enmity with the United States.” (09:45)

Why No Compromise, Even at the Cost of His Life?

• Khamenei’s “survival instincts” clashed with his ideology of resistance. His failure of operational security—attending meetings despite open threats—may have reflected an acceptance or embrace of martyrdom in the twilight of his rule (09:07–09:45).

---

SECTION 2: IRAN’S FUTURE—STRONGMEN, NATIONALISM, AND THE PROSPECTS FOR DEMOCRACY

What Comes Next? Archetypes and Power Factions

• Short-term: The regime, battered but not broken, is in “existential fighting mode.”

• Medium-term scenarios (12:05): Iran’s post-Khamenei arc could resemble Soviet Russia, China, North Korea, Pakistan, or Turkey.

• Karim’s View:

  • The next strongman likely won’t be a cleric.
  • Nationalism, possibly “pluralistic Iranian nationalism” that accommodates multiple ethnicities, is poised to grow as the dominant political narrative—potentially for good or ill.

• Quote:
  “The winning argument in Iran in future years is going to be nationalism... you can have the negative nationalism of Vladimir Putin… or a positive patriotism…” (12:51)

Will Security Forces Take Over? Where Is the Democracy Movement?

• Revolutionary Guards are not monolithic; senior insiders are “revolutionaries” but many in the rank-and-file prefer pragmatism (17:22).
• Violence by security forces has splintered hopes for quick democratic change. Most Iranians long for a “normal life”—opportunity and dignity, not overt democracy per se (19:22).
• “It’s not to say that people don’t want democracy… what people say increasingly in Iran is we want a normal life [zendegiye normal]...” (19:40)

Public Sentiment Amid War and Repression

• As many as 80–85% of Iranians oppose the regime; the populace is “furious,” “shell-shocked,” and reeling from both government violence and collateral damage (21:27).
• Many wanted U.S. intervention, “a magic bullet only hurting oppressors,” but the realities of military escalation (e.g., bombing of a girls’ school with 200 casualties) shattered these hopes (22:30).
• “There hasn't been any winners here. The regime is obviously flailing. The Iranian society is in shell shock. Hasn't been good for America… lose, lose, lose all around.” (47:20)

---

SECTION 3: U.S. & ISRAELI STRATEGY—ARE THEIR INTERESTS ALIGNED?

Supporting Ethnic Militias: Dangers of “Playing with Fire”

• Speculation is rife that the U.S. or Israel may arm Kurdish or separatist factions within Iran to destabilize the regime. Sadjadpour strongly cautions against this (“opposite of sound,” 27:59):
  “If it's perceived that the U.S. Government or any outside power is... trying to factionalize the country, that is going to badly backfire…” (29:32)

Shared Goals, Diverging Tactics

• U.S. and Israeli ideal outcomes are similar (pre-1979-style Iran neutral or friendly to the West)—but Israel could accept protracted chaos and targeted assassinations, while the U.S. must weigh regional stability (34:17).
• Iran is seeking to “regionalize” the conflict, using attacks on Gulf States to pressure U.S. public opinion (35:25).
• “Israel may want to continue this campaign of targeted assassinations indefinitely… an Israeli prime minister is more willing to live with [the fallout] than an American president.” (35:07)

---

SECTION 4: SUCCESSION & THE FUTURE OF IRANIAN LEADERSHIP

The Rise (and Likely Short Reign) of Mojtaba Khamenei

• Mojtaba Khamenei, the Ayatollah’s son, is reclusive and even more hardline; likely interim, not transformational (39:56).
• As security becomes the regime’s sole pillar, hereditary succession undermines the very revolutionary identity that overthrew the Shah.
• “If Mujtaba succeeds him, he may not rule for 37 days.” (42:36)

Gulf States’ Calculus

• Iran’s regionalization of violence is beginning to unite Gulf neighbors with each other and the U.S., despite efforts to remain neutral (43:38). Pakistan’s nuclear defense pact with Saudi Arabia heightens the stakes (46:14).

What Should U.S. Policy Be?

• Sadjadpour advocates clarity and strategy, not improvisation:
  “It's absolutely critical for United States and the US President to be clear in his own mind, okay, this is what we're actually trying to achieve. This is the viable objective we have. And then how do we fulfill that objective?” (47:34)

---

SECTION 5: THE PENTAGON VS. ANTHROPIC—WAR, AI, AND GOVERNANCE

Contract Breakdown: Mass Surveillance & Lethal Autonomy

• Pentagon cancels Anthropic’s contract over disagreements on limits, notably on fully autonomous weapons and mass surveillance (55:20). OpenAI steps in and wins favor by (initially) accepting Pentagon’s terms.
• The Pentagon threatens to designate Anthropic a “supply chain risk,” an unprecedented and likely illegal move (60:57).
• “The threat by the Pentagon to impose a supply chain risk designation on Anthropic is outrageous. This is a patriotic American company. Ironically… they're the most forward leaning on these issues.” (59:14)

The Precedent and the Principle

• Anthropic’s principled objection is commended—yet the hosts worry about the risks if tech CEOs can impose their own policy priorities in warfare contracts (63:00).
• OpenAI, under pressure, later partially amends the contract with new guardrails against domestic surveillance (63:34).
• “In this particular case I'm very sympathetic to Anthropic… but the universalizing of this case makes me a little bit more nervous. I think the right answer… should be a responsible government... not a technology CEO.” (62:42)

Wider Implications

• The dispute weakens the U.S. in strategic technology sectors:
  “It is not comforting… to have a precedent where the government essentially tries to destroy a company… in a totally disingenuous way.” (66:36)
• Unlike previous defense industries, the government lacks a deep bench of expertise on LLMs, heightening the role and leverage of private innovators (68:37).
• U.S. government has already begun phasing out Anthropic—increasing strategic vulnerability and undermining trust (71:00).

---

MEMORABLE QUOTES & MOMENTS

• “There are few nations in the world with whom the United States has more common interests and less reason to quarrel than Iran.” – Henry Kissinger, quoted by Karim Sadjadpour (31:13)
• “If this is the country I'm not even saying a democratic government, but a government who prioritizes its own national interests, the US and Iran should be natural partners.” – Sadjadpour (32:04)
• “In some ways, I think probably for your very intelligent listeners, those archetypes are pretty self explanatory.” – Sadjadpour (13:07)
• “Almost certainly illegal as well… The Pentagon probably knows that, which is why they haven’t actually imposed it." – Sullivan, on the supply chain risk threat (61:10)
• “We've reached the bite off our nose to spite our face portion of the programming here.” – Sullivan, on U.S. agencies dropping Anthropic (71:37)

---

TIMESTAMPS FOR KEY SEGMENTS

• Khamenei’s Succession & Regime Futures: 00:58–25:02
• Democracy, Nationalism & the Public’s Mood: 16:19–25:02
• US/Israeli Policy Dilemmas & Ethnic Faction Armament: 27:11–35:25
• Who Succeeds Khamenei? Mojtaba & Regime Cohesion: 39:33–43:38
• Gulf States' Position & Regionalization of Conflict: 43:38–46:36
• What Should US Policy Be? 46:36–49:20
• Pentagon vs. Anthropic & AI, National Security: 55:20–72:00

---

CONCLUDING REFLECTIONS

The hosts and guest converge on a note of frustration, even sorrow:
Despite dramatic upheaval in Iran, every side appears to be a loser so far—regime, people, America, the region—due to both structural traps and muddled strategy. In parallel, the U.S.'s handling of breakthrough technologies risks shooting itself in the foot even as great power competition intensifies.

As Sadjadpour puts it:
“There hasn't been any winners here... lose, lose, lose all around.” (47:20)

---

FOR FURTHER READING

• Sadjadpour’s Foreign Affairs essay (referenced at 12:05; see show notes)
• Ongoing reporting on Pentagon–AI company conflict

---

This summary captures all essential insights, character, and memorable turns of phrase. For direct quotes and detailed analysis, consult the episodes’ timestamps.

Transcript

[Jake Sullivan]

AI can Fix healthcare. I'm Henry Blodgett and this week on my show Solutions, I had a fascinating conversation with Dr. Bob Wachter, author of A Giant How AI is Transforming Healthcare and what It Means for our future. Dr. Wachter was not expecting to be an AI optimist. What convinced him? Follow Solutions with Henry Blodgett wherever you get your podcasts to hear more. Until now, mobile phone companies have worked

[Kareem Sagapour]

very hard to ensure their phones do not start fires. But we found one company that dared

[Jake Sullivan]

to go in a different direction and

[Kareem Sagapour]

make a phone with fire. Starting as a feature on the Vergecast,

[Jake Sullivan]

we talk about all the greatest and

[Kareem Sagapour]

weirdest phone concepts from Mobile World Congress in Barcelona. Plus, after years of legal battles, Google and Epic are now best buddies, contractually obligated to not say mean things about their app stores anymore. That's this week on the Vergecast.

[Jake Sullivan]

We Wherever you get your podcast,

[Kareem Sagapour]

the big question is, if there is going to be a next strongman in Iran, what kind of strongman will that person likely be? It's not likely going to be someone who is wearing a turban. I don't think that there's going to be another powerful cleric supreme leader. I do think that the winning argument in Iran in future years is going to be nationalism.

[Jake Sullivan]

Welcome back to the Long Game. I'm Jake Sullivan.

[John Finer]

John and I'm John Finer. Today we are grateful to be joined by Kareem Sagapour, who is an expert on Iran, a senior fellow at the Carnegie Endowment, and a contributor to the Atlantic magazine. He's also someone who we have both known, Jake and I, and learned from for more than 15 years, which is about the same amount of time that we've been wrestling with the vexing issue of Iran. And there's no one whose views we more often seek out on this confounding country and area of policy. And we're not the only ones. Part because of both his deep analytical expertise and also his gift for memorable turns of phrase. Just one that always sticks with me, Kareem, which is you once called Iranian elections unfree, unfair, and unpredictable. Karim's wisdom is sought out by journalists, by TV bookers, by conference organizers, and editors of magazines, journals, and op ed pages. So candidly, we're very lucky that he made time for us. And Karim, welcome to the Long Game.

[Kareem Sagapour]

Thank you so much to both of you. You know, there's a Persian expression when you want to give someone compliments and kind of puff them up. You're putting watermelons under their arms So I appreciate those watermelons. Thank you, John.

[Jake Sullivan]

We're going to have to borrow that in the future.

[Kareem Sagapour]

I can tell you.

[Jake Sullivan]

I never put watermelons under John Finer's arms. Ever.

[John Finer]

Not even once. Not even oranges?

[Kareem Sagapour]

Maybe some. Yeah, yeah, exactly.

[Jake Sullivan]

So, Karim, before we dig into the substance, and there's a lot of substance to cover, I wanted to start with a personal perspective. You've been anticipating, weighing the prospects of considering the implications of dramatic events like these for much of your career. So could you just talk a little about what it's like on a personal level in this moment to see such momentous change come so quickly after so much time. And how are you personally processing this, Jake?

[Kareem Sagapour]

I'm a University of Michigan guy, and every year during March Madness, I fill out two brackets, one with Michigan winning it all and one with what I think might happen. What I've learned over the years on Iran, and I've learned this the hard way because oftentimes I'm of Iranian heritage. So I would love to say I see Iran evolving into Denmark or Norway, but I try to not conflate my own hopes and analysis in the case of the Islamic Republic. I was doing some research last summer in the aftermath of the 12 Day War. In anticipation of this essay, I was writing in Foreign affairs, and there was a statistic that jumped out at me, which is that around four out of five authoritarian transitions lead to another form of authoritarian government. And when that transition is triggered by either internal or external violence, the likelihood of a democratic outcome is much, much lower. And so I started with the premise that I believe, on one hand, this is a very unique country. It has a very rich history and civilization. It has the human capital and the natural resources to be a G20. And I think it is a society which is ripe for representative government. But empirically, we also know that that is not usually the outcome after five decades of authoritarianism. And so I think that the way this authoritarian transition happened with the United States or Israel assassinating Ayatollah Khamenei, I think at least in the very near term, made it even less likely that the country is going to. To peacefully transition to stable, representative government. Because as both of you have experienced throughout your careers, when you introduce a power vacuum into a society, it's not the intellectuals and the poets that prevail in that kind of atmosphere. It's men who are able and knowledgeable about mobilizing violence.

[John Finer]

Kareem, there's been one question I've really wanted to ask you since all of this started to go down. I mean, you're somebody who's studied, written about, probably thought more than you would have liked about the Supreme Leader. Now late Supreme Leader Ayatollah Khamenei. What is your best explanation as to why, even though he, I think, felt like it was probably likely that he could face incredible violence, was unwilling in the end to make a deal? He had done that once before. He had done that famously also to end the Iran Iraq war in the 1980s. And then relatedly, maybe why did he not bother to hide himself better, given everybody on the planet was anticipating that there could be military action directed at him personally.

[Kareem Sagapour]

So there's that wonderful Isaiah Berlin essay, which I know you're both familiar with the Hedgehog and the Fox. The fox knows many things and the hedgehog knows one big thing. For me, there was no more prototypical hedgehog in the geopolitical landscape than Ayatollah Khamenei, basically, as had one big idea over the last four decades, and that's been resistance against the United States, resistance against Israel. And there was a study I did when I first joined Carnegie. I spent about six months reading all of his major speeches over his 34 decade career. And what I found remarkable in that is that there was essentially, it was essentially variation of one speech that he'd consistently given, and that is obviously that concept of resistance, but also this view that when you're under pressure, whether internal or external, if you take a step backward in the hopes of alleviating that pressure, that's actually going to project weakness and embolden your adversary. So I think that on one hand, John, this is someone who was arguably the longest serving dictator in the world. He served for 37 years, and you don't get that title if you're a reckless gambler. So on one hand, he had these great survival instincts, and I think those survival instincts were in tension with those resistance instincts. And I think in the case of Donald Trump, he probably felt that over the last decade or so that he's been dealing with Trump. But in particular after the aftermath of Trump pulling out of JCPOA, Trump killing Qassem Soleimani in 2020, and then last summer's Operation Midnight Hammer, he probably felt that he had not and Iran had not really exacted sufficient cost to Trump for those actions. And Trump viewed Iran as a paper tiger. And for that reason, I think he was more in resistance mode. On your question of why they were so complacent about operational security, I don't have a Great answer to that question. Because last summer, the bombing started just a couple days before another round of negotiations was meant to happen. So they knew that there was a likelihood of military action. And they knew. If I knew, they, I'm sure, knew that Ayatollah Khamenei was thought to be a target. It may be that they thought they won't do it in broad daylight. But that was an enormous, enormous failure of operational security. And part of me thinks maybe at age 86, that's how he also wanted to go out as a martyr.

[Jake Sullivan]

I was going to ask you that, Karim, because it just struck me like you have the biggest buildup in the region by the American military since the Iraq war in 2003. You have the US telegraphing that military action was likely. You have the possibility of a direct strike on the Supreme Leader plastered all over the newspapers of the world. And the guy just shows up to a meeting. So is it that in fact, he thought, you know what? Come what may, I'm going to put myself out there, and if I'm in his view, if I'm martyred, I'm martyred. You think that that's at least a possibility?

[Kareem Sagapour]

I hope one day. It's been two decades that I've not been able to go to Iran, but I hope one day I can get to Iran and look at the archives of these typ of major decisions. What on earth were they thinking? Back to Khamenei's worldview. So when I did the study years ago, I'd kind of reached this conclusion that Ayatollah Khamenei, we, the United States, are never, ever going to be able to reach a modus vivendi with him. It may be possible we can kind of get a tactical nuclear compromise, tactical deal. But he essentially wanted to have the United States as an adversary for his own internal legitimacy. Not longer after I concluded that study, I was invited to a meeting in Oslo. And the former President of Iran, Mohammed Khatami, was invited there as the keynote speaker. And they sat me next to Khatami at dinner. And Khatami, in his keynote speech, said, off the record to a group of people, he said, there are those in both capitals, both Washington and Tehran, who don't want to normalize relations because it's not in their personal interest to do so. And after his speech, I went back to that point And I said, Mr. Khatameh, whom are you referring to in Tehran when you say it's not in their personal interest to restore relations? And I was shocked he said to me, said, when I was president, Mr. Khamenei used to tell me that we need enmity with the United States, the revolution needs enmity with the United States. And that was kind of a conclusion I had reached independently from reading his speeches. But to hear the former President of the Republic say that so candidly was for me, really an eye opening experience.

[John Finer]

Wow.

[Jake Sullivan]

Wow. That is, I think that's the literal definition of saying the quiet part out loud. So the Ayatollah is dead, which is a remarkable sentence to utter. You've said you don't think that this necessarily leads to the outbreak of Jefferson and Jeffersonian democracy in Iran. Where do you see this going? What does happen with this regime? And what are you looking at in terms of the factors that will dictate whether it takes one course or another?

[Kareem Sagapour]

So I think, Jake, we can talk about the immediate term and then let's say the medium term, three, four, five years out. In the immediate term, this is a regime fighting for its life, both internally and externally. What I found maybe three, four days into the war is that it's a regime which it can fight. It's proven it can continue fighting, it can kill, it can kill its own people, but it really can't govern. And so it's really in this existential fighting mode at the moment. I think President Trump, in his head, was hoping for Venezuela redux, where you take out the top guy and then you do a swift deal with his successor. In the case of Venezuela, Delcey Rodriguez, that hasn't happened at least yet. In the Iranian context, there's been 40 years of this ideological culture, revolutionary ideology of death to America, death to Israel. And at least in the first four days of this war, we've seen continuity. Even though numerous senior officials have been assassinated at the moment, there's no one who's emerged who has the legitimacy, the authority and the will to make a break with that 47 year culture. And in a way, it's almost like the system is more prepared to break than they are to bend. And their future really in some ways depends on how far the United States and Israel are willing to go on this campaign. And we can talk about what their strategy is right now to counter that. But in the medium term, let's say five years out, the foreign affairs I wrote, I focused on five possibilities and as a shorthand, I'll just list them here. One is Iran as the Soviet Union and then post Soviet Russia. A second is Iran as China. A third is Iran as North Korea. Fourth is Iran as Pakistan and fifth is Iran as Turkey. And in some ways, I think probably for your very intelligent listeners, those, those archetypes are pretty self explanatory. And in some ways, I don't know if you guys agree with Kissinger's observation. He said when he was a professor at Harvard, he thought that history was driven by impersonal forces. And after he served in government many years, he reached the opposite conclusion, that the individual makes a profound difference in history. And I'm one of those that I know academics don't like the great man theory of history, but I think about this as kind of almost biographical archetypes in the Iranian context. And that this is a country which over the last century has really been only ruled by four men, Reza Shah, Mohammad Reza Shah, Ayatollah Khomeini and Ayatollah Khamenei. And so the big question is who, if there is going to be a next strongman in Iran, what kind of strongman will that person likely be? In my view, it's not likely going to be someone who is wearing a turban, even though they may choose Mojitava Khamenei or another cleric as the next supreme leader. I don't think that there's going to be another powerful cleric supreme leader. I do think that the winning argument in Iran in future years is going to be nationalism. Now, nationalism can take many different forms. You can have the negative nationalism of Vladimir Putin. You can have a positive patriotism of I'm trying to think of, well, you could argue like Mohammed bin Salman or Mohammed bin Zayed in the uae. I think that's obviously an outcome that many in Iran would prefer. This process from where we are now to where the country could go in the coming years is going to be fraught with enormous drama. And obviously it's not Iranians themselves who are shaping their current history.

[John Finer]

Not to try to pin you down too much or to make you explain the entirety of your excellent essay, we will link that for our listeners and viewers. But I did want to ask is what you're saying basically that we are not likely to have clerical leadership in Iran, suggestive that we might have the security forces take charge? They have not shown much indication that they are cracking, at least not yet. Maybe their ability to fight back and launch missiles and rockets continues, maybe diminished over time. Does that suggest to you that they are likely to take over? And I just want to also get you to comment on the prospects you seem to be almost ruling out democratic transition. Where does that leave the democracy movement that kind of kicked all of this off earlier this year that led President Trump to actually first pledge to intervene militarily in Iran. Does that leave them sort of in the wilderness at this point?

[Kareem Sagapour]

Well, let me start by saying that I really hope I'm wrong and that the country will transition to secular, Jeffersonian, accountable democracy. That would be wonderful. The issue, as I said earlier, is that especially when this transition is triggered by external violence, there's a moment now for these security forces who think it's kill or be killed. They rightly believe it's killer be killed. And so for that reason, they're willing to inflict enormous violence on their citizens. And I do think that the internal battles within the Revolutionary Guards, within the security forces are going to be determinative for Iran. We use the term Revolutionary Guards often. And I want to emphasize for your listeners this. They are not a monolith. There are 150,000 men with somewhat different worldviews. I think the very senior leadership because of the fact that they were hand chosen by Ayatollah Khamenei and he was very careful about shuffling them with some frequency to prevent them from getting too comfortable and attaining their own power base. Those folks, I would say, are revolutionaries. They still believe in the ethos of what I call Vision 1979. I think a lot of the rank and file folks would actually prefer to prioritize economic and national interests before revolutionary ideology. And so I think that that is likely going to be a battle that is going to be with us for some time. I think the latter argument is if you're someone who's trying to appeal to popular aspirations, appealing to people's aspirations is more popular than continuing to try to appeal to grievances. But that's a fight which I think is going to be very important for Iran's future. On the democracy front, what's interesting is that in the couple decades that I've been following Iran, it's not to say that people don't want democracy. Absolutely. I think they want political freedoms. But what I find interesting, especially in the aftermath of both our failures in Iraq and Afghanistan and the failures of the Arab Spring, is that I think the model people have in their heads for what constitutes a good life are places that are palpable to them because they visit often. Places like the uae. Turkey is a place which is, I would argue, kind of the closest cousin to Iran. It's a large non Arab Muslim country with a very proud history. And what people say increasingly in Iran is we want a Normal life, Zendegia normal. And that is like life that gives you economic dignity, economic opportunities, social freedoms. You can watch what you want, go out with your boyfriend, girlfriend, if you want to drink alcohol. You don't have a police state micromanaging every aspect of your life. But it's not the same way. You may have heard, I remember when I was first beginning, you would hear the word democracy two decades ago, in a way more than you hear right now.

[Jake Sullivan]

So, Kareem, those people who went out onto the streets who were so ruthlessly gunned down in just these incredible and disturbing numbers, and their compatriots who are now sitting at home kind of looking at all of this unfold, what are the Iranian people? Obviously, they're not monolithic, but broadly speaking, what are the Iranian people thinking about this bombing campaign? How are they reacting to it, and where does it leave them in terms of whether they go back out into the streets at some point here? Do you have a view on that?

[Kareem Sagapour]

So it's always very difficult to do polling on authoritarian regimes. But my estimate, based on there is one poll coming out of a group in Europe called Gammon, which has done pretty good polling, and then just kind of based on under indicators, including just anecdotal evidence, is that probably about 80 to 85% of Iranians are opposed to this regime. And I think their popularity is so low because of the fact that they are brutally politically authoritarian, they're socially authoritarian, and they're doing a terrible job managing the country's economy. And they're also a theocracy which rules from a moral pedestal. And that is particularly insulting to people when you're elbow deep in repression and corruption. But you have these moral pretensions of ruling, of carrying out God's will. And so for that reason, it's a society which is really furious, and they have a deep loathing for their government. And I think at the moment, Jake, they are shell shocked. They're shell shocked at the violence of January of 2026. People I spoke to said it was as if they were in a war zone. The regime was just opening fire on unarmed civilians with weapons of war. And for that reason, I think that many of those, again, there's no scientific polling, but based on talking to people, based on everything I read, I would argue of the 80% of society that opposes the regime, it seemed like a majority of those people did want President Trump to make good on his promise to help. As you guys know, he said on nine occasions that he incited people to the streets and said, help is on the way. Go seize your institutions. And I think a majority of people did want him to make good on that promise. And as you guys both know, there was a former general friend of mine who served in Iraq and he said all societies living under tyranny understandably want a magic bullet which is only going to hurt their oppressors and not cause any civilian damage or harm innocent people. And of course, that war never works out like that. And so four or five days into this, I think that some of the scenes are horrific. There was a girls school in southern Iran that was inadvertently bombed, killing almost 200 young schoolgirls. I have to say, I don't have a sense right now. What does the average person think? Do they want this campaign to continue? Do they want the campaign aborted? I think they are in a very terrible position right now and my heart goes out to Iranians.

[Kara Swisher]

Hey, Kara Swisher here. I want to let you know that Vox Media is returning to south by Southwest in Austin for live tapings of your favorite podcasts. Join us from March 13th through the 15th for live tapings of Today Explained Teffy Talks, Prof. G Markets, and of course, your two favorite podcasts, Pivot and On with Kara Swisher. The stage will also feature sessions from Brene Brown and Adam Grant, Marques Brownlee, Keith Lee, Vivian Tew and Robin Arzon. It's all part of the Vox Media Podcast stage at south by Southwest, presented by Odoo. Visit voxmedia.comsxsw to pre register and get your special discount on your innovation badge. That's voxmedia.com SXSW really, you should register. We sell out and we hope to see you there.

[Jake Sullivan]

This is what President Trump had to say about why the United States is at war with Iran.

[John Finer]

We sought repeatedly to make a deal. We tried. They wanted to do it, they didn't want to do it again. They wanted to do it, they didn't want to do it. They didn't know what was happening.

[Jake Sullivan]

Not the best explanation for a war of choice, sir. I'm personally a do my own research kind of guy, but let's ask AI why we're at war with Iran. Chat the United States attacked Iran in 2026 because it claimed Iran posed an imminent threat, particularly due to Iran's advancing nuclear program and missile capabilities, and aimed to reduce Iran's ability to project power in the region.

[Kara Swisher]

Wow.

[Jake Sullivan]

That was a better explanation. Thanks, Chat. Fitting that AI was more clear than the President of the United States because it turns out the United States is using AI to fight the war in Iran. The future of war is AI, and that future is now. Here you can find out whether or not you should be freaking out over in the Today explained feedback.

[John Finer]

So you alluded earlier to US Strategy, Israel's strategy at this point, which does also seem to sort of evolve almost by the day or even within the day by the answer from different senior officials. One topic about which there's now some reporting and a lot of speculation is the possibility that armed groups outside Iran, like Iranian Kurdish factions based in Iraq, possibly armed by the United States or Israel, could enter the fray or alongside perhaps separatists inside Iran like the Baluch, taking up arms against the regime. I'm wondering what you make of these reports. Would this be a sound strategy and how would this likely play out? If this is what happens, my view

[Kareem Sagapour]

would be the opposite of sound. This is really playing with fire to start to think about arming different factions inside Iran. This is a country which is different than a lot of the other countries we see in the Middle east that were post World War I creations or post Ottoman creations. This is a country with over 2000 years of history as a nation state. In my view, there is a cohesive national identity. It's true that only around 50% of Iranians are Persians. And then you have perhaps a quarter of the country who are Azeri. And then you alluded to, John, the Kurds, who constitute about 10%. And the Kurds, that's the news reports of arming Iranian Kurds. And listen, I love the Kurdish community in Iran. They're wonderful people. They have been disproportionately repressed by not only this regime, but they've suffered disproportionate repression even before 1979. And they have very legitimate and profound grievances. The challenge is, in my view, that the antidote to this regime, the antidote to kind of Shiite revolutionary ideology, is, I would argue, a form of nationalism, patriotism in its best form. In my view, that nationalism would be best if it were pluralistic. So it's not an ethnic nationalism. It's a pluralistic Iranian nationalism that encompasses the country's various ethnic groups, Persian, Azeri, Baluch, Kurdish, Turkoman, Arab, et cetera. But if it's perceived that the US Government or any outside power is playing around with the country's territorial integrity or trying to factionalize the country, that is going to badly backfire because it's something which people feel very sensitive about. And there is a danger that even the opponents of the regime will grudgingly either cease protesting or perhaps even, I don't want to say side with the regime, but it's such a profound issue for people that if they feel the goal of this entire effort is to break Iran up, that's something that most people don't want any part of.

[Jake Sullivan]

Karim, you've said this is the opposite of sound. To pursue a strategy, to pursue arm up a bunch of ethnic factions. Do you think that the Israelis and the US are serious about this? Does this concern you as a real possibility that you actually could see ethnic militias of one flavor or another actually trying to bring pressure on the regime, or do you think this is more speculative?

[Kareem Sagapour]

I can't tell, Jake. Based on talking to people here and based on the reporting, I'd say a couple of things. Number one, if you're Israel and this is your worst adversary in the world, they've been supporting these proxies against you, these militias against you for many decades. You may say, I don't care if this causes havoc inside Iran or even state collapse because we want or it's better for us to have this weakened adversary. That may be your calculation if you're Israel, which is frankly different than the calculations of a lot of the neighboring Gulf countries because they do worry about that spillover and refugees and things like that, which Israel wouldn't have to deal with. I think the United States has always recognized that that would be an unwise strategy. Certainly I think you guys probably did when you were both in government, that the goal here is not to cause a civil war in Iran. The goal is for Iran to essentially become a country whose organizing principle is their national interest, not revolutionary ideology. I always quote Henry Kissinger here. This is a verbatim quote from Kissinger. He said, there are few nations in the world with whom the United States has more common interests and less reason to quarrel than Iran. So if this is the country that I'm not even saying a democratic government, but a government who prioritizes its own national interests, the US and Iran should be natural partners. And so I think that's, in my view, long been the goal of successive US governments from Jimmy Carter to the present. And there's so little strategic planning in this current administration that I'm skeptical it's actually a concerted plan to arm ethnic factions to bring about civil strife. It just seems more to me that there improvising as they go along. It's kind of regime change.

[John Finer]

By jazz improvisation Karim, you described a goal that one could have for the future of Iran. I'm wondering if you think the United States and Israel share a common objective at this point. They've obviously worked in lockstep in conducting the military operation up till now. May have even been a high degree of coordination when it came to the diplomacy, some degree of subterfuge. There's at least some reporting that suggests that. But in terms of a future kind of end state for Iran, do you think that that is shared between the US Administration and Prime Minister Netanyahu?

[Kareem Sagapour]

I think if you gave President Trump and Prime Minister Netanyahu, if they wrote on a paper for you independently, their ideal goals vis a vis Iran, they probably would be similar. They would say, we want an Iran, which is, they would probably say similar to how it was pre1979 nationalist government that at that time there was a good partnership between Iran and Israel. The goal of the government was to advance the economic welfare and security of its people. And there was not really a conflict between Iran and Israel. Israel is a tech superpower. Iran is an energy superpower. And so their interests were complementary. And obviously, I think President Trump would like that outcome as well. But absent that perfect option, it may be that the interim options between the two countries are different because I think for Israel, they may want to continue this campaign of targeted assassinations indefinitely. And the fallout from that, which is Iran reacting against Gulf countries attempting to close the Strait of Hormuz, spiking the price of oil, those are all consequences that perhaps an Israeli prime minister is more willing to live with than an American president. And so we'll see in the coming weeks. I think in some ways, the Iranian strategy here is pretty clear, which is what they Learned in the 1980s when Hezbollah, backed by Iran, bombed the US Marine barracks in Beirut. Was that one of the best ways of restraining the regional ambitions of American presidents to impact public opinion in the United States? And they did that also in the aftermath of the 2003 US invasion of Iraq. And I think, in a way, that's what they're trying to do here. When Americans turn on the television sets and they see explosions and smoke everywhere, service members being killed, their gas prices are higher at the pump. Their hope is that President Trump watches that, too, and says, you know what? Let's pull the plug on this. And I think Israel's calculations are different there.

[Jake Sullivan]

Kareem, do you have a view about how this ends at this point? I mean, that's really putting you on the spot, since it's also unpredictable. But what is your take on the most likely endgame for the military action. Given what you just said.

[Kareem Sagapour]

I've often joke with my Georgetown students that you're better off having a psychology degree than a political science degree to assess the futures of Middle East. And that's also true now about Washington because we're essentially trying to get inside the mind of the President. We do know that he doesn't like long drawn out operations. We do know that his base is opposed to, I would say much of the MAGA base is opposed to getting sucked into another Middle east war. I was writing an Atlantic piece earlier today and I was thinking of how virtually every US president since 1979 has had their presidency in some ways consumed by the Iran issue. The hostage crisis basically ended Jimmy Carter's presidency. Iran Contra tainted Reagan's presidency. 911 obviously wasn't Iran, but the aftermath of 9 11, the Iraq war, Iran sabotaged Bush's efforts to democratize Iraq. You guys, the second part of the Obama administration was consumed by JCPO and the bitter partisan fights October 7th to 2023 when you were both in government, Iranian proxy Hamas and then the wars that consumed the period you guys were there. And now Trump, he criticized all previous US Presidents for getting sucked in. And it seems to be happening with him as well. I think there's two big questions. It's the resolve of the president. On any given hour, it seems to change. Sometimes he says he's ready to pursue this for four or five weeks. Other times he says he wants to do a deal.

[John Finer]

Hegseth said eight weeks earlier today.

[Kareem Sagapour]

Exactly. And so they're on one hand signaling that kind of resolve. But that could change by tomorrow. And for the Iranians, it's okay, how long can they endure this? How cohesive are the security forces? How many missiles do they have left? That's going to be the question for them. And who is going to emerge in Iran? Who is going to start to fill this enormous vacuum left by the death of the Supreme Leader?

[Jake Sullivan]

By the way, on that you mentioned earlier, Mojtiba Khamenei, the son of the Ayatollah who was just killed. There's these reports that maybe he's going to be the new Supreme Leader. A, do you have a reaction to that? B, who is this guy? And can he be a figure who kind of holds together the cohesion of the regime through the continuing military pressure of the US And Israel?

[Kareem Sagapour]

So he's very much his father's son. He's a 56 year old cleric. He doesn't really have a public profile. I was trying to watch some speeches, for example, that he's given just to get a sense of even what his voice sounds like, what does he think? And there's very little out there. He's operated in the shadows over the last few decades. He was long thought to be his father's right hand and kind of a liaison between his father and the Revolutionary Guards. And so. So if indeed he replaces his father, we can expect continuity. Although some say he's even more ruthless than his father. And given the fact that he'll have even less legitimacy, it means he'll have to rely even more on security forces. I don't think he's going to be a powerful figure. And at best, I think he'll probably be a transitional figure. It's going to be the Revolutionary Guards, who. Forces within the Revolutionary Guards, I think, who shape his thinking. So in 1979, when Khomeini and the revolutionaries came to power, one of the things that they said was that hereditary monarchy is un Islamic, and that was their critique of the Shah's government. And so the fact that they're now potentially becoming a hereditary theocracy, it undermines the ethos of 1979. But after the death of Ibrahim Raisi, if you remember, he was Iran's previous president who was killed in a helicopter crash a couple years ago, there was only really two people in that conversation for a succession, Raisi and Mojtabal. And it may be that ultimately they go with a different choice, because there's been rumors that Mojabaz, not only was his father killed in those strikes, but his wife was also killed. There's rumors that he's been injured and he's convalescing. We saw that Israel actually bombed the meeting of where the assembly of Experts, this body of 88 clerics, my joke about them is their average age is deceased. You know, Israel bombed that location. And so every.

[John Finer]

It's become less of a joke these days.

[Kareem Sagapour]

It's true. And so every senior official in this regime has an Israeli bullseye on their back. And, you know, Ayatoll Khamenei ruled for 37 years. If Mujtaba succeeds him, he may not rule for 37 days.

[John Finer]

Karim and one other way we've talked about how this war could end, one other way in which it could expand, maybe even escalate, is if other countries joined the Gulf. Countries who you've alluded to a number of times, have really borne the brunt of Iran's retaliation. I think possibly even the majority of munitions that Iran has fired, have landed, or been intercepted by the uae. Just one of these Gulf countries, obviously. Saudi Arabia, Qatar, Kuwait, other countries have been shot at as well. They wanted to stay out of this conflict. I think that's quite clear. Some of them, I think, lobbied the administration quite hard against going to war, although there's been some conflicting reporting on that. At this point, though, the pressure is mounting on them. These are countries that spend a lot of money on having relatively advanced militaries to defend their territory, their populations. So they face this sort of fateful choice. How do you think that's going to play out?

[Kareem Sagapour]

So I did a Fulbright scholarship in Beirut a couple decades ago, and one of my big takeaways from that year in Beirut is that it takes decades to build things and it takes weeks to destroy things. I have great admiration for some of these Gulf countries. UAE is a good example. In 1978, Ayatollah Khomeini and Sheikh Zayed, the ruler of the UAE, went to the same elevator, and Sheikh Zayed pushed up and Ayatollah Khomeini pushed down this country, United Arab Emirates, which was really a backwater five decades ago. It's amazing how they've become this global hub of finance and transport and technology, and Iran is this global pariah. And what those countries, Saudi Arabia and Qatar, are trying to follow in the footsteps of the uae, and those countries have a lot to lose because they know that what has taken them many decades to build, Iran can attempt to destroy. And so I think of them as kind of like the battle between falcons and vultures. One team wants to build these soaring cities, societies, economies, and. And Iran and its proxies are vultures. They feed off chaos and misery. And so big question now is, how do these countries react? Because Iran did telegraph that it was going to regionalize this war. But I think many of those countries didn't anticipate that that was going to mean attacks on civilians and civilian outposts. So far, my sense is that it's really rather than having the effect of them going to the United States and saying, please stop this war, it's had the opposite impact that they've gone to the United States and said, how can we be helpful now? This is kind of, at the moment, a tangential point, but a very interesting one, which is that you guys both know that Saudi Arabia and Pakistan signed a mutual defense treaty not that long ago. And the Pakistanis had to recently remind Iran that we have a mutual defense treaty with Saudi Arabia. And so if you decide to continue lobbying strikes against Saudi Arabia. We may have an obligation to come and defend them.

[Jake Sullivan]

I missed that, Karim, that the Pakistanis went to the Iranians and said that

[Kareem Sagapour]

there was a statement from Pakistan's. I believe it was the deputy prime minister, the foreign minister, who. Because in all of the news that happens, we forget about some of those developments. And that was a big story for 24 hours, and then we forgot about it. And for that reason, the Pakistanis made a public reminding the Iranian.

[Jake Sullivan]

Yeah, exactly. No, what's remarkable about that mutual defense agreement that was signed between Pakistan and Saudi Arabia, I think it was, last September, is that it also has a nuclear dimension. Actually, it's extending the Pakistani nuclear umbrella to Saudi Arabia, which is quite remarkable. It just goes to show you how fast things have evolved here, that you've got Pakistan having to go out publicly and talk about what effectively is an article 5 commit to Saudi Arabia, which is currently being attacked by Iran. So we'll get you out of here on this question. Not an easy one, but let's say, Karim, that President Trump called you in tomorrow to the Oval Office and said, we're six days into this thing trying to figure out where to go from here. With all that you know about this country, what would the right US Strategy be from this point forward to try to produce the best outcome possible? How would you answer that question?

[Kareem Sagapour]

It's so difficult to answer it right now because so many of the things that he's already done I would have advised against. So we're kind of inheriting a quagmire. I think what's so critical is for President Trump to be clear in his own mind about his endgame and his objectives. At one point during the negotiations, I think I heard from Secretary Rubio that there are four main objectives here. It was to limit their nuclear program, limit their missile program, limit their support for proxies, and limit their ability to brutalize their population. I think those are four very defensible objectives. But even within them, you have to kind of define what those objectives are. The problem we have now is that in the President's head, it's not clear what his objectives are, and he's just been kind of improvising as we go along, and that is a recipe for absolute disaster. And what I felt in the first four days of this war is that everyone has been a loser. There hasn't been any winners here. The regime is obviously flailing. The Iranian society is in shell shock. Hasn't been good for America, hasn't been good for Gulf countries. Israel is receiving incoming fire. And so this has been a lose, lose, lose all around. And I think it's absolutely critical for United States and the US President to be clear in his own mind, okay, this is what we're actually trying to achieve. This is the viable objective we have. And then how do we fulfill that objective?

[John Finer]

Kareem I'm pretty confident the president is not one of our listeners. I don't know that he reads 5,000 word essays in the Atlantic, but he does watch TV. So if you happen to see an unknown number pop up on your phone, who knows, he could certainly he would struggle to do better, I think, for somebody to get advice from in this moment than you. And I don't think we could have had a better guest try to help us make sense of the madness happening in the Middle East. So grateful that you were with us today.

[Kareem Sagapour]

Thank you guys so much. I enjoyed it.

[Jake Sullivan]

Amen. Thank you, Kareem.

[Kareem Sagapour]

Thank you, Jake.

[Kara Swisher]

Hi, this is Kara Swisher. And this week on my podcast on with Kara Swisher, I talked to California Governor Gavin Newsom. While he hasn't officially announced or running run for president yet, he's telegraphing it all the time. It's exhausting. He's also got a new book out, which is what you do when you're running for president. It's called Young man in a Hurry. I recently interviewed him live in San Francisco. Have a listen.

[Kareem Sagapour]

The problem with the Democratic Party so

[Jake Sullivan]

often is we appear weak and we've

[Kareem Sagapour]

got to be stronger and we've got to be more assertive. And so that's, you know, it's the

[Jake Sullivan]

spirit, I think, that is required of this moment.

[Kara Swisher]

I've known Gavin Newsom since he was mayor of state San Francisco a million years ago, a million hair gels ago. And he's a really interesting and compelling politician. He's done a lot of things in his career and this one, this run for presidency which is going to happen is among the most interesting. You can find a full conversation wherever you get your podcast and on YouTube, obviously. Be sure to follow and subscribe to on with Kara Swisher for more.

[Jake Sullivan]

Are Democrats their own biggest problem?

[Kareem Sagapour]

You know, a party becomes defined by

[John Finer]

who their central figure, who their quarterback becomes.

[Jake Sullivan]

Democrats haven't really anointed a effective quarterback

[John Finer]

since Barack Obama pretty much.

[Jake Sullivan]

And this week, the Atlantic staff writer Mark Leibovich joins me to discuss the state of the Democratic Party and which races to keep an eye out for this midterm election. The episode is out now. Search and follow Stay tuned with Preet wherever you get your podcasts. So, John, always illuminating to have Karim, who's got, I think, just a unique set of insights, particularly around Iran and how they're thinking about all of this, both the people and the regime.

[John Finer]

Yeah. And he told us off the air that he's been doing some more writing on this topic that'll be coming out soon, so certainly be looking for that. He's a great writer, in addition to obviously being quite eloquent on this topic. I don't know, I felt like he sounded almost a little bit wistful about the situation. This is a moment of some promise, but also just massive, massive suffering and peril and not clear at all that it's trending in the right direction and does seem to be weighing on him.

[Jake Sullivan]

And he's a dignified and polite person. But you could hear, I think, a very consistent subtext of frustration about the lack of coherence, as he put it at the end, the lack of planning such that that promise is not likely from his perspective, to be converted into real, meaningful political change in Iran. And as he put it right at the end, I thought was really interesting that everybody's been a loser in this. It probably sums up my view right now that everyone has been a loser, including us, because of the inability of this administration to really answer fundamental questions about why and what and what then. And so I thought it was striking that through the sober analysis, that good turns of phrase that really came through as a theme of his remarks over the course of the podcast.

[John Finer]

Yeah, I mean, so much of what happens when you start a war ends up not going the way you planned. But it seems in this case, like there are just many things that haven't even been contemplated, like the scramble right now to try to help Americans who were stranded in some of these places, attacks on embassies and other diplomatic facilities, just events that I think are going to change how this is viewed in the United States in a pretty negative way. But we'll see how it plays out.

[Jake Sullivan]

We will see how it plays out. Obviously, things are developing very rapidly. We've heard Pete Hegseth, as you noted, say this could go on for eight weeks or longer. He also said it could go on for three weeks or shorter. So we actually have very little idea of what lies before us. Meaning you and I could be back on an emergency pod announcing who knows what in the next few days. And in any event, by the time we're back with our regularly scheduled programming in the next podcast next week, we'll certainly be talking about this issue again because of its significant implications, not just immediately in the region, but beyond as well. Should also note one thing which is that if it goes eight weeks or if it even goes five weeks, at the end of this month, President Trump is supposed to go to Beijing for a major summit. And the question of how that all plays into this, it's going to be quite interesting, something that we will watch over the course of March. But for now, we wanted to take the remainder of our time in this podcast the next 10 minutes or so to pick up on a topic we've talked a bit about before, this engagement between Anthropic and the Pentagon ultimately OpenAI coming in as well, what lessons we draw from that, where we see things going. So John, why don't I turn it over to you to tee up the conversation and how you're seeing things right now given the fast moving developments of the last several days.

[John Finer]

So I think when we last spoke about this, the issue had not been fully resolved, by the way. Still probably not fully resolved the way this administration operates. But we'll get to that. But, but at the time Anthropic, which is the only large language model that had been approved for use in US Government classified networks and had a contract with the Pentagon to do that, was in a standoff with Pete Hegseth and the Pentagon over terms that Anthropic was insisting on in that contract that the Pentagon had decided it no longer supported. One of those terms was not allowing their model to be used for fully autonomous weapons that do not have a human decision maker in the loop when conducting lethal targeting. And the other was for the mass surveillance of Americans. The Pentagon's position was we'll make the decisions about what policies we will follow as long as they are consistent with the laws constitution of this country. This led to this protracted and pretty nasty it sounds like negotiation that spilled into the public. And when we last spoke about this issue, the Pentagon had given Anthropic a deadline of last Friday essentially to accede to the Pentagon's proposal for how to resolve this situation, which was less than, I guess, a full ironclad promise to abide by the terms that Anthropic wanted that all completely broke down. And in advance of the deadline, the Pentagon had also threatened, as we talked about, two possible ramifications for Anthropic or really three. One was canceling the contract, which they have now since done, and the other was essentially punishing Anthropic by either using the defense production Act, a sort of expansive authority that allows the government to order private companies to do certain things, use the Defense Production act to insist that Anthropic provide the sort of model that the Pentagon is seeking or in the alternative. And these goals, as you pointed out, are somewhat intention declare that Anthropic is actually a risk to the US Supply chain, which would have mass massive implications for Anthropic as a business. Obviously it would not allow them to do business with the U.S. government anymore. But also any other entities that do business with the US Government would have a hard time doing business with Anthropic, which would include many of their largest customers, hyperscalers and the like. So that all erupted. The Pentagon said that it was going to declare Anthropic a supply chain risk, has not, I think, formalized that process. But in the meantime, as you said, OpenAI, anthropic, anthropic's probably bitterest rival of the very few large language models that really operate at the frontier, are leading kind of crown jewels in this area, essentially negotiated behind the backs of Anthropic a separate agreement with the Pentagon to replace them as the contractor. Have I got that basically right? There's more to the story.

[Jake Sullivan]

I mean, bottom line, Anthropic insisted on additional limitations on mass surveillance and on fully autonomous lethal weapons. The Pentagon said, nope, we're going to have a clause that says all lawful purposes. OpenAI basically came in and accepted that, although it has subsequently added some clauses to it. And now OpenAI has the contract. Anthropic does not. The Pentagon has said it's going to designate Anthropic as supply chain risk, but actually hasn't done so yet and we will see if it does. By the way, you said that OpenAI and Anthropic are the bitterest of rivals. There's a great. I think it's on video. I've just read about it. So actually I haven't seen the video yet.

[John Finer]

It took place in India where I'm sitting tonight. Actually, I think what we're about to

[Jake Sullivan]

talk about, Sam Altman, the head of OpenAI, and Dario Amadeh, the head of Anthropic, were standing next to each other at this AI summit that India hosted a couple weeks ago. And as often happens at these summits, they have these kind of cheesy moments where they have all of the participants kind of gather on stage together and then and hold hands and raise their hands. And Sam and Dario refused to join hands. Everyone else did, but it was a image which very much reflected the bitter coldness between these two people and these two companies. So look, I think for me, there's kind of three big takeaways here. One, and the threat by the Pentagon to impose a supply chain risk designation on Anthropic is outrageous. This is a patriotic American company. Ironically, the whole reason why Anthropic is the only AI company that has a classified contract with the Pentagon until OpenAI just came along is because they started very early going to the Pentagon and saying, we want as good patriots to support the US Military and we're prepared to do the work to build this capability with and for you and put a lot of effort into that. So ironically, they are the most forward leaning on these issues. And for the Pentagon to turn around and say there's somehow a risk to the American supply chain, it holds no water. You pointed out to me actually the crazy reality that Anthropic could be deemed a supply chain risk when no Chinese LLMs like deep seq are currently designated as supply chain risk. So what a bizarre.

[John Finer]

And I believe, by the way, no American company has ever been deemed a supply chain risk. Only foreign entities have gotten that particular designation. So this is quite precedent setting and

[Jake Sullivan]

almost certainly illegal as well. We've seen, I think, a number of our former colleagues lay out legal analyses

[John Finer]

that say we're just fake lawyers, but we sometimes offer legal analysis.

[Jake Sullivan]

Well, we stayed at a Holiday Inn last night, Holiday Express last night. So we can do it with our law degrees, but basically no good solid legal experience. So it's almost certainly illegal. The Pentagon probably knows that, which is why they haven't actually imposed it, as we've noted. It may just be that they want to put a cloud over Anthropic to make companies that work with the Pentagon less likely to use them. And that tells you a lot about the gangster like tactics at work here. The second point that I would make though is that the substance of this so on process, it's just ridiculous what the Pentagon has done. Substance, it's a little more complicated. Let me explain why I think so in this particular case. I think Dario is absolutely on the right side on the issue of mass surveillance, the issue of lethal autonomous weapons, and the demand that there be a higher set of guardrails put in place for these two things because of the capabilities of the AI systems. I think he's totally, totally right and he's right to stand on principle on that. But if you generalize this principle so that it's just like for other cases where a technology company might come along and say, I don't like current US policy. So I'm going to create kind of my own US policy that feels like it could take us in a more problematic direction if it became the norm. So this particular case I'm very sympathetic to Anthropic Andario, but the generalizing of this case makes me a little bit more nervous. I think the right answer on who decides should be a responsible government with the input of the public and not a technology CEO. Of course, we don't really have a responsible government or a responsible Pentagon right now as we've watched them play this thing out. So that's what makes all of this so challenging. And then the third point is that it's been interesting to me to see OpenAI basically accept the assurances of the US government that all lawful purposes covers their concerns around mass surveillance and fully autonomous weapons. But after a lot of pushback over the weekend, OpenAI actually got the deal amended to say, I think this is a quote that ChatGPT shall not be intentionally used for domestic surveillance of US persons and nationals. And it had other quotes around deliberate tracking, surveillance or monitoring of US persons or nationals. So what that suggests to me is that internal pressure and external pressure on OpenAI got them to firm up their position to come closer to where Dario was, but also that the Defense Department, for them this was as much about score settling with a company that they didn't like Anthropic, for reasons I don't understand entirely, as it was about the substance, since they've now agreed to do things with OpenAI that they had not agreed to do with Anthropic. So a kind of, sorry, I think experience all around from the point of view of how the Pentagon has approached this, a principled and quite brave stand by Anthropic. But then this underlying tension and question about the relationship between technology CEOs and the US government and the US National Security Enterprise that I think we will have to puzzle through in the months

[John Finer]

and years ahead look bad, I think for the country and the sector all around. This is one of, as we've said, the three leading companies developing what is maybe the most important technology in human history. And in some areas it is the world leading large language model for things like coding. It's the dominant player in kind of enterprise business of large language models. And now because of this dispute is also jumped to the very top of the Apple Store rankings for all applications. And it is essentially being put on a US government blacklist, not being celebrated, but maybe a step that could ultimately badly damage, if not destroy the business that is not good for the United States that that is happening. This is also freaking out industry players in the United States and around the world because as you said, it's pretty unprecedented for the US Government in a punitive way over a dispute that really does feel political. They clearly don't like Anthropic. It's got this in this right wing ecosystem woke AI label that's been put on it by people like Elon Musk. And I think largely because of that, they were willing to go further for OpenAI in the contract negotiations than they were willing to go for Anthropic. But it is not comforting to anybody to have a precedent where the government essentially tries to destroy a company for that reason and destroy it in a totally disingenuous way with this supply chain risk designation. I'm in India, I hear tech players in the Indian industry and in the Indian government saying things like we're making a big bet on the US technology stack. And many of them have made a big bet on, on Anthropic given that it's only one of three top end players. And they're now wondering do we have to totally rip and replace Anthropic's models out of our systems and if so, how much can we rely on other US technology companies? Given if you get in the crosshairs of this administration, you might get kind of obliterated off the map too. It has created massive uncertainty in an area where we are in a, as we've discussed many times outside this podcast and on it, just very intense competition with our main competitor in the world, China, which is racing ahead with these technologies too. The last thing I was going to say is, I really do agree with you though about universalizing this principle that companies can come to the US Government when they're contracting and insist on a bunch of policy, policy changes or restrictions. I think about what somebody like Elon Musk could do or could have done to an administration we served in SpaceX, for example, has a even more stark monopoly on heavy lift rocket transport for the US government for satellites and other things. And he could come to the United States government and say, look, I'm very willing to keep doing this, it's a good relationship, it's a big contract for us. But I just want to make sure we're all on the same page about one of his pet issues, for example, dei. He could say, anybody who I consider to be a DEI hire, obviously we would agree can't possibly be working on this program because that would put at risk everything that we're trying to do it make it less safe. And I'm not willing to do that. How would an administration handle that demand? I think these sorts of precedents are dangerous from both sides, which is why I wish that the two parties had been able to work it out. That would have been a much better outcome for the country, for the U.S. government, for the industry. But that's not where it landed.

[Jake Sullivan]

No. And I'd say one more thing on this that is unique to the AI context. If you think about most forms of military technology or military relevant technology, a Raytheon or a Boeing or a Lockheed making patriot interceptors or F35s or you name it, that's typically done in close partnership between the US military and the company, such that the technology is very well understood by both the military and the defense contractor. These LLMs, the U.S. government is a total spectator. They didn't have anything to do with creating them. They don't have the depth and breadth of expertise to understand the capabilities and how they would apply an understatement, if any. Yeah. In the context of mass surveillance or lethal autonomous weapons. And part of the argument Dario was making, which resonated with me, was, guys, I get this thing and what it can do and what its risks are. And unlike a lot of other weapons systems, there isn't a counterpart on the US Government side with the depth of knowledge of that because they didn't have a hand in building it. And so I think we're at the front end of a lot of tension in this space because the military is going to want to make use of these capabilities, and yet the companies are the ones that really understand them, both the upsides and the risks of them. And it's interesting to me that OpenAI, as part of its understanding of the Pentagon, is actually going to have OpenAI personnel go in to fit alongside the US government in the the application of these systems. So that's relevant to this particular issue of the contract dispute, but it has much broader implications. And I think as we look at the issue of AI and the national security enterprise down the line in this podcast, we're going to have to return to this theme that unlike space or nuclear or the Internet or particular weapons systems, where the US Government was at the center of their creation and scaling, the LLM revolution has been entirely private sector driven. And that just makes this a unique beast.

[John Finer]

By the way, this has already had real world implications for the US Government. The government gave essentially its entities a six month period to wind down the use of anthropic. But we learned today, at least it was reported today that the State Department, the Treasury Department, Department of Health and Human Services have already essentially excised anthropic from their systems. The State Department in a time of war is now using ChatGPT 4.1, which is an 11 month old model at a pretty critical period for diplomacy for providing consular services to Americans overseas and doesn't exactly put us on our best footing to not be using our best tools.

[Jake Sullivan]

Yeah, we've reached the bite off our nose despite our face portion of the programming here.

[John Finer]

Exactly.

[Jake Sullivan]

All right, well, more to say on this. This is going to be an unfolding issue and obviously with a broader set of implications, but for now we'll call it there. We'll be back next week and maybe even before next week, depending on what happens in the evolving war in Iran.

[John Finer]

Well, that's all for today. We'll be back next week with a new episode of the Long Game.

[Jake Sullivan]

In the meantime, send us your questions and comments@long gameoxmedia.com and find us on [e-mail] the links are in the show Notes.

[John Finer]

That's it for this episode of the Long Game.

[Jake Sullivan]

If you like the show, please follow, share with friends and leave a review. It really helps listeners find us for

[John Finer]

updates and more analysis in your information. Join the [e-mail] the long game is

[Jake Sullivan]

a Vox Media Podcast Network Production Executive Producer Tamara Sepper Lead Editorial Producer Jennifer Indig Deputy Editor Celine Rohr Senior Producer

[John Finer]

Matthew Billy Video Producers Nat Weiner and

[Jake Sullivan]

Adam Harris Supervising Producer Jake Kaplan Associate Producer Claudia Hernandez Marketing Manager Liana Green

[John Finer]

Music is by Nat Weiner we're your hosts John Finer and Jake Sullivan.

[Jake Sullivan]

Thanks for listening.