What this covers

McChrystal, a retired four-star general, speaks directly to Tyler about the habits of thought that lead nations and institutions astray in assessing risk. The conversation ranges across military strategy, institutional design, geopolitics, and technology, anchored by a single claim: that people routinely misjudge danger because they treat the present moment as typical of all time. He argues that without a serious grounding in history, the mind defaults to Linear Extrapolation Bias—the assumption that because national boundaries, great-power alignments, and technological balances have held steady since World War II, they will continue to do so. This backward-looking myopia, he contends, leaves policymakers blind to plausible shifts in China's behavior toward Taiwan, the risk calculus of Putin's Russia, and the true nature of modern conflict.

The conversation then unfolds into three overlapping domains. First, McChrystal diagnoses how the structure of American military institutions—narrow recruiting into a warrior caste, frequent rotation of officers in foreign postings, promotion systems that penalize any stumble, and technological education that lags private-sector competition—weakens the force's resilience and cultural wisdom. He contrasts this with examples of how the British built durable advantage through long tenure and cultural acuity. Second, he reframes modern warfare itself. Wars are "war among people"—societal struggles, not clean battlefield victories—meaning the first Gulf War's apparent triumph over Saddam left the larger problem unsolved. Clean military wins rarely translate to political resolution. Third, he surveys emerging and near-term risks: drone assassination in the continental United States, surveillance systems more powerful than weapons, the Democratization of Technology enabling small actors to inflict mass harm, and the Afghanistan Taliban's forced evolution beyond their 1990s template. Throughout, he emphasizes Deciding Under Uncertainty as the mark of effective leaders, and the costs of treating Shared Sacrifice as optional.

Sharpest takeaway

McChrystal argues that effective risk thinking requires taking a long historical view rather than extrapolating from the recent past, and that modern military and societal challenges are too complex to be solved by simple battlefield-style solutions or narrow, insular institutions.

  • People wrongly assume national boundaries and current conditions are permanent because little has changed since WWII
  • Modern war is 'war among people,' not weapon-on-weapon, so clean military victories rarely solve the underlying problem
  • A narrow warrior caste and self-reinforcing recruiting weaken the military's perspective and resilience

The argument · threads13 threads · 50 claims
0.61

Military recruitment has become dangerously narrow and self-reinforcing, creating an insular warrior caste that fails to reflect the nation.

4 pointscentrality 5/5
  • Because recruiting has become narrow and self-reinforcing into certain regions and segments of society, the military risks becoming an echo-chamber reflection of those groups, so we should widen recruiting and even institute a draft—not to grow the force, but to bring in a broader cross-section and channel non-military draftees into other civilian national service.

    “because it’s gotten self-reinforcing into certain areas of the country, in certain parts of American society, the danger is that the military becomes an echo chamber reflection of that. I would widen recruiting. I personally think we ought to have a draft.”

  • Common defense is every American's responsibility, not the duty of a warrior class; if you hold a mirror up to the face of the US military you ought to see the nation reflected, and if you don't, you have a long-term problem.

    “If you hold a mirror up to the face of the US military, you ought to see our nation. If you don’t see our nation, then you have a problem long term.”

  • The all-volunteer force's reliance on recruiting children of veterans has become counterproductive by creating an insular warrior caste; you don't want a soldier class in America, you want military service spread across the nation geographically and sociologically, because narrow recruiting single-threads the force, and real diversity is about different perspectives, experiences, and talents rather than race or gender.

    “you don’t want a soldier class in America. You want military service to be spread across the nation geographically and sociologically. I actually think that that aspect of the all-volunteer force has weakened us a bit. We have created a bit of a warrior caste, and it’s insular.”

  • The post-WWII era in which most American males experienced military camaraderie and trauma was net positive because near-universal service meant nobody could claim special status for serving and trauma was experienced collectively rather than as victimhood; today's tiny share who serve produces unhealthy dynamics—non-veterans feeling guilty and veterans feeling self-righteously more patriotic—neither of which is healthy.

    “so many people were veterans or have been involved heavily in war sacrifices that nobody felt they could pound their chest and say, “I served and therefore, I’m special.” Everybody thought, “That was a responsibility, and I met my responsibility.””

0.58

Modern warfare requires investing in adaptable people—cyber talent, linguists, cultural operators—over expensive capital platforms like carriers.

2 pointscentrality 5/5
  • If forced to choose, the military should accept risk by cutting maneuver-warfare ground capability and even the number of aircraft carriers (big capital things), and instead invest in really good people—crafty cyber-warfare talent, innovators, and people with cultural acuity and language skills—because those capabilities matter most in the modern era.

    “I think where we can’t afford — and therefore, I would invest — is in really good people... we are going to need very crafty people at things like cyber warfare. We’re going to need very innovative people. We’re going to need people with cultural acuity, which means language skills”

  • Because cyber and drone talent can earn far more in the private sector, the military must reverse its model of bringing in already-skilled people and instead become a training ground that teaches high-tech skills—just as it did in the aerospace age—so that access to hard-to-get technology and training becomes a recruiting tool, attracting talented people for several years before they move to tech firms.

    “The military’s got to think about creating an education system for talented people so that becomes a recruiting tool. If you come in and work in the military, you get access to technology and training that is hard to get on the outside.”

0.47

Strategic leaders must study history and adopt far-seeing perspective to escape cognitive traps of recent-past extrapolation and linear thinking.

4 pointscentrality 4/5
  • Reading a great deal of history is not merely useful but essential for thinking about risk, because without it you constrain your mind almost entirely to personal experience.

    “I think it’s not only useful; it’s essential. If you don’t know history, you constrain your mind almost into personal experience only.”

  • People assume national boundaries don't change much, but that is only because few have changed since World War II; historically boundaries and identities change routinely, so assuming the current global map is permanent runs counter to historical experience.

    “they think that national boundaries don’t change much, but that’s because after World War II, not a lot of national boundaries have changed. Some have, but if you go before that, they actually change in history pretty routinely.”

  • US policymakers' most common probabilistic mistake is oversimplifying war by modeling it on WWII (crush the enemy then rebuild), when conflicts like Korea, Vietnam, Afghanistan, and Iraq are wars among and of people, involving whole societies rather than army-on-army combat.

    “from Korea, Vietnam, Afghanistan, and Iraq, we often don’t admit the complexities because you’re dealing with war among people and of people. So, you’re dealing with societies, not straight weapon on weapon or an army on army.”

  • To train yourself out of the habit of extrapolating from the recent past, you must step back into an almost out-of-body perspective and look at things historically—asking what has happened before and why it wouldn't happen again in some new permutation—which is a discipline that Henry Kissinger exemplifies by soaring to 30,000 feet on any issue.

    “you have to step back — almost an out-of-body experience — and look historically at things and say, “Okay, what has happened in the past? And why wouldn’t that happen again, or some new permutation of that?” It’s a discipline.”

0.44

China is more likely to act on Taiwan than Western linear thinking acknowledges because of Beijing's historical self-conception and containment anxieties.

2 pointscentrality 4/5
  • We are probably underestimating the risk China takes serious action on Taiwan because we think linearly—assuming that because something hasn't happened for a long time it never will—whereas China's view of itself as the Middle Kingdom, its goal of avoiding containment, and its sense that Taiwan is a legitimate part of China make such action plausible.

    “We tend to think in terms of linear. If something hasn’t happened for a long time, it never will, and then we’re always shocked.”

  • China is not a black box—they signal their intentions and aspirations fairly carefully in their writings and statements—but we don't grasp their calculations because we lag the reality of their ambition to be a major world player by quite a lot, still picturing the poor, struggling China of past decades.

    “I don’t think they’re a black box if you read their writings and if you listen to what they say. In fact, I think that they signal pretty carefully what their intentions, or at least their aspirations, are.”

0.44

Effective military advising requires cultural fluency and long tenure, not rapid rotations, as the British model in India demonstrated.

2 pointscentrality 4/5
  • Effectively training foreign armies requires cultural acuity and long tenure, as shown by the British model in the northwest provinces of India where East India Company officers served 10-year tours and became fluent in language and culture, whereas the US rotates people so frequently it lacks a cadre with deep regional experience, dooming its training efforts.

    “The British used a tremendously effective technique of just small numbers of people seeded, but they were sent for long periods of time... their first tour was 10 years... By that time, they had become completely fluent in the language and fluent in the culture.”

  • The US trained roughly 5,200 Americans to become fluent in Japanese in the first year or two of WWII but trained less than a tenth as many for Afghanistan and Iraq despite being there longer, because the military celebrates generalists and fears that language specialists will lose their wider service skills—and this aversion must change by making fluency in at least one language a requirement for officers.

    “During World War II, in the first year or two, we trained something like 5,200 Americans to become fluent in Japanese, a very difficult language. In the period in Afghanistan and Iraq, I don’t think we trained one-tenth that many, even though we were there much longer.”

0.44

Military commanders' training to achieve battlefield victory blinds them to larger political and social complexities that victories alone cannot solve.

2 pointscentrality 4/5
  • Military commanders are shaped to win battlefield victories, but in the modern era such victories rarely solve the larger problem; the first Gulf War is held up as a clean win, yet in its aftermath Saddam Hussein used the chaos to brutalize his population and cement power, so the war solved Kuwait but not the bigger problem.

    “the first Gulf War, which is held up as an example of a very clean victory — in reality, in the aftermath of that, Saddam Hussein used the chaos to brutalize part of his population and to cement his hold on power.”

  • Individual combatants are trained to be warriors, so to a man with a hammer everything looks like a nail — they are spring-loaded to use that one tactic, which fails when they encounter complex problems in villages, towns, and cities.

    “to a man with a hammer, everything looks like a nail, so they are spring-loaded to want to use that particular tactic. That’s what they’ve been trained, and then they run into complex problems in villages and towns and cities, and they find it’s not simple.”

0.43

The military must actively purge seditious violent extremism through clear communication and removal to prevent cascade of ideology down chain of command.

1 pointcentrality 4/5
  • The seditious violent right wing in the military is a serious problem that must be addressed by clearly communicating that the military stands for the Constitution and cannot be perverted by extreme views, and by actively ferreting out and removing or prosecuting (via UCMJ) those reflecting such views—because a single sergeant or lieutenant can cascade extreme views down to young privates, making it extraordinarily dangerous.

    “a sergeant can take young privates and have an extraordinary effect on them, and a lieutenant can do the same with people. The ability to cascade down extreme views is extraordinarily dangerous.”

0.34

Military service should be opened to far broader physical and mental range of people because most jobs do not require traditional warrior physiology.

4 pointscentrality 3/5
  • The vast majority of military jobs do not require a strapping, barrel-chested male and could be done by people including those with significant handicaps, so recruiting should be opened up to a far broader range of people willing to do the job.

    “the vast majority of the jobs in the military don’t take a strapping 6-foot-2 barrel-chested male. They require somebody willing to do your job and a smart person, so I would open that up.”

  • A functional way to work and contribute is the great cure for many of life's problems because it is where most people derive their satisfaction and feeling of self-worth, so for disabled people leaving the military we should fight to decrease the barriers to employment, especially since technology now lets even significantly disabled people do many jobs.

    “The great cure for so many problems in life is some kind of very functional way to work and contribute because that’s where most of us derive our satisfaction and, say, feeling of self-worth.”

  • If a draftee is ignorant but not stupid—able to learn—the military can remediate them through programs like the old Basic Skills Education Program, but it should avoid repeating McNamara's 100,000 by bringing in people with very low scores who lack the mental ability to learn, because that limits the force's capability.

    “I would avoid McNamara’s 100,000, when he brought in people with very low scores, because you limit the force’s capability. If somebody doesn’t have education, if they’re ignorant but not stupid, you can address that.”

  • Providing disabled veterans state-vouchered access to sex workers, as the Netherlands does, is problematic because sex workers are often badly exploited and such a program could create people forced into sex work as the only job they can get, so one should not force people into that kind of life.

    “I’d be really frightened that we would create a bunch of people who end up having to be sex workers because that’s the job they can get. You don’t want to force people into that kind of life.”

0.32

Russia will play its weak hand aggressively to reclaim superpower status because Putin has consistently been willing to absorb significant pain for that goal.

3 pointscentrality 3/5
  • Russia is willing to suffer a fair amount of pain to get back into the superpower position they believe they should occupy, though whether Putin personally is risk-averse with respect to personal loss has never really been tested because the West hasn't directly confronted him.

    “Russia is willing to suffer a fair amount of pain to try to get back into a position that they think they should be in as a superpower.”

  • Russian risk thinking is easier to read than China's because it can be focused on Vladimir Putin, who for 20 years has been a good proxy for the Russian psyche; Russia, holding far worse cards than China including a real demographic problem, will play those cards more aggressively than expected to reclaim superpower status, as seen with natural gas leverage over Europe.

    “they’ve got a real demographic problem. They’ve got other issues, but the reality is, in the short term, they’ve got some aspirations to be powerful in the Mideast... They’re going to play the cards they’ve got more aggressively than we might expect them to.”

  • Germany is sufficiently worried about Russia, but because it is geographically close and dependent on Russian natural gas and trade, it rationally favors a realistic, somewhat accommodating approach, whereas the distant and stronger United States can poke Russia with relative impunity—so the apparent risk miscalculation is a function of geographic position, not German naivety.

    “The United States, from a long way away, can put a finger in the eye of Russia with relative impunity. Germany’s a strong country but not as strong as the United States and not as far away.”

0.30

Rigorous Cold War nuclear strategy thinking was sound methodology; combat trauma should not automatically be pathologized; and history's dynamics reshape boundaries and governance.

20 pointscentrality 1/5
  • Despite nuclear war seeming unthinkable, the deep work on nuclear strategy, game theory, and deterrence represented some genuinely careful and rigorous thinking about actual risk, better than common caricatures suggest.

    “In reality, game theory and whatnot was actually some pretty careful thought on actual risk and then things like deterrence.”

  • The likelihood of nuclear war during the Cold War was probably closer than people thought, and having lived near the precipice so long we now take such risks for granted, while threats like cybersecurity are a greater risk than we actually admit.

    “I actually think that the likelihood of nuclear war was probably closer than we thought. Now we’ve lived near the precipice long enough where we take it for granted. Things like cybersecurity, I think, are a greater risk than we actually admit.”

  • Anything that forces you to problem-solve against an opponent trying to create threats—including war board games and chess—is very good for thinking about risk because it builds muscle memory in the mind for handling adversarial uncertainty.

    “Anything that forces you to problem-solve is very good about thinking about risk because you’ve got an opponent who is trying to cause you problems, to create threats and put you at risk.”

  • The two qualities that mark someone for high promotion are self-discipline—defined not as making your bed but as treating people the way they should be treated and doing hard, inconvenient, or frightening things—and the ability to make decisions under uncertainty, since people who try to drive uncertainty to zero keep asking for more information, while good leaders accept imperfect knowledge and still decide decisively.

    “a person who’s trying to drive uncertainty to zero will keep asking for more information... Some people just have the ability to live with not having perfect knowledge, and yet, they can accept that and still make decisions decisively.”

  • Starting a company after retirement was an attempt to recreate the comradeship of military life, and it revealed that what he loved about being a soldier was less about carrying a weapon than about being part of a team—the source of the satisfaction he sought to replicate.

    “It made me realize that what I loved about being a soldier was less about carrying a weapon and being a soldier than it was about being a part of a team.”

  • Al-Qaeda's psychology, like a player who hits a grand slam and wants another home run, led them to seek big dramatic operations rather than thinking realistically about achievable effects, but other organizations—most likely via cyber—may take a different approach and achieve significant effect through many smaller attacks.

    “I just don’t think that they thought realistically enough about creating the effect that they could do. They were looking for the big dramatic operation.”

  • Military fiction is in temporary decline because the population lacks broad experience with military service—the genre appeals mostly to those who served or are close to someone who did—so without that connection it is hard for readers to appreciate or judge whether a work captures things compellingly.

    “One of the reasons it’s in temporary decline is you don’t have a broad experience in the population with military service. Military fiction is more apt to be interesting to people who either served earlier in their lives or they are close to someone who served”

  • The emotionally hardest part of military command was not putting people in harm's way—since everyone enters with that expectation—but coping with debilitating combat injuries that ended soldiers' ability to serve, because while the healthy move on to the next operation, the wounded are left behind, and making that separation was most difficult.

    “the healthy ones go on to the next operation, and you leave the wounded and dead behind, and they try to band together. I found it most difficult to make that separation from the wounded.”

  • Talent doesn't always reflect itself in a degree; what you really want in officers is the right core values combined with experiences and enough innate talent to be developed into good leaders with sound judgment, rather than everyone holding a PhD.

    “talent doesn’t always reflect itself in a degree. Sometimes talent reflects itself in a combination of values and experiences and native intelligence.”

  • The US Army's one-strike-and-you're-out, no-blemish promotion habit prevents officers who got scuffed up from rising, so high ranks fill with people who never had blemishes—meaning they may have taken too few risks and lived too conservatively—whereas in WWII relieved division commanders were given second commands and often succeeded, making the no-blemish norm a real negative for the officer corps.

    “We have a no-blemish habit in the US Army... what you tend to get higher level is people who’ve never had any blemishes, which means maybe they’ve not taken enough risks; maybe they’ve lived a little too conservatively. And that’s a real negative for producing a better officer corps.”

  • At elite schools like Yale, most students have no real contact with the military, so when active-duty officers and senior NCOs attend for graduate degrees they have a disproportionate, eye-opening effect on young people, replacing two-dimensional movie stereotypes of soldiers with real understanding—and Yale's 40-year exclusion of ROTC limited its diversity and open-mindedness.

    “they have a disproportionate effect on young people because it’s the first time any of these young people have been up close to anybody who’s a soldier... they have a view that’s often very narrow of what a soldier is. They’ve seen some movies, and they’ve got a two-dimensional, stereotypical opinion of it.”

  • When advising a young soldier on writing a will, the key is not to be superficial—think about who could genuinely benefit from your insurance or assets rather than jokingly signing wills to a bunkmate—because it is one of the first times many young people confront realistically that they are not immortal and that what they do can impact others.

    “It’s one of the first times in life many young people can think very realistically about the reality they’re not immortal and that what they do can have an impact on other people.”

  • What made military living quarters special was shared experience—everyone had the same quarters, nobody had a bigger house, and everyone made the same money—and the main improvement to make is changing where and how quarters are used, such as housing unmarried sergeants in barracks so that multiple generations live in close proximity rather than segregating young and older people.

    “You want people of multiple generations in some close proximity. You don’t want to have all the young people boxed off here and the older people completely somewhere else.”

  • America's failure against COVID-19 was a failure to view public health as part of the common defense, reflecting the broader loss of a sense that we all have common requirements and shared responsibility to society and to each other.

    “I would argue that our failure against COVID-19 is a failure to view public health as in the common defense.”

  • Suppression of combat trauma is unhealthy and some people have real trauma, but if you go looking for trauma long enough you'll find it—telling people they must be traumatized can make it self-fulfilling—and trauma doesn't always produce weakness; combat may instead produce more resilience and maturity, so veterans should not all be treated as victims or casualties.

    “if you go looking for trauma long enough, you’ll find it — meaning, whether it’s there or not, you can almost create it by looking for trauma and telling people, “You must be traumatized because you went through this.””

  • Afghans had genuine but flawed expectations of America in 2001, having helped fight the Soviets at a cost of 1.2 million dead; their huge hopes for political and economic progress went partly unmet due to both their own failure to wrestle out corruption and America's shortfalls, creating bitterness, and crucially they never believed America would stay, so they saw us as unreliable—another source of frustration—even though most preferred Americans to the Taliban.

    “there was an increasing gap between what they were getting from the relationship with the United States, the experience, and what we were able to provide or did provide... they didn’t believe we were going to stay, so they didn’t feel like we were reliable.”

  • The next phase of work is applying the ideas of risk to civilian organizations, including states and cities, to help them find a path to become more resilient and capable as entities, because that is where the future of organizations must go.

    “We’ve worked with a number of states and cities to try to let them find a path to become more resilient, more capable as entities because that’s where I think the future has to go in organizations.”

  • While high technology empowers terrorists and homemade precision strikes (offense), the rise of a surveillance-based society in which authorities can know everything about anyone and watch them almost every minute is a form of defense that could easily become more powerful than the ability to attack, and more nations will go the Chinese surveillance route faster than expected.

    “we’ve got a surveillance-based society now. We actually can know everything about Tyler Cowen, and we can watch you almost every minute... if you consider that as defense — knowing everything all the time — then that could easily get more powerful than the ability to attack.”

  • If current trends continue, in 20-30 years a huge number of capable people will refuse to become public figures, leaving public space to 'Teflon celebrities' who don't mind the limelight; the danger is that politicians come from that latter group—people who in earlier times would have been superficial celebrities—which is very dangerous and will require some new accommodation around social media and scrutiny.

    “The danger is that our politicians come from that latter group — people who are willing to go into very public spaces like that, or actually the people who in previous times would’ve been superficial celebrities.”

  • Afghanistan has changed too much since 2001—too much social progress and education—for the Taliban to simply reimpose their 1990s rule, so after near-term friction one of two things will happen: the Taliban will be forced to mature, accommodate, and evolve into a different kind of government, or they won't maintain control, with the former being more likely.

    “I don’t think Afghanistan is anything like it was in 2001. There’s been too much progress socially, too much education. I think the Taliban are going to find that they caught something they really aren’t prepared for.”

0.29

Women need a generation of affirmative promotion to senior command to provide necessary mentoring and break historical exclusion from competitive career paths.

1 pointcentrality 3/5
  • Women advanced slowly to senior command partly because they were earlier given peripheral jobs (like protocol officer) that left their experience records uncompetitive, so the nation must accept a period of affirmative action—promoting senior women who will be less prepared than peers—for about a generation, because providing senior-level examples to younger female officers is essential and it is too risky not to do so.

    “you have to have a period of affirmative action. You have to promote some senior women, and you need to accept risk in that because they will be less prepared than some of their peers. You say, “Well, that’s too risky for the nation.” I think it’s too risky not to.”

0.18

Emerging drone and cyber technologies at commercial scale now enable attacks domestic and foreign actors could execute but surprisingly have not yet attempted.

3 pointscentrality 2/5
  • The absence of major foreign terror attacks on US soil since 9/11 is partly due to improvements in American intelligence and counter-terrorist operations that made a 9/11-scale attack far harder to pull off, but it is genuinely puzzling why al-Qaeda never attempted many simpler, hard-to-stop attacks like bombing shopping malls.

    “the improvement in American intelligence and, to a degree, counter-terrorist operations was part of that. It became far harder for an organization like al-Qaeda to pull off a 9/11-scale attack.”

  • The Armenia-Azerbaijan war shows that drones and advanced technology now dominate, and that even relatively less advanced nations—not just superpowers—can field smart weapons, drones, and quick intelligence-sharing to mass artillery, as Russia demonstrated in Ukraine.

    “you’re finding that even relatively — I don’t say backward, but less advanced nations can put out really advanced technology. You’re not talking about that being limited to superpowers to have smart weapons and whatnot. Everybody’s got them now.”

  • The first drone assassination within US borders is probably closer than we think and hard to stop, since off-the-shelf technology from places like Best Buy is enough to do it; the bigger surprise is that mass attacks—such as targeting aircraft taking off—haven't already happened, and such attacks could easily come from domestic players.

    “It is probably closer than we think. I don’t think it’s hard to do. You can go and buy off-the-shelf technology at Best Buy or RadioShack or anything like that, and you can get enough to do it. It’s hard to stop.”

0.17

Smartphones and foreign applications pose serious operational security and intelligence exploitation risks that require military-controlled trusted networks.

2 pointscentrality 2/5
  • Smartphones create real problems on deployment because they let soldiers call home every 20 minutes and live their home problems while trying to operate, and because they can be tracked and exploited by foreign intelligence, so the military likely needs work-purpose phones with severe limits on leverageable functions.

    “One of the problems with the smartphones is, people are able to call home every 20 minutes if they want to. When a person’s deployed, they live all the problems at home... There’s also the ability to track it.”

  • Servicemembers should not be able to download Chinese apps like TikTok or WeChat; the country needs to look upstream at all such applications, clean that up, and build trusted clean networks—not just on phones but across wider networks—an improvement that is achievable but has not been done.

    “You’ve really got to take a look upstream at all of these, and we’ve got to start cleaning that up. We’ve got to get trusted, clean networks.”