Three physicists on cooperation between rivals, from fusion to nuclear restraint

The Short Version

Three physicists, asked in separate conversations about a dangerous and divided world, each reach for examples of rivals working together. Dennis Whyte, who directs MIT's Plasma Science and Fusion Center, describes fusion as one of the few fields where East and West collaborated during the Cold War, and the ITER reactor as a project Reagan and Gorbachev picked as a symbol when their other talks had stalled [1]. Carlo Rovelli argues that the Cold War's leaders knew how to step back from the edge, and that the lesson was "living together on the planet is more important than overpowering your enemy" [2]. Max Tegmark tells the stories of the people who prevented nuclear war and of the biological weapons ban, then warns that "relying on luck is not a good long-term strategy" [3]. The record they describe is real, and so are its limits: the seven-partner fusion project has been slow, and several of the escapes from nuclear war came down to one person [1][3].

Table of contents

What They Said

Dennis Whyte: fusion was shared when almost nothing else was

Whyte's account begins with secrecy. Fusion research started in the 1950s behind closed doors on both sides [1]. Then the scientists ran into the physics. Everyone was stuck at low temperatures and could not get the hot fuel to stay confined, and they realized how hard the problem was and how little they knew about what they were doing [1]. So the work was declassified, and fusion became what Whyte describes as one of the few places where the West and the Soviet Union collaborated [1].

The first test of that openness came in the late 1960s. The Soviets claimed that their device, the tokamak, a Russian acronym for magnetic coils in a donut shape, had reached temperatures around 20 times higher than anyone else had managed [1]. Western scientists were skeptical. What convinced them was a measurement: a British team used a then-new laser to measure the plasma temperature in the Soviet machine [1]. Once the result was confirmed, the world rushed to build tokamaks [1]. The design that came out of a Soviet laboratory is still the most mature approach to magnetic fusion, and the one that has come closest to the point where a plasma gives off as much fusion energy as was used to heat it [1].

The second episode is political. In the mid-1980s, Whyte says, a summit between Ronald Reagan and Mikhail Gorbachev had stalled on other negotiations [1]. The two leaders could not get further on the details of a treaty, but they wanted some kind of symbol to show they would keep working toward something important to everyone [1]. Whyte tells it as a question and an answer: "What did they pick? A fusion project" [1]. They signed an agreement to collaborate on a machine that would show large net energy gain and point toward commercial fusion [1]. Japan and Europe joined soon afterward, and South Korea, India and China came in later [1]. Whyte notes that the coalition has held together despite political turmoil [1].

That project is ITER, now under construction in southern France. It aims for a plasma gain of around 10 and about 500 megawatts of fusion power, enough for a city [1]. Whyte spent much of his career on it and says he was deeply excited by it in the early 1990s [1].

He does not present it as a simple success. Repeated delays, and the difficulty of steering a project with seven partners, which he likens to seven chefs in the kitchen, led him and his students to ask whether there was another path [1]. He also found it strange that the whole planet had essentially one attempt at demonstrating fusion power, when several attempts with different technical, schedule and scientific risks would improve the odds [1].

The path he took was SPARC, a tokamak being built by MIT and the spinout Commonwealth Fusion Systems. It uses the same principles as ITER with real deuterium-tritium fuel, but much stronger magnets make it about 40 times smaller in volume, while still aiming for about 140 to 150 megawatts of fusion power and a gain around 10 for roughly 10 seconds [1]. The key was a new generation of high-temperature superconductors that Whyte and his students realized could reach about 20 tesla, which by tokamak scaling would allow a machine that a company and a university could build instead of a seven-nation treaty [1]. MIT and Commonwealth Fusion Systems demonstrated such a magnet in September 2021, and a team of about 20 to 25 people designed SPARC in roughly two years [1].

Whyte is careful to say where that speed came from. He credits the tools and knowledge built for ITER, and compares the relationship to the way the government-funded Human Genome Project laid the foundation for faster private sequencing [1]. In his telling, the big international collaboration did not fail. It produced the science that let a small team move quickly later.

He carries the same attitude into rivalry between approaches. When the National Ignition Facility's laser experiment passed scientific breakeven in December 2022, a different method from the magnetic one he works on, his reaction was that this is no occasion for rivalry or jealousy: "fusion is really hard. So anything that keeps pushing the needle forward is a good thing" [1]. Lex Fridman draws the wider conclusion for him, saying that giant international projects are "a really powerful way to ease some of the geopolitical tension," even military conflict, and names the war in Ukraine and the tension with China as the present-day setting [1].

Carlo Rovelli: the Cold War knew how to step back

Rovelli is a theoretical physicist, and his case for cooperation is made about nations, with the Cold War as his evidence [2].

His fear is specific. "I'm not scared. I'm scared by humans," he says of artificial intelligence [2]. He points out that most AI money now goes toward military uses, that there are 12,000 nuclear warheads that could end humankind in minutes, and that nations are racing to overpower each other instead of cooperating [2]. Joe Rogan adds that autonomous fighter jets now beat human pilots in simulated dogfights and that AI war games tend to escalate to nuclear strikes [2]. Rovelli worries that deterrence depends on humans ultimately refusing to kill millions, a restraint an AI might not share [2].

Against that he sets what the rivals of the last century managed. Kennedy and Khrushchev corresponded [2]. The Cuban Missile Crisis ended with both sides stepping back: Soviet ships turned around, and the United States removed missiles from Turkey and Italy [2]. Reagan and Gorbachev's START treaties cut warheads dramatically [2]. He holds up Kennedy's 1963 American University speech as a model [2]. The lesson he takes from all of it is the line quoted above, that living together matters more than overpowering the enemy [2].

He then argues that the choice has come round again. Thirty years ago, he says, the United States and its allies held about 80% of the world economy and now hold about 40%, not because they got poorer but because others got richer [2]. The United States faces a historic choice between defending its military dominance, at the risk of a third world war, and turning rivals into collaborators [2]. He offers Italy and the United Kingdom as countries that learned to live well after losing their empires [2]. He adds an estimate of what is at stake in ordinary terms: even a 5% cut in global military spending would free enormous resources for poverty and climate [2].

Rovelli knows the record includes failures. When the United Nations was founded, he says, "eliminating poverty was declared an interest of humanity by all the countries together," and then "special interest and power games prevail" [2]. He is also cautious about predictions, including hopeful ones about technology. Every era reads the universe through its newest invention, he says, whether books, machines or computers, and the future usually surprises us [2]. But he shares Rogan's hope and calls pushing toward the common interest everyone's responsibility [2].

His physics supplies the frame. He argues that quantum mechanics shows there is no single objective reality beneath our perspectives, and that the properties of a thing are how it shows up to everything it interacts with [2]. He applies the same idea to nations, which he believes must learn to see one another's perspectives instead of hunting for enemies [2]. He pairs it with humility. His father grew up believing his name was Franco until his birth certificate showed a different first name, and Rovelli's moral is that you can be wrong about everything, even your own name [2]. Certainty, he argues, is what blocks further understanding [2].

Max Tegmark: the people who did not launch, and the ban that held

Tegmark, an MIT physicist who moved his research into machine learning, gives the most detailed history of how close the rivals came, and of what has worked since [3].

The Future of Life Institute, which he co-founded, gives an award to people most of the world has never heard of [3]. The first went to Vasili Arkhipov, whom Tegmark credits with the greatest positive contribution of anyone in modern history [3]. During the Cuban Missile Crisis, American ships dropped practice depth charges on a Soviet submarine that, unknown to them, carried a nuclear torpedo and had authority to launch it [3]. The crew was running out of air and battery power in extreme heat. The captain believed World War III had begun, and the political officer agreed [3]. Arkhipov was the third required vote, and he refused [3]. He had noticed that the explosions alternated left and right and concluded the Americans were signaling the submarine to surface, not trying to sink it [3]. Tegmark gives his reasoning in plain words: "I don't think the Americans are trying to sink us. I think they're trying to send us a message" [3].

Tegmark offers two explanations for that calm. Arkhipov had earlier helped contain a reactor accident on another submarine, which ultimately cost him his life to cancer, so nuclear catastrophe was not abstract to him [3]. And he was level-headed in the way Tegmark says the best pilots are, comparing him to Neil Armstrong landing on the moon while almost out of fuel without changing his tone of voice [3]. Arkhipov never bragged about it. The institute found his family living in relative poverty outside Moscow and flew his daughter to London for the award [3].

Later awards went to Stanislav Petrov, who judged a Soviet warning of incoming American missiles to be false, largely on gut instinct, and to Viktor Zhdanov and Bill Foege, who led the eradication of smallpox [3]. Tegmark notes that smallpox killed about 30% of those infected and says an estimated 200 million deaths have been prevented since it was wiped out [3].

Another award went to Matthew Meselson of Harvard, and his story is the one Tegmark treats as a working model. Meselson persuaded President Nixon that America did not want a 500-dollar weapon of mass destruction that any enemy could afford, and the biological weapons ban followed [3]. On the obvious objection, that a rival will cheat, Meselson told Tegmark that verification does not have to be perfect: "You don't need 100 percent. He said, one percent is usually enough" [3]. A rival weighing even a small chance of being caught, and the embarrassment and backlash that would follow, has little reason to cheat [3]. Tegmark says that in 50 years of the ban he knows of only two bioweapon incidents, both self-inflicted by the superpowers, which is why people associate biology with cures and not with fear [3].

He wants the same thing for artificial intelligence. He calls 2020 the best marketing year ever for autonomous weapons, because drones proved decisive in Libya's civil war and in Nagorno-Karabakh [3]. He is not a pacifist and supports strong defense, but he draws a line: "I think algorithms should never make life or death decisions" [3]. He suggests a ban built on stigma, perhaps allowing such weapons only above a high cost threshold, so that they never become as cheap as smartphones [3].

Tegmark is the one who refuses to let the good stories settle the matter. He says an accidental nuclear war between Russia and the United States is something "we almost have about a dozen times" [3]. Counting on the right person being in the right place fails eventually, because with repeated near-misses the odds of survival drop exponentially over time [3]. And the stakes have changed. Easter Island could collapse without ending humanity, but technology keeps enlarging the damage one mistake can do, while, as he puts it, "our planet is not getting bigger anymore" [3]. When Fridman offers his hope that brilliant people tend to be good and that humanity rallies at the last minute, Tegmark says he hopes so but warns that such faith breeds complacency, like skipping a smoke detector [3].

Where They Agree, and Where They Don't

All three point to the same period and reach a similar conclusion: bitter rivals with nuclear weapons still found ways to cooperate. Whyte's fusion summit and Rovelli's START treaties involve the same two leaders, Reagan and Gorbachev [1][2]. Rovelli and Tegmark both go back to the Cuban Missile Crisis, one for the leaders who stepped back and the other for the officer on the submarine who refused to fire [2][3].

They also agree on why it is necessary. Each describes a danger that no country can solve alone and that punishes everyone if it goes wrong: nuclear arsenals for Rovelli and Tegmark, engineered pandemics and autonomous weapons for Tegmark, and for Whyte an energy problem he says we owe future generations a solution to [1][2][3]. Whyte's worry about the Fermi paradox, that civilizations may have only a brief window before destroying themselves or falling back to 1800-era technology, is close to Tegmark's own bet on the same question, and both take it as a reason to try hard not to get this wrong [1][3].

And all three treat humility as part of the method. Whyte tells of the late 19th-century physicists who thought physics was nearly complete [1]. Rovelli says certainty blocks understanding [2]. Tegmark says the first step in being a scientist is never to fool yourself into thinking you understand something when you do not [3].

The differences are about how much weight the record can bear.

How well does big cooperation work? Whyte is the only one who has worked inside a large international project, and he is its most practical critic. ITER holds together politically, but it is slow, hard to steer and a single bet [1]. His response was a small team, private capital and parallel attempts [1]. Rovelli speaks of collaboration between nations as a choice of direction and does not get into how joint projects are run [2]. The two are answering different questions, one about whether to cooperate and the other about how to organize it, and Whyte's answer to the second is that cooperation should build shared knowledge and then leave room for several independent tries [1].

Is restraint a system or an accident? Rovelli reads the Cuban Missile Crisis as proof that leaders can step back [2]. Tegmark's version of the same crisis turns on a single vote in a submarine the Americans did not know was nuclear-armed [3]. Rovelli draws encouragement from the outcome. Tegmark draws a warning, because the outcome depended on who happened to be aboard [3]. Tegmark prefers arrangements that do not need a hero, such as the bioweapons ban, where the incentive not to cheat is built in [3].

Is cooperation natural? Rogan presses Rovelli on this. Rogan thinks war is hundreds of thousands of years old, pointing to chimpanzee wars, Comanche raids and cannibalism among tribes, and says this long history is why living in peace is so difficult [2]. Rovelli leans toward the view that war is fairly recent, arising once agriculture created stored resources worth stealing, and notes that bonobos, unlike chimpanzees, do not fight when groups meet [2]. He then accepts Rogan's premise for the sake of argument and turns it around. If our ancestors were that violent and we no longer are, then civilization and education have had "a very strong pacifying effect on us," and the learning "that there is an advantage in collaboration" can go further [2]. Tegmark adds a point from a different direction: he believes most people are fundamentally good, and that the bad things they do come mostly from the information they are fed [3].

Does sharing technology make things safer? Rogan raises the hard case. When nuclear power was distributed to other countries, he says, people did not realize how easily it could be converted to weapons, and he cites India and Pakistan [2]. That makes it complicated to offer nuclear power to developing countries at all [2]. Whyte's answer, from his own conversation, is that fusion is different: it has no chain reaction and physically cannot release its energy explosively, so it breaks the pattern in which each new energy source brings a more powerful weapon [1].

Why It Works

The sources give several reasons that cooperation between rivals has held where it has.

The problem was too hard for one side. Fusion was declassified because everyone was stuck [1]. Whyte calls it perhaps the most multidisciplinary field imaginable, and his refrain that it is really hard is his explanation for why progress anywhere is welcome [1].

There was little military advantage to protect. Whyte explains that so-called hydrogen bombs are at heart fission weapons that use fusion to boost their yield, and that this does not translate into an energy source. A fusion plant stays hot like an oven instead of running on a chain reaction, holds very little fuel at any moment, and cannot run away [1]. Science with no direct route to a weapon is easier to share.

Claims could be checked by measurement. The Soviet tokamak result was accepted after a British team measured it with a laser [1]. Meselson's rule for arms control is the political version: a modest chance of detection is enough to deter cheating [3].

A shared project gave leaders something to agree on. Reagan and Gorbachev chose fusion when the treaty details were stuck, as a signal that they would keep working on something important to both [1].

Stigma lasts. The bioweapons ban has held for 50 years by Tegmark's count, and he attributes that to the embarrassment and backlash a violator would face [3].

Experience of the danger steadies people. Tegmark believes Arkhipov's earlier reactor accident made nuclear war real to him in a way it was not for others on the submarine [3].

Cooperation is learned. Rovelli's reading of history is that civilization and education have already reduced human violence, which is evidence that the process can continue [2]. In Tegmark's conversation, Harari is cited for the idea that empires and money were themselves forms of collaboration that aligned the incentives of people who did not know each other [3].

Where It Applies, and Where It Doesn't

The pattern is strongest for problems that are open, slow and shared: basic science, a disease, an energy source [1][3]. It is strong where results can be verified and where a violator would be shamed [1][3].

It is weaker where speed matters. Whyte's point about S-shaped adoption curves is that fusion needs pilot plants on the grid by the early 2030s to matter for climate by 2050 or 2060, and a seven-partner project was not moving at that pace [1]. His preferred mix is public, international science as the foundation, with smaller focused teams on top [1]. Tegmark makes a parallel argument for AI safety by pointing to the Manhattan Project, which pursued all four uranium enrichment ideas at once [3].

It is weaker again where the technology itself can be turned into a weapon. Rogan's example of nuclear power leading to weapons in India and Pakistan is the caution [2]. Tegmark's autonomous weapons are harder still, because they could become cheap and because remote-controlled drones can be jammed, which creates pressure to remove the human entirely [3].

And it does not apply automatically to moments of crisis. The record there includes at least one case where the vote of a single officer decided the outcome [3]. Rovelli's concern about putting AI into the loop is that the machine may lack the human reluctance that saved those moments [2]. Tegmark makes the same point about a killer drone that would carry out anyone's definition of an enemy with none of the basic human reluctance to kill children and civilians [3].

What To Do With It

1. When a problem is too hard for one country, open it up. Fusion moved once the secret programs admitted they were stuck and declassified the work [1].

2. Check a rival's claims by measurement before dismissing or accepting them. The Soviet tokamak result changed the field after a British team measured the temperature themselves [1].

3. When talks stall, find a shared project. Reagan and Gorbachev used a fusion reactor as a symbol that they would keep working together, and the coalition it started is still building [1].

4. Use big collaborations to build common knowledge, then run several attempts. Whyte credits ITER's tools for SPARC's speed and argues that one planetary attempt is too few [1].

5. Keep teams small where speed counts. SPARC was designed by 20 to 25 people in about two years, and Whyte favors smaller, focused teams and splitting a problem into parallel parts [1].

6. Do not demand perfect verification before agreeing to a ban. Meselson's rule, as Tegmark reports it, is that a small chance of being caught is usually enough [3].

7. Keep a human in decisions to kill. Tegmark's line is that algorithms should never make life or death decisions, and Rovelli's fear is deterrence without human restraint [2][3].

8. Replace luck with preparation. Tegmark suggests setting aside a small share of GDP for systematic scanning of future risks, and points to South Korea's COVID response, which he attributes to preparation after SARS [3].

9. Keep communication open with adversaries. Rovelli's examples are the Kennedy and Khrushchev correspondence and the mutual withdrawals that ended the Cuban Missile Crisis [2].

10. Honor the people who chose restraint. Tegmark's institute found Arkhipov's family and gave his daughter the award, on the view that the examples worth copying are mostly unknown [3].

Sources

1. Dennis Whyte on Nuclear Fusion: Plasma, the 2022 Ignition Breakthrough, SPARC and the Road to Fusion Power · Lex Fridman Podcast
2. Carlo Rovelli on Relational Reality, Quantum Weirdness, AI Consciousness and the Case Against War · The Joe Rogan Experience
3. Max Tegmark on AI and Physics: Intelligible AI, Alignment, Autonomous Weapons and Consciousness · Lex Fridman Podcast