Jamie Metzl Puts the Lab-Leak Odds at 85 Percent

Lex Fridman asks for a number. Jamie Metzl, the author and former National Security Council official who has spent two years assembling a public dossier on COVID-19’s origins, does not flinch. He will stand by what he has said since the middle of 2020: it is more likely than not that the pandemic stems from an accidental lab incident in Wuhan. Is it 90 percent? 65? He calls those figures arbitrary. When he stacks every available fact, and when he admits that every fact on both sides is still circumstantial, the pile leans hard toward a laboratory accident.

He has no leak, no secret source, no classified annex. The method, he says, is logic and deductive reasoning. Before he argues for a lab incident he steels the other story, because the virus itself is natural in the deep sense: SARS-CoV-2 sits on a horseshoe-bat coronavirus backbone that has been evolving for a very long time. Viruses have been on Earth for some 3.5 billion years. They last because they change. The natural-origin hypothesis says this particular lineage jumped from bats into an intermediate mammal — a pig, a raccoon dog, a civet, a pangolin; all have been nominated — adapted in that new host, and then entered humans at a farm or a wet market, through a cut or an aerosol of blood.

That story is biologically plausible. The problem, Metzl says, is that nobody has found the evidence it predicts. Recent outbreaks usually leave a messy trail of Darwinian trial and error as a virus learns a new host. The first SARS did exactly that: genome sequences show it experimenting in early patients, like a startup trying product after product. SARS-CoV-2 arrived already adapted to humans, unusually stable, with no recovered history of that experimental phase. Chinese labs have sequenced more than 80,000 animal samples and still have not produced the intermediate-host story the natural-origin model requires.

The geography makes the coincidence worse. Horseshoe bats live more than a thousand miles from Wuhan. The natural-origin screenplay, in the version Metzl finds least credible, needs a single infection in an isolated southern village, a driver who covers that distance without infecting anyone, and a delivery that somehow reaches only the Huanan seafood market — a market that does not even specialize in the animals most often named as intermediate hosts. Strange things happen, Fridman reminds him. Metzl agrees. Probability still applies.

Fridman steels the market paper. Michael Worobey, writing in *Science*, had just argued that the cluster of early December 2019 cases around Huanan cannot be dismissed as ascertainment bias: if the spark were elsewhere, other clusters should have shown up. Metzl calls Worobey a thoughtful scientist and the work contributive. He does not think it is as decisive as the headlines. Worobey himself places the first human infection in October or November, before the December cluster. The viruses in those market patients were already mutating, which means they were not the original SARS-CoV-2. The infected people were not even in the part of the market that sold the candidate intermediate animals.

Early Chinese case-finding asked whether patients had been to the market, which baked in recall and sampling bias. Most important, Beijing still refuses to open the blood-bank records, early charts, and contact-tracing files that would settle much of this. Fridman supplies the image both men keep: it is like grasping at straws in the dark with gloves on. Metzl’s own reading is that Huanan was extremely likely as a point of *amplification*. That answers a different question from where the pandemic began.

Then there is the database. The Wuhan Institute of Virology held a password-protected collection of some 22,000 viral samples and sequences, a tool built for exactly this moment: when an unknown virus appears, compare it with what you already have. Collection of some of those samples had U.S. support through EcoHealth Alliance, on the order of $600,000 from NIH — not huge, Metzl says, and not nothing. In September 2019, months before the outbreak was public, the database became inaccessible. A few months later the whole thing disappeared. Shi Zhengli, the institute’s best-known bat-coronavirus researcher, said it had come under cyberattack.

Metzl does not believe that explanation, and he does not think she believes it either. The timing is too neat. If someone inside the institute or the Wuhan government noticed that a researcher was sick, or that something had gone wrong, the first move would be the Homer Simpson shuffle: back slowly out of the room and start hiding things. He reaches for a relationship joke. It is the boyfriend whose girlfriend wants to see his phone, who answers that he is currently under attack by Russian hackers. Fridman, whose family lived through Chernobyl, zooms out from Wuhan officials to human nature. Metzl’s working guess is that local officials squashed the news on day one, the national government then went all-in on a cover-up because the system cannot punish concealment without collapsing its own logic, and a small group of Western virologists stepped into the data vacuum and called lab-leak talk a conspiracy theory.

After SARS, France and China agreed to build China’s first and largest BSL-4 laboratory in Wuhan, plus mobile BSL-3 units, with French standards and fifty French experts on site. Wuhan, Metzl notes, is China’s Chicago: eleven million people, wealthy, educated, not a backwater of nightly bat dinners. The contractors swapped materials. Standards slipped. Alain Mérieux, vice-chairman of the project, refused to sign off, saying France could not support a lab whose safety it could not vouch for. When the new facility opened, one French expert remained. In 2018, French intelligence privately warned Washington, and a U.S. embassy team in Beijing sent a scathing cable, later obtained by Josh Rogin of *The Washington Post*: dangerous bat-coronavirus work was being done in conditions where a leak was possible.

Shi Zhengli — “Batwoman,” director of the center for emerging infectious diseases — was the scientific heroine of the first SARS hunt and the main Chinese partner of Peter Daszak, president of EcoHealth Alliance. Metzl’s most generous reading of Daszak is that he genuinely believed this was another natural spillover, feared that political attacks would wreck the collaborations needed to fight the next one, and tried to steer the public conversation toward cooperation. The indictment is harsher. In February 2020 Daszak drafted the *Lancet* letter that branded lab-leak questions a conspiracy theory, recruited signatories, and emailed Ralph Baric and Linfa Wang to keep their names off so the letter would not look like an inside job. He then organized Nobel laureates to protest after Trump cut EcoHealth’s funding.

What Daszak knew, and the public did not, was a March 2018 application to DARPA. EcoHealth, the Wuhan institute, and others asked for $14 million to collect the most dangerous southern-Chinese bat coronaviruses and genetically engineer them to insert a furin cleavage site — the spike-protein feature that helps a coronavirus enter human cells. DARPA declined. A year and a half later a bat coronavirus with a furin cleavage site unseen in that SARS-like family appeared in the city that held the world’s largest collection of those viruses and had already built chimeric versions of them. The proposed wet-lab work was to have been done in Baric’s lab at UNC; Metzl believes that lab did not do it without the money. He does not know what Wuhan did with a playbook it already had the skills, the viruses, and the motive to run.

Alina Chan, co-author with Matt Ridley of *Viral*, put the common-sense version in one image: someone applies to add horns to horses, the grant is rejected, and a unicorn shows up where they proposed to work. Either this is the first wild unicorn anyone has found, or the organization that had the equipment, the history, the funding, and the application went ahead. Metzl’s 85 percent is that stack, plus the Casablanca line he used on *60 Minutes*: of all the gin joints in all the towns in all the world, the outbreak of a SARS-like bat coronavirus happened not in the bats’ habitat but in the one Chinese city with the first and largest BSL-4, the biggest bat-coronavirus collection, a record of making scary viruses scarier, a rejected plan to add a human-entry cleavage site, a virus already primed for people, and a cover-up from day one.

Did Anthony Fauci fund gain-of-function research at the Wuhan Institute of Virology? Metzl’s answer is the one nobody wants on television: it depends on the definition. Under the common-sense meaning — lab work that gives a pathogen new abilities, especially ones that could seed a human pandemic — Fauci probably did. Under a narrow reading of a 2017 NIH document, a credible case can be made that he did not. Reports EcoHealth later filed, some late or missing, describe work on MERS that might qualify even under that narrow language. Rand Paul can be right and Fauci can be right, Metzl says. The binary gotcha is the wrong question. The right ones are: what was actually being done in Wuhan, what role if any did U.S. money play, and what right do taxpayers and the rest of us have to the files?

The scientific fight is older than this pandemic. In 2011, labs in Wisconsin and the Netherlands used serial passage to turn a deadly but poorly transmissible bird flu into a virus that was both. Marc Lipsitch and Richard Ebright argued that you cannot invent an unlimited number of monsters in order to study monsters; eventually they get out. Daszak’s camp, with support from Fauci and Francis Collins, argued you can safely collect the world’s most dangerous viruses and poke them to predict pandemics and build vaccines. A U.S. funding pause ran from 2014 to 2017. It did not stop the EcoHealth awards. Metzl’s operational rule is simple: do not send a dollar to a virology institute whose records and databases you cannot audit.

Fridman is angrier about a smaller sentence. Fauci has said he represents science. Metzl knows what Fauci means in a political brawl, and he will not erase the HIV work or the vaccines; Metzl himself took Moderna. Words still matter. Fauci is one scientist among millions, not the method. Great scientists, Fridman says, live in the humiliation of how little they know. Metzl wants the same posture from NIH: here is what we know, here is the uncertainty, here are the mistakes. In an information world this fragmented, a material omission will be found, and the finding will wreck the rest of the message. Trust-me-because-of-who-I-am, he says, does not fly anywhere anymore.

He launched his origins site in April 2020. When Matt Pottinger, then deputy national security adviser, wanted allied officials to look at the question, he sent them Metzl’s website, not a government memo. Listening to Collins on Fridman’s show, Metzl heard a welcome call for an investigation that still assumed the answer: find the intermediate host in the wild. That is a 90/10 tilt toward nature, the mirror image of Metzl’s own. The sentence he wants from everyone is narrower. We do not know conclusively. The circumstantial evidence, in his view, favors a lab incident. Additional information could flip it. And whatever the origin, the Chinese government’s cover-up — still running, most intense in the first months — disgraces the memory of the dead. Using *The Economist*’s count, he puts that number at 15 million.

Joe Rogan told Fridman that asking hard questions is not his strong suit. Fridman says Rogan is 100 percent right, and that the skill is still worth learning: sometimes a devastating claim is smuggled through as an aside, and you have to stop the tape without turning the moment into a takedown. COVID, he thinks, pulled the worst out of people who should still be talking, including Brett Weinstein and Sam Harris. Metzl, who was among the handful raising origin questions early — Weinstein was another — agrees that disagreement is not a reason to exit the conversation. His own split life makes the point: criticism of Beijing is popular; “let’s build a better future” gets three likes, one of them his mother’s. He still goes on Fox several times a week because speaking only to people like himself is comfortable and, he thinks, a contribution of nothing.

Does he think Shi Zhengli is lying? Yes. Forced to lie? Yes. He respects her as a scientist; he has never met her. He does not buy the cyberattack story for the database, and he does not think she buys it. At a poorly locked-down Rutgers Zoom forum in March 2021 he sat at home in a neon-green fleece, discovered that anyone could chat anyone, and ended up asking the last question of the night: did she know all the work at the institute, not just her own, and did she deny any Chinese military involvement? She called the military claim crazy and said absolutely not. Metzl thinks that was a lie. He also thinks she could not have said the true alternatives — yes but I cannot discuss it; there is much here I do not know; staff go through classification training — because those sentences are not available to her.

He describes two keys. The science key, used alone, treats Shi as a colleague in Switzerland or Norway, assumes the Chinese state is not in the room, and trusts the papers. The suspicion key, used alone, treats anything from China as guilt. He needed both: WHO work and books on one side, Asia expertise and NSC time on the other. Fridman wonders whether what lives in her head is fear or habit, the Chernobyl momentum of wearing clothes because that is what one does. Metzl’s larger claim is colder. A significant share of the 15 million dead, whatever the origin, died because alarm-raisers were suppressed. Our fates, he says, are tied to everybody else’s. He does not think Shi, or anyone in the Chinese government, would sit for an uncontrolled interview. There is a gag order on Chinese scientists. There is still no whistleblower address at WHO.

Xi Jinping’s project, in Metzl’s reading of Rush Doshi’s *The Long Game*, is to make China the leading country by the hundredth anniversary of the modern Chinese state, which means subverting the postwar order rather than sharing it. Censorship has to be petty to work. The Winnie-the-Pooh meme, Peng Shuai’s disappearance after accusing a Party elder, Jack Ma’s vanishing after a mild regulatory critique, Fan Bingbing’s erasure for insufficient loyalty: the lesson is that rank will not save you. The goal is to move the censor into your own head. Metzl has seen the end-state in North Korea, where a six-person delegation advising on special economic zones ended with him telling a hall of 500 that the plan could not work, then teaching a circle of officials the hora to a hummed *Hava Nagila*. If they had YouTube, he says, it would have been a hit. He does not expect that clip, or a Shi interview, anytime soon.

Patient zero and infectee zero may not be the same person. The first human infection could have been asymptomatic and never recorded; the first diagnosed hospital case is a different event. Metzl’s list of infectee-zero candidates includes a remote villager (almost impossible, not quite), an infected animal shipped only to Huanan (too many coincidences), a Wuhan Institute or Wuhan CDC researcher who collected samples in the south and came home silent, a waste-disposal failure — the Wuhan CDC had put out a bid to fix its waste system and, in early 2019, moved across the street from the market — or a lab animal sold at market, a practice China only recently banned. The whole laboratory-related category is, for him, the most likely bucket.

In 2012, six miners sent into a copper mine in Mojiang, Yunnan, fell ill with COVID-like disease; three died. Their blood went to Wuhan. Later collecting trips brought back RaTG13, about 96.2 percent similar to SARS-CoV-2, plus eight or nine other viruses from the same mine. That similarity is still a large genetic gap; Metzl is not claiming RaTG13 is the backbone. Bat viruses recombine constantly, so a mine that held RaTG13 likely held cousins, and EcoHealth documents show collection plans for Laos and Cambodia with Wuhan as the destination. Peter Daszak has cited the Laotian virus BANAL-52 as proof against a lab origin. Metzl hears the opposite: those viruses were being brought to the institute. If Wuhan held SARS-CoV-2 or a reasonable precursor before the outbreak, that would prove the lab-incident hypothesis. He thinks that is almost certainly why there has been no real investigation. China’s virus-hunters are among the world’s best and they sit in Wuhan. When WHO arrived, officials said they had not even tested the blood bank. Best case, the government is blocking its own scientists. Worst case, someone already knows.

WHO, founded in 1948, is in Metzl’s view essential and structurally weak. It has no independent surveillance network; member states will not allow one. It controls only about 20 percent of its own budget. Tedros Adhanom Ghebreyesus, elected with Chinese backing, repeated Beijing’s claim of no human-to-human spread — as, Metzl notes, Donald Trump also repeated Xi’s private assurances. China sat on the sequenced genome for two critical weeks until a scientist leaked it and was punished; Moderna had a vaccine formulation 48 hours after the sequence was out. Beijing blocked a WHO team from Wuhan for more than three weeks, destroyed samples, jailed questioners, and imposed a gag order. When Australian prime minister Scott Morrison called for a real investigation, China attacked him personally and hit Australia with trade sanctions, a message to every other capital.

The World Health Assembly then passed, with Chinese support, a joint study of *zoonotic* origins — one hypothesis, not an investigation of the pandemic. China vetoed the international roster; the Trump administration’s three lab-incident experts were rejected; the one American who got in was Daszak. The team spent two of its four Wuhan weeks in hotel quarantine, then about ten working days on a bus, including a propaganda exhibit on Xi’s success. They could not audit the institute or see raw data. On 9 February 2021, Peter Ben Embarek called a lab origin “extremely unlikely” and not worth investigating, a show-of-hands vote taken with Chinese officials in the room, and floated frozen-food spread that Chan mocked as “Popsicle origins.” Months later, on Danish television, Ben Embarek said he had been lying: Beijing wanted no lab language in the report, and he traded the phrase for that silence. Metzl messaged Tedros that night; Tedros wrote back that WHO was determined to do the right thing, then publicly refused to treat the joint study as final and, later, created SAGO over the governing board’s head. The Paris group of roughly two dozen experts, Metzl among them, pushed the list more open. Marion Koopmans, who had called lab leak a “debunked conspiracy theory,” disappeared from it.

Fridman asks the incentive question. If a lab accident on your soil has already cost the world trillions and millions of lives, why would any government hand over the proof? Metzl splits the problem. What China should have done is easy: what South Africa did with Omicron — announce an outbreak, ask for help, let scientists speak, share raw data, let WHO in on day one. What would happen if the lab story were proved is ugly. Xi has enemies and unlimited terms. Congress would talk about wiping out U.S. debt held by China. India and others would demand reparations. The global financial system could take a hit. Chernobyl played a role in the end of Soviet power, and Beijing knows it. A few weeks of fog, Fridman suggests, is human. Two years of systematic concealment is not.

Metzl writes a fantasy press conference anyway. We were doing this work because of SARS. We worked with France and the United States. We covered early because we believed the spillover story and feared the investigation would be used against us. We were wrong. The accident was ours. We will fund a stronger WHO, accept a treaty that can pierce absolute sovereignty in a pandemic, and put a vast sum — he floats five trillion dollars as a starting number — toward victims. He knows it is a speech no one will give. He still thinks you cannot build a future prevention system on top of this obfuscation, especially while China plans BSL-4 labs in every province and, in his view, tries to leapfrog in AI, autonomous weapons, genome editing, and animal experiments by skipping the intermediate steps.

Would the United States behave better? Metzl’s thought experiment: a “Texas horseshoe bat” exists only in Texas; the Montana Institute of Virology holds the world’s largest collection of those viruses, including lineages from a prior pandemic; people in that Montana town fall ill with a genetically related syndrome; there are no horseshoe bats in Montana; the government says nothing to see here. Every government has the same instinct. The American difference is the annoying machinery — Congress, an opposition, a free press, people like Rogan and Weinstein and Rogin — that makes a systematic cover-up harder. Metzl is a Democrat who worked on Clinton’s NSC. One of his closest friends in the Senate is Kansas Republican Roger Marshall. They disagree on almost every policy and talk constantly about origins. When he was in the executive branch, congressional oversight felt like a pain. Looking back, that pain was the system working.

Fridman tries to name the more important conversation: if it leaked, how likely is a cover-up? Metzl’s answer is 100 percent. There is a legitimate debate about origins; he respects people such as the British virologist Stuart Neil who will have it in good faith. There is, in his mind, no credible argument that there has not been a cover-up. Citizen journalist Zhang Zhan went to Wuhan, posted video, was jailed for “picking quarrels,” and by the time of this recording was on a hunger strike near death. That concealment, he argues, is responsible for a significant percentage of COVID’s total deaths. The origin can wait for evidence. The cover-up is already in the regulations, the missing files, and the people who disappeared for asking.

The conversation pivots to Metzl’s book *Hacking Darwin*. A single ejaculation can carry up to about 1.2 billion genetically unique sperm; each is a different possible person. Earth has hosted perhaps 100 to 110 billion humans. Every one of those unused genomes is an alternate life, an alternate Hitler or Einstein, the kind of forked city Italo Calvino imagined in *Invisible Cities*. IVF already moves conception into a lab. Pre-implantation genetic testing, after five days, sequences a few cells from each early embryo and lets parents screen out deadly single-gene diseases before choosing which of, say, ten embryos to implant. As whole-genome databases grow and machine learning reads the patterns, selection will not stop at Tay-Sachs. Height, eye color, and, within ranges, personality and IQ-related traits will come with probabilities attached.

Shinya Yamanaka’s 2012 Nobel work on induced pluripotent stem cells points further. In animal models you can take a skin cell, wind it back to a stem cell, push it toward an egg, and, given the surplus of sperm, fertilize thousands of eggs, run PGT on all of them, and choose. Metzl’s *New York Times* sketch of a 2050 clinic is not “do you want the child with the lethal mutation.” It is a menu. Genome editing — CRISPR is only a sliver of that pie — then lets you alter the embryo you picked. He had predicted the first CRISPR babies would be born in China; He Jiankui’s 2018 announcement forced him to pull *Hacking Darwin* from production and add a few sentences. The hardback had been right about the where. The goal, HIV resistance, he still thinks was misguided and dangerous.

Fridman worries that IQ-on-a-menu hardens a genetic ruling class. Metzl steps into the quagmire carefully. IQ is a perfect measure of IQ, not of all intelligence, but it correlates with income, longevity, and stable relationships, and the various IQ-like tests correlate with one another. The deeper objection is Darwinian. Diversity is not a preference for moths of different colors; it is the sole survival strategy of a species, and you cannot know which diversity the next environment will reward. Ask a T. rex what to select for and you get teeth and a roar. From Earth’s point of view the winning move was to be a cockroach, an alligator, or a shrew. There is no better or worse in evolution, only better or worse for a given environment. When the environment changes, last year’s optimum is next year’s liability. Sickle-cell carriers and malaria are the textbook warning that even “disease resistance” is not a free lunch.

The equity story cuts both ways. If a society really ranked people by genomic IQ prediction, Metzl guesses the people in those seats would not be Harvard legacies. They would be kids from slums and refugee camps whom the current filters have already thrown away. Genes do not follow zip codes. He is an Ashkenazi Jew whose father fled Austria; most of that side of the family was killed. He knows what it means when someone else’s definition of “good” does not include you. He still thinks a new way of procreating is coming, and that the choices will be about values, not techniques. Mating is already not random — Harvard is a filter. These conversations, he kept pushing inside WHO’s expert committee on human genome editing, are too important to leave to experts. Denmark’s public-engagement machinery is a model. Sex will not disappear, because it is enjoyable; sex-for-reproduction may recede as people conclude that the lab is safer. The danger he names is selecting for social norms so internalized they feel like eternal truths — IQ today, compliance tomorrow.

Jon Stewart’s Hershey-factory bit on Stephen Colbert’s show — a chocolate-scented outbreak next to the plant, and you want us to call it nature? — is, for Metzl, the lab-leak case in a form everyone could hear. In late January and early February 2020 he had laid the same logic on his website in too many words: a bat coronavirus more than a thousand miles from the bats, in the city with the largest collection of those viruses and a research program to make them more aggressive, arriving already primed for people. Stewart distilled it. Colbert’s discomfort, Fridman thinks, stood in for the scientific community’s. Metzl writes science fiction for the same reason: you have to reach people where they are. If you only ever say the proper thing, you lose the freedom to think in public. That is part of why he likes Elon Musk’s willingness to say ridiculous things. Scientists, both men agree, have a lot to learn from comedians.

Sanjay Gupta, a close friend of Metzl’s, went on Rogan’s show to lower the temperature, and Metzl thinks that part worked. He also thinks Rogan won the debate, for a structural reason that has nothing to do with who was right: Gupta brought averages and studies; Rogan brought a guy whose sister’s cousin had a bad reaction. Anecdote beats statistic on arrival. Gupta could have answered with a counter-anecdote and then the numbers. He was torn between winning and de-escalating. Collins, on Fridman’s show, had used a wrestler who supposedly almost died of COVID and recanted his anti-vaccine stance. Metzl says that story is wrong: the wrestler was fine. If you are going to use an anecdote, use a true one. If you are a scientist, you should not be leading with anecdotes at all.

Fridman’s rule for disagreement is steal-manning. You have not earned the eye-roll at a flat-earther, or at a vaccine skeptic, if you have not read their strongest case. Rolling your eyes is laziness, and lazy people do not deserve opinions. After the Gupta conversation Metzl asked Columbia’s Daniel Griffin to help collect vaccine concerns in the skeptics’ own words, then answer with evidence and links. Griffin sent the CDC’s Q&A. The questions were written by people already convinced. Metzl supports vaccines and masks; he still thinks a page that never lets people hear their own voice will not break through, even when it is right. A reasonable listener to Rogan and Gupta, he fears, left thinking shots are good but optional if you are young and healthy. That was not the case he wanted made. The virus, in his view, is much scarier than the vaccine.

Fridman keeps pushing the other side: mandates as a Trojan horse for control, ivermectin crowded out by a vaccine-only story, American suspicion of even beneficial overreach. Metzl grants the caution and the need to study ivermectin properly — Collins, even while mixing up the wrestler, had promised a high-quality trial, and Gupta was right to concede that calling the drug “for livestock” was disingenuous. He will not call the mRNA shots a sideshow. *The Economist*’s death count is 15 or 16 million; many more would be dead without the vaccines, a public-private push he traces through DARPA and NIAID to an 11-month working shot. Government underreach, he says, is a danger too. The American genius is that messy individual refusal is allowed. The American job is still to find the balance.

Metzl has finished 13 Ironman races. People say that is amazing. He says one Ironman is impressive; 13 means something is wrong with you and we should find out what. The lesson is always the same. At mile 120 of 140, after the swim and the bike, six miles into the run, he is sure he has nothing left and cannot imagine 20 more miles. There is always more. The ability to call on that reserve has to be cultivated. His friend Charlie Engle went from addiction and prison to running laps of the yard, then ultramarathons and books; the next plan is the Dead Sea to the summit of Everest, lowest point to highest. Metzl told him to start at the ocean floor and get Musk or Bezos to finish in the stratosphere. The point is not sport. It is how easily a guessed range becomes the actual range.

His advice for a first ultra is unromantic: pick a date six months or a year out, register, and work backward. Build strength so the miles do not break you — he plugs his brother Jordan Metzl’s Iron Strength work, burpees and functional load at the Hospital for Special Surgery. Increase mileage steadily, take rest days, listen when the body talks, especially if you are a mind-over-matter type. The 50-kilometer mountain races he actually runs can be gentler than a road marathon: you walk the steep bits, you eat blintzes and boiled potatoes every few miles. Then he registered for something in Taiwan called the Taiwan Beast, assumed it was another 50K, packed the required wilderness kit while mocking the drama, and started at 4:30 a.m. with a headlamp. It took 19 hours: fifteen cliff faces on a dangling string in the rain, twenty torrential river crossings, slopes you climbed and slid down, Sahara-500 veterans sprawled on the trail. Quitting, he told himself, was not an option unless the universe was trying to kill him.

He has never done drugs and does not drink. He does call himself a chocolate shaman, a title with no accrediting body, which he discovered after a cancelled Berlin keynote sent him to a cacao ceremony of hippies who spun in a circle for 45 minutes. He lasted ten, left, and the next day saw photos of the mostly naked chaos that followed. A saner ceremony two days later sold him on the idea: drink chocolate, and people become free. Back in New York he looked for a credential, found none, trained himself, and declared the job. He has since led ceremonies at tech conferences, Soho House, and Rancho La Puerta. He always starts the same way. There is no such thing as sacred cacao, or sacred mountains, or sacred people. Nothing is sacred unless we ascribe sacredness to it. If we noticed that everything is, we would live differently.

The core of the ritual is gratitude. Someone planted the seed, watered it, harvested, fermented, roasted, made the cup. His friend A.J. Jacobs spent a year spreadsheet-tracking every person behind one New York cup of coffee, then traveled to thank them, including the worker who painted the yellow line so the truck would not crash. Fridman has a stuffed hedgehog on the table; Metzl turns it into a short course in globalization, design, factories, and families fed by a toy. You would go insane if you did this every hour, with every microphone. Once in a while, with one object, the world feels connected again. That feeling, he thinks, is not decoration. It is the psychological precondition for solving problems that no single country can touch.

Early in the pandemic, Singularity University asked Metzl to talk about whether genetic and biotech tools could match the outbreak. He woke up sure that was the wrong speech. Why had we been unprepared? If we somehow beat this virus and not the next one, or beat pandemics and not climate, or beat climate and then had a nuclear war, each victory would be Pyrrhic. The meta-category is a mismatch: the biggest problems are now global and shared, and our ability to solve that class of problem has not kept up. The Peace of Westphalia gave us the modern sovereign state; the late nineteenth and early twentieth centuries showed how unstable a system of jockeying sovereignties is; 1945 added an overlay — the UN, WHO, Bretton Woods — that pools some sovereignty without erasing it. Our reach has outgrown even that overlay. Unless the collective-action problem is faced, he told the audience, we will extinct ourselves.

The talk went viral. He stayed up the next night and drafted a Declaration of Global Interdependence, posted it, and watched that go viral too. A meeting drawn from his own email list pulled in people from 25 countries. In 35 days they broke into eight working groups, put the declaration into 20 languages, wrote a work plan, and founded One Shared World. The site is oneshared.world. Participants now come from 120 countries; public events draw millions of viewers. The point is not to abolish states or the UN. It is to build a constituency that demands leaders balance wider interests against the national and corporate mandates they were hired to maximize. For formerly colonized societies, Metzl is careful to say, sovereignty is not a Westphalian abstraction. It is the thing Europe took and they had to seize back. Asking them to pool it is a different request than it is in Brussels.

The method is to grow out of movements that already exist rather than invent a new brand. With Model UN, the Aga Khan Foundation, and the India Sanitation Coalition, they ran multilingual debates on a concrete goal: safe water, basic sanitation and hygiene, and essential pandemic protection for everyone on Earth by 2030. Young people wrote the plans; friends including Hans Corell, former chief legal counsel of the United Nations, helped turn them into something that looks like a UN resolution; the text went to every permanent representative. Language drawn from that process made it into a G20 communiqué in Riyadh and, with German and Spanish support, was on the verge of a UN General Assembly resolution. Passing a resolution, Metzl admits, changes little by itself. It tests a model: give people outside traditional power a structured way to speak, not just a concert and a waving phone.

Greta Thunberg’s “blah blah blah” after Glasgow is, he thinks, a merited critique of empty process and a missed instruction. Young people have never been more educated, connected, or able to organize. Complaint without a draft, a coalition, and a delivery mechanism is not enough. Fridman offers crypto as another on-ramp: number-go-up as the gateway drug to the feeling that a young person can challenge a monetary system and, maybe, a nation-state. Metzl is more interested in the layer under the coins, the distributed ledgers and the habit of coordinating without waiting for permission. The nation as the only serious actor, both men suggest, may be a twentieth-century tool that still has uses and has already shown, in the COVID response, how secrecy and censorship can cost trillions and countless lives.

What gives him hope is not a mood. It is the young people he works with every day through One Shared World, the ones who did not know a UN resolution was supposed to be impossible. He calls himself a Gramscian in the narrow sense that frameworks and norms decide which leaders can emerge, and that the job is to make it easier for the ones who call on our better instincts. Given half a chance, he thinks, people want to be good. We are a species that learned to slaughter at scale and also wrote philosophies of love. Enough infrastructure — not control, just enough that people feel like stakeholders in a common story — and he believes the sky is the limit. Advice to a kid in high school or college is the same advice he gives about science: the knowledge base will keep shifting, so learn how to learn, but the core is knowing who you are and what you stand for. That is the path to a life that does not feel alien as you age. Love yourself through the wrong turns. At the Ironman, when you are sure you cannot go on, you can. When you are sure nothing better will happen, it will. You just have to keep going.

The 15 percent he withholds from the lab-incident number is not a hedge for the cameras. Natural origin remains a live possibility. If an animal sample from a Southeast Asian market turned up a virus close to SARS-CoV-2 with the same furin cleavage site, he says he hopes he would be open-minded enough to move. Two years on, 80,000 Chinese animal sequences have not produced that virus, nor has the rest of the world’s search, nor has an intermediate host. Meanwhile Wuhan held the largest bat-coronavirus library, had already engineered viruses in that family, had applied to insert the cleavage site, occupied a BSL-4 the French would not sign, took a database offline in September 2019, and ran an information blockade that has not stopped. Each fact can be a coincidence. Together they are the gin joint. All of it is still circumstantial. If the truth is zoonosis, he will take it. Until then the number is 85.