The Dollar Paper Strip That Could Have Turned Covid's Curve

Lex Fridman opened by asking Michael Mina, a Harvard professor of infectious disease and immunology, for the most beautiful or surprising idea in the biology of humans and viruses. Mina went not to Covid but to the work that had occupied him before the pandemic: how one infection or vaccine changes susceptibility to pathogens that look unrelated. The system that still fascinates him, he said, is measles.

Measles enters a child's body and the immune system does what the textbook says it should: it grabs the virus and carries it into the lymph nodes so the body can learn it. The virus plays a trick. Instead of being eaten, it takes over those cells, uses them as a Trojan horse, proliferates in the very cells meant to kill it, and spreads into bone marrow. In monkey studies, Mina's group watched it move through the body. He told Fridman you can watch it “just destroy and obliterate people's immune memories,” the way a parasite might destroy a brain.

His research conclusion is stark: before measles vaccination, that one virus may have been responsible for as much as half of all infectious-disease deaths in children, because it wiped immune memories of influenza, pneumococcus, and everything else a child had already survived. From measles's point of view, Mina thinks the amnesia is coincidence, not strategy—it simply found a good place to replicate. Influenza takes a different shortcut. Flu proteins cleave sialic acids on airway cells and expose new attachment sites; bacteria that want the same receptors piggyback, because viruses replicate a hundred or a thousand times faster than they can. That is why a flu infection is so often followed by severe bacterial pneumonia.

Asked whether that elegance fascinates or terrifies him, Mina said he is a scientist first: it fascinates him far more than it scares him. What does scare him is how easily this pandemic could have been worse. The 2009 H1N1 outbreak was, in his telling, another near miss. It takes one wrong mutation, one rearrangement, one new host. SARS-CoV-2, he argued, landed in an awkward place. It is not quite serious enough at the individual level to force every human into maximum urgency—plenty of people still called it a hoax—yet it is bad enough to collapse economies, fill hospitals, and kill on a vast scale.

Fridman noted that the virus is not intelligent, but that social media, politics, mask refusal, and vaccine hesitation have become entangled with its biology. Mina agreed. Choosing not to counter the virus in an organized way, he said, simply gives it more room to escape. The world was about to run one of the most aggressive vaccination campaigns in history, and it was doing so at the peak of transmission, with millions still getting infected.

Inside a single infectious person there may be trillions of viral particles. Multiply that by millions of infections and the planetary pool is measured in quadrillions. Applying incomplete immunological pressure to that pool, when every vaccine still aimed at the same target, means it takes only one escape mutant and the variant has the whole globe. If the same campaign had waited until virus was scarce, Mina said, there would simply not have been as many chances to sneak around immunity. In December 2020, that warning sounded like a memo written for the year that followed.

Fridman turned to the work that had made Mina a public voice: a Time essay from November 17, 2020, titled “How We Can Stop the Spread of Covid-19 by Christmas.” Mina's answer was not a new drug. Viruses spread because people spread them. Vaccines might eventually block onward transmission—nobody yet knew if the new shots would—but another tool had been sitting on the table since May, June, July, August, September, October, and November. The United States, he said, still chose not to use it.

He opened a cartridge on the table. Inside was a paper strip. Used twice a week at home, he said, it could tell someone with on the order of 99 percent sensitivity and specificity whether they had infectious virus in them. Manufacturing 10, 20, or 30 million a day was, in his view, not a moonshot: “We make more bottles of Dasani water every day.” He had used one that morning before sitting down to record. It was not perfect. It cut his already low odds of being infectious by about 99 percent.

Then he walked through the arithmetic slowly. For every 100 people infected at that moment, they went on to infect perhaps 130 more. A couple of days later those 130 became about 165; over three weeks, 100 became 500. That is exponential growth, familiar in theory and rarely run in the head. Reverse it, and you do not need to go from 100 to zero. If those 100 infect 90 instead of 130—if you remove 30 to 40 percent of new onward spread—R drops below one and growth becomes decay. Do that for a few weeks and the curve that ran from 100 to 500 runs from 100 down toward 20.

The program is not a lockdown. It is information: if the strip is positive, stay home and do not visit Grandma. People will skip tests, tests will miss some cases, some positives will still go out. Mina's point was that perfection is the wrong target. Enough people learning they are infectious on their most contagious days fills the 30-to-40-percent gap. Fridman loved the politics of it: unlike a mandate, a test empowers you and then leaves the ethics to you. Mina said that is exactly why it grated on officials who wanted the data in their own hands.

Why, then, had the tests not shipped? Mina blamed a paternalistic public-health culture that lets perfection block action, and a medical industry that only knows how to see a test as a diagnostic device. Rapid antigen strips kept being compared with laboratory PCR that can cost around a hundred dollars and take days to return. In every practical scenario, he said, the paper test is the more effective public-health tool. He was blunt on the microphone: he was “so fucking pissed that these tests don't exist.” In the same months the country had designed a new vaccine, run three trial phases, scaled production, and pushed freezers through UPS and FedEx. A paper strip was dismissed as too simple, too cheap—unless billions of dollars sat on the other side of the problem.

Cost is not the constraint. Mina held up a strip he said costs less than a dollar to make, perhaps a dollar fifty with the swab, profitable at three dollars and richly so at five. The Abbott BinaxNOW on the table was the one the U.S. government had bought: 150 million tests at five dollars each, $750 million total, with 99.9 percent sensitivity and specificity in the figures he cited, and Abbott not losing money. Anyone who claimed these tests were expensive, he said, was looking at the object in his hand. The real bottleneck is a nitrocellulose membrane, not the paper, and he wanted the Defense Production Act used to build that supply.

The FDA, however, was authorizing the tests only as medical devices. Once a company has that label, insurance will pay. A strip that costs fifty cents to make can be billed at thirty or a hundred dollars. Rational firms then make fewer tests at high margin instead of tens of millions a day at a dollar of profit. Mina wanted the government to buy 20 million a day the way it had bought the Abbott lot, and get them into homes. Covid had already cost the country trillions; restaurants and main streets were shuttered. Twenty to forty million tests a day, he said, would be enough to push R below one. Household pooling—two people, one tube, one strip, split only if positive—could buy a two- or three-fold efficiency gain in schools, offices, and testing lines.

That volume, he insisted, was already being built—just not for Americans. A California company he named Enova was producing five million tests a day and shipping them to the United Kingdom. SD Biosensor was making tens of millions a day for Europe, Southeast Asia, and East Asia. None of that flow was coming to the United States, because without federal authorization as a public-health tool the domestic market did not exist. Fridman asked what the monthly goal should be. Mina's working number, matched to what could be produced “today if we wanted,” was 20 million a day.

Mina spread several cartridges on the table and walked through how a lateral-flow antigen test actually works. You swab the front of the nose, dunk the swab in buffer, and put a few drops on a nitrocellulose strip. Capillary action pulls the liquid along, the way paper drinks water from a cup. Antibodies printed in a line grab viral protein if it is there; gold colloid darkens the line. One line is negative, two is positive, in five to fifteen minutes—the same form factor as a pregnancy test, and the same one several of these companies had already used for malaria, with the antibodies swapped for SARS-CoV-2. He preferred the versions that skip the plastic shell entirely, because tens of millions of cartridges a day would become a mountain of waste.

The testing landscape, he said, has three families. Serology looks for antibodies, a record of past immune response, usually from blood. Everything else on the table looked for the virus itself, either its RNA sequence (PCR and PCR-like tests) or its protein structures (antigen tests). He likened PCR to reading someone's DNA and an antigen test to recognizing a face. Detect, a rapid RNA device from a company associated with Jonathan Rothberg, sat in the first camp: more expensive, heater required, useful as confirmation. Antigen tests sit in the second. They generally turn positive only while virus is actively replicating.

That distinction is the whole fight over “sensitivity.” After infection, virus explodes to trillions of particles in a few days, then the immune system beats it down. What remains is wreckage: RNA fragments floating in the airway like blood at a crime scene after the people are gone. Ultra-sensitive PCR will keep finding that RNA for weeks or months. Antigen tests go dark as soon as replication stops. Mina had tried to publish on this and to get it into the *New York Times*; doctors and agencies did not want the headache. Conspiracy accounts then grabbed the observation and called the extras false positives. They are not, he said. They are “late positives,” no longer transmissible.

He draws a line between diagnostic sensitivity and contagiousness sensitivity. A physician reconstructing why a patient was sick two weeks ago wants every shred of RNA; PCR is the right tool, a detective at a crime scene. Public health only looks forward. It needs the people who can infect someone today, who may be infectious for five days and PCR-positive for three to five weeks. Evaluate an antigen test against all PCR positives and you will be told it is 50 percent sensitive, because most PCR-positive people at any moment are already post-infectious. Evaluate it against people who can still transmit and Mina's figure is north of 99 percent—about 99.8 percent of currently infectious people in the numbers he cited. The people the strip “misses” are not the people who need isolation.

They ran a BinaxNOW live on the show. Six drops of buffer, a few circles in each nostril, the swab pushed into the cartridge and turned three times, the adhesive peeled, the card closed. A purple line crept up the window and the control line converted from blue to pink. The package says wait fifteen minutes; Mina noted that a truly high-viral-load positive often appears in a minute or two. Fifteen minutes later there was still a single line. Fridman was negative.

False positives on that Abbott card, Mina said, run around 0.1 percent—one in a thousand, perhaps one in five hundred. His protocol for a positive, especially at an airport or other entrance screen, is not an automatic ten-day isolation. Confirm immediately on an orthogonal test. Detect looks for RNA rather than antigen, so there is no biological reason both should be falsely positive. If the antigen card is positive and Detect is negative, treat the first as a false alarm. If someone is highly infectious, he added, you do not wait the full quarter hour to know.

Then he dissected Ellume, a polished antigen test with LEDs, a fluorescent reader, a circuit board, a battery, and Bluetooth that pairs with an iPhone and can auto-report to a health department. He called it a genuinely nice medical device and the wrong public-health tool. You do not manufacture twenty million of those a day and throw them away. The paper strip, he said, is not less accurate; it is simply paper. The rest of the world already had the simple version at scale. The United States, in his phrase, could not get out of its own way.

In November 2020, Elon Musk tweeted that he had taken four rapid tests in a day on a BD Veritor instrument, gotten two positives and two negatives, then a PCR that was a low positive. The episode was widely cited as proof the tests were junk. Mina read it the other way. Musk, he guessed, had tested at the tail of his infectious window, when viral load sat near the antigen limit of detection. Results there become stochastic: two and two is what you expect. Two days earlier the same cards would likely have been “screaming positive.” PCR would stay positive for weeks because RNA debris remains.

That is a confusing signal if you are sending astronauts to a space station that morning. It is a useful signal if you are one person in a country of hundreds of millions: even one positive is a reason to stay home for a couple of days. Because the strips can be made at huge scale, Mina did not want a single positive to mean ten days off unpaid. Test daily, he said; once you have been negative for 24 or 48 hours, go back to work. The people the cards catch most reliably are the people most infectious—the ones who actually move the epidemic.

He had already had a brief Twitter exchange with Musk and wanted a real conversation. A lot of the production machines are bought from South Korea or Taiwan; they are not exotic. Someone with Musk's engineers and capital, Mina argued, could build the machines from scratch in two weeks, even without the federal government, and the return to the country would be in the hundreds of billions if society could reopen. He framed the whole problem as engineering, not medicine. Physicians were making public-health decisions in a pandemic that needed engineers. He wanted to found a field he called public-health engineering, ideally as a department at MIT.

Vaccines rolling out, he warned, were not a reason to drop testing. Uptake was uncertain—Fridman cited a figure that 40 percent of Americans might refuse the shot—and efficacy had only been measured in the first two or three months, with biological reasons it might fall. Nobody knew if the vaccines would block transmission, which is what herd immunity actually requires. Some shots needed deep cold chains and had short lives outside them; hospitals were saying they lacked staff and were getting smaller allocations than promised. Right now, besides several months ago, was the best moment to flood the country with rapid tests. A presidential order, or an NIH and CDC stamp outside the usual FDA device pathway, could have moved them “tomorrow.”

Fridman asked how you change the policy. Mina had been talking to senators, governors, mayors, and, informally, to both the sitting president's and the president-elect's teams. Almost everyone except the FDA was receptive. In July or August he had sat in a roundtable with Los Angeles mayor Eric Garcetti and the manufacturers and asked how to get the tests out. It went nowhere, because the tests still were not authorized as public-health tools. The irony, he said, is that this was one of the few truly bipartisan ideas he knew: conservatives and liberals alike just wanted to know their status without a four-hour line and a result the following Saturday.

Lockdowns, in his view, are an emergency bandage, not a public-health solution, if public health includes psychological and financial well-being. Half-measures were destroying small businesses while the simple tool sat unused. Why would Cuomo, Newsom, Baker, DeWine, or Garcetti choose that? Because when a mayor asks what options exist *today*, unauthorized tests do not exist. No mayor was going to pick a fight with the FDA. Other agencies deferred too. What they were missing, Mina argued, is that the FDA does not even have a mandate to evaluate tests as public-health tools. Tim Stenzel, who headed in vitro diagnostics there, was doing his job: scoring medical devices.

The CDC, in Mina's account, never drew the distinction. He would tell Stenzel the strips were public-health tools; Stenzel would say, fine, go use them—then ask whether a result went back to a person. If yes, CMS, the Centers for Medicare & Medicaid Services, defines that as a medical device requiring medical authorization. Mina then asked Seema Verma, who headed CMS, whether the tests could be carved out as public-health tools. She said CMS had no jurisdiction over point-of-care rapid devices, only laboratory devices. Nobody owned the category, so the tests stayed in purgatory for the better part of a year.

His fix was a presidential order, or a CDC stamp against clear performance metrics. What the FDA was doing instead, he said, was eroding medicine to fake a public-health pathway: officials asked why anyone needed the tests without a prescription when a doctor could already write one prescription for an entire college campus. Repeat testing of a campus is not the practice of medicine. It is a workaround that empties out the legal reasons prescriptions exist, rather than admitting the strips are a population tool. “We have a regulatory landscape which was developed totally for good,” he said later, “getting in the way of us deploying a tool that could stop deaths.”

Testing was only half of Mina's engineering brief. In his lab he builds methods that take a drop of blood or saliva and profile antibodies against huge numbers of pathogens. The field-building project he wants is a weather system for viruses: open a phone, type a zip code, and get probabilities for why a child has a runny nose—Covid, rhinovirus, adenovirus, flu. People do not carry umbrellas every day, and they do not board up the house every day against a hurricane. They check a forecast, act in a burst, and go back to life. Pandemic response, he argued, could work the same way instead of a year-round wartime posture. Every flu season, by contrast, still begins blindfolded.

The central problem is samples and trust. Getting specimens, he said, is not actually that hard. Tens of thousands a week were already flowing into his lab. A Massachusetts company makes about 80 percent of the world's plasma-collection instruments; that introduction led him to Octapharma, which pays donors at sites across the country. He collects fully anonymous tubes—date and a zip code or county ID, no names—and had been tracking Covid antibodies from January through December, watching the virus enter the United States day by day. He planned to put a rough public map online. Offer people something back, as 23andMe does—an immunological history of infections, even allergies—and the same technology becomes both a consumer report and a sensor network.

Privacy is the knot. Samples can stay anonymous tubes. With consent, more fields could be attached. Mina also described a more uncomfortable property of the data: an immune profile is a biological fingerprint. Antibody generation is partly random; against one coronavirus a person may make hundreds of antibodies to different bits of the virus, and the combination across many pathogens is unique and stable for years. He can take two unlabeled blood spots a year apart and say they came from the same person at about 99.999 percent accuracy without knowing who that person is. DNA cannot reliably split identical twins; he said this tool is one of the few that can. Fridman has an identical twin brother. The same uniqueness is both a re-identification risk across databases and a way to keep names off the tubes.

He had been trying to fund a Global Immunological Observatory for years. Governments called it a cool idea and declined. He had written in 2015 about why it would matter. Had it been running on 2017 technology in 2019, New York would not have been blind for a month and a half or two months after SARS-CoV-2 started spreading—the lag before the first PCR tests. Andrew Cuomo, Mina said, had hints but no data with which to close the city. The observatory would not even have needed a Covid-specific assay. People's antibody patterns would have looked like a known coronavirus and yet not quite, with gaps and odd shapes. “You could have put it in an Excel spreadsheet,” he said. Algorithms would have helped; a chart would have been enough. He thinks that one difference could have saved tens of thousands of lives in New York City.

Fridman asked whether a future natural virus could kill on a scale civilization has not seen. Mina does not think human extinction by a virus is likely, partly because a 60 percent killer would trigger lockdowns faster than SARS-CoV-2 did. The darker scenario is a respiratory virus that transmits like this one but wrecks hearts a month later, or a pathogen so bad that grocery and shipping workers cannot stay at their posts. Ebola, he noted, used to burn in disconnected villages; 2014 showed what happens when roads connect those villages to cities. Mobility is new. Global aviation raises the odds even if lethality has a natural check.

His own ranking puts influenza above coronaviruses. Coronaviruses have one long genome and mostly accumulate point mutations. Flu has eight segments; two strains in one cell can swap whole cassettes and produce antigenic shift, not drift. “Any day of the year, you can have, boom, a whole new virus that didn't exist yesterday.” Industrial livestock has widened the human–animal interface while air travel has accelerated spread. He thinks it is a matter of time. Man-made viruses, he added, should sit “right up there with the greatest concerns for humanity,” in the same conversation as nuclear weapons. The tools to build a virus already exist. Most living people were never vaccinated against smallpox; a few years more and the planet is a fully susceptible pool, no novel design required.

Measles is the subtler case. Researchers are trying to build measles strains that evade immunity so the virus can be used as an oncolytic agent against cancer—a legitimate medical goal. Measles has an R0 around 18, orders of magnitude more transmissible than SARS-CoV-2. Loading immune-evasion into that chassis, even by accident, is a combination Mina thinks the species should treat as an alarm. His Harvard colleague Marc Lipsitch, director of the Center for Communicable Disease Dynamics, has been outspoken against gain-of-function experiments that accelerate deadly viruses in the lab so the world can be “prepared” for something that may never arise naturally. If the cost of a leak is catastrophic, Mina asked, is the rehearsal worth it?

Fridman raised AlphaFold's protein-folding breakthrough and the dual use of machine learning to design proteins—and, downstream, viruses—for cancer or for harm. Mina called AlphaFold amazing and treated it as proof that the world has rules even when humans have not written them down. Once you can go from sequence to fold, the next iteration is the reverse: choose a fold, emit a sequence, engineer from the ground up instead of tweaking nature. The first customers will be good ones, funded to save sight or kill tumors. A small tweak, he said, is also how you would imagine a virus that blinds a population. Search engines and social platforms already show how a tool that is “wholly good” gets leveraged by people who are not. He does not know how a species balances that. He does know the same pattern is already visible in medical regulation: a system built for good, now blocking a tool that could stop deaths.

Fridman asked what he would tell students who want to solve large problems. Mina hesitated: advice drawn from his own life may not transfer. What he does not do is treat barricades as final. In January 2020 he was a medical director helping oversee molecular virology diagnostics at Brigham and Women's Hospital. Wearing an epidemiologist's hat, he judged that the outbreak in China was already a pandemic, even if CDC was not ready to say so. Hospital leadership asked why they would build a test, and who he was. He kept pushing. A couple of weeks later the Biogen conference became a superspreader event, and the hospital's posture changed.

He then argued that hospital capacity would not be enough and that the Broad Institute should retask its high-throughput genomic robots for Covid testing. Everyone in his hospital orbit called it impossible. A new faculty member, his superiors told him, was too naive to know that. The Broad went on to become the highest-throughput testing laboratory in the country. “Most people in the world don't see solutions. They just see problems,” he said, because listing how a plan will fail does not require inventing anything. The Broad conversion, to him, was “the most dead simple thing to do.”

The same pattern showed up in April, when he published that laboratories should report PCR CT values—the quantitative readout—not just a binary positive. Directors told him the numbers were not accurate enough. Months later Florida mandated reporting them, and Mina said they were becoming headline news because they help both clinical care and public-health interpretation. His education is deliberately broad: engineering, physics, a stretch as a Buddhist monk, then medicine, immunology, mathematics, policy. He does not claim to be the best in any of those rooms. “My superpower is bringing them all together.”

The recommendation he would actually give is to get out of problem-only mode, even if that sometimes pisses people off, and to treat the word *barrier* as evidence that a solution exists. If there were no solution, it would not be a barrier; it would be a law of physics. Most people will keep handing you barriers until you start to believe that is all the world contains. Rapid testing, he said, is the most dead-simple solution in the world, and countries that took it up were already changing course while the United States still was not. Think big anyway. Big ideas attract people with fire in their eyes, and they compound.

The quietest stretch of the conversation was the monastery. Mina was at Dartmouth, studying engineering, already interested in under-resourced countries and in Buddhism's sense that present actions ripple further than anyone can see—rebirth included, a doctrine he never fully accepted. Rowing at a national level had felt meditative; so had pottery. He went to Sri Lanka for two months, shadowed a rural clinic whose aging physician had not updated his training in decades, and watched a girl with shrapnel in her hand opened without anesthetic so the wound could “air out.” Mornings in that clinic, evenings trying to teach monks English, he was, by his own account, a terrible English teacher. They sat him down to meditate instead.

He did not come home. He moved into a monastery built into rock on a mountainside, because living there was free, and meditation went from hours to fifteen or eighteen hours a day. He did not want to sleep; the day's work felt unfinished, even though the work has no end. The first stage is simply to sit, watch the breath, and keep other thoughts out—a gate he says most people, including present-day Mina, never get through. Past that came a trippy, slightly euphoric stretch, and then the stage he still misses: mindfulness so fine-grained that picking up a glass of water stops being one action. It becomes a chain of small engineering feats and sensations—cold glass, lift, drag, swallow. Life, he said, is a language you have not learned. A foreign tongue sounds like an unbroken roar until spaces appear between the words and the stream slows. The same thing happens to a day.

It ended with the 2004 Indian Ocean tsunami. On the Sri Lankan coast, about 40,000 people died in a few minutes. Mina went to the shore. What he could not process was the expectation around him: people wanted him to remain a monk, keep meditating so they could bring offerings and attend to karmic duty, while most monks had already fled the coast. He took off the robes, moved into a refugee camp for some six months, raised money from the United States, and did work he admits he did not know how to do. He was 22. That turn, he said, is what pushed him into public health.

Fridman asked what any of it means, given that Mina studies things that could end a civilization. Mina offered no theological cushion. Western faiths treat this life as the precious thing and aim at forever; Buddhism treats existence as suffering and aims at leaving the cycle—poof, not returning. He looks at people as proteins, pluses and minuses, charges and conformations that have become complex enough to make rich feeling and consciousness, with no non-biological ingredient required. If Earth blew up tomorrow and nothing existed the day after, that would be “just another blip in the universe.” Since we are here, “we may as well make the most of it while we're here blipping.” Fridman answered that it is one hell of a blip. Mina, still in the arena over tests nobody would authorize, had spent two hours acting as if that were true.