Elon Musk is good at making electric cars, flamethrowers, and rocket ships; he is bad at making music, choosing friends, and forming opinions in real time. I know these latter, more personal facts about Musk because he has a Twitter account—one of the most popular accounts in the world, with more followers than CBS News, NBC News, and ABC News combined. In 2018, when twelve young soccer players were trapped in an underwater cave in Thailand, Musk, on Twitter, mused about building a rescue vessel. His help wasn’t needed, as it turned out, but he did take the opportunity to get into an ugly Twitter spat with a cave diver who was involved in the rescue. (For reasons too complicated to explain here, Musk ended up calling the diver a “pedo guy,” and the diver sued him, unsuccessfully, for defamation.) A month later, Musk tweeted, about one of his companies, “Am considering taking Tesla private at $420. Funding secured.” This was probably nothing more than a dumb joke—420, in extremely outdated slang, refers to marijuana—but the S.E.C. took it seriously, and the tweet resulted in Musk paying a fine of twenty million dollars, stepping down as chairman of the company, and agreeing to get preapproval from Tesla’s lawyers before tweeting anything similar again.

Another kind of person—especially a person with several companies to run and many billions of dollars to manage—might have taken any of these incidents as a perfectly good cue to delete his account. Yet here we are in the grip of a global pandemic, and Musk cannot help himself. Last week, he tweeted, “danger of panic still far exceeds danger of corona imo.” (The initialism is sometimes rendered “imho,” for “in my humble opinion,” but Musk is apparently self-aware enough to omit the “h.”) A few days later, one of Musk’s followers, a guy whose display name was “bill lee” followed by a smiling-pile-of-poop emoji, tweeted a link to an article called “Stanford Professor: Data Indicates We’re Severely Overreacting to Coronavirus.” The article was from the Daily Wire, a right-wing blog full of partisan clickbait. “Imo, this professor is correct,” Musk replied. In the meantime, he added, he would devote some of his factories’ capacity to building ventilators, “even though I think there will not be a shortage by the time we can make enough to matter.”

<a aria-label=”The New Yorker’s coronavirus news coverage and analysis are free for all readers.” class=”external-link external-link-embed__hed-link” data-event-click=””element”:”ExternalLink”,”outgoingURL”:”https://www.newyorker.com/tag/coronavirus”” href=”https://www.newyorker.com/tag/coronavirus” rel=”nofollow noopener” target=”_blank”>

The New Yorker’s coronavirus news coverage and analysis are free for all readers.

 

The professor in question was John P. A. Ioannidis, a well-regarded Stanford epidemiologist who is known for precisely this kind of bubble-bursting argument. (His most widely cited publication is called “Why Most Published Research Findings Are False.”) Despite the Daily Wire headline, though, Ioannidis wasn’t exactly arguing that we are overreacting to the virus. He was arguing that we may be overreacting, because we don’t yet have enough evidence to know whether our “draconian countermeasures” will do “more good than harm.” This much is clearly true. Until we’re able to test a broad sample of the population, for example, we won’t know whether the fatality rate from COVID-19 is closer to five per cent or to 0.05 per cent. It’s possible that we’ll look back in a few months and agree that the worst part about the coronavirus was our panicked response to it. We should be so lucky. For now, I would contend that we have no choice but to act decisively, even without complete evidence.

So far, this might seem like an anecdote about the Internet as a basically functional marketplace of ideas. Professor writes provocative analysis (for STAT, an online publication about the life sciences); extremely online industrialist amplifies said analysis (or, at least, a tweet-size synopsis of it); and here I am, online, lodging my critique. What’s not to like? Steven Levy, in Wired, recently wondered whether the coronavirus would “kill the techlash”—whether Americans under lockdown, convening on Zoom and stocking home bars via Drizly and socializing distantly on Instagram Live, would start to feel less indignant about our Silicon Valley overlords and more grateful for all the nifty apps they’ve bestowed on us. But one problem with the concept of the techlash is that it’s always been about too many things at once: surveillance capitalism, anticompetitive practices, phone addiction, Mark Zuckerberg’s uncanny-valley smile. It’s possible to ameliorate one of these problems without broaching the others. Facebook can be the death knell of consumer privacy and also a fun place to share baby photos. Amazon can be a rapacious monopoly and also the most reliable way to get light bulbs in a time of crisis. Twitter can keep us informed (and anxious) about the pandemic, but this doesn’t obviate concerns about its long-term effects on our public discourse.

After asserting that Ioannidis “is correct,” Musk shared his reasoning: “growth rate of confirmed C19 cases is dropping every day,” he tweeted, linking to a bar graph on a C.D.C. site that he thought corroborated this view. In another tweet, Musk predicted that there would be “close to zero new cases in US too by end of April.” A Twitter user called Hopeful Pope of Muskanity, referring to Musk as “my liege,” asked whether this would happen essentially by magic, in the absence of social distancing and other public-health measures. Musk replied, “Kids are essentially immune.”

These claims are wrong, and dangerously so. Most children with COVID-19 do seem to experience milder symptoms than adults—but “most” is not “all,” and, besides, a metascientist like Ioannidis would caution us against clinging too dogmatically to this preliminary finding, or to any finding, about what is still a novel virus. What we do know is that children can get infected with the virus, and can pass it on to others, which makes them, in the most relevant sense of the word, very much not immune. The C.D.C. bar graph that Musk linked to did seem to imply that cases were dropping—but only if you squinted at the graph without bothering to read the words hovering above it, in large type: “Illnesses that began during this time may not yet be reported.” One portion of the graph, helpfully shaded gray, clarified that “this time” referred to the most recent stretch of six days—the very days in which cases appeared, misleadingly, to have gone down. As far as anyone can tell, and as dozens of less muddled bar graphs attest, American cases of COVID-19 are trending exponentially upward.

I don’t mean to imply that Elon Musk is the main problem. There have always been tycoons and celebrities with bad opinions; if this crisis teaches us to pay less attention to them, at least during public-health emergencies, so much the better. And if Musk does end up helping to stave off a shortfall of ventilators, the people whose lives he saves will owe him an incalculable debt of gratitude, and will not particularly care about his shoddy reasoning. Still, Musk is a useful case study, because his combination of characteristics—blithe self-assurance, poor reading comprehension, a proclivity toward the contrarian and controversial, and an apparent willingness to spend ten minutes boning up on a new topic before explaining it to the world—are hardly unique. Rather, they are precisely the characteristics that make him, in a phrase that should only be used in scare quotes, “good at Twitter.”

In 2018, Jack Dorsey, the C.E.O. of Twitter, vowed that the company would work to “increase the collective health, openness, and civility of public conversation.” (Dorsey, unlike Paul Elie, does not seem to have brushed up on his Sontag recently; he uses the metaphor of “conversational health” almost as a mantra.) Since then, the company has made modest improvements. Harassment, threats, and bots are still rampant on the platform, although their proportions seem to have diminished. But, in this backlash-to-the-techlash moment, Twitter has been enjoying a rare bit of good press. The pandemic, Ben Smith argued in the Times, is showing that “Twitter, Facebook, YouTube and others can actually deliver on their old promise to democratize information and organize communities, and on their newer promise to drain the toxic information swamp.” Dorsey, in a direct message to Smith, wrote, “Public conversation can help the world learn faster, solve problems better and realize we’re all in this together.”

In the run-up to the 2016 election, social media was more or less overrun by junk. In the current crisis, social-media companies have been more proactive about preventing the most overt liars and chaos agents—financial scammers, Russian spies, the President of the United States—from monopolizing their platforms. And yet this is where the distinction between disinformation and misinformation, which might normally seem pedantic, becomes relevant. Disinformation means intentional deception (for example, the false insinuation that the coronavirus was created in a Chinese lab). Misinformation is a broader category. Some of it is intentional; some of it isn’t. When social-media executives are asked what they’re doing to combat misinformation, they often respond by describing what they’re doing to combat disinformation, because disinformation lends itself to simpler, sharper answers. But most people who spread misinformation on social media are not Macedonian teen-agers hoping for a quick payday or Iranian spies trying to meddle in a foreign election. Most people spread misinformation because they are misinformed. This is a much broader problem, and the solution to it, if there is one, is far less obvious. In February, the World Health Organization took to Twitter to debunk a few urban legends (“there is no evidence from the current outbreak that eating garlic has protected people from 2019-nCoV”; “Sesame oil is delicious but it does not kill 2019-nCoV”). Earlier this month, Twitter announced that it would take down tweets containing “denial of established scientific facts about transmission . . . such as ‘COVID-19 does not infect children because we haven’t seen any cases of children being sick.’ ” But, when Elon Musk tweeted almost exactly these words, Twitter reviewed the tweet and decided that it “does not break our rules.”

Some would argue that the solution here is simple: more aggressive enforcement. Find all the bad tweets and remove them; repeat until all the bad tweeters are gone. But this can’t be the whole solution. Enforcement is necessary, but it’s not sufficient. Banning the worst of the worst is a relatively easy call, but many of the less egregiously bad tweets—tweets that do not appear to violate any of the platform’s rules but nonetheless sow unnecessary fear, or exacerbate distrust, or cause confusion regarding matters of life and death—come from people who are merely trying to be “good at Twitter.” Social media was always designed to give us what we want, not what we need. For years, it has incentivized controversy, outrage, and half-baked contrarianism. Now its administrators are starting (inconsistently, half-heartedly) to punish some of the people who have correctly internalized those incentives. This is better than nothing, and it may be cause for one or two celebratory news cycles, but the problem is too systemic to be reversed overnight. A bad tweet, morally speaking, is often a good tweet, judging strictly by the numbers. We will not wake up tomorrow and find that all the bad tweets are gone. In the short term, at least, all we can do is flatten the curve.