So there is no way to cast a net around Covid-19 and tighten it to nothing; its animal hosts will always provide it an escape hatch. However, it’s not actually useful to start thinking about alternate hosts until all of a disease’s potential human victims have been protected by vaccination—and so far, we are not remotely close. As long as people somewhere in the world are still waiting for their first shots, Covid-19 will have human hosts to reproduce in. And also, potentially, to mutate in, creating the kind of variants that are now appearing across the globe.

That raises the possibility that, as the virus changes, we’ll need to keep tinkering with vaccines to keep up with it. “I think most people feel that this will be something where likely for the next several years we’ll be getting a Covid-19 shot,” Alex Gorsky, the CEO of Johnson & Johnson, said earlier this month at a CNBC event. “Exactly what that shot is going to be comprised of, I don’t think we know today.”

If Covid cannot be a disease we try to squelch quickly—the way, for instance, we roll out vaccines to counter Ebola outbreaks—it has to become a disease we plan for, such as measles and influenza. With measles, we begin vaccinating in childhood. With the flu, we revaccinate annually, while tuning the vaccine’s contents to keep up with viral evolution. We vaccinate against those because they take such a toll. In the past 10 years, influenza has killed anywhere from 12,000 to 61,000 people per year in the US; globally, measles kills 140,000 each year.

We have no guarantee whether Covid-19, if it becomes endemic, will be as ferocious as measles, or mellow into something mild. Before the pandemic began, there were six coronaviruses known to infect humans: the original SARS from 2003; MERS, which arose in 2012; and four that cause seasonal illnesses. Those last four, which are now considered endemic, are responsible for about 25 percent of the colds we contract every winter, and they demonstrate that some coronaviruses can become something that we dislike, but don’t need to fear. (They have not always been mild, though. One of them has recently been linked to a worldwide epidemic in 1889 and 1890 of respiratory illness and neurological problems; it came down in history as the “Russian flu”—but that name was a guess at its cause, since flu viruses weren’t identified til 40 years later.)

A recent paper modeling the potential future of the novel coronavirus, written by postdoc Jennie Lavine of Emory University, attempts to forecast the ways that Covid-19 might behave in the future, based on data gathered from the four endemic coronaviruses, plus SARS and MERS. It finds that Covid-19 could reach the state that the four endemic strains now occupy, of causing mostly mild disease on a regular basis—but that outcome will depend on how the circulating disease behaves in children during their first infections, since it’s those first infections that set the immune system up to respond down the road.

That is the same function that vaccines perform, of course. Our bodies create multiple types of immunity in response to pathogens; it’s too soon, Lavine says, to gather the long-term data we’ll need in order to know whether Covid-19 vaccination and childhood infection both protect in such a way that any subsequent infections produce only mild disease.

But assume, for the moment, that the virus doesn’t become a mild infection like a cold, but remains an unpredictably dangerous one. That prospect makes it more urgent to defuse vaccine nationalism and to distribute doses worldwide as fast as possible, not just to protect people from illness, but to deprive the virus of hosts in which it can mutate.