Last week, I wrote about the measures we may need to take to persuade the unvaccinated to sign up for shots. This week, I want to explore the other side of the question: How much danger does the Delta variant pose to the vaccinated? In particular, how does it compare with the seasonal f

I’ll be honest about the question behind my question. I want to know if there’s an endgame here. In San Francisco, where I live, 70 percent of residents are fully vaccinated, and 76 percent are partially vaccinated. These are the kinds of numbers we were once told would carry us to herd immunity. Now the hope of herd immunity appears to be gone, and even in San Francisco we’re back to universal, indoor masking. I’m exhausted, and frustrated, and everyone else is, too. Is the future just a spike protein stamping on a human face, forever?

No life lived fully can be lived perfectly safely. There’s much we do that endangers us. And so only part of the answer here revolves around the absolute risk the coronavirus poses to the community. The rest depends on the level of risk that communities are willing to live with. Which brings me to the flu.

[Get more from Ezra by listening to his Opinion podcast, “The Ezra Klein Show.”]

According to C.D.C. estimates, seasonal flus infect between nine million and 45 million Americans annually, depending on the year. They hospitalize between 140,000 and 810,000 of us. They kill between 12,000 and 61,000, mostly infants and the elderly. If vaccinations turn the coronavirus into a flu-level threat, that doesn’t mean they leave us immune to disease or even death. It means they leave us at a level of risk we routinely accept.

Here’s the good news: As of now, if you’re an adult vaccinated with a double dose of an mRNA vaccine like Pfizer or Moderna, most experts I talked to believe the Delta variant is no more likely than the flu to hospitalize or kill you. (The Johnson & Johnson vaccine is another story, and while I do not give medical advice from the confines of this column, all the doctors I spoke to told me they would get an mRNA shot if all they’d gotten was Johnson & Johnson, and San Francisco General Hospital has made that official and so that’s what I did.)

“If you’re a fully vaccinated person in America, your risk of something bad happening to you from Covid is as bad or lower than in a normal flu season,” Ashish Jha, dean of the Brown University School of Public Health, told me.

Celine Gounder, an infectious disease specialist at Bellevue Hospital Center in New York, agreed. “If you are vaccinated and get Covid, your risk of death is lower than if you just wander around and get the flu as somebody who doesn’t get the flu shot, which is unfortunately a lot of people,” she told me.

You can see this even in the case study that kicked off the current panic. Provincetown, Mass., hosts epic gay party weeks, and over July 4 approximately 60,000 people began crowding into the small town. Dance clubs and restaurants were full, contact between the vaccinated and unvaccinated was close and constant, rain drove people indoors, and few were wearing masks. This is the kind of party the coronavirus would plan, if the coronavirus could plan parties.

The vaccinated revelers weren’t being irresponsible. Partygoers were overwhelmingly vaccinated, and they’d been told that the vaccine was overwhelmingly protective against infection. And against earlier strains of the coronavirus, that was true. But Delta can generate roughly 1,000 times the viral load of its predecessor, and indoor parties are the perfect petri dish. “This was an exceptional circumstance in an exceptional location,” wrote Ingu Yun, a doctor who was present at the festivities and analyzed the data in the aftermath.

But of the nearly 1,000 cases that were tracked back to the Provincetown parties by the end of July, there were only seven hospitalizations and no deaths. “The Provincetown numbers tell me that the vaccines are working,” Yun concluded.

All of this is to say: The data we have suggests the vaccines can turn even Delta into a flu-level nuisance, or better, in terms of the risks of hospitalization and death. There is some worry that Delta is modestly worse for children than the original strain, but the absolute risk for young kids is still quite low, and the best firewall for them is vaccinated adults. The big unknown here is the possibility for long Covid or other lingering consequences. But it’s worth noting that this is true with the flu, too. A number of chronic diseases seem to trace back to the body’s reaction to viral infections.

“Do I wish anybody long Covid? No,” Gounder told me. “Do I want to get long Covid? No. However, we run the risk in our everyday lives of getting one of these viral infections that for most people are very mild, but that can very rarely set off something like chronic fatigue syndrome or an autoimmune disease, but that’s a risk we tolerate.”

All of this made me feel a bit better. And then I talked to Bob Wachter.

Wachter is the chair of the department of medicine at the University of California at San Francisco. His main point was simply this: The numbers aren’t stable. He’s concerned that the immunity people got from past coronavirus infection is waning more quickly than we’d expected. And he thinks the same is true for vaccine-based immunity. “I think the best estimate now is the vaccines begin to lose some efficacy after six months and your immune response loses some mojo too,” he told me.

This is why Wachter supports the masking mandate in San Francisco. In his view, it’s not just the unvaccinated who benefit from masking. There may be some vaccinated people, particularly older people who got vaccines early, who will need a booster shot soon. “Two doses of the vaccines provided a healthy person a huge amount of immunologic cushion, but for a 75-year-old not so much,” Wachter told me. “Their immunity needs to wane only a modest amount before they cross the curve, before they are susceptible to a more serious infection.”

If that’s true, it suggests a future where even in highly vaccinated places, it will be a continuous fight between the immunity offered by vaccines, an always mutating virus and the speed of our booster campaigns. As Wachter dryly put it, “It’s nontrivial to go give boosters to 200 million people.” Even worse, much of the rest of the country, and even more of the rest of the world, isn’t vaccinated, which is giving the virus vast opportunities for evolution. Delta is by no means the final form the coronavirus could take.

The more optimistic outlook came from Jha. He thinks that in highly vaccinated places, we’re going to see Delta slam into a wall of vaccination, and hospitalizations and deaths won’t follow cases the way they have in the past. “Even in highly vaccinated states, unvaccinated people cluster,” he told me. “So you will see the initial rise, but once that cluster starts bumping into immunity, it won’t be able to sustain itself. We’ll find that out in the next couple of weeks in places like San Francisco and Boston.” If that happens, it’ll also be a powerful argument for vaccination in the parts of the country that have lagged and that will watch the virus tear through their communities even as more vaccinated areas are largely spared.

Another argument for optimism comes from Britain, which saw a surge in Delta cases, and then they mysteriously burned out. “The U.K. data is the most reassuring thing out there and I’d feel better if we had a clue about what happened,” Wachter told me.

Me too. But uncertainty is a good reason for caution. I began the week upset about the return of mask mandates and depressed about the possibility that the vaccines were beginning to fail. Now I’m convinced that the revived mask mandates make sense, cheered by how well the vaccines have performed and worried about whether they’ll continue to hold up. I wish I could tell you we know how this ends, even just for the vaccinated, but I can’t.

But let’s say that the data shakes out as I hope it will and that vaccinations can turn the coronavirus into a merely flulike menace. The fear is likely to linger — particularly in communities, like mine, that have become risk averse as both a matter of public health and political identity.

“It’s hard for me to imagine people saying, ‘OK, I will go back to normal because the flu kills 30 or 40,000 people a year, and that’s where we are with Covid,’” Wachter told me. “The flu is background noise to most people and it’s hard to imagine this becoming background noise. At least for the foreseeable future, it feels like every blip or surprising event or congressman infected or person on their deathbed saying, ‘I wish I’d been vaccinated,’ will become a story or a social media phenomenon.”

I suspect we’re headed for a two-tiered society (or maybe a many-tiered society) built not just on the risk the coronavirus poses to the local population, but on the sensitivity to that risk. In a place like San Francisco, the absolute risk is relatively low, in part because the population’s sensitivity to coronavirus risk is quite high. We will be hair-trigger in reimposing restrictions when cases rise. In Texas and Florida, where the politics have settled nearer to a live-free-and-die-coughing approach, the absolute risk is higher precisely because the sensitivity to the risk is lower. So there may be no one endgame here, only constant management of the risks we face and are willing to bear.