And when outbreaks do occur, the vaccines still provide good protection. A second CDC study examined an outbreak in a Kentucky nursing home where just half the staff were fully vaccinated. The outbreak, which began with an unvaccinated staff member, led to 46 covid-19 infections. Out of 71 vaccinated residents, 18 (25%) became infected, two were hospitalized, and one died. The staff fared better. Of the 56 vaccinated employees, four (7%) became infected. Most of those infections were asymptomatic. Only 6.3% of the residents and staff who had been vaccinated developed symptoms, compared with 32% of unvaccinated individuals.
During a nursing home outbreak, “staff and residents are continuously encountering the SARS-CoV-2 pathogen over and over again,” says Meagan Fitzpatrick, who models infectious diseases at the University of Maryland School of Medicine. So seeing such a small number of infections in this type of setting is encouraging.
Tracking the variants
New studies also hint that variants may be to blame for some of these breakthrough infections. Viral variants are “one of the wild cards,” said Anthony Fauci, chief medical advisor to the president of the United States, in a briefing on April 12. Though there is little real-world data, lab studies suggest that at least some of the variants are less susceptible to vaccine-induced antibodies than the original SARS-CoV-2.
In the Kentucky study, the researchers found that the outbreak was fueled by a variant known as R1, which had not previously been identified in the state. This virus had several important mutations that had also been identified in other variants. For example, the E484K mutation, also found in the B.1.351 variant first identified in South Africa, seems to help the virus evade the antibody response. And the D614G mutation might increase transmissibility. The authors note that although vaccination decreased the likelihood of infection and symptomatic disease, the virus still managed to infect more than a quarter of vaccinated residents and about 7% of the staff. That suggests the vaccine might not work as well against this variant, but the authors caution that the study was small. (The authors of the Chicago study didn’t sequence the virus.)
A New England Journal of Medicine study tracked infections in staff at Rockefeller University in New York. Between January 21 and March 17, the researchers tested 417 employees who had received a full course of the Pfizer or Moderna vaccine. Two women tested positive. When the researchers sequenced the viruses, they found that each was a slightly different variant, and these weren’t exact matches for any that had been previously identified.
One woman, for example, had a variant with mutations found in B.1.1.7, which originated in the UK, along with mutations common to B.1.526, which originated in New York City. “She had variants somewhere between the two,” says Robert Darnell, a physician and biochemist at Rockefeller and lead author of the study.
When a breakthrough infection occurs, the assumption is that the patient failed to mount a strong immune response to the vaccine, Darnell says. But that didn’t seem to be the case with this woman. Darnell managed to get a blood sample soon after she tested positive. He and his colleagues found high levels of antibodies able to neutralize SARS-CoV-2. Because she was newly infected, the antibody response was likely due to vaccination, not her recent infection. Antibodies take some time to develop.
Why her immune system didn’t protect her from infection isn’t entirely clear, but one possibility is that the variant managed to dodge her response. “For this particular patient, that’s probably the best explanation for what we saw,” says Stephen Kissler, an epidemiologist at of Harvard’s T.H. Chan School of Public Health. “It’s not surprising to me that a lot of these breakthrough infections that we’re seeing are from variants,” he adds. As more and more people get vaccinated, “there’s an evolutionary selection pressure that’s being applied.”
On the other hand, as more people get vaccinated, we’ll see fewer infections and the virus will have fewer opportunities to mutate. And Fitzpatrick points out that even if immune escape explains the woman’s infection, it’s just a single case. And there’s no evidence that she passed the infection to others who also had been vaccinated. The phenomenon is worthy of future studies, but “I don’t yet see this as alarming,” she says. “There’s not yet a public health crisis.”
And even when breakthrough infections occur, it doesn’t necessarily mean the vaccine has failed, says Monica Gandhi, an infectious disease doctor at the University of California, San Francisco. Antibodies are only part of the immune response. T cells play a huge role too, by ramping up other parts of the immune system and clearing out the virus once it has infiltrated the body. They don’t prevent infection, but they can curb the virus’s spread. And some research suggests that the body’s T cell response will be much tougher to evade. “You may actually get a mild infection, but hopefully you’ll still have protection against severe disease,” Gandhi says.
Still, it’s important to track breakthrough infections to look for unexpected changes. A rising number of infections in vaccinated people might mean waning immunity or the emergence of a new variant that can dodge the immune response. The vaccines might need to be tweaked, and we might need booster shots. But over time, “our bodies will develop a more complete immune response,” Kissler says. “And even if we do get reinfected, we’ll be protected from the most severe outcomes. In the long term, the outlook is good.”
A new training model, dubbed “KnowNo,” aims to address this problem by teaching robots to ask for our help when orders are unclear. At the same time, it ensures they seek clarification only when necessary, minimizing needless back-and-forth. The result is a smart assistant that tries to make sure it understands what you want without bothering you too much.
Andy Zeng, a research scientist at Google DeepMind who helped develop the new technique, says that while robots can be powerful in many specific scenarios, they are often bad at generalized tasks that require common sense.
For example, when asked to bring you a Coke, the robot needs to first understand that it needs to go into the kitchen, look for the refrigerator, and open the fridge door. Conventionally, these smaller substeps had to be manually programmed, because otherwise the robot would not know that people usually keep their drinks in the kitchen.
That’s something large language models (LLMs) could help to fix, because they have a lot of common-sense knowledge baked in, says Zeng.
Now when the robot is asked to bring a Coke, an LLM, which has a generalized understanding of the world, can generate a step-by-step guide for the robot to follow.
The problem with LLMs, though, is that there’s no way to guarantee that their instructions are possible for the robot to execute. Maybe the person doesn’t have a refrigerator in the kitchen, or the fridge door handle is broken. In these situations, robots need to ask humans for help.
KnowNo makes that possible by combining large language models with statistical tools that quantify confidence levels.
When given an ambiguous instruction like “Put the bowl in the microwave,” KnowNo first generates multiple possible next actions using the language model. Then it creates a confidence score predicting the likelihood that each potential choice is the best one.
The news: A new robot training model, dubbed “KnowNo,” aims to teach robots to ask for our help when orders are unclear. At the same time, it ensures they seek clarification only when necessary, minimizing needless back-and-forth. The result is a smart assistant that tries to make sure it understands what you want without bothering you too much.
Why it matters: While robots can be powerful in many specific scenarios, they are often bad at generalized tasks that require common sense. That’s something large language models could help to fix, because they have a lot of common-sense knowledge baked in. Read the full story.
—June Kim
Medical microrobots that travel inside the body are (still) on their way
The human body is a labyrinth of vessels and tubing, full of barriers that are difficult to break through. That poses a serious hurdle for doctors. Illness is often caused by problems that are hard to visualize and difficult to access. But imagine if we could deploy armies of tiny robots into the body to do the job for us. They could break up hard-to-reach clots, deliver drugs to even the most inaccessible tumors, and even help guide embryos toward implantation.
We’ve been hearing about the use of tiny robots in medicine for years, maybe even decades. And they’re still not here. But experts are adamant that medical microbots are finally coming, and that they could be a game changer for a number of serious diseases. Read the full story.
We haven’t always been right (RIP, Baxter), but we’ve often been early to spot important areas of progress (we put natural-language processing on our very first list in 2001; today this technology underpins large language models and generative AI tools like ChatGPT).
Every year, our reporters and editors nominate technologies that they think deserve a spot, and we spend weeks debating which ones should make the cut. Here are some of the technologies we didn’t pick this time—and why we’ve left them off, for now.
New drugs for Alzheimer’s disease
Alzmeiher’s patients have long lacked treatment options. Several new drugs have now been proved to slow cognitive decline, albeit modestly, by clearing out harmful plaques in the brain. In July, the FDA approved Leqembi by Eisai and Biogen, and Eli Lilly’s donanemab could soon be next. But the drugs come with serious side effects, including brain swelling and bleeding, which can be fatal in some cases. Plus, they’re hard to administer—patients receive doses via an IV and must receive regular MRIs to check for brain swelling. These drawbacks gave us pause.
Sustainable aviation fuel
Alternative jet fuels made from cooking oil, leftover animal fats, or agricultural waste could reduce emissions from flying. They have been in development for years, and scientists are making steady progress, with several recent demonstration flights. But production and use will need to ramp up significantly for these fuels to make a meaningful climate impact. While they do look promising, there wasn’t a key moment or “breakthrough” that merited a spot for sustainable aviation fuels on this year’s list.
Solar geoengineering
One way to counteract global warming could be to release particles into the stratosphere that reflect the sun’s energy and cool the planet. That idea is highly controversial within the scientific community, but a few researchers and companies have begun exploring whether it’s possible by launching a series of small-scale high-flying tests. One such launch prompted Mexico to ban solar geoengineering experiments earlier this year. It’s not really clear where geoengineering will go from here or whether these early efforts will stall out. Amid that uncertainty, we decided to hold off for now.
