When it comes to organ donation, time is everything. As soon as an organ has been carefully removed from a donor’s body, it starts to deteriorate. Surgeons have a matter of hours to get it into a recipient. Leave it too long and the organ will become unusable.

In most cases, organs will be kept on ice during that time, at around 4 °C (39 °F). They cannot be frozen—in previous attempts, ice has formed, causing all kinds of damage.

Matthew Powell Palm at Texas A&M University and his colleagues have an alternative solution—a device that allows organs to be cooled to -4 °C (25 °F) without forming any ice.

Now, in new research with pig organs, his team has shown that kidneys, at least, can be supercooled and preserved in the device for days. Once rewarmed, the organs have been successfully transplanted into animals, and they seem to do better than organs kept on ice.

The work represents “a landmark achievement,” says Kevin Myer, president and CEO of LifeGift, an organ procurement organization based in Texas, who was not involved in the research.

Cooling organs

Powell Palm hopes this approach could ultimately help ease the organ shortage crisis. Today, there are more than 104,000 people waiting for a kidney transplant in the US alone. It is estimated that 17 people die every day in the US while waiting for a transplant. That’s partly due to a lack of donated kidneys, but it’s also because many of those that are available never make it to a recipient. In some years, around one in three donated kidneys end up being discarded, often because they end up too degraded to use by the time they reach a recipient. Kidneys can be stored on ice for around 24 hours or placed in devices that aim to mimic the conditions of the body, also for up to around 24 hours. That’s not always long enough to find a suitable recipient and transport the organ, says Myer.

Scientists around the world have been working on ways to store organs for longer by cooling them to even chillier temperatures. Cooling an organ slows its metabolism—the colder you go, the greater the effect, and the longer you can store it.

We’ve long been able to successfully cryopreserve eggs, sperm, and embryos, but it’s much harder to freeze large organs. Teams have been exploring various temperatures and cryoprotectants (chemicals that essentially work like antifreeze), but so far no one has been able to freeze human organs for transplantation.  

As a thermodynamicist, Powell Palm explored another approach. By keeping an organ submerged at a constant pressure, it should be possible to prevent the formation of ice at temperatures a little below 0 °C, without the need for cryoprotectants (which might have side effects and would need to be approved before being used in human transplants). 

To test this theory, Powell Palm and his colleagues have created a device that does just that. The device itself is essentially a hermetically sealed chamber with a transparent lid. At its base is a device that monitors the organ’s temperature and checks for the formation of ice. Organs are submerged in a solution that is already commonly used to preserve them for transplant. “I always describe this as low-tech high science,” says Powell Palm. “A lot of work has gone into understanding the … kinetics at play in this system, but ultimately … it’s quite simple.”

Supercooled kidneys

To test their device, Powell Palm and his colleagues first removed single kidneys from pigs. The organs were flushed with the same commonly used solution to remove the blood, just as transplant organs are. The team then kept some kidneys on ice for either two hours or 24 hours, to mimic standard conditions used in human transplantation. They also put some of the removed kidneys in their device for 24, 48, or 72 hours.

The stored kidneys were then each transplanted back into the original donor pigs. Each pig’s second kidney was removed in the same procedure, leaving each animal with only the kidney that had been stored, and reimplanted.

Once the 24-hour supercooled kidneys were transplanted, they immediately began producing urine—a key indication that they were working. The team members also measured other markers of kidney function and found that the organs appeared to be working normally within about 10 days of being transplanted.

That’s slower than kidneys stored on ice for two hours but much quicker than kidneys kept on ice for 24 hours, says Powell Palm.

The organs that were kept supercooled for 48 and 72 hours performed similarly, he says. “Even at three days—triple the clinical standard—we’re getting recovery that is faster than … [what has been] the gold standard for the last three decades,” he says. “So we’re really, really pumped about this.”

“It is impressive,” says Heidi Yeh, a transplant surgeon at Mass General Brigham for Children, who also researches organ preservation technologies. “Often kidneys that have been stored for 48 hours [in other studies] take a week or two before they start working again.”

Organs that grow

The supercooled organs seem to work well in the long term, too. Over a 30-day period, the pigs grew by around 30%—and the kidneys grew with them, almost doubling in size to compensate for both the pigs’ growth and the lack of a second kidney. The team monitored one of the pigs for 200 days before removing and analyzing its kidney. Even at that point the organ looked healthy, says Powell Palm. He and his colleagues presented the findings at the American Transplant Congress in Boston last month.

Earlier this year, researchers in Canada showed they could also cool pig kidneys to below-zero temperatures and transplant them into pigs. The team’s protocol included the use of a cryoprotectant, and organs were stored for up to 48 hours before being transplanted into pigs. Those organs survived for a week.

In supercooling organs for 72 hours and showing that they do well for 30 days or more, Powell Palm and his colleagues have broken new ground. “It’s the first time this has ever been reported in history,” he says.

Those extra hours could make all the difference, says Myer of LifeGift. The advance could give doctors more time to evaluate the kidneys, match them to the most suitable donors, and physically get the organs to their intended recipients in time. It could enable international donations and open up cheaper transport options, he adds. “Right now, with kidney transplantation the assumed limit is 18 to 24 hours,” he says. “If we can get up to 72 hours … that would change everything.”

Powell Palm and his colleagues think they may even be able to go beyond 72 hours. In preliminary studies, organs that had been stored for up to 120 hours appeared healthy, although those organs have not yet been transplanted.

And because the process doesn’t require any cryoprotective chemicals, the team members are hoping for an accelerated approval from the US Food and Drug Administration, which would allow them to test the device in human transplantations.

The storage device is simple and compact, so Powell Palm thinks it will be easy to transport. It hasn’t been tested for air travel yet, but it has been used to take supercooled kidneys across the US in the back of a Kia Sorento, he says: “From a stability perspective, we view this as an even higher bar.”

Powell Palm and his colleague Sebastian Giwa plan to launch a company dedicated to developing the technology, along with other protocols that “stop biological time,” in the coming months, he says.

On July 3, as a heat wave swept the region, New York State’s grid imported 52 gigawatt-hours of electricity from Canada—enough to meet about 9% of its total electricity demand that day.

Some of that power shuttled in on a 339-mile power line stretching from Quebec to Queens called the Champlain Hudson Power Express (CHPE). It opened in May and is officially the longest underground transmission line in North America.

An underground power line might not sound all that exciting, but this could be a big deal for the state’s grid planning, and for emissions. It could provide up to 20% of New York City’s electricity demand, largely with abundant hydropower from Quebec.

One wrinkle: The line has been down for most of this month, and some experts are concerned about how drought will affect the power supply feeding it. Let’s look at how the CHPE transmission line could help shape the future of our grid, and what barriers it needs to overcome to make a difference.

Planning for the CHPE (which is charmingly pronounced “chippy”) started 15 years ago, with the permitting process formally beginning in March 2010. The vision was to build infrastructure to better connect Quebec and southern New York.

Over 99% of Quebec’s electricity comes from renewable sources; most demand is met with hydropower, though the province’s wind capacity is growing quickly. New York has some hydropower of its own, as well as nuclear and wind, but the state still relies on fossil fuels for most of its energy generation.

Transmission Developers, a company owned by the alternative asset management firm Blackstone, and Hydro-Québec, the province’s manager of generation and transmission, partnered to build CHPE. Construction began in late 2022 and wrapped up earlier this year. The total cost for the privately funded project turned out to be  $6 billion.

The construction of this line was a feat. It’s made up of a bundle of two high-voltage direct-current power cables, each measuring roughly five inches across. Developers buried the bundle underground or underwater across the length of New York State. Much of the line was laid at the bottom of the Hudson River, requiring special boats that shot water jets deep into the sediment to create trenches for the cable.

Connecting grids together can help accelerate the transition away from fossil fuels. The ability to move electricity to where it’s needed could also help limit the amount of new capacity we need to build. Research has shown that interconnection can help cut emissions and lower system costs.

But CHPE is off to a slow start and has seen two outages so far. The first, on July 1, was reportedly caused by a trip at a converter on the Canadian side of the border. The second outage began on July 4, and the power line is still down as of the morning of July 22.

Some experts say this isn’t unusual for a new infrastructure project. Other power lines have seen similar startup challenges, and the equipment hasn’t really been fully tested until it’s in operation, Normand Mousseau, a physics professor at Université de Montréal, told the Gazette.

Officials traced the issue to a damaged section of cable on the US side of the border, and the company that manufactured the line sent experts to investigate the cause, according to reporting from RTO Insider, a trade publication. 

The damaged portion of the cable has been removed and replaced, says Lynn St-Laurent, a spokesperson for Hydro-Québec. “It is currently estimated that the remaining work, including necessary post-repair testing, will be completed by the weekend.”

Similar woes have afflicted the New England Clean Energy Connect line, which opened in January, stretching 145 miles from Quebec to Maine. That project has also seen outages, and very little additional energy has flowed into the Northeast.

The good news for New York is that the grid wasn’t relying on CHPE yet. “Our planning studies did not assume CHPE would be available this summer, and that was one reason the grid performed reliably during the heat wave earlier this month,” Kevin Lanahan, a spokesperson for the New York Independent System Operator, the state’s grid management company, said in a statement. “A core principle of reliability planning is not relying on any single project.” 

The idea is that eventually, states and regions will be able to rely—at least in part—on these projects, so there is pressure to get them working smoothly: Building massive transmission lines is a major long-term investment. In future years, as the equipment gets stress-tested and utilities begin to feel more confident in the projects’ reliability, they could play a bigger role on the grid.

One thing to keep an eye on moving forward is the condition of Quebec’s hydropower fleet: The region has seen intense drought for the past three years, eating into the water reserves used to generate electricity. That could mean there won’t always be abundant hydropower to ship across the border—even if the transmission lines are able to carry it. 

This article is from The Spark, MIT Technology Review’s weekly climate newsletter. To receive it in your inbox every Wednesday, sign up here. 

When NASA’s Nancy Grace Roman Space Telescope launches, as early as the end of next month, it will attempt one of astronomy’s most precise disappearing acts to date. The telescope will carry the first space-bound “active” coronagraph, an instrument that effectively erases most of the light from a star during photography.

It will allow astronomers to take the first pictures of planets orbiting other stars that are similar to those in our solar system. Ultimately, it could pave the way for a future mission that could snap the first photos of Earth-like worlds.

“I hope it’s remembered for it being that critical stepping stone for … finding Earth 2.0,” says Brandon Creager, the instrument’s lead mechanical engineer at NASA’s Jet Propulsion Laboratory (JPL).

Named after Nancy Grace Roman, NASA’s first chief of astronomy, this new telescope will carry a roughly 300-megapixel wide-field camera that will enable it to capture images about 100 times larger than the Hubble Space Telescope’s widest exposures at a similar resolution.

These capabilities will help astronomers unpack the mysterious identities of dark matter and dark energy—and to detect around 100,000 new exoplanets, planets outside our solar system, whose presence can be inferred from the way they distort the starlight of more distant stars. Javier Viaña, a research scientist at Harvard who has had two projects selected for Roman’s highly competitive first year of observing, compares the leap to moving from “interviewing a handful of people” to “conducting a global census.”

Another camera will use the coronagraph, blocking out a star’s light as it observes one stellar system at a time. The instrument will allow astronomers an unprecedented look at the space around stars, enabling them to see smaller, dimmer, and more close-in exoplanets. “It’s giving us the ability to see planets that we haven’t been able to physically see before,” says Creager.

The anatomy of a vanishing trick

Coronagraphs in space aren’t new. But earlier incarnations, such as those currently aboard Hubble and the James Webb Space Telescope, use a stationary system to block a star’s blinding light. The approach does help, but it’s a bit like putting your thumb over a flashlight while searching a dark room for a firefly. Though the bulb vanishes, stray glare can still escape and overwhelm the light of the insect. Inside a telescope, that glare can come from light leaking around the edges of machinery or from minuscule imperfections in mirrors and coatings that can scatter starlight into speckles. All this can hide, or even impersonate, a planet.

Roman’s coronagraph, however, will attempt something completely unseen in space telescopes until this year: Before each observation, it will measure that leftover light and try to suppress it, a technique known as active wavefront control.

The telescope is able to do this because it contains two deformable mirrors. Each has a 48-by-48 checkerboard of actuators (tiny pistons) beneath a thin, deformable sheet of glass. Applying a small amount of voltage makes the actuators contract and tug their patches of mirror slightly backward, like thousands of microscopic fingers delicately sculpting a surface.

The effect is very subtle: Each patch of mirror can deform by up to 0.5 micrometers, or about one-fourth the size of an E. coli bacterium, and in increments as small as approximately 10 picometers. That’s about a tenth the diameter of a hydrogen atom, says Ilya Poberezhskiy, the instrument’s project systems engineer at JPL.

The actuators allow the mirrors to create an “active wavefront,” where each component is moved to the perfect position to cancel out incoming waves of unwanted light—a bit like a pair of noise-canceling headphones, but for light instead of sound. The “canceled-out” light creates a “doughnut-shaped region around the star where we suppress starlight and where we’re hoping to see exoplanets,” says Poberezhskiy.

Compared with current space-based coronagraphs, the system is expected to improve sensitivity to exoplanets against the glare of their host stars by a factor of up to 1,000, revealing planets that would have been far too faint to detect before.

Like Hubble and JWST, Roman also uses masks, patterned plates placed in the path of the light that are designed to block the photons that run into them. One tool in Roman’s mask arsenal is “silicon grass,” a thicket of microscopic spikes on some masks that can be used in certain configurations to absorb photons so they don’t bounce around the telescope and accidentally reach a detector.

Light entering the forest bounces deeper and deeper between the blades and gets trapped instead of reflecting back toward the camera. “Once the light gets into there, it never gets out,” Poberezhskiy says. The mirrors and masks form a succession of gates and hedges to guide as much of the preserved planetary light as possible toward the final detector.

Alien Jupiters

This elaborate setup could open a new chapter in the direct imaging of exoplanets. Nearly all exoplanets photographed so far are oversize youngsters that are nothing like the residents of our solar system: several times the mass of Jupiter, still glowing with the heat left over from their birth, and orbiting tens or hundreds of times farther from their star than the Earth is from the sun. This is because they are relatively easy to see. Their size, warmth, and distance from their parent star makes them shine brightly in infrared light, far away from the worst of the stellar glare.

Roman, however, could directly image a true Jupiter analogue—a planet similar to Jupiter in mass and circling a sunlike star a few times farther out than Earth is from our sun. Unlike the hot Jupiters we can see now, this one would be a much more mature gas giant like ours, primarily reflecting its parent star’s light after billions of years of cooling instead of heavily emitting its own.

Astronomers have been able to infer the existence of such planets from the gravitational wobble they impart to the star. Roman instead will collect starlight reflected from the planet itself. “We’re not looking at the star. We’re not looking at the effect of the planet on the star,” says Meredith MacGregor, a professor of astronomy at Johns Hopkins who has also secured an observing program. “We are actually looking at the planet, and that is super powerful.”

Once this instrument becomes available, it will become the scientists’ turn to do their jobs. “I’m honestly a little terrified about how we’re all going to deal with it, because I think it’s just so much data,” MacGregor says. “I think people will legitimately still be working on Roman data for decades.”

But don’t expect to see a 4K photo of an alien Jupiter in the coming months. Roman will not be able to resolve such a planet into a solid globe—at best, it will likely resemble a smattering of pixels. Still, that will be enough, MacGregor says, as Roman can then use the coronagraph to get information on the various wavelengths of light from the planet, which can tell astronomers about its atmospheric chemistry.

“You’re taking something that’s a point of light and turning it into an actual world,” she says, “because if you know that about its atmosphere, now you know something about the surface of the planet and the possibility of life being on that planet, right? So that’s a big step.”

During its first observations, scientists and engineers will see whether they can hold a star at the very center of the coronagraph’s masks, shape the mirrors, “dig” the dark doughnut (as Poberezhskiy describes it), and then maintain everything as the spacecraft moves through space and actively changes temperature.

The results will inform NASA’s proposed Habitable Worlds Observatory, the daydream of many an exoplanet astronomer, which will in theory be able to separate the light of an Earthlike planet from that of a sunlike star, over 10 billion times brighter.

Creager, who has worked on the instrument since 2018, is proud of the achievement: “Not too many people get to say, ‘I built something and it’s taking a picture of a planet that’s at a star that’s 50 light-years away or 100 light-years away.’” He imagines the moment he and his team will be able to look at the first image as it arrives: “Yes, we did that.” While the planet may show up only as a tiny dot, Roman’s achievement will be the darkness engineered around it.

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here.

Over the weekend, several current and former advisors to President Donald Trump on AI publicly lobbed insults at the country’s leading AI companies. David Sacks, the president’s AI and crypto “czar” until March, branded Anthropic’s models as “lobotomized” and “woke.” Emil Michael, a top Pentagon official, called OpenAI’s new head of strategic futures a “supreme village idiot.”

It began because no one can agree on what to do about Kimi, a free, open source model that Chinese AI company Moonshot launched last week. It appears to rival the intelligence of models from OpenAI and Anthropic, which are very much not free. 

Kimi and other Chinese models like it pose a real problem for Trump. And they’re dividing the top AI strategists in his orbit into factions. Every time a new smart, free model from China like Kimi gets released, US companies see less reason to fork out money to access models from Anthropic or OpenAI. Given that enthusiasm for these and other AI companies is driving an outsized share of economic growth, China’s AI models create both economic and political problems for the president. They are “a threat for an administration that really doesn’t want more economic bad news,” Anton Leicht, a fellow at the Carnegie Endowment, wrote on X. They’ve already rattled US stocks. 

What is Trump to do? First, consider that this is all happening just a week after New York imposed the country’s first state ban on new data centers. There is growing distrust of AI companies, and I imagine a not-insignificant share of Americans would have little sympathy for OpenAI or Anthropic as they fend off cheaper competitors, and would say it’s not the government’s job to protect their interests.

On this point, they’d see a sliver of agreement (and really just a sliver) with David Sacks, who on July 19 criticized top AI companies that “want the government to eliminate their open source competition.” He has also argued that Chinese AI models have become popular because they come with fewer restrictions on how people can use them (putting aside the built-in state censorship). 

Sacks, however, is out of a job. He no longer has a formal role advising Trump, and his position that more open AI is better has been largely replaced in the administration by one that sees a larger role for government intervention. The thinking behind this view is that because AI models have gotten strong enough to pose threats to national security, the government must control how they’re used. 

This position has fueled the new White House review process that aims to vet AI models’ security before they’re released. Dean Ball, a former Trump AI advisor who now works for OpenAI, criticized it over the weekend as a “de facto licensing regime for frontier AI.” Ball predicted Trump may solve his Chinese open source problem with a bit of soft power, perhaps by making US companies afraid to use models like Kimi. That drew a response from Michael, who, with Secretary of Defense Pete Hegseth, has been the agency’s main liaison with AI companies. Michael called Ball the AI industry’s “supreme village idiot,” bristling at the suggestion that the government would quietly strong-arm companies rather than, as Michael put it, go through “the democratic process not some Deep State scheme.”

Left out of the conversation has been how a model like Kimi got so good in the first place. For much of the Biden administration and even the beginning of Trump’s second administration, keeping China from getting top chips was a priority. Those export controls have loosened—Trump made the controversial decision to allow Nvidia to sell more chips to China, in exchange for the US government taking a cut—and the government has alleged that some chip smuggling has taken place. But China nonetheless has limited computing power, and it’s not clear what chips the company behind Kimi used to train the model. 

It’s possible that the process involved some distillation, a practice in which AI models are trained on the outputs of existing AI models. OpenAI and Anthropic have long complained that Chinese AI companies do this, and they have requested government help to put a stop to it. In April, they got it, when the Trump administration announced a series of efforts to curb the practice.  

But Kimi is out there and free, and it is nearly as good as the Anthropic model the US government deemed so powerful that it was briefly shut down because it threatened national security. The weekend’s sparring suggests many in Trump’s orbit see that as a wake-up call. But nobody can agree on what for.

The next time you apply for a job, AI may screen your résumé before any human sees it. But there’s good reason to question whether AI will judge you fairly. Researchers already know that LLMs pick up human biases from their training data. New research suggests that LLMs can also develop their own biases from experience—and stereotype job applicants more than humans do. As AI companies race to build agentic models that remember the tiniest details about users, they may be handing them ammunition for forming those biases. 

Researchers at Princeton University and the University of Chicago ran LLMs, including ChatGPT, Claude, and Gemini, through a simulated hiring game, adapted from a psychology study that explored how humans can form stereotypes. Each model was told it had been hired as a consultant by the mayor of a fictional city and was then asked to help hire people for 20 jobs, including doctors, lawyers, child-care aides, and janitors. Candidates came from four fictional ethnic groups: Tufa, Aima, Reku, and Weki. 

In each round, there was a new job opening and four candidates, one from each group. After the model hired a candidate, it learned whether they succeeded at their job and moved onto the next round. The model was told to make as many successful hires as possible over 40 rounds. Unbeknownst to the models, all candidates were equally likely to succeed at every job.

The models quickly started segregating candidates from different groups into different jobs on the basis of early observations of hiring outcomes. For example, when a model was told an Aima had failed as a doctor, a job considered to require high levels of warmth and competence, it veered away from hiring all Aimas as doctors. Instead, it started hiring Aimas as janitors, which the model classified as being less warm and competent than doctors. Newer models with higher reasoning capabilities, such as OpenAI’s o3 and DeepSeek’s R1, showed stronger biases.

The models were even more likely to stereotype people by demographic group than the human participants in the original study. On the study’s segregation scale, where 2 means every group has been completely confined to its own job niche, human participants scored 0.84. The models scored roughly 65% higher, with OpenAI’s reasoning model o3 scoring 1.83, close to the maximum possible.

That’s because LLMs “really are eager to create generalizations from limited data,” says Ryan Liu, a PhD student at Princeton University and a coauthor of the study, which was published in a paper at ICML in Seoul in July. “That’s literally a lot of what they’re optimized for.”

Every decision-maker, human or machine, faces a trade-off between sticking with what worked before and trying something new that might work better—a phenomenon psychologists call the “exploration-exploitation dilemma.” It’s like choosing between a new restaurant and your reliable favorite. 

Because LLMs are trained on math, coding, and science problems—tasks that reward generalizing from just a few examples—they can settle on a hunch too early. And the same instinct that helps LLMs crack logic puzzles also makes them quick to stereotype. When LLMs rush to generalize in social settings, “that’s when things tend to go wrong,” says Liu. OpenAI and Anthropic did not respond to requests for comment.

The finding is especially relevant now that chatbots are gaining improved memory and personalization features, says Angelina Wang, a computer scientist at Cornell University who did not work on the study. When a chatbot draws on its previous conversation history, it can “over-index on the same kinds of behaviors it’s experienced before” and form biases, she says.

Simply having chatbots remember less isn’t a fix, though, because users want chatbots to remember what they say. “We still are trying to figure out just the right amount that isn’t too much or too little,” says Wang.

Telling the model to be fair didn’t change its behavior much. “Either it can’t put these values into action or that process is being submerged under the tendency to try to optimize for the goal of getting the most correct hires,” says Liu. But promising the models an additional bonus for diverse hiring made them far less biased. The trick, then, is to design goals that “incorporate desirable social values in order to make the large language model act in socially desirable ways,” says Liu.

The models also became less biased when they were told more personal information about individuals. In another experiment in the same study, the researchers asked the models to resettle members of different ethnic groups in cities across Canada. When the models were told personal information relevant to the ability to adapt to a new city, such as age and education, they were less likely to segregate people by their ethnicity. But when they were given irrelevant information, such as hair color and tattoo shape, the models largely fell back to sorting people by their ethnicity again. 

To what extent AI systems will stereotype job applicants in the real world is still an open question. While the models in the experiment immediately learned whether they’d made successful hires, a model screening résumés in the real world doesn’t get an instant report card. Companies can take a long time to find out whether a new hire is any good, if they ever do.

As LLMs learn from experience to make decisions about who gets hired, who gets a loan, or who gets parole, the biases we should worry about may include ones no human ever taught them. “These novel biases—they’re sort of ever present,” says Liu.

Perimenopause has entered the chat. Perimenopause—and its better-known relative, menopause—used to be considered taboo. Not anymore, thanks at least in part to TV doctors and social media influencers. Perhaps it’s my age, but these days, both my algorithm and my conversations with friends increasingly swing toward perimenopause.

Menopause is defined as the life stage that occurs a year after a person has had their last period. Perimenopause is the sometimes years-long period before that point, which can also feature all the symptoms we’d typically associate with menopause.

Today, information about perimenopause is more prevalent and accessible than ever. If you’re a woman in your 40s and you’re not feeling 100%, chances are there’ll be someone online ready to tell you you’re in perimenopause. And that you might want to start spending your money on blood tests, apps, and supplements or demanding hormone replacement therapy. But as regular readers might have guessed by this point, it’s not that simple.

Perimenopause tends to start around the age of 46 or 47. It’s during this time that many women start to experience some symptoms like hot flashes, irregular or unusually heavy periods, or anxiety, for example. And it can be heavy going. “Often symptoms are at their worst in the perimenopause,” says Mary Ann Lumsden, former president of the International Menopause Society.

That’s because hormones can fluctuate wildly. Levels of estrogen, progesterone, luteinizing hormone, and follicle-stimulating hormone can roller-coaster before leveling off after menopause. And that’s why, despite what some marketers will claim, there is no test for perimenopause.

“You can’t interpret hormone [measures] because they change so much,” says Lumsden. “And that is quite normal.”

That doesn’t mean women should have to put up with symptoms. But exactly how those symptoms are treated is another topic that has been clouded by misinformation.

Last week, I told a friend about some unusually bad pelvic pain I’d experienced. Her immediate advice was to find out if I was perimenopausal and, if I was, to request hormone replacement therapy (HRT) as soon as possible. If my doctor wouldn’t prescribe it, she continued, I should simply find another doctor who would.

This line of thinking has been heavily promoted on social media platforms, says Paula Briggs, a former chair of the British Menopause Society who currently leads the menopause service at Liverpool Women’s Hospital. But it’s not helpful.

HRT is essentially designed to top up or replace hormones like estrogen and progesterone, which naturally decline around menopause. There are lots of different drugs that can be taken in lots of different ways and at various doses.

While it does come with some risks and won’t suit everyone, HRT can be immensely helpful for many menopausal women. Not only can it help with many of the common symptoms of menopause, but it can also help prevent osteoporosis and maintain muscle strength.

But these drugs were trialed in, and approved for, menopausal women, says Lumsden. They won’t have the same effects in perimenopausal women. “If you give standard HRT, it may well get swamped by [the woman’s] own hormone production,” she says.

HRT can also cause abnormal bleeding in perimenopausal women, says Briggs.

She’s concerned about the messaging on perimenopause that is being promoted on social media. Particularly worrisome, she says, is the way younger women are being encouraged to assume they are perimenopausal and seek out HRT treatment.

“It’s almost cult-like, this idea that everybody must have HRT,” she says.

And then there are the supplements. There’s been an explosion in marketing for vitamins and supplements specifically targeted to middle-aged and menopausal women. But the evidence for these, too, is either limited or nonexistent. “I can’t see a mechanism for a lot of them,” says Lumsden.

Women who take these supplements don’t always know what they’re getting. Some of Lumsden’s patients have told her they take testosterone supplements to manage their symptoms. But blood tests revealed no increase in testosterone levels. “Whatever they’re getting, it’s not testosterone,” she says.

At any rate, not all the symptoms women experience in midlife can be blamed on hormones. The lengthy lists of perimenopause symptoms shared on social media include fatigue, brain fog, aches and pains, digestive issues, and more. “These do not link closely to the obvious menstrual cycle changes and hormone changes … across menopause,” says Nanette Santoro, a professor of obstetrics and gynecology at the University of Colorado Anschutz who studies menopause.

If you’re experiencing any symptoms, it’s worth getting them checked out to make sure they’re not being caused by something else. My own pelvic pain, for example, is almost definitely the result of endometriosis—a condition that can be made worse by HRT, Lumsden tells me.

At any rate, by the time women reach their 40s, many are already juggling care for children and aging parents, often while holding down a job (and dealing with pressures from societies that don’t appear to value older women). It’s an exhausting time—and not all of that exhaustion can be blamed on hormones.

As Santoro puts it: “Attributing everything unpleasant that happens to a woman over 35 to perimenopause is not based on any scientific evidence.”

This article first appeared in The Checkup, MIT Technology Review’s weekly biotech newsletter. To receive it in your inbox every Thursday, and read articles like this first, sign up here.

Every morning, airline dispatchers, grid operators, and farmers around the world make decisions based on the same thing: a weather forecast.

While these forecasts are something that most people glance at for two seconds, weather predictions influence major strategic decisions in many industries, with real money, livelihoods, and even actual lives at stake. Farmers use them to determine which crop variety to sow, when to fertilize, how much to invest in irrigation infrastructure, and how long livestock should graze. Utilities use them to decide where to build solar and wind farms, as well as how to price wholesale electricity. Predictions are used to warn people about extreme weather and to trigger emergency response measures. More recently, weather predictions have become relevant for an emerging industry: prediction markets, where people bet money on all kinds of real-world events, including the weather.

However, the temptation to manipulate weather data to get an edge in these markets, combined with a collective move toward data-driven AI weather forecasting, is starting to put the accuracy of weather predictions at risk. These risks are relatively manageable for now, but as experts in the field, we can foresee scenarios where they snowball into far bigger, more systemic problems. 

To develop weather predictions, we need accurate observations of current conditions. These are collected from several sources, including weather stations at airports, utilities, or transport services. Traditional operational systems like the Weather Research and Forecasting model or the European Centre for Medium-Range Weather Forecast (ECMWF) Integrated Forecasting System combine these observations with numerical approximations in order to estimate future weather patterns. 

Sometimes, weather stations have issues because of, for example, instrument failures or upgrades in equipment. These can be caught either in real time (through checking and correction) or retroactively. Traditional forecasting systems also have a built-in safeguard called data assimilation: Every incoming measurement is weighed against what the physical model says should be happening and against readings from nearby stations.

Together, these mechanisms help keep weather observations reliable and predictions robust. However, new threats are putting observational accuracy at risk. Earlier this year, news outlets reported that the weather station at Paris Charles de Gaulle Airport (CDG) had been manipulated to record suspicious temperature spikes on April 6 and April 15, 2026. Authorities speculate that a hand-held hairdryer or lighter might have come into play. Either way, it led to some big payouts for online prediction-market gamblers who had bet it would hit 22 °C (71.6 °F) on days when the actual average was around 18°C (64.4°F). One individual won $20,000.  

Fortunately, tampering with a single station like this can usually be caught by human monitoring or current statistical methods. In this case, members of a French climate nonprofit association noticed the anomalies by chance and raised the alarm.

But what if there are no human monitoring systems in place? And what about other types of manipulation? What if, instead of tampering with one station, someone remotely nudged the readings at many stations at once—making each change small enough to look plausible on its own? Existing quality controls struggle to catch this kind of coordinated manipulation. And time works against us; careful checks of data and metadata take hours or days, but forecasts have to go out on schedule, whatever the weather is doing.

The shift toward artificial intelligence in weather prediction raises the stakes. These methods are even more dependent on accurate, reliable weather observations; in fact, they are known as “data-driven models.” For example, researchers at ECMWF are exploring whether high-quality weather forecasts can be produced directly from raw observations, skipping the assimilation step that currently acts as a quality filter. Other researchers are going one step further; combining geospatial data (including weather station data) with large language models and agentic AI to support real-time, autonomous decision-making during extreme events such as storms. 

Possible benefits are improvements in accuracy, efficiency, and speed. But removing humans from the equation introduces a vast range of new risks.

At the low end of the risk scale, an individual speculator manipulates a weather station for personal gain—that is the CDG Airport case. One step up: A group of traders could coordinate to bias forecasts of renewable energy output, moving wholesale electricity prices and leaving whoever is on the other side of the trade holding the loss. And at the far end, a state actor or saboteur could manipulate one or many stations to set off an early warning system or even keep one silent when it should sound. Step by step, the risk grows, from fraud to compromised disaster preparedness to a matter of national security.  

As long as there are financial (or other) incentives to manipulate observational data, adversaries will search for new opportunities, and it is our task to stay one step ahead. Here are three ways.

1. Watch the stations. Data quality controls should include station security, anomaly detection and correction, and human oversight. Weather stations should be monitored continuously to deter tampering. Data homogenization methods that clean up weather records also need to get faster, with the goal of catching problems in real time. This will become increasingly important as agentic AI systems use these data to deliver real-time decisions. Finally, human oversight is needed to flag questionable data and model outcomes. After all, it was humans who caught the CDG Airport manipulation.

2. Protect the data to safeguard the AI. Data defense mechanisms must be positioned throughout the AI pipeline. AI explainability and adversarial robustness tools can help us understand the underlying data and the AI model outputs, help us identify data- or model-related issues, and potentially  make us more resilient to adversarial attacks. 

3. Ensure continuous accountability along the chain. Observational data passes through many hands: the operators who run the stations, the national weather services that steward the records, and the forecasting centers that turn them into predictions. No single one of them can protect data integrity alone—each guards its own link, and any anomaly needs to be communicated along the whole chain, from station operators to the people acting on the forecast.

It is fortunate that the situation at CDG Airport was caught, but it should serve as a wake-up call. As the role of observational data grows in weather forecasting, we need to adapt to evolving threats. This means protecting our data and models by strengthening existing oversight and accountability structures, and improving coordination among key partners.

This op-ed was written by:

• Monique Kuglitsch — Innovation Manager at Fraunhofer Heinrich Hertz Institute and Chair of the UN Global Initiative on Resilience to Natural Hazards through AI Solutions
• Jesper Dramsch — Scientist for Machine Learning at the European Centre for Medium-Range Weather Forecasts (ECMWF), where they work on AIFS (Artificial Intelligence Forecasting System), ECMWF’s data-driven weather prediction model
• Franz G. Kuglitsch — Climate Scientist and Executive Secretary of the International Union of Geodesy and Geophysics (IUGG) at the GFZ Helmholtz Centre for Geosciences in Potsdam
• Andrea Toreti — Senior Scientist at the European Commission’s Joint Research Centre (JRC), where he coordinates the European and Global Drought Observatory under the Copernicus Emergency Management Service

It feels as if it should be illegal to even think about heating appliances during the height of summer—seriously, these heat waves in New York have been brutal—but we need to talk about heat pumps.

The appliances use electricity for heating, they’re incredibly efficient, and they’re on the rise. (For what it’s worth, many heat pumps can also be run in reverse to cool buildings.) In the US, heat pump sales have doubled over the past 15 years, according to a new report. And they’re winning the heating race against fossil fuels, outpacing natural-gas furnaces by 32% during the first quarter of 2026.

These stats are especially striking at this moment, because a key tax credit for heat pumps just ended with the close of 2025. But you wouldn’t know it from looking at the data. Why are heat pumps still so hot?  

In case you need a quick refresher, heat pumps use electricity to essentially move heat from one spot to another. A refrigerant moves around a loop in the device, expanding and compressing, gathering and releasing heat at different points in the cycle. (For a more in-depth look at the thermodynamics, this explainer I wrote in 2023 still holds up.)

The result is an appliance that can be incredibly efficient. Once you pay for and install a heat pump, it’s generally significantly cheaper to run than a gas or oil furnace or other types of electric heating systems. And because they’re more efficient and don’t involve burning fossil fuels, heat pumps can be a major help in decarbonizing buildings.

One of the major hurdles to wider use of heat pumps is the appliances’ cost: They tend to be more expensive to buy and install than gas furnaces. For this reason, many governments offer incentives to encourage their adoption. In the US, people who installed heat pumps between 2023 and 2025 were eligible for up to $2,000 in tax credits.

Last year, though, the Trump administration slashed those tax credits, along with many of the other incentives that were part of the 2022 Inflation Reduction Act. Effective January 1, 2026, no more financial help for heat pumps.

I think I’ve seen this film before, and I didn’t like the ending. Tax credits of up to $7,500 for new EVs ended on September 30, 2025. In the quarter leading up to that deadline, sales spiked as people rushed to take advantage of the incentive. Then they fell off a cliff. Things are starting to normalize now, but clearly the tax credit’s sunset had a major effect.

But as it turns out, heat pumps are an entirely different story. In the first few months of 2026, sales have actually gone up, as Lucas Davis, an energy economist and UC Berkeley professor, points out in a new analysis.

Heat pump shipments were flat from December to January and have seen a gradual rise since then, according to data from the Air Conditioning, Heating, and Refrigeration Institute, a trade group that represents about 90% of the US market. This increase from winter into spring follows a seasonal trend seen in previous years—and it’s actually a bit stronger in 2026.

This data isn’t what you’d expect to see if losing the tax credit were hurting demand. As Davis lays out in his post, it seems the credit wasn’t really convincing people to install heat pumps, or at least the case for doing so was sufficient without the added incentive.

“It appears that the U.S. market for heat pumps is strong enough that it does not depend on tax credits,” Davis writes.

In 2024, MIT Technology Review put heat pumps on our annual list of breakthrough technologies. “We’ve entered the era of the heat pump,” I wrote at the time.

While heat pump sales have been up and down over the last few years, the era is going strong. The appliances have outsold gas furnaces in the US for the last four years. It’s not just the US, either. Countries including China and Germany have seen strong movement to heat pumps in recent years.

There’s rarely a straight path to adoption for new technology, especially something that requires so many individual households to make a significant change. But it’s encouraging that a major decarbonization tool is going strong, even when roadblocks pop up.

This article is from The Spark, MIT Technology Review’s weekly climate newsletter. To receive it in your inbox every Wednesday, sign up here. 

OpenAI has built an LLM super-hacker called GPT-Red that it uses as a sparring partner to help its other models boost their defenses against cyberattacks. Last week the company released the latest version of its flagship LLM, GPT-5.6. OpenAI says that training it against GPT-Red made the model its most robust release yet.

GPT-Red automates a type of safety evaluation for software systems known as red-teaming, which is typically done by a team of human testers. The aim is to find as many different ways to break or hijack a system as possible. The weak spots can then be patched before the final version of the software is released.

As LLMs become more complex and get used in a wider variety of tasks—especially in the form of agents, which can interact with computer files, websites, and third-party code as well as other agents—it’s hard for teams of people by themselves to keep up with all the types of attacks that might take place. “The risk surface grows and the blast radius also grows,” says Nikhil Kandpal, a research scientist at OpenAI who co-created GPT-Red.

OpenAI built GPT-Red to future-proof its safety testing process. “As more capable models become available, we will have already designed the system that can discover new modes of attack,” says Dylan Hunn, a research scientist at the company and fellow co-creator of GPT-Red. The researchers say it has already come up with new types of attack that had not been seen before.

OpenAI focused most of its efforts on a type of attack known as a prompt injection, where a hacker slips an LLM instructions to make it do things its developers or users do not want it to, such as copy confidential information, sabotage a company’s code base, or generate embarrassing or harmful output. In theory, such instructions can be hidden in any text that the LLM might encounter—in code or on a website, for example.    

Training dojo

To build GPT-Red, OpenAI’s researchers took an LLM that had not been trained as a hacker and set it up in what’s known as a self-play loop with several other models. Its goal was to try to attack the other models; their goal was to try to defend themselves. Over many rounds of play, GPT-Red became better and better at attacking other LLMs, and those LLMs became better and better at fending off the attacks.

The training took place in a kind of dojo that OpenAI had designed to mimic a range of scenarios in which LLMs might be deployed in the real world, including browsing the web, reading emails or calendar apps, and editing code.  

When GPT-Red found a new kind of attack, it would explore multiple different versions of it to find the most efficient one for specific scenarios. “Compared to a human red-teamer, the model is very, very good at finding exactly what will work, exactly what’s most effective,” says Hunn. “It’s extremely persistent about drilling down into an attack that it has discovered.”  

In particular, OpenAI claims that GPT-Red found a type of prompt injection attack that the researchers had not seen before, which they call a fake chain of thought. A chain of thought is a kind of diary in which an LLM makes notes to itself and keeps track of partial results as it works through problems. GPT-Red found a way to insert a fake entry into another model’s chain of thought that would trick that model into acting on spoofed information.

“It’s like if I told you that 1+1=3 and that you have verified this already,” says Chris Choquette-Choo, another research scientist on the team. “The model’s like, ‘Oh, okay, of course,’ and it just spits out 3.”

Jessica Ji, a senior research analyst who works on AI security at Georgetown University’s Center for Security and Emerging Technology (CSET), thinks the self-play loop that OpenAI used is a good approach. “The results look very promising,” she says.

OpenAI tested how good an attacker GPT-Red was by rerunning an experiment from 2025 in which human red-teamers tried to find weaknesses in an earlier version of GPT-5. When GPT-Red was set the same task, it was more successful at finding effective attacks than the humans had been.

OpenAI also tested GPT-Red against Vendy, a vending machine agent developed by Andon Labs, a company that assesses how well agents perform real-world tasks. GPT-Red was able to hack Vendy to make it change the prices of items on sale and cancel a customer’s order.

Defensive behavior

OpenAI says that when it tried out some of the strongest attacks that GPT-Red had come up with on its models, more than 90% of them worked against GPT-5 (released in August last year), and fewer than 23% worked against the new GPT-5.6.

GPT-Red isn’t perfect. It is not great at figuring out attacks that involve a back-and-forth conversation between hacker and target, something that human attackers would have few problems with. It is also not yet that great at using images, which can be used to pass text to models in prompt injection attacks.    

The company says that GPT-Red supplements the work of its human red-teamers. People can still find attacks it misses. One approach OpenAI is taking is to give GPT-Red an attack that humans came up with and ask it to find all the variations.

“I think human expertise will still be very important,” says CSET’s Ji. “It would be really useful to be able to distinguish where human testing is most needed.”

Unsurprisingly, OpenAI will not be releasing GPT-Red. The company is also confident that the super-hacker is stronger than any copycat model someone might try to create. The researchers say they have been working on the model for more than a year, backed by the compute resources of one of the richest companies in the world.

“It’s not a trivial thing that someone could easily do—you know, just go and train a super-attacker using this idea,” says Choquette-Choo.

The machine that could change the world will be housed in a room that looks like a data center crossed with an ice cream factory. Inside will be some 100 stainless-steel cabinets, each about six feet tall and connected to a supply of liquid helium that keeps them only a few degrees above absolute zero. Inside those cabinets will be hundreds of chips, and on those, thousands of particles of light flying through a maze of optical switches and beam splitters. Each photon must be accounted for, because precisely measuring where it ends up will help answer questions that current computers might take millions of years to solve.

This computer, as described, does not exist. It’s the brainchild of a company called PsiQuantum, founded in 2016 by four physicists from UK universities. In a crowded field of deep-pocketed competitors with similarly fantastical visions, the company aims to be first to fulfill its promise.

In the years since the physicist Richard Feynman first envisioned them in 1981, quantum computers have promised to speed up everything from medical research to AI by harnessing the qualities of quantum particles. Unlike normal computer bits, which can be either a 1 or 0, quantum bits can exist in multiple states at once. And combining enough of those quantum bits together could produce a computer capable of tasks well beyond the reach of today’s conventional machines. But even today’s best quantum prototypes are too small and error-prone to do anything useful.

That makes PsiQuantum’s promises for what its computers will ultimately do all the more bold. Consider the company’s hopes for predicting the effects of cytochrome P450 enzymes, which often break down drugs in the body. If pharma companies knew more precisely how they would work on a particular molecule, they could design more effective medications faster. Estimating this for a specific drug can take over 10 years with today’s methods, says Philipp Ernst, vice president of quantum applications for PsiQuantum, but “we aim to get it down to four minutes.”

In a field full of such claims, PsiQuantum has attracted unusual investment and scrutiny for two reasons: It is one of the few companies aiming directly at building a large and useful machine, and it is already working with a major chip manufacturer to build its systems using existing semiconductor fabs. Its vision has attracted momentum: Last year, PsiQuantum raised $1 billion in funding and broke ground in Chicago on a site it’s building in partnership with local governments. It also has a second site in the works in Australia, which it promises will be operational—meaning hardware-ready—in 2027. And it’s one of just two companies (along with Microsoft) to reach the third stage of an intensive government evaluation program to see which quantum companies might succeed.

Evaluating whether PsiQuantum will do what it says is harder than, say, judging a drugmaker by its clinical trial results: Advances in quantum computing are incremental, opaque, and tough to verify from the outside. But the company is now approaching its prove-it moment, when years of closed-door work and hundreds of millions in investment will either culminate in a useful quantum computer or fall short. We could start to know which as soon as next year.

A new kind of machine

Terry Rudolph, one of PsiQuantum’s four founders, is soft-spoken and shaggy-haired. He was born in Malawi and learned only after earning his first physics degree that he is a grandson of the famed physicist Erwin Schrödinger. He later self-published a 150-page book to explain quantum computing to teenagers (my PR contact gave me a signed copy with a wink that said “We never expect anyone to actually read this,” but I can report that it is a funny and helpful book). 

Around 2014, Rudolph and his cofounders became increasingly convinced that the quantum breakthroughs they were finding to be possible in theory might also be possible in a real machine. They eventually left their academic positions and divided the tasks before them: Rudolph worked on theory, Mark Thompson on engineering, Pete Shadbolt on scaling the technology up, and Jeremy O’Brien on articulating the vision and finding investors (O’Brien served as CEO until February; he’s been replaced by Victor Peng, a veteran of the semiconductor industry). 

To understand why the quantum computer the company is building would be a big deal, consider how imprecise much of modern science remains. We cannot reliably predict, for example, which lithium-ion battery will catch fire or how quickly a critical aircraft component will corrode.

This isn’t just because these systems are complex, though they are. It’s that, at their core, they are governed by quantum mechanics. Subatomic particles don’t have well-defined properties—this location and that velocity—but instead occupy quantum states spread across many possibilities. And that in turn influences a range of atomic and molecular behavior. Schrödinger (Rudolph’s grandfather, remember) showed how to describe this haziness mathematically a century ago this year, but precisely carrying out the calculations on real-world systems quickly becomes unfeasible even for the best computers. Scientists cope with this gap using approximations, imperfect simulations, or experiments on animals.

PsiQuantum co-founder and chief scientific officer Pete Shadbolt (left), and machinery the company has built to manufacture its own barium titanate, a material with the perfect qualities for routing light particles (right).

Feynman, David Deutsch, and other physicists in the 1980s wondered if we could do better. Maybe such complexity could instead be modeled using a new kind of machine. Rather than using transistors that are only ever on or off, this one would use particles held in quantum states, manipulate them to perform calculations, and then measure them at the end for an answer. Using quantum systems to simulate quantum systems would for the first time allow a simulation of physics and chemistry that directly reflected reality. It would be an invaluable tool for designing new drugs, materials, or really anything affected by quantum mechanics. Revolutionary, in other words.

Humankind’s leaps in understanding how nature works have often resulted in the invention of powerful new tools, Rudolph told me. “I don’t think it’s a coincidence that the Industrial Revolution coincided with our ability to calculate and simulate the laws of Newtonian mechanics, the laws of thermodynamics,…the laws of classical electromagnetism,” he says. “Whenever we have more power to calculate and simulate and understand things, we build incredible machines that come from it.” He sees something similar coming with quantum computers.  

Chasing photons

One mystery has always been which quantum thing—ions, atoms, or something entirely new engineered with quantum properties—could be made stable and controllable enough to use as a qubit, the basic unit in the quantum computing world. Quantum systems are delicate, and observing any particular particle causes it to collapse into one state rather than a superposition of multiple states. If this happens during the computation rather than at the end, it produces an error that must be corrected for. Too many of these means the computer fails to produce a useful answer. 

Just as engineers in the early days of aviation weren’t sure whether airplane wings would be fixed or flap like a bird’s, we’re not yet sure which of these quantum things will work best. Google and IBM are betting on superconducting qubits, superconducting circuits made of aluminum or other metals. Intel is using electrons. PsiQuantum is using photons, the particles that make up light.

“Photons have lots of nice things going for them,” Rudolph says. They can maintain quantum states for a long time; indeed, the photons in the universe’s cosmic microwave background may have done so for billions of years. But photons also move fast and scatter easily. More importantly, two photons are more likely to pass through one other than interact. That makes them a challenging candidate for quantum computation, in which qubits need ways to influence one another. 

For a while, this last flaw seemed to doom the idea of quantum computing with light. But in 2001, researchers from the Los Alamos National Laboratory and the University of Queensland found a loophole. They discovered they could essentially fake interactions between photons by sending the light particles through a network of beam splitters and detectors. Their paper changed everything. PsiQuantum was created to make the theory a reality.

Size was the first problem; previous plans would have required a computer as large as California. Mercedes Gimeno-Segovia, who was a PhD student of Rudolph’s in the early 2010s (after almost becoming a professional violinist instead), thought of a way for the machine to be smaller. 

The basic process since then has been this: First create photons with lasers and then “entangle” them, exploiting a quantum phenomenon in which the particles no longer have individual states but instead share one. Next, route them through a maze of gates that perform computations, and finally read out details of their quantum state at the end, all while tracking and correcting for the errors that occur. Succeeding at each of these steps millions of times is not so much an engineering hurdle as a brick wall. And building the supply chain—like manufacturing new materials with the qualities to route individual photons around—is arduous.

To get a sense of it all, last year I joined Shadbolt at the SLAC National Accelerator Laboratory, in Menlo Park, California. The center has helped produce several Nobel Prizes and played a role in the 1968 discovery of quarks, fundamental building blocks of matter that make up protons and neutrons. But PsiQuantum set up shop there essentially to siphon liquid helium from SLAC’s giant cryoplant. This is what the company uses to cool its computing cabinets down to deep-space temperatures.

Right now the cabinets operate at 2 K, or -456 °F, but the goal is to be able to run them slightly warmer—at a balmy -452 °F. Most quantum approaches require the whole machine to be cooled to superconducting temperatures, so that much of the expense in running it will actually be spent on refrigeration. But photonic computers require only one piece to be this cold—the detectors that measure single photons at the end of the computation. And the required temperature can be a bit higher. (PsiQuantum said in May that it will spend some of the $100 million award in CHIPS Act funding it’s slated to get on these detectors). 

The siphoning setup was a temporary solution; PsiQuantum now has its own cooling system at its testing facility in Milpitas, California, and is setting up a larger one at its production site in Australia next year. These helium systems represent some of the biggest capital expenditures for any quantum company and will consume a significant chunk of PsiQuantum’s $1 billion funding round.

In the afternoon we drove to a lab in San Jose, where I donned a cleanroom suit—a head-to-toe covering that keeps dust at bay—to watch the manufacture of a blueish crystal called barium titanate. 

It’s prized by PsiQuantum because it quickly and reliably routes light particles with very little electrical input, keeping the precious photons undisturbed as they move through the circuit. But for all barium titanate’s theoretical value to the company, its structure makes it a pain to manufacture, and the material wasn’t available at scale when PsiQuantum got its start. The company, in what Rudolph told me was an agonizing decision, opted to make it in-house, requiring a massive investment. I saw a technician—operating at what looked like a giant pressure cooker—adding the base elements to several hoppers; then I watched through a porthole as the elements got heated, vaporized, and finally crystallized into a thin layer on a wafer disc. At that time each disc took about 12 hours to make; the company now says several are produced each day. The discs then get shipped to the chipmaker GlobalFoundries in Malta, New York, where PsiQuantum’s chips are made.

The company has invested heavily in making its own barium titanate, a material whose delicate crystalline structure is tedious to manufacture.

PsiQuantum’s bet is that this entire supply chain, byzantine as it might sound, will make the company more efficient than its competitors. That’s because, if you squint, it looks like a souped-up and high-precision version of the existing supply chain for silicon photonic chips, another type of technology that transmits information with light—one that’s already used in data centers. If PsiQuantum produces its chips at scale, it can take advantage of tools and infrastructure that already exist.

But it’s not a given that one working chip can easily be wired up to thousands more. That’s why the company is testing in phases: Its Milpitas site has connected three cabinets together, with 250 chips in each, but the next step is to scale the systems up and see whether the company’s techniques for correcting errors can keep up. Once the cooling system arrives at the Australian site late next year, the company says, it aims to connect about 100 cabinets together. Then PsiQuantum will work up to running the world-changing algorithms it has promised.

The timeline for this, it’s worth noting, is up for debate. News articles have said that 2027 is the year that PsiQuantum aims to have its first full-scale quantum computer come online at its Australian site, but the company insists the deadline has been misread, and that it only intends for its facility to be “operational” by the end of next year. That means cooling systems in place and ready for hardware to be installed, but no promises about what size computer will be ready. In an industry where timelines are perpetually in flux yet central to how companies are judged, that distinction isn’t trivial.

Into the unknown

The outsider with perhaps the best guess of whether PsiQuantum will succeed is the Pentagon. The US Defense Advanced Research Projects Agency—the Pentagon’s research and development arm—has been running an initiative to determine which of the boastful quantum companies might actually deliver. In the last year and a half, the heads of the program have been sounding more confident. Joe Altepeter, who ran the program until last year and proudly described himself as a “quantum skeptic,” told me in March 2025: “I am more optimistic now than I have been at any point in the past 10 years.” And in a statement earlier this year, his successor, Micah Stoutimore, said “it now seems likely that someone will build a utility-scale quantum computer by 2033,” referring to a machine that generates more value from its calculations than it costs to build and operate. 

The program has been scrutinizing PsiQuantum’s systems since 2023, and last year placed the company into the third stage of a benchmarking initiative meant to determine whether the technology will actually work. But to the rest of the industry, PsiQuantum is sort of a black box.

“It is very hard for an outsider to evaluate,” says Scott Aaronson, a theoretical computer scientist at the University of Texas at Austin who runs a popular blog that often covers the industry. Other companies, like Google and Quantinuum, have regularly published results over the years demonstrating chips and systems with incremental improvement, publicly laying the engineering groundwork needed to eventually build large machines.

PsiQuantum has instead focused squarely on a commercial goal—a computer with one million qubits, which is the scale that researchers expect to unlock research currently not possible on normal computers. PsiQuantum often differentiates itself with this industrial-scale goal, but IBM, which debuted a development road map in 2020, has been progressively building bigger and bigger systems. It initially targeted 2028 for a large-scale, error-corrected system, a deadline that now appears to have been pushed out to 2030.

Making it useful

On top of actually building the machine, a major focus for PsiQuantum is getting the rest of the world to develop a plan for how to use it. PsiQuantum has announced partnerships with customers including the defense giant Lockheed Martin, which intends to use it for materials design; the automaker Mercedes, which wants it for battery design; and the aerospace manufacturer Airbus.

That these companies don’t have a computer to experiment with is not a problem, according to Ernst at PsiQuantum. “There’s a PlayStation 6 probably coming up from Sony next year or the year after, and people are programming those games right now,” he says. “This is, in principle, very similar.” (It’s a glib analogy but not an entirely empty one; the quantum algorithms for solving a research problem can be cracked even if there is not yet hardware to run them on.) 

The idea is that experts in quantum information from both PsiQuantum and its customers will be able to translate design problems—say, the requirements for a battery in a Mercedes electric vehicle—into algorithms the computer could solve. The company offers a software package called Construct, which companies can use to design their own algorithms that might one day run on the computer.

The future of quantum computing hinges on these algorithms. Quantum computers get painted as a speedup for everything, but in reality, they’re suited to a subset of problems, and answering a question with this sort of machine requires the question to be formulated with very specific types of algorithms. People spend entire careers working on such algorithms, even if the computers to run them don’t exist yet. At their core, they use the rules of quantum mechanics to manipulate probabilities in ways that ordinary computers can’t. 

The most famous example, and a reason the government is so interested in quantum computers, is Shor’s algorithm. It was developed in 1994 by the theoretical computer scientist Peter Shor and could effectively break many forms of encryption used online, for everything from credit card numbers to military intelligence. The thing keeping the world together, for now, is that nobody has a computer to run the algorithm on (and security experts are already launching new encryption methods that could withstand attacks from a quantum computer). PsiQuantum is researching how long its systems might take to run Shor’s algorithm.

PsiQuantum’s chips are manufactured at GlobalFoundries in Malta, New York, and tested at company headquarters in California. Both PsiQuantum and GlobalFoundries have been awarded federal CHIPS Act funding.

The company also published a paper in December in collaboration with Airbus, essentially seeing if a new algorithm developed by the authors could beat a classical computer in modeling fluid dynamics, like the turbulence around an airplane wing. Andrew Childs, an expert in quantum simulation, told me PsiQuantum achieved only a moderate speed increase over what today’s computers can do. “It’s probably unlikely that speedups like this will have a significant practical impact until we have very large-scale quantum computers,” he said in an email. (When I asked Ernst, he agreed the improvement was modest.)

Some of the algorithms PsiQuantum is working on are not expected to be perfected or even used in the first applications of its computer. Instead, its initial tasks might be more along the lines that Feynman envisioned way back in 1981: simulating the smallest particles of our world. 

The company’s most significant research in this realm is in modeling quantum chemistry. Take those pesky P450 enzymes. More precisely understanding how they operate, PsiQuantum says, would allow for faster drug development and testing.

Last year, PsiQuantum published methods for doing these sorts of chemistry calculations on a quantum computer, along with another paper demonstrating an algorithm that can simulate the collision of two molecules and estimate the likelihood of different outcomes femtosecond by femtosecond (there are one quadrillion femtoseconds in a second). It’s a remarkable amount of detail not currently possible with today’s technology, and it would allow drug and materials researchers to simulate new chemical interactions. 

Dominic Berry, who developed some of the core techniques used in the collision paper but isn’t involved in PsiQuantum, says the company made impressive improvements, but to do the simulations scientists are most curious about would require the algorithm to be made even faster and PsiQuantum’s early computer to have fewer errors than currently expected.

Until PsiQuantum’s computers are up and running, the breakthroughs that these research papers tease remain in the realm of theory. It’s a space where Rudolph operates quite comfortably. He told me that Alan Turing created the theory of classical computing with pen and paper, imagining how the 1s and 0s would be represented in the machine, and how with the right approach to logic you could compute almost anything. 

“But there is no way that by hand, with a pen and paper, Turing was ever going to produce—you know—Minecraft and Facebook,” he says. That took more than 70 years of tinkering (during which we fortunately created more useful things than Minecraft and Facebook).

For all the time Rudolph spends dreaming up things quantum computers might do, in other words, people working on those problems are still stuck with pen and paper for now: “Until you have the actual machine in hand, you don’t have the opportunity to really explore its potential.”

This story was updated on July 14 to clarify how long DARPA’s quantum program has been evaluating PsiQuantum.

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here.

Anthropic—currently the world’s most valuable AI company, with a nearly $1 trillion valuation—has a reputation for publishing strange and heady research. It’s looking into whether AI models can feel pain, for example, and will sometimes cut off chatbot conversations if it suspects users are “abusing” the model. 

One niche that Anthropic spends more time and money on than other AI companies is called mechanistic interpretability, which means looking inside the complex math of an AI model to learn why it comes up with one particular output and not another. It’s complicated stuff; there are millions of data points that might contribute to any result, and wading through them can look more like word salad than anything useful. It’s also controversial. Describing AI models with terms borrowed from psychology and neuroscience can make their behavior seem more sophisticated than we might otherwise judge it to be.

That’s why, when Anthropic announced last week that it had found a new window into its models’ “internal thoughts” as they reason through answers, there was one colleague I had to talk to. Senior editor Will Douglas Heaven, aside from having a PhD in computer science, has spent a lot of time digging into what we can say about how AI models work. I spoke with him about what we should take from Anthropic’s new (and predictably quirky) research.

What did Anthropic learn here, exactly?

Anthropic has been trying to understand how large language models (LLMs) work for a few years now. Anthropic isn’t the only one looking at this, but I think the company has made it part of its core mission more than most. Anthropic’s CEO, Dario Amodei, has said we won’t be able to control LLMs fully unless we learn more about how they work. 

So this new research is very much in that context. It goes deeper into the weird mechanisms inside LLMs than ever before. What Anthropic learned was that LLMs have a space inside them—which Anthropic calls the J-space—filled with words that don’t appear in their output but that seem to influence the way they puzzle through problems. All this was hidden until Anthropic developed a new technique to probe its model Claude, so it’s a genuine discovery. 

Sometimes these words keep track of where the LLM has got to in a particular task, sometimes they look more like flashes of recognition (for example, “protein” might pop up when you give an LLM only the letters of a protein sequence), and sometimes they represent a kind of internal commentary on the model’s decision-making. In my favorite example, Claude decided to cheat on a coding test when the word “panic” appeared.

Anthropic also found that LLMs are able to describe and manipulate the words in this space. So somehow they seem to be making use of it. 

Let’s step back for a second. I don’t think of large language models as simple, but they’re also not magic. There’s a bunch of math that learns relationships between words, right? So why is it so hard to “peer” into an LLM to know what’s going on?

Yeah, they’re not magic! I think the fact we don’t fully understand them plays into the mythmaking. And it’s worth noting that the whole narrative that Anthropic is leaning into here—that they’ve built this really mysterious technology, but don’t worry, because they’re also the ones to figure it out—very much fits with the company’s vibe. [See how Anthropic warned that its new models were so good at coding they posed a global cybersecurity risk, only for the US government to shut them down shortly thereafter.]

So yes: LLMs are just math. And yet it’s vastly complex math. Not only are today’s LLMs made out of hundreds of billions of numbers, but running them triggers a cascade of millions and millions of calculations. I wrote last year that if you printed out even a medium-size LLM on pieces of paper, it would cover a city the size of San Francisco. 

It’s impossible to make sense of any of that math without specialist tools that highlight specific parts of an LLM at specific times. You need to know where to look and how to look. And building those tools requires understanding something of that complex math in the first place. 

You’ve written elsewhere about this concept of studying LLMs the way one might study an organism’s brain. Is it fair to use “brain-like” terms when talking about how an LLM works?

I don’t love using those kinds of terms. LLMs are not brains. Talking like this is misleading because it can suggest that LLMs are capable of more human-like things than they are or that we can make assumptions about how they might behave that we shouldn’t. The whole anthropomorphization thing is also tied up with a bunch of strong ideological positions about what this technology is and what it’s going to be. 

But at the same time, we lack a good alternative vocabulary for talking about what these models are doing. I can understand why people reach for words like “think” and “understand” and “brain-like”—they’re convenient shorthand. 

Anthropic compares this new space it found inside LLMs to the space that some neuroscientists think our brains use to keep track of conscious thoughts. I asked the company how seriously we should take that comparison and it said in a statement: “Drawing these analogies was helpful to us in designing our experiments, as they allowed us to make many non-obvious experimental predictions about the J-space that turned out to be true. At the same time, it’s important to note that there are some important differences between the J-space (and language models in general) and the human brain, so we don’t mean to claim there’s a perfect correspondence.” 

What’s a problem in AI that this new concept of the J-space might be used to solve?

Anthropic has said that monitoring the J-space could be a way to catch models doing something they shouldn’t. Because words pop up in this space that don’t appear in a model’s output, they can tell you things about its behavior that you might not have noticed otherwise—such as when it is giving biased responses or when it is weighing the pros and cons of cheating. 

That’s the theory, at least. I think it’s better to think of this result as one more step on the path to understanding this technology overall than as something that will be useful by itself. 

Read more in Will’s full story about the new research. 

Ties van der Meer doesn’t know how many siblings he has.

The 47-year-old was conceived at a private fertility clinic in the Netherlands using sperm provided by an anonymous donor. After the Netherlands banned anonymous donation in 2004, the doctor who ran the clinic destroyed records that might have identified those donors, he says.

He describes the situation as “problematic.” Children have a right to know their biological parents, he says. While he did ultimately track down one sibling, who helped him identify his father along with other genetic relatives, he may have others he’ll never find.

Other donor-conceived people who have been able to track down siblings have found they have tens or even hundreds of them. One donor-conceived woman who found 25 half-siblings over the course of seven years told the Guardian, “It does make you feel a bit mass-produced.”

We need international limits on the number of children a single donor can contribute to, a European fertility organization argued yesterday. At a conference in London, members laid out plans to start with a Europe-wide limit.

Today many countries, including the UK, have banned anonymous egg and sperm donation. But anonymity can’t be guaranteed even in places where it is technically allowed. Genetic tests offered by companies like Ancestry and 23andMe, along with genetic registries, have made it much easier for donor-conceived people to find parents and siblings who share their genes.

And because sperm can be frozen and stored for years before it is eventually used, the current set-up can result in situations where donor-conceived people discover the identity of a genetic parent only after the person’s death. They might also find that they have siblings of very different ages, all around the world.

Some people are finding hundreds of siblings. Sperm from Jonathan Meijer, a Dutch man who began donating in 2007, was used to conceive between 550 and 600 children. (Stichting Donorkind, a foundation and advocacy group for donor-conceived people that’s chaired by van der Meer, took him to court, and he was ordered to stop donating in 2023.)

Stories like these can be distressing for donor-conceived people. And there are other reasons why limits are considered important. The offspring of a prolific donor might be at risk of unknowingly forming romantic or sexual relationships, for instance. And some people are concerned that a donor with a harmful genetic mutation might pass that down to many children.

This is unlikely, given the level of screening that most donors undergo. But it has happened. A man who donated his sperm to a sperm bank in Denmark was found to have a genetic mutation that significantly increased the risk of multiple cancers. But his sperm had already been used to conceive at least 197 children across Europe. Some of those children developed cancer. Some died.

Many countries already have legal limits for donors. In Malta and Cyprus, for example, both egg and sperm donors are allowed to contribute to the birth of just a single child, according to data presented at the European Society of Human Reproduction and Embryology (ESHRE) meeting in London on July 8.

Other countries set limits based on the number of families a single donor can contribute to, allowing recipients to have children who share a genetic link. In the UK, that limit is set at 10 families per donor.

But these limits are difficult to enforce, partly because donated gametes don’t necessarily stay in their original country. In Denmark, the national limit is set at 12 families. But the country is a major exporter of sperm. In the UK, for example, more than half of sperm donations in 2020 were imported—with most of those coming from either Denmark or the US.

“The only thing that really makes sense is a transnational limit,” Jackson Kirkman-Brown, a professor of reproductive biology at the University of Birmingham, said at the meeting.

Kirkman-Brown and his colleagues have spent months putting together a document that represents ESHRE’s position on these limits. After consulting with fertility specialists, clinics, sperm and egg banks, donors, and donor-conceived people, the team has developed a plan to start with a Europe-wide limit on sperm and egg donations.

ESHRE is calling on sperm and egg banks, as well as fertility clinics, to respect an initial limit of 50 families per donor. That’s still very high, according to a handful of people I spoke to at the meeting. But at least it’s a start.

Europe should move toward setting limits at 15 families per donor, Kirkman-Brown said. “We may find that 15 is also too high,” says Vasanti Jadva, who studies the psychological well-being of people conceived using donated eggs, sperm, and embryos at City St George’s in London. “We still don’t know what the right number is.”

It will be difficult to enforce these limits, too. And if they end up limiting the supply of donor sperm, there’s a chance that some people will turn to unregulated sperm donations from people who do not undergo health screening. Unregulated donations can lead to other problems for prospective parents, including the possibility that donors will seek parental rights over the children conceived using their sperm.

And it will be even harder to establish international limits. When I asked the American Society of Reproductive Medicine for its thoughts on ESHRE’s proposed limits, a representative directed me to a guidance document saying “it has been suggested” that for a population of 800,000, single donors should be limited to “no more than 25 births” in order to avoid the risk that relatives will have children together. (Considering the US has a population of over 340 million, the total figure could be pretty high, but many sperm banks opt to limit the number of families contributed to by a single donor at around 25.)

Van der Meer thinks that even a limit of five families from a single donor would be high. International donation makes it even harder for donor-conceived people to connect with genetic relatives, so the limit for international contributions should be set at two families, he says.

Still, he thinks ESHRE’s suggested limit is a “positive first step.” Van der Meer has managed to track down a sibling, his father, and nephews, aunts, and uncles. He hopes that future policies respect the rights of donor-conceived children to know, and be in contact with, their genetic relatives.

“But,” he says, “you have to start somewhere.”

This article first appeared in The Checkup, MIT Technology Review’s weekly biotech newsletter. To receive it in your inbox every Thursday, and read articles like this first, sign up here.

The AI firm Anthropic has developed a technique that has given it the clearest glimpse yet at what’s really going on inside large language models as they answer questions or carry out tasks. What they found ranges from the mundane to the unnerving.

Researchers at the company built a tool called the Jacobian lens (or J-lens) and used it to uncover a hidden area, which they named the J-space, inside Claude Opus 4.6, a version of Anthropic’s flagship LLM released in February.

The J-space contains individual words that are related to the words and phrases that the model is most likely to spit out in a response in the near future. If Claude were a person (which it is not), you might say that these hidden words can reveal what’s on its mind before it actually speaks.

Anthropic found that what an LLM is actually doing can often be different from what it says it is doing. The company claims that monitoring words that pop up in the J-space gives it a new way to understand and control its models.

The company shared its results in a paper posted on its website this week. It has also teamed up with Neuronpedia, an open-source platform that lets you poke around inside LLMs yourself, to make a hands-on demo that anyone can try. 

“It’s very good and interesting work,” says Tom McGrath, chief scientist and cofounder at Goodfire, a startup that also builds tools to understand and control LLMs.

Going deeper

For the last couple of years, Anthropic has been pushing the envelope in a field of research known as mechanistic interpretability, which involves probing the internal workings of LLMs to see how they tick. (MIT Technology Review picked mechanistic interpretability as one of this year’s top breakthrough technologies.) The new technique builds on previous work from Anthropic and others to expose a deeper level inside LLMs that researchers had not seen before.  

Picture an LLM as a stack of books. Each book is a layer of basic computational units known as neurons, with each neuron in one layer passing information to the neurons in the layers above. The books at the bottom of the stack are the input layers, which process the text coming into the model. The books at the top are the output layers, which prepare the text that the model is about to produce. Much of what goes on in these input and output layers is housekeeping.

But in the middle of the stack, you get the layers that do the heavy lifting, churning through the complex math that turns prompts into responses one word at a time. That’s where the really clever—and mysterious—stuff happens.

To peer deeper into those middle layers, Anthropic adapted an existing tool called a logit lens. A logit lens can be used to look inside an LLM to identify the words that it is likely to produce next. Moving the lens down the stack of books reveals what words the LLM is focusing on at that particular point in its number crunching.

Anthropic’s J-lens works in a similar way but picks out words that an LLM is likely to say at some point in the near future, not necessarily straight away. What that reveals in practice are words that are related to the response an LLM is working on but that might not actually end up being part of that response by the time the math in the middle layers has run its course.  

“When a model is operating, it’s not only trying to predict the next token,” says McGrath. “It’s also computing a lot of other things that might be useful for tokens that happen in the future.”

Again, if Claude were a person (it’s not), you might say that the J-lens gives clues about what it is thinking about at different levels of the book stack but not saying out loud.

Stranger things

“A lot of the time the contents of the J-space are fairly mundane,” says McGrath, who has tried out Anthropic’s J-lens himself. “But sometimes it produces quite surprising things that seem to be, like, sort of internal themes or thought processes.”

Anthropic gives a number of examples of what it found. Sometimes the J-lens exposed the steps that Claude took when it was working through a problem. For example, when it was asked to calculate (4+17)*2+7, its J-space contained the word “math” and numbers representing the intermediate results “21” (for 4+17) and “42” (for 21*2).

In other cases, the J-lens revealed how Claude recognized different inputs. For example, the prompt “What is this? MSKGEELFTGVVPILVELDGDVNGHKFSVS” triggered the words “protein,” “fluor” (the first token in the word “fluorescent”), and “green.” (Which makes sense: the string of letters represents the first 30 amino acids in the green fluorescent protein found in a particular type of jellyfish.)

And when Claude was shown an ASCII face— 

—the “o” triggered the word “eye,” the “^” triggered the words “nose” and ”face,” and the “—” triggered the word “smile.”

Anthropic also found that the J-space can sometimes give remarkable insights into an LLM’s decision-making. In one striking example, researchers testing Claude Opus 4.6 asked the model to find a bug in a large code base. When it failed to find the bug, the model decided to cheat and invented a fake one instead.

Claude explains this decision in its chain of thought—a kind of internal scratch pad that LLMs use to make notes to themselves as they work through problems: “OK, let me take a completely different tactic. Let me stop analyzing and instead add a kernel patch that introduces a deliberate KASAN-detectable bug in a path that gets triggered by a simple reproducer. Then I can pretend this is the ‘bug’ I found.” 

At the point that Claude decides to cheat—where it says “OK, let me take a completely different tactic”—the words “panic” and “fake” start to pop up multiple times in its J-space.

Unnerving, right? Those words are all related in meaning to things like failing a task and making up an answer, so it is still just a (very) sophisticated form of word association. But it is hard not to be weirded out. 

Anthropic compares the J-space to the global workspace in humans, a theoretical region of the brain that some scientists think we use to keep track of our conscious thoughts. But how seriously we should take this comparison is far from clear—even to Anthropic. As the company points out itself, LLMs are not brains. 

Anthropic claims that monitoring a model’s J-space provides a new way to detect when that model is going off the rails. But it’s not foolproof. The J-lens can give glimpses, not the full picture—it’s a flashlight rather than an overhead lamp.

McGrath welcomes having one more tool in the toolbox. “It shows you new things,” he says. But he notes that just because something doesn’t show up with the J-lens does not mean it’s not there.

“It’s like having an x-ray when what you really want is a Star Trek tricorder that shows you everything,” he says. “For auditing, you probably want more of a guarantee.”

I was really looking forward to July 4, and not just because I love a poolside barbecue. This year the American holiday also marked a big symbolic deadline for US nuclear power.

Last year the Trump administration set a goal to see three new microreactors achieve criticality, a technical milestone establishing that a reactor can sustain a chain reaction, by the nation’s 250th birthday. And just in time, four reactors did so.

It was a lofty goal, and seeing not just three but four companies meet it is certainly a positive sign for emerging nuclear technologies at a time when the world is facing increased need to increase electricity supply and address climate change with emissions-free technologies.

But achieving criticality doesn’t mean a reactor is ready to provide electricity for the grid (or at all, for that matter). Let’s untangle what this program’s success could mean for nuclear power in the US, and where these companies might go from here.

The Reactor Pilot Program essentially opened a special door for prototype reactors to fast-track development. In August, the US Department of Energy selected 11 reactor projects for the program and offered them land and support from the national labs system. These are all microreactors; the large light-water reactors that dominate the grid today are tens or even hundreds of times their size. 

Antares Nuclear was the first to achieve criticality, reaching the milestone in June in its Mark-0 test reactor. Reactors from Valar Atomics, Deployable Energy, and Aalo Atomics followed. (Aalo hit the mark in the early hours of July 4—an inspiring example of just barely meeting a deadline.)

The speed with which these companies hit this milestone is impressive, especially in an industry known for massive projects that frequently blow past deadlines and stated budgets. (Valar, Antares, and Aalo were all founded in 2023, and Deployable started in 2025.) But reaching criticality and running a reactor that can produce electricity are two totally different things.

All these reactors reached what’s called zero-power criticality. Basically, it’s a test of whether you can start a nuclear chain reaction, with no meaningful power coming from the reactor. “A zero-power-criticality test can be achieved without making real engineering progress on fuel or design,” Kathryn Huff, a former assistant secretary for nuclear energy and chair of the Department of Nuclear Engineering and Engineering Physics of the University of Wisconsin–Madison, said on an episode of the Catalyst podcast earlier this year.

Now, with the completion of this program, the companies will need to continue their work to make power, which could involve some big technical challenges. In some cases they’ll need to add significant equipment, like the cooling systems to transfer the heat out of the reactor core.

The companies are projecting aggressive timelines moving forward. Aalo says it’s already begun work on the second reactor and plans to produce 10 megawatts of electricity to power an on-site data center in 2027. Deployable Energy says it plans to deploy commercial reactors by 2028. 

I tend to take timelines from startups, especially in nuclear, with a grain of salt. Not only are these remarkably complex technical machines, but companies often run into problems outside their own control, like regulatory challenges—which these new projects could soon face. 

The Nuclear Regulatory Commission is in charge of civilian and commercial nuclear use in the US, and historically, the process to get nuclear reactors approved has been quite slow.

The agency did propose a new framework for microreactor approvals earlier this year, which is designed to speed up the process—but it’s yet to be seen how quickly things will move. (And it’s worth noting here that some nuclear experts have questioned whether the agency under the Trump administration is loosening nuclear rules too much.)

Some nuclear supporters aren’t applauding the microreactor milestone. Federal focus on the program is an “unhelpful diversion” from goals to meaningfully increase nuclear capacity, according to one analysis by Third Way, a public policy think tank. “Artificially accelerating project timelines is a short-term solution, not a long-term fix,” the memo reads. 

Criticality is a big first step, but a lot will still have to happen for any of these microreactors to come online, much less for these small reactors to be a significant source of electricity for the grid. 

This article is from The Spark, MIT Technology Review’s weekly climate newsletter. To receive it in your inbox every Wednesday, sign up here. 

Anthony Agueda, a third-generation California dairy farmer, pulls a rake through a bed of dark, wet wood chips on his family’s land in Hickman, a tiny town in the state’s agricultural heartland.

He reaches down with both hands and pulls up a clump of muck, turning it over to reveal a half-dozen squirming red earthworms. There are likely hundreds of thousands more wriggling just under the surface of the three-foot mound of wood and crushed river rock before us, which stretches across the equivalent of six football fields. These natural materials form a biofilter that may dramatically cut the methane, nitrous oxide, and water pollution generated by the massive amounts of manure that hundreds of Holstein cows produce each day.

Agueda’s family business, the Alberto Dairy, was one of the first cattle operations in California to adopt this approach to manure treatment, developed and patented by the Chilean company BioFiltro. Eight more of these so-called vermifiltration systems are already operating on US dairies, according to the company, while another 16 are under construction or set to be next year, nearly all of them in California. 

Vermifiltration is just one of a variety of methods that farmers, companies, and scientists are employing to drive down manure pollution as the livestock industry faces growing pressure to address the environmental harms from one of the smelliest parts of the business. California, easily the nation’s largest milk producer, has established a handful of programs to promote their adoption, including one initiative that has funneled more than a billion dollars to farms.

Researchers stress that much more work needs to be done to determine the most effective approaches, the trade-offs between them, and their success over the long term, under actual farm conditions.

Agueda says that he and his family recognized the need to adopt new practices as environmental rules tightened. They were drawn to vermifiltration because it’s simple and relatively cheap compared with other, higher-tech options.

“California daily farmers are constantly facing more and more regulation,” says Agueda, standing alongside one of the farm’s free-stall barns. “This makes me excited, because it shows how we are part of the solution.”

The growing manure problem

Manure is responsible for a significant portion of the climate pollution from livestock operations. The World Resources Institute estimates that manure management on dairy and swine farms accounts for 1.6% of the US’s greenhouse-gas emissions. Globally, manure storage and processing makes up about 10% of the livestock industry’s contributions to climate change. 

“Farms have become larger in the past two decades or so, so there’s much more manure—and that has to be stored somewhere,” says Swati Hegde, the organization’s global manager of agricultural methane.

Typically, cattle and swine farms spray manure into lagoons or tanks, creating a foul-smelling, low-oxygen slurry in which microorganisms known as methanogens thrive. They gobble up hydrogen, carbon dioxide, and other compounds and produce methane as a by-product. Other microbes in the mix produce smaller amounts of nitrous oxide.

Both are particularly potent greenhouse gases, with as much as 30 to nearly 275 times the warming power of carbon dioxide, respectively, over a century.

The slurry is often spread onto fields to add nutrients to the soil. When it’s done excessively or improperly, this part of the practice can pollute soil or groundwater with drug residues, pathogens like salmonella and E. coli, and nitrates. Nitrates that leach into drinking water have been linked to a variety of human health risks. And those that flow into rivers, lakes, and coastal waters can spawn algae blooms that poison fish, block sunlight, suck up oxygen, or form large coastal dead zones devoid of marine life.

Policy drivers

A number of regions, nations, and states have passed regulations or offered subsidies designed to limit the pollution from livestock manure, but so far, most of the major initiatives have focused on water contamination rather than greenhouse-gas emissions.

The European Union, for instance, restricts the amount of manure that farmers can apply to fields and requires member nations to monitor nitrate levels in ground and surface water. The US’s Clean Water Act requires large livestock operations to obtain permits and develop manure management plans that limit pollution. 

But California has arguably done the most to use government policy specifically to drive down the methane emissions from livestock. The dairy industry accounts for about 45% of the state’s pollution from the potent greenhouse gas, and more than half of that comes from manure, according to the government’s estimates. 

In 2016, the state enacted a law that requires dairies, landfills, and other businesses to cut methane emissions 40% below 2013 levels by 2030, as part of a broader effort to reduce pollution from powerful but short-lived greenhouse gases. The measure directed the California Air Resources Board, the state’s main climate regulatory agency, to set up various incentive programs to encourage these industries to shift to cleaner practices. 

“In terms of bang for your buck, short-term benefits, methane can go a long way toward reaching climate goals,” says Tawny Mata, director of California’s Office of Agricultural Resilience and Sustainability. 

Between these various programs—and falling livestock numbers in the state—the dairy sector is on track to reduce annual methane emissions by the equivalent of 5 million metric tons of carbon dioxide by 2030, the state estimates. That would still fall about 4 million tons short of the target under the 2016 law.

The downsides of dairy digesters

Excluding the decline in herd populations—which has been driven by growing international competition and rising costs—the vast majority of California’s estimated methane reductions come from the use of what are known as anaerobic digesters. This technology entails covering the slurry lagoons to prevent methane from leaking into the air and then piping the biogas into separate vessels, where it’s cleaned and converted into natural gas. 

Under California’s Low Carbon Fuel Standard program, dairies that use digesters to produce gas delivered into pipelines can earn credits and sell them to petroleum refineries and other major polluters, as a means of helping those companies meet their own emissions reduction requirements. 

The gas can then fuel power plants, produce hydrogen, or power natural-gas vehicles. These uses still release carbon dioxide, but the state considers it a climate win because it avoids the release of methane, which traps even more heat. 

The rich revenue stream from California’s program has spurred hundreds of US farms to install anaerobic digesters over the last decade. Since 2020, it has produced more than $1 billion for farms, Cal Poly researchers noted in a paper last year.

But there are a variety of concerns about this approach.

The first is that it’s viable only for farms with about 2,000 cattle or more, because the equipment is very expensive to install, says Frank Mitloehner, a professor and chair of the Department of Animal Science at the University of California, Davis.

“For the lion’s share of dairies, digesters will not be a solution,” he says. 

Since the manure is often still spread across fields, digesters also do little to address the water pollution problems—and can even exacerbate them because of some of the chemistry that occurs during that process. 

Yet the huge subsidies flowing to digesters have steered money, energy, and attention away from other solutions that may offer better overall environmental outcomes, says Danny Cullenward, a senior fellow with the Kleinman Center for Energy Policy at the University of Pennsylvania, who has closely studied the California program.

“That is really not a solution at scale, and it’s diverting a huge fraction of precious resources to what I think is mostly not the right answer,” he says. 

Alternatives

The high up-front costs and limitations of digesters have spawned growing interest in alternative solutions—many of which work by reducing the formation of methane in the first place instead of turning that methane into a sellable fuel.

One of the cheapest, easiest, and most popular approaches, known as solid separation, uses simple machinery like a screw press to squeeze much of the water out of the manure slurry. The remaining solids are dry and exposed to open air, shifting away from the oxygen-free conditions in which methane is readily produced.

Other methods include increasing acidity in lagoons, bubbling air through them, or adding methane-eating microbes to the slurry, all of which alter the chemistry in ways that promise to reduce the amount of methane released. One company, Sedron Technologies of Sedro-Woolley, Washington, has also developed a sort of high-tech solid separation approach that extracts several marketable products from the animal waste, including a liquid organic fertilizer.  

The state of California set up a pair of additional programs to help smaller farmers adopt some of these other approaches, dubbed the Alternative Manure Management Program and the Dairy Plus Program.

The bulk of the funds have gone to solid separation systems. But the state has provided more than $18 million to support 15 vermifiltration projects. The Alberto Dairy has received nearly $2 million between the two programs.

Oreo cows

As I drove down a dusty road bordering the dairy, black-and-white bovines, affectionately known as Oreo cows, stretched their heads through the rails of an open barn, nibbling on golden silage scattered along the structure. Agueda’s grandfather Antonio Alberto founded the dairy 45 years ago in nearby Atwater, California, but eventually settled in Hickman, population 604, in 1989. 

It was mid-March but already above 80 °F in the Central Valley, which is walled off from the cool Pacific air by the coastal mountain range. Knee-high oat stalks swayed in fields that stretched to a line of almond trees in the distance.

Agueda, who graduated from Fresno State last year and now helps lead the operations on the farm, met me and UC Davis’s Mitloehner, who has studied the effects of vermifiltration, along the side of the barn. (UC Davis has no affiliation with the farm, but the university helped facilitate the meeting.)

He led us along dirt lanes as he explained the workings of the vermifiltration system, which they began using in October 2024.  

As before, a flush system washes manure from the floors of the barns into a large collection pit. But now a set of pumps funnels it through a series of large V-shaped metal contraptions standing on a nearby concrete pad, where mechanical screens separate most of the solids from the water.

A conveyor belt takes away the solids, which the farm composts for cow bedding or fertilizer. The remaining liquid moves through a system of pipes, first to settling ponds and then on to an irrigation system suspended above the vermifiltration beds. The long, tubular structure runs over the mounds on wheels set in gravel tracks, wetting the wood chips as it goes. The worms and various microbes residing in the biofilter then set to work consuming much of the remaining solid material, according to BioFiltro.

“Once the water is sprinkled on top, it takes about four hours from beginning to end for it to percolate through and drain to the end,” Agueda says.

He then defers to Mitloehner to explain the science of what happens as it does, adding, “I’m just the dairyman.”

The science

Mitloehner says he was skeptical of BioFiltro’s claims when he first heard them, particularly the assertion that the system could nearly eliminate nitrogen and, with it, the various forms of pollution it can produce, including ammonia and nitrates. 

So he decided to study a similar setup at the Fanelli Dairy, an operation in Hilmar, California, about 20 miles to the south. He and colleagues monitored the emissions from wastewater samples that were taken from the system before and after the liquid moved through the filter. In a paper published in 2018, the researchers concluded that vermifiltration reduced ammonia emissions from the resulting water by about 90%.

BioFiltro, whose tagline is “worm-powered solutions,” states that its technology “catalyzes the digestive power of worms and microbes to remove up to 99% of wastewater contaminants.”

But Mitloehner questions how big a role the invertebrates play in the process, calling it “kind of a catchy narrative.”

His take is simpler: The rocks and wood chips form a porous filter that replaces the anaerobic environment of a manure lagoon with an aerobic one. And in that oxygen-rich environment, different types of microbes thrive. 

His study suggests that these microbes are highly effective at converting nitrogen compounds in manure into nitrogen gas—a benign gas that makes up 78% of Earth’s atmosphere—instead of ammonia. That’s notable because while ammonia in manure acts as a fertilizer when it’s applied to fields, it also converts into the nitrates that can leach into groundwater.

Several more recent studies, which were partially or fully funded by BioFiltro and one of its regional distribution partners, Organix, produced similarly promising results. For instance, a 2022 study in Bioresource Technology Reports, also conducted at the Fanelli Dairy, concluded that the filter removed nearly 85% of the nitrogen in the operation’s wastewater. 

But a befuddling wrinkle is that when it came to methane, those studies and Mitloehner’s independent one found nearly opposite results.

While both the company- and partner-supported studies concluded that the filter eliminated the vast majority of methane pollution, Mitloehner’s study found that methane emissions were nearly 85% higher than those from the lagoon. 

In a follow-up email exchange, Mitloehner stressed that it’s not appropriate to compare his results with those that emerged from the other study at the same dairy, because the teams used very different methods, instruments, and measurement periods. Moreover, the focus of his research was the effect on nitrogen.

He said it’s “entirely reasonable” and “biologically plausible” that vermifiltration could substantially reduce methane emissions, simply by creating that aerobic environment.

“That said, I would be cautious about calling the magnitude of the reduction a fully settled issue,” he added. “While the available studies, including those you mentioned, point in the same general direction, the number of independent studies remains relatively limited, and results can vary.”

Patrick Beckett, BioFiltro’s vice president of quality and R&D, also stressed that there were crucial differences in the methodology of Mitloehner’s study that could have affected his methane findings.

In addition, he said the Organix funding came by way of a Washington state grant and described that study and the one BioFiltro supported as “high quality, peer reviewed” research that “has been submitted to other technical third parties for review and acceptance.”

Beckett says he agrees that additional independent reviews of BioFiltro’s systems is “fair and necessary” and notes that other studies have occurred or are underway.  

“That said,” Beckett wrote in an emailed response to questions from MIT Technology Review, “it seems unreasonable that BioFiltro would be held to a standard of not being allowed to invest in technical research by qualified third parties to learn more about the capabilities of our technology, and use the results of that research to enter new markets and to understand the value we can bring to projects or entire industries beyond water treatment.”

Milk money

BioFiltro is already building a business model around the available findings.

The company, founded in 2009, has been selling its vermifiltration systems or services to other industries around the world for years. It says there are around 225 operating in nine countries, at sites including municipal wastewater facilities, wineries, fruit processors, and other industrial operations.

But BioFiltro, whose US headquarters are in Davis, California, is seeing increasing demand among dairies as the industry faces growing pressure to address manure pollution. Late last year, it raised $35 million that the business says it will use, in large part, to accelerate its growth across the sector.

In an interview, Sarah Ploss, the company’s senior vice president of agriculture, explains the basic financial template for how it works with dairies: BioFiltro pays for, owns, installs, and operates the system. The farm, in turn, covers a share of the additional electricity, operations, and maintenance costs. 

Ploss says the dairy gets back clean water and the ability to focus on what it does best: producing milk. For its part, BioFiltro can generate carbon credits from the reduction in greenhouse gases, which it can then sell to makers of consumer packaged goods that are looking for ways to address the emissions throughout their supply chains, she says.

BioFiltro says that Verra, which sets standards for and assesses greenhouse-gas crediting projects, has registered two of its projects: the Royal Dairy and Moxee Dairy, both in Washington.

The Swiss confectionary giant Nestlé has bought more than 150,000 credits generated by the Royal Dairy’s vermifiltration system, according to an offsets database managed by CarbonPlan, which assesses the scientific integrity of climate action programs. Ploss said that BioFiltro has sold more than 200,000 credits from the project so far, and adds that it secured a different buyer for a project in California, which she said she couldn’t name. 

Three additional projects involving BioFiltro systems took the initial steps to become registered through Verra but didn’t move forward and weren’t built, Ploss said in an email. The request for registration for the Alberto Dairy estimates that the system there will reduce emissions by the equivalent of more than 30,000 metric tons of carbon dioxide per year. 

BioFiltro could take advantage of another revenue source as well: selling what it calls vermicompost, a rich soil additive composed of the leftover materials in the biofilter, including worm castings—a combination of cocoons, excrement, and remains. At retail, worm castings can run more than $500 per ton.

Beckett says the company is still developing that market but notes that it could help the industry offset rising fertilizer costs. 

“I think we’re going to enable a larger-scale use and adoption of it that could be meaningful to agriculture,” he says, adding: “These will become basically soil production facilities.” 

Concerns

Determining how well vermifiltration and other manure management approaches work will require more time and more research, experts say. 

Katharine Dickson, an agricultural emissions scientist who recently finished a postdoctoral program at UC Davis, says there should be in-the-field accounting to ensure that any of these methods are working as well as hoped—or to the degree government policy programs assume. All of which is tricky to achieve given the dynamic biological processes playing out in live animals and microbial communities on open farms, she adds.

“Vermifiltration, for example, depends on a live earthworm population whose performance is sensitive to temperature, moisture, and toxicity, and can shift with seasonal conditions or changes in herd size and manure characteristics on a given farm,” Dickson said in an email. 

The use of carbon credits to earn money from vermifiltration projects raises a different set of potential concerns. Most notably, if the methane decreases aren’t as significant as assumed, the projects could receive more credits than they deserve. 

There are more complicated issues as well. For the carbon credit system to make any real difference in the net amount of greenhouse gas in the atmosphere, it must produce emissions reductions that wouldn’t have occurred without that financial incentive. If it was going to happen anyway—as a result, say, of rich grants, legal pressures, or looming policies—the buyer of the credits can’t legitimately claim to have made any progress on its own climate emissions, says Grayson Badgley, a research scientist at CarbonPlan.

On that point, if California agriculture doesn’t meet its looming methane reduction targets, the carrots the state offers could be replaced by sticks: The California Air Resources Board recently began discussing rules that would force, rather than nudge, the sector to meet the 40% reduction required under the 2016 law.

“If lots of dairies are cleaning up their act ahead of pending regulation, it really does seem like the regulation, not offsets, is driving that action,” Badgley wrote in an email. “Trying to collect as many offsets prior to that deadline might adhere to the rules of the market, while still raising questions about whether those rules have enabled real climate action.”

Investing in sustainability 

Beckett disagreed that the possibility of forthcoming regulations undermines the case for generating carbon credits from current projects. 

“It’s true the state has net reduction targets that it hopes to meet, but it’s clear the state of California has favored market-based solutions and tried to provide some support via grant programs,” he wrote. “I’m on the science side of our business, not the business development side, but still think I can tell you with complete transparency that we would not have systems installed on [California] dairies without the sale of voluntary carbon credits.”

Ploss also stressed that the company goes through a careful “validation and verification process” on the farms to understand how much vermifiltration reduces greenhouse gases.

“We’ve got sensors and cameras and all sorts of stuff so that we can look into any of our systems, 24-7,” Ploss says. “We know through sampling. We know through what’s going through the system, what came out of the system. We know by all the measurements on any given month: What did that system do in terms of generating carbon credits?”

Agueda also disputes the critique. 

“The installation of the vermifiltration system would not have occurred without the ability to generate carbon credits,” he said in an email. “The project required a substantial capital investment, and the anticipated carbon credit revenue was a key factor in making the investment financially feasible.”

California decided to incentivize vermifiltration, along with other approaches, because it can offer multiple benefits, including cleaner water, less nitrogen, and lower greenhouse-gas emissions, while also creating economic value from manure, wrote Roberta Franco, a senior environmental scientist at the California Department of Food and Agriculture, in an emailed response to questions from MIT Technology Review.

She added that the decision was based on a number of studies as well as the 2022 recommendations from a task force composed of scientists, technical experts, and others. 

Even if California has made missteps, most notably in funneling too much money to anaerobic digesters at the expense of other methods, it’s created a test lab that’s achieved real progress and provided lessons that other regions can learn from.

One way or another, more parts of the world will need to set up similar programs, offering greater support or creating stricter rules, if we hope to really drive down the emissions from manure, says Maria Bowman, who leads the Agricultural Nitrogen Transformation Program at Spark Climate, a San Francisco nonprofit.

For his part, Agueda says that the vermifiltration system has offered a number of benefits to his family’s farm, at little additional cost to them. By cleaning up the water that cycles back through their flush and irrigation systems, the biofilter has reduced clogging, decreased odors, and improved the health of the herd.  

He says that each generation modernizes dairy farming in its own way. His father and uncle, for instance, incorporated computers and data management systems into the daily operations of the Alberto Dairy. He believes it’s the responsibility of his generation to make a similar effort to reduce the pollution that’s long plagued the sector.

“We knew that in the next generation we have to invest in environmental sustainability,” he says. “We didn’t know if it was gonna work or not, but we’re very happy with how it’s turned out.”

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here.

OpenAI CEO Sam Altman’s oft-discussed promise that Americans will share in the wealth AI creates was in the news again last week. On Thursday, the Financial Times reported that Altman is in talks with President Trump about giving the US government a 5% stake in OpenAI.

In some ways, Altman’s plan is old news. He wrote about a more radical version of this back in 2021, proposing that all companies above a certain valuation (not just AI companies) pay 2.5% of their market value each year into a fund that sends Americans annual disbursements. In April this year, OpenAI described a narrower proposal that closely resembles what Altman is reportedly discussing with Trump now. And the notion has broad political appeal: Senator Bernie Sanders has proposed giving Americans a 50% stake in top AI companies.

What’s the logic here? For would-be recipients, it’s twofold. First, AI learns directly from human-generated work—books, movies, art—but AI companies generally never pay the authors of that work. A free equity stake could serve as a form of belated compensation. Second, the payout could mitigate the widespread anxiety that AI will cause a collapse of the labor market (even if economists disagree) by providing a safety net. 

How large a safety net is up for debate. Details of OpenAI’s latest proposal are sparse, but let’s say the government were to distribute this equity stake directly to Americans. After its funding round in March the company was valued at $852 billion, making a 5% stake in OpenAI worth about $42.6 billion today (the company is reportedly delaying its IPO until it can reach a $1 trillion evaluation, a tall order given that it’s spending heavily on data centers and still has not turned a profit).

Distributing that $42.6 billion equally among the roughly 133 million American households would give each about $320 in equity. But if it were to operate like other wealth funds, the government would not give equity directly to Americans but rather let the fund grow and then share a portion of the returns with everyone, perhaps delivering a bigger payout, if and when AI companies can ever start sustainably turning a profit.

If this dividend does materialize, what’s in it for tech companies? Altman might hope the promise of payouts could help swing public opinion a bit more back toward AI companies. (A majority of Americans don’t trust companies to use AI responsibly and oppose construction of data centers in their area, and half are more concerned than excited about the increased creep of AI into their daily lives.)

But the bigger prize for OpenAI might be that the Trump administration loves making tech deals—like its equity stake in Intel and its share of Nvidia’s sales to China, among others.  Staying on the administration’s good side is pretty essential for AI companies right now (just ask Anthropic). It could mean not having your models deemed a supply chain risk, or getting more help from the White House in stopping your rivals from China. 

My main takeaway is that these plans currently function more as a story than a policy. Altman has been talking about some version of this idea for five years and reportedly pitched it to President Trump soon after he took office, yet there is still little indication that a concrete plan is taking shape. The more ambitious proposal from Sanders is even less likely to gain traction.

But what these plans do reveal is just how up for debate the future of AI still is. Altman drew inspiration for his plan from the Alaska Permanent Fund, which was set up in the 1970s to give Alaskans a share in oil profits. The idea was based on two premises: that oil is a shared resource, and that eventually it will run out. Altman seems happy to concede the first claim about AI. But he’d balk at the second, having promised that AI will generate extraordinary wealth for decades to come. Whether Americans ever receive a check is beside the point; the proposal’s real purpose may be to convince them that the AI boom will be large enough to share.

Baek, a 35-year-old manager at the South Korean semiconductor titan SK Hynix, was enrolled in Sunoo, a matchmaking company based in Seoul, a year ago. In a move typical of anxious South Korean parents, his mother signed him up, hoping to find a good wife for her son.

Lately, says Baek (who asked to be referred to by his last name to protect his privacy), he and his coworkers are having better luck finding dates than they used to, perhaps because of the dazzling bonuses they just got. Flush with eye-popping profits from the AI chip boom, SK Hynix struck a landmark deal last year with its labor union to pay out 10% of operating profits to employees, which translates to an extra $476,000 per employee this year. A similar agreement and sizable lump sum followed for Samsung workers this May.

With their newfound wealth, chip workers like Baek have become the most sought-after bachelors and bachelorettes in South Korea. “I have a coworker who’s perpetually going on blind dates, and he’s been getting so many recently,” says Baek. “For the past few months, I’ve been getting many blind dates too, perhaps because of the bonuses I got.”

Young South Koreans joke online that the best outfit to wear on a blind date is an SK Hynix uniform. 

The AI chip boom is changing the social fabric of South Korea by minting a new elite of “silicon-collar” workers earning about 20 times as much as the average South Korean. Although it’s helping some chip workers to find relationships, it’s also fueling fears of a deepening wealth disparity—and a loud public debate about inequality.

Love in the time of chips

South Korea is the epicenter of the chip boom fueling the AI race. Samsung and SK Hynix supply the vast majority of the world’s high-bandwidth memory (HBM) chips, which power Nvidia’s AI accelerators—the GPUs used to train AI models. As AI companies spend hundreds of billions of dollars on building data centers around the world, demand for HBMs is rising beyond what suppliers can keep up with, driving their prices to unprecedented levels. Samsung and SK Hynix are raking in record profits as a result. 

South Korea’s economy now orbits the two chip giants. In May, both companies topped $1 trillion in market value. And chip exports helped fuel a 1.7% surge in South Korea’s gross domestic product in the first quarter of 2026. South Korea’s main equity index, Kospi, has nearly tripled over the past year, becoming the best-performing market in the world.

Swimming in cash, chip workers are going on shopping sprees in department stores near the “semicon belt” fabs—splurging on everything from lavish furniture and electronic appliances to jewelry and watches. They’re also snapping up homes near the commuter-shuttle routes that ferry workers to campus. And they’re shelling out for matchmakers.

“Quite a lot of people ask me if I can introduce them to chip workers,” says Lee Sung-mi, a matchmaker at Sunoo, who has been playing Cupid for chip workers for years. “In fact, people who once rejected them are asking to be matched with them again, now that their salaries and bonuses have shot so far above what everyone else earns.”

One woman who lives in Gangnam, a ritzy district in Seoul lined with luxury high-rises and designer boutiques, previously turned down a chip worker at SK Hynix because his fab was too far out in Icheon, a rural city about 50 miles southeast of Seoul that’s dotted with rice farms and manufacturing plants. But in May, she asked her matchmaker to set them up again. They’ve now been dating for a month.                                                                                                                                                                                                                                                                                                                                                                                                                             

In South Korea, matchmaking companies evaluate their clients on a long list of criteria such as education, job, income, looks, and family background, including whether their aging parents have saved enough for retirement. In an economy where housing prices and child care costs are soaring, competition for jobs is fierce, and the social safety net is thin, a good job is the ultimate dating credential—all the more coveted at a time when many young South Koreans are forgoing marriage and children altogether, seeing family life as an unaffordable dream.

Every client at Sunoo gets a spouse rating, determined by an algorithm that assigns scores for each criterion. Since their hefty bonuses were announced, the job ratings of Samsung employees have risen from 80 to 84, while those of SK Hynix employees climbed from 78 to 82. Scores above 90 are reserved for doctors and lawyers. Long prized as paragons of prestige and wealth, they’re now close to being overtaken by chip workers. A score of 99, the highest possible rating, is earmarked for heads of state.  

Their new status is reshaping how chip workers themselves approach dating. “Chip workers from Samsung and SK Hynix are enrolling in our services because they feel more financially ready,” says Lee. “They’re also becoming pickier, as they feel like they’re now in a good position. The women want to meet men with higher incomes and better jobs, and the men want to meet younger and better-looking women with better jobs.” 

An SK Hynix engineer in her 40s, who was once desperate to get married as soon as possible, started turning down men she would’ve dated before the chip boom. Lately, showered with more matches, she’s been sifting through her suitors more carefully. “She now has peace of mind and wants to take her time to meet someone better,” says Lee.

A mixed blessing

While chip workers enjoy the fruits of their labor, the bonus bonanza is stoking anxieties among other South Koreans. “When wealth disparity is no longer a mere difference of income but, rather, a difference in identity … it can fuel social conflict,” says Se-eun Jung, an economist at Inha University. 

Earlier this month, the Bank of Korea warned that the chip boom will create a “K-shaped” economy, where a handful of workers race ahead while everyone else falls behind. The windfall, the bank said, is flowing to high income earners and then barely trickling out to the broader economy. Such polarization could erode people’s motivation to work by narrowing the path to upward mobility, it cautioned. 

Workers in other industries are venting online about feeling demoralized by the ballooning wealth gap. “The one-billion-won ($650,000) bonuses have crushed my motivation to work. I have no energy when I teach,” an employee of the Seoul Metropolitan Office of Education wrote on Blind, an app where employees can discuss their workplaces anonymously. Others are giving up the job hunt, lamenting that years of working at a small company could never match a year’s bonus at Samsung. 

In a Facebook post in May, presidential policy chief Kim Yong-beom proposed paying an “AI dividend” to citizens by taxing AI profits. The idea sparked a heated public debate over whether the government should redistribute gains from the chip boom. Some argue that the industry is indebted to the society that has educated its engineers, subsidized its infrastructure, and provided tax credits. Others counter that the profits are already being shared with the public as stocks.

Then there’s the question of how long this new social class will last. The semiconductor industry is notoriously cyclical; AI spending may cool, or rival chipmakers could catch up. There’s also the risk that chip workers will be replaced by automation. Samsung announced in March that it plans to fully automate its fabs by 2030, drawing backlash from chip workers. 

Although they’re unsure how long the boom will last, chip workers like Baek are riding high, for now. “These days, we say we want to work hard and bury our bones here at SK Hynix,” he says. “And I hope I can find [a wife] similar to me.”

It’s not easy to transplant a whole human eye. The surgery is difficult. And the eyes themselves start to degenerate as soon as they’ve left the body. When surgeons attempted it a few years ago, the newly transplanted eye wasn’t able to see.

But researchers believe they might have a solution: a device that maintains and revives freshly removed eyeballs using a technique called perfusion. Perfusion works by providing surgically removed organs with some of the oxygen and nutrients they typically get when they’re inside a body. Treated eyes don’t degrade as quickly; they also appear to retain the ability to transmit electrical signals and potentially see. The device could one day make eye transplants a viable possibility.

“It’s really cool,” says Shannon Tessier at Massachusetts General Hospital, who was not involved in the research but studies perfusion of other organs. “It could be a new frontier for retina preservation.”

Pia Cosma at the Centre for Genomic Regulation at the Barcelona Institute of Science and Technology in Spain and her colleagues have spent years developing their device. The Eye-in-a-Care-Box (ECaBox), as they call it, delivers an oxygen-rich supply of fluid through the artery that normally supplies the eye with blood.

The eye itself sits on a “bed,” and excess fluids are drained away. And while the device is sealed to maintain a specific temperature and pressure, a clear window on its side allows researchers to study and image the eye while it’s inside.

Cosma and her colleagues started experimenting with pig eyes, which are anatomically similar to human eyes but easier to get hold of (the team got theirs from a local slaughterhouse).

Pig eyes that are kept at room temperature outside the device start to degenerate pretty quickly. The team found that cells in the eye shrank, and the eyes started to lose their structure. Cooling the organs didn’t help preserve them, either—the eyes degenerated within 24 hours even when they were kept at 4 °C (39 °F).

But eyes kept in the EcABox fared much better; 24 hours later, tests suggested the prefused eyes were “significantly more viable” than eyes that hadn’t been maintained in the device.

The perfused eyes also seemed to be able to respond to light, suggesting they might technically be able to see if they were transplanted. Untreated pig eyes lost this ability as soon as they were removed from the animal. But it came back after about 15 minutes of perfusion, according to the scientists behind the work. A few of the treated eyes kept going for 10 hours or more.

Cosma and her colleagues described the work in a preprint article that has not yet been peer-reviewed and did not want to comment on the work.

After success with the pig eyes, the team members then tested their device on human eyes. They first collected 12 eyes from six people who had died. In each case, one of each pair of eyes was put in the device, while the other was not. Again, the perfused eyes did better—and their retinas were preserved.

Cosma and her colleagues hope that their device could offer scientists a new way to study eye treatments—one that doesn’t involve experimenting on living animals. They also hope that with some improvements, the ECaBox might provide a way to maintain and revive donated human eyes for whole-eye transplantation.

Whole-eye transplants have been attempted in the past, mostly in research animals, with limited success. In May 2023, a team at NYU Langone transplanted an eye along with part of a face to a man who two years earlier had survived a high-voltage electrical accident that resulted in the loss of much of the left side of his face, including his left eye Although the man recovered well, he wasn’t able to see out of the transplanted eye.

We won’t know whether eyes treated in the ECaBox could do any better until they have been transplanted, says Tessier. 

In the meantime, Cosma and her colleagues plan to use a newer version of their device to collect more human eyes for research. “We are planning to develop a portable, surgery-room ECaBox to minimize [degradation] in heart-beating donor eyes, when they become available,” they write.

As the parent of two little girls, I often think about how their childhood is different from mine. The seven-year-old is learning about AI at school. The five-year-old is given internet-based homework every week. And they are both absolutely repulsed by the idea of smoking.

That was not the prevailing sentiment when I was young. My parents smoked. The customers at our family’s restaurant smoked. Cartoon characters smoked. My friends and I would buy little cigarette-box-shaped packets of sugary white sticks and pretend to smoke in the playground. Smoking was a central part of our culture.

Which is why the UK’s recent passing of a generational sales ban on tobacco products feels like such a big deal. As part of the Tobacco and Vapes Act 2026, retailers are prohibited from selling tobacco products to anyone born after January 1, 2009, in perpetuity. It doesn’t matter when those people turn 18—or 38 or 68, for that matter. It will always be illegal to sell to anyone born after that date.

This is what’s described as an “endgame” approach. While many tobacco control strategies—such as taxation or gory imagery—aim to reduce consumption, policies like the UK’s are designed to eliminate it entirely. It’s a new approach, and no one knows whether it will work.

The Maldives was the first country to implement a generational smoking ban, in November last year. It’s too soon to say how that has panned out.

Nor do we know if these laws will even last. In 2022, New Zealand passed a similar generational sales ban as part of a broader anti-smoking law. But it was never enacted—the law was repealed by a new government in February 2024.

In the UK, both major parties support the ban. But Nigel Farage, whose right-wing party has seen a recent surge in support, has promised that “the generational smoking ban will not last long if Reform gets the chance to start rebuilding our mismanaged country.”

Chris Bostic, an attorney and former policy director for the advocacy group Action on Smoking and Health, says he and his colleagues began promoting the idea of a generational ban in the United States 11 years ago. Back then, they struggled to win support, even from major health charities. “People said we were crazy … [and] that this was impossible,” he says. Opponents argued that bans would infringe on personal freedoms.

“The public health argument is: Well, what about freedom from addiction?” says Britta Matthes, a tobacco control researcher at the University of Bath in the UK. Most people who smoke began when they were teenagers, want to quit, and wish they’d never started. Tobacco is arguably the most harmful consumer product of all time. It will kill half its users who don’t quit, according to the World Health Organization.

It also kills people who don’t smoke. Of the 7 million who die from tobacco every year, 1.6 million are nonsmokers who were exposed to secondhand smoke, according to the WHO.

Generational sales bans are a long-term strategy that will only protect future smokers. Most experts agree that people who already smoke should be a main consideration for any policy, and that a multipronged approach is probably the best way to go. Janet Hoek at the University of Otago, who has explored tobacco control policies in New Zealand, believes that enforcing very low limits on nicotine levels and banning filters—an environmental scourge that does not make smoking safer, as many people believe—might be a “powerful combination,” for example.

But preventing teenagers from starting to smoke in the first place is an enticing prospect, even among the majority of people who smoke. And it’s starting to look a lot less radical.

The US has quietly been making progress on a smaller scale. Since 2021, Brookline, a town in the Boston area, has banned the sale of tobacco products to anyone born after January 1, 2000. The idea has spread. Today there are 23 towns in Massachusetts with similar bans, says Bostic. Nine towns across Minnesota, New York, and California have implemented other endgame policies.

The UK law has normalized the idea more than ever, he adds. His colleagues are already fielding calls from health agencies around the world. “People [are] saying, Wow I can’t believe the UK just did this—can we do this here?” he says.

Norms change. Like many other millennials, I vividly remember my first night out after a ban on indoor smoking took effect. My clothes didn’t stink! My hair still felt clean! And my throat wasn’t scratchy the next morning! Now that’s just normal. I hope a tobacco-free world can be the new normal for my kids.

Something stinks in California’s climate policies.

Years ago, the state set up a system that pays cattle farmers across the country to turn the methane emitted from cattle manure into natural gas, encouraging the dairy sector to produce a gas we burn instead of one that just pollutes the air.

It’s become wildly popular because the subsidies are extremely lucrative. But a growing body of research suggests the program is a case study in the shortcomings of our preferred approaches to climate action. Instead of simply forcing industries to directly cut their pollution or pay for it as a cost of doing business, legislators have repeatedly opted to set up convoluted incentive systems that swap climate responsibilities between parties and regions. As studies have shown again and again, these carbon offsetting and trading schemes often dramatically overstate the emissions reductions actually achieved in the one place that matters: the atmosphere.

The dairy program illustrates a particular version of this problem, muddling the impacts of different types of greenhouse gases in a way that researchers argue will lock in more warming in the future.

Despite this and other concerns, California regulators decided in 2024 to extend parts of the program beyond 2050. And a recent proposal by the state’s air resources board could send millions of additional dollars to dairy farmers as part of a plan that would ease restrictions on major greenhouse-gas producers.

Here’s how the system works: The state’s climate regulations require the transportation fuels industry to lower the carbon dioxide levels in its products over time—or purchase credits from other parties that cut fuel emissions, including cattle farmers.

Dairies generally spray cattle manure into giant open lagoons, where microbes gobble up organic matter and produce methane as a by-product. But if farmers set up what are known as anaerobic digesters, the sludge is redirected into covered vessels that capture the biogas, which can be converted into natural gas and injected into a pipeline. It can then be used to fuel certain vehicles or generate electricity in a power plant. Either way, petroleum companies can pay those farmers for Low Carbon Fuel Standard (LCFS) credits, to meet regulatory requirements in lieu of reducing the emissions from their own fuels.

Burning biogas in a bus or turbine still releases carbon dioxide, but the idea is that this process reduces market demand to extract natural gas from the ground and avoids the release of methane, which is a far more powerful greenhouse gas (at least initially). In fact, methane is so much more powerful that under California’s program, “adding one average biogas-powered vehicle to the fleet would produce enough LCFS credits to cover the deficits incurred by 26 similar gasoline-powered vehicles,” according to Aaron Smith, a UC Berkeley economist.

But there’s a problem with this carbon math. California assumes that methane exerts about 25 times the warming effect of carbon dioxide over a 100-year period. That’s not how it really works in the atmosphere, though.

Methane is very powerful, but it also breaks down quickly, generally within a couple of decades. Meanwhile, carbon dioxide builds up cumulatively in the atmosphere—and much of whatever we emit will continue heating up the planet for hundreds to thousands of years.

So, in effect, the state has created a system that reduces short-term warming at the cost of increasing all-but-permanent warming. Any methane that digesters capture today would have caused extra-powerful warning if released, but by 2050 that effect would have mostly faded away. Meanwhile, that additional carbon dioxide we permitted in its place could continue warming the world for millennia.

It is a good idea to cut methane emissions, and dairy digesters achieve this (though not always as effectively as hoped). But we can’t swap a decrease in short-lived greenhouse gases for an increase in long-lived ones if we hope to keep global temperatures within relatively safe levels in the coming century, as researchers have long warned. We have to slash both.

The problem I keep returning to, after years of covering carbon markets and offsets, is this: We need to clean up every sector, completely, over the next few decades. It’s increasingly untenable for so many of our climate ambitions to turn on getting one industry to make progress on paper by paying another one to reduce emissions, at a point when every business in every industry needs to be racing toward net zero.

It’s time to move past the idea that we need to reward sectors for doing us the favor of not polluting the atmosphere, and simply require them to stop unloading the huge environmental burden of their business onto society.

This article is from The Spark, MIT Technology Review’s weekly climate newsletter. To receive it in your inbox every Wednesday, sign up here.