nuclear forces. The fourth force, gravity, requires a separate description known as general relativity. Physicists have good reason to trust Yang-Mills theory. They use it to make predictions about particle behaviour which are borne out to a staggering degree of accuracy. Mathematicians, however, being of a more rigorous bent, would like to confirm it stands on firm mathematical footing. The first part of the Yang-Mills problem aims to do just that. That involves proving the theory breaks no rules and has no exceptions, all without making any simplifying approximations. The committee which OpenAI convened has confirmed that, in their judgment, the model has successfully cleared that hurdle. To do so, the physicists checked that the AI’s formalisation of its solution in Lean, computer code confirming the proof’s internal logic, corresponded to the problem at hand. Although the committee has published the formalisation, which runs to around 1m lines of computer code, the AI’s 1,147-page “natural language” proof—which, the committee members said, “we do not claim to fully understand”—has not yet been shared. No reason for this secrecy has yet been given. What is known for now is that the proof builds on techniques developed in the 1980s by Tadeusz Balaban, a researcher at Rutgers University. “There were no obvious showstoppers in his programme, but [it] was very complicated,” says Edward Witten, a mathematician and physicist at the Institute for Advanced Study in New Jersey, who helped formulate the official Yang-Mills problem statement. “I actually believe if there had been ten Balabans, this would have been done by the 1990s.” Dr Witten calls OpenAI’s proof a “very important milestone”, but reckons it is unlikely to change how physicists work. A solution to the second part of the Yang-Mills problem, which involves showing that the theory predicts no massless particles (and on which OpenAI’s latest results shed no light), would be a different story, however. Such a proof may shed light on why subatomic particles called quarks bind together to form the protons and neutrons found in atomic nuclei. Though the AI had Dr Balaban’s work to draw upon when proving Yang-Mills’s solidity, Dr Witten believes no such handholds exist for the mass gap. “At the risk of possibly sounding foolish,” he says, “my guess is that it’s pretty far off.” ■
This article was downloaded by zlibrary from https://www.economist.com/science-and-technology/2026/11/11/ai-has-proved-that-a- theory-underpinning-fundamental-physics-is-trustworthy
Science & technology | Hide and seek Cameras in space could help spot invasive plants on Earth The trick lies in how the leaves of different species reflect light Nov 12th 2026 AS ITS NAME implies, the Earth Surface Mineral Dust Source Investigation (EMIT) exists to analyse terrestrial dust—and to do so from the International Space Station. It does this by recording the wavelengths of light that dust absorbs and reflects, measurements that are important for modelling how Earth’s temperature changes over time. Unexpectedly, though, as a paper in Biological Invasions by Susan Meerdink at the University of Iowa has now shown, EMIT may bring a second environmental benefit: an ability to track invasive plant species. Such tracking has real value. Invasive plants compete with native crops, infest livestock pastures, clog waterways and destroy native ecosystems. In
America alone their spread cost over $190bn between 1960 and 2020. But spotting them is not as easy as might be supposed. Until recently the best way to find invasive plants was to go into the field and look for them. That changed with the invention of hyperspectral imaging. This does not record colours in the way that ordinary cameras do. Instead, it records how surfaces both absorb and reflect different wavelengths of incoming light. The result is a visual fingerprint that permits surfaces that appear identical in colour to the eye—such as the leaves of different plant species—to be distinguished from afar. To test whether such fingerprints might be visible from space, Dr Meerdink and her colleagues decided to turn EMIT’s camera onto the Everglades, a network of swamplands in Florida that is rife with alligators and challenging to survey in person. They first took high-resolution photographs from a helicopter of 242 locations to see where Brazilian peppertrees—an invasive species that grows rapidly and can endure brackish conditions—were present. The team then ran a computer analysis of EMIT images of the same locations, looking for the peppertree’s hyperspectral fingerprint. EMIT, they found, supported the conclusions of their helicopter survey. Satellites less sensitive to the wavelength of incoming light, by contrast, confused native plants with the invasive target. Since EMIT has already snapped pictures of most of Earth’s land-surface at least once, and the catalogue is expanding (and is freely available from NASA), it should now be possible to use it to track invasives more widely, and take appropriate action. ■ This article was downloaded by zlibrary from https://www.economist.com/science-and-technology/2026/11/11/cameras-in-space-could- help-spot-invasive-plants-on-earth
Science & technology | Well Informed Is grip strength a useful indicator of health? It is an attention-grabbing claim Nov 12th 2026 “Get a grip” goes the idiom. Lately, the fashion for exercises including “dead hangs” and “farmer’s carries”, and for squeezing hand grippers, suggests that some are taking the advice literally. The likely trigger was a headline-grabbing study in the Lancet in 2015 that associated grip strength with longer life. Following 139,691 people in 17 countries for an average of four years, the researchers linked every 5kg decrease in grip strength with a 16% higher risk of death from any cause. For cardiovascular death, the figure was 17%. Grip strength, they found, was a better predictor of both than systolic blood pressure. Alas, regular grip exercises alone are unlikely to lengthen your life. They may, however, be a good way of monitoring your overall health. Muscles, notes Brian Clark, a physiologist at Ohio University, are “puppets” at the
end of a neurological chain of command. To generate force, electrical signals must be continuously generated by the brain and transmitted to the muscles by nerves. If any part falters, strength suffers. What makes a squeeze especially revealing is its complexity. To produce a firm grip, the brain must co-ordinate a system of 39 muscles controlling the hand and wrist. Such orchestration is neurologically taxing, which helps explain why fingers clenching together produce less force than the sum of what each can muster alone. Tellingly, in a small experiment with older people, published in 2024 in Ageing Clinical and Experimental Research, this shortfall in force was greater among participants with lower scores on a cognitive test. This raises another intriguing possibility. A weakening grip, says Ivan Bautmans, a researcher on ageing at Vrije University in Brussels, can also foreshadow cognitive decline. Dr Bautmans believes that easy and inexpensive strength tests could provide a valuable early warning. Consider a meta-analysis he co-wrote, published in June in Neuroscience & Biobehavioural Reviews, that examined 60 observational studies involving around 380,000 people. Across studies with follow-up periods ranging from one to 22 years, participants with stronger grips were 19% less likely to experience cognitive decline. Grip asymmetry may be another useful measurement. Having a dominant hand that is roughly 10% stronger than the weaker hand is common, but larger gaps can be telling. A meta-analysis published in April in Archives of Gerontology and Geriatrics pooled nine cohort studies involving around 900,000 people, some followed for many years. Participants whose hands differed markedly in strength had a 29% higher risk of death from cardiovascular disease, perhaps because problems with the nervous system can also impair control of the heart and blood vessels. Comparing both hands, the researchers argue, can reveal subtle neuromuscular problems that testing just one might miss. The brain’s division of labour may be responsible. Each hand is chiefly controlled by the brain’s opposite hemisphere, and neurological deterioration often progresses unevenly. This can produce a striking imbalance. When one
hand weakens to half the strength of the other, Dr Bautmans warns, that decline may be accelerating. A word of caution is in order: a weak or lopsided grip is no diagnosis. It can, though, suggest that further checks are warranted. As for enthusiasts for grip exercises, research has shown resistance training of muscles across the body can help keep the mind sharp. Hang in there. ■ After a free, evidence-based guide to health and wellness? Sign up to our weekly Well Informed newsletter. This article was downloaded by zlibrary from https://www.economist.com/science-and-technology/2026/11/06/is-grip-strength-a- useful-indicator-of-health
Plague, gruel and bad teeth: why do people idealise the olden days? The curious case of the surgeon who inspired Sherlock Holmes At 50, “Rocky” isn’t as good as you remember. It’s better Why small dogs are so easy to love—and hate How X-rated content shaped the internet “Ramayana”, India’s biggest film, turns a Hindu god into a star
Plague, gruel and bad teeth: why do people idealise the olden days? A rip-roaring rebuttal of the idea that progress is a myth Nov 12th 2026 Many people yearn for the good old days. In a global Ipsos poll, respondents given a choice of being born half a century ago or now picked the past by nearly two to one. A startling number think even the distant past was not so bad: in a YouGov poll, 40% of Britons were unsure that life was worse in the 1600s. Anti-capitalists on social media grouch that workers in Amazon warehouses are more oppressed than feudal farm labourers. “You have less free time than a medieval peasant,” and this is “frankly outrageous”, complained one on X. P.J. O’Rourke, an American satirist, once offered a one-word rebuttal of paeans to the pre-industrial past: “dentistry”. Chelsea Follett, a scholar at the Cato Institute, a free-market think-tank in Washington, expands on this