Different parts of Mars have different advantages and disadvantages when it comes to their available resources, just like Earth. The polar caps are likely the most valuable in terms of their water content, which will be critical to any early stage crewed mission to the Red Planet. But to really unlock the fully potential of Mars, geologists think we’ll need to look to the volcanoes, where there is likely to be easily accessible valuable materials like nickel, titanium, and chromium, that were placed there when the volcanoes were active. Reaching those deposits on the side of some of the largest mountains in the solar system safely is a challenge, and one that is tackled in a new paper by Divij Gupta and Arkajit Aich, where they look at the necessary requirements to set up an effective mining operation on the slopes of Olympus and Elysium Mons.

According to a report released today by CopernicusECMWF, in 2025, the Antarctic ozone hole formed earlier than usual and has already reached about 21 million km² in the first half of September.

If you are curious about the history of the ozone hole and how satellites are helping us track its recovery, here's an interesting video recap:
https://t.co/XxJbcEQs6k

📹CAMS

The new image of M87* just released today comes to life in this stunning animation, revealing its incredible evolution over the years!

Watch closely and see how this mysterious supermassive black hole changes over time.

Credit: EHT Collaboration | Animation: Saurabh (MPIfR)

OpenAI’s latest research paper diagnoses exactly why ChatGPT and other large language models can make things up – known in the world of artificial intelligence as “hallucination”. It also reveals why the problem may be unfixable, at least as far as consumers are concerned.

The paper provides the most rigorous mathematical explanation yet for why these models confidently state falsehoods. It demonstrates that these aren’t just an unfortunate side effect of the way that AIs are currently trained, but are mathematically inevitable.

The issue can partly be explained by mistakes in the underlying data used to train the AIs. But using mathematical analysis of how AI systems learn, the researchers prove that even with perfect training data, the problem still exists.

The way language models respond to queries – by predicting one word at a time in a sentence, based on probabilities – naturally produces errors. The researchers in fact show that the total error rate for generating sentences is at least twice as high as the error rate the same AI would have on a simple yes/no question, because mistakes can accumulate over multiple predictions.

In other words, hallucination rates are fundamentally bounded by how well AI systems can distinguish valid from invalid responses. Since this classification problem is inherently difficult for many areas of knowledge, hallucinations become unavoidable.

It also turns out that the less a model sees a fact during training, the more likely it is to hallucinate when asked about it. With birthdays of notable figures, for instance, it was found that if 20% of such people’s birthdays only appear once in training data, then base models should get at least 20% of birthday queries wrong.

Sure enough, when researchers asked state-of-the-art models for the birthday of Adam Kalai, one of the paper’s authors, DeepSeek-V3 confidently provided three different incorrect dates across separate attempts: “03-07”, “15-06”, and “01-01”. The correct date is in the autumn, so none of these were even close.

The evaluation trap
More troubling is the paper’s analysis of why hallucinations persist despite post-training efforts (such as providing extensive human feedback to an AI’s responses before it is released to the public). The authors examined ten major AI benchmarks, including those used by Google, OpenAI and also the top leaderboards that rank AI models. This revealed that nine benchmarks use binary grading systems that award zero points for AIs expressing uncertainty.

This creates what the authors term an “epidemic” of penalising honest responses. When an AI system says “I don’t know”, it receives the same score as giving completely wrong information. The optimal strategy under such evaluation becomes clear: always guess.
[...]
In short, the OpenAI paper inadvertently highlights an uncomfortable truth: the business incentives driving consumer AI development remain fundamentally misaligned with reducing hallucinations. Until these incentives change, hallucinations will persist.

Climate models are complex, just like the world they mirror. They simultaneously simulate the interacting, chaotic flow of Earth’s atmosphere and oceans, and they run on the world’s largest supercomputers.

Critiques of climate science, such as the report written for the Department of Energy by a panel in 2025, often point to this complexity to argue that these models are too uncertain to help us understand present-day warming or tell us anything useful about the future.

But the history of climate science tells a different story.

The earliest climate models made specific forecasts about global warming decades before those forecasts could be proved or disproved. And when the observations came in, the models were right. The forecasts weren’t just predictions of global average warming – they also predicted geographical patterns of warming that we see today.

These early predictions starting in the 1960s emanated largely out of a single, somewhat obscure government laboratory outside Princeton, New Jersey: the Geophysical Fluid Dynamics Laboratory. And many of the discoveries bear the fingerprints of one particularly prescient and persistent climate modeler, Syukuro Manabe, who was awarded the 2021 Nobel Prize in physics for his work.

Manabe’s models, based in the physics of the atmosphere and ocean, forecast the world we now see while also drawing a blueprint for today’s climate models and their ability to simulate our large-scale climate. While models have limitations, it is this track record of success that gives us confidence in interpreting the changes we’re seeing now, as well as predicting changes to come.

What are the stripes in the images of M87*?

They show the direction of polarized light. This animation shows how M87 would look over the years through a polarizer- like the filters in reflex cameras or 3D glasses.

Animation: I. Martí-Vidal (Univ. Valencia)

At our hearts, we are seekers, and today we celebrate a scientific milestone!

Nearly 30 years after the first discovery of a world orbiting a Sun-like star beyond our solar system, NASA’s Exoplanet Archive reached 6,000 confirmed exoplanets. 🎉

Read more: https://www.nasa.gov/universe/exoplanets/nasas-tally-of-planets-outside-our-solar-system-reaches-6000

The official number of exoplanets—planets outside our solar system—tracked by NASA has reached 6,000. Confirmed planets are added to the count on a rolling basis by scientists from around the world, so no single planet is considered the 6,000th entry. The number is monitored by NASA's Exoplanet Science Institute (NExScI), based at Caltech's IPAC in Pasadena, California. There are more than 8,000 additional candidate planets awaiting confirmation, with NASA leading the world in searching for life in the universe.

"This milestone represents decades of cosmic exploration driven by NASA space telescopes—exploration that has completely changed the way humanity views the night sky," said Shawn Domagal-Goldman, acting director, Astrophysics Division, NASA Headquarters in Washington.

"Step by step, from discovery to characterization, NASA missions have built the foundation to answering a fundamental question: Are we alone? Now, with our upcoming Nancy Grace Roman Space Telescope and Habitable Worlds Observatory, America will lead the next giant leap—studying worlds like our own around stars like our sun. This is American ingenuity, and a promise of discovery that unites us all."

The milestone comes 30 years after the first exoplanet was discovered around a star similar to our sun, in 1995. (Prior to that, a few planets had been identified around stars that had burned all their fuel and collapsed.)

Although researchers think there are billions of planets in the Milky Way galaxy, finding them remains a challenge. In addition to discovering many individual planets with fascinating characteristics as the total number of known exoplanets climbs, scientists are able to see how the general planet population compares to the planets of our own solar system.

This animation shows the locations of the exoplanets that had been confirmed as of 2022, in the order they were discovered. Sound on! 🔊
https://t.co/KtZEDIDNVB

#PPOD: Solar Cells ☀️

The National Science Foundation (NSF) Daniel K. Inouye Solar Telescope has captured the highest-resolution image of the Sun's surface ever recorded. The image, produced by the National Solar Observatory (NSO), reveals a turbulent, "boiling" pattern of gas covering our star. The "boiling" appearance consists of cell-like structures, each approximately the size of Texas, which are formed by convective currents that move heat from the Sun's interior to its surface. Hot plasma rises in the center of the cells and sinks in the dark lanes at the edges. Tiny, bright markers of magnetic fields are visible in the dark lanes. It is thought that these specks of light funnel energy into the corona, the Sun's outer atmosphere, which could explain its high temperature. This image was part of the Inouye Solar Telescope's first-light observations, and its high resolution allows for the viewing of features as small as 30 kilometers for the first time.

Credit: Inouye Solar Telescope, NatSolarObs AURADC NSF

The first-ever measurements of the ethanol content of fruits available to chimpanzees in their native African habitat show that the animals could easily consume the equivalent of more than two standard alcoholic drinks each day, according to researchers at the University of California, Berkeley.

It's not clear whether they actively seek out fruit with high ethanol levels, which are typically riper fruit with more sugars to ferment. But the availability of ethanol in many species of fruit that they normally eat suggests that alcohol is a regular part of their diet and likely was a part of the diets of our human ancestors.

"Across all sites, male and female chimpanzees are consuming about 14 grams of pure ethanol per day in their diet, which is the equivalent to one standard American drink," said UC Berkeley graduate student Aleksey Maro of the Department of Integrative Biology. "When you adjust for body mass, because chimps weigh about 40 kilos versus a typical human at 70 kilos, it goes up to nearly two drinks."

A "standard drink" in the U.S. contains 14 grams of ethanol, irrespective of the consumer's body size, although in much of Europe the standard is 10 grams.

The 21 species of fruit Maro sampled at two chimp study sites—Ngogo in Uganda and Taï in Ivory Coast—had an average alcohol content of 0.26% by weight. Primatologists who have studied chimps at these sites estimate that the animals consume about 10 pounds (4.5 kilograms) of fruit per day, on average, and that fruit makes up about three-quarters of their diet. The researchers have also recorded for each site the approximate proportion of each fruit species in the chimp diet. This information allowed the Berkeley biologists to calculate an average rate of dietary ethanol consumption.

"The chimps are eating 5 to 10% of their body weight a day in ripe fruit, so even low concentrations yield a high daily total—a substantial dosage of alcohol," said Robert Dudley, UC Berkeley professor of integrative biology. "If the chimps are randomly sampling ripe fruit as did Aleksey, then that's going to be their average consumption rate, independent of any preference for ethanol. But if they are preferring riper and/or more sugar-rich fruits, then this is a conservative lower limit for the likely rate of ethanol ingestion."

Fruit consumption occurs throughout the day and the chimps show no overt signs of intoxication, Maro said. In fact, to get a buzz on, a chimp would have to eat so much fruit its stomach would bloat. But chronic low-level exposure suggests that the common ancestors of humans and chimps—our closest living relative among the apes—were also exposed daily to alcohol from fermenting fruit, a nutrient that is missing from the diets of captive chimps and many humans today.

"Chimpanzees consume a similar amount of alcohol to what we might if we ate fermented food daily," Maro said. "Human attraction to alcohol probably arose from this dietary heritage of our common ancestor with chimpanzees."

Scientists at Hiroshima University have created a practical and highly sensitive method for detecting the Unruh effect, a long-anticipated phenomenon that lies at the intersection of relativity and quantum theory. This new strategy not only advances the study of fundamental physics but also opens the door to future technological applications.

The work was recently published in the journal Physical Review Letters.

The Fulling-Davies-Unruh effect, often referred to simply as the Unruh effect, is a profound theoretical concept linking Albert Einstein’s Theory of Relativity with Quantum Theory. “In quantum theory, even the vacuum seethes with tiny energy fluctuations, where particles and antiparticles briefly appear and vanish. Remarkably, the Unruh effect shows how these ‘vacuum ripples’ are perceived depends on the observer’s motion. A stationary observer sees nothing, but an observer undergoing acceleration perceives them as real particles with a thermal energy distribution—a ‘quantum warmth’,” explained Noriyuki Hatakenaka, professor emeritus at Hiroshima University.

This surprising prediction highlights the deep relationship between two cornerstones of modern physics. Experimentally verifying the Unruh effect would not only unite aspects of relativity and quantum mechanics but also offer valuable insights into the very structure of spacetime. However, achieving such verification has remained one of the most persistent and difficult challenges in physics.

A new trend is emerging in psychiatric hospitals. People in crisis are arriving with false, sometimes dangerous beliefs, grandiose delusions, and paranoid thoughts. A common thread connects them: marathon conversations with AI chatbots.

WIRED spoke with more than a dozen psychiatrists and researchers, who are increasingly concerned. In San Francisco, UCSF psychiatrist Keith Sakata says he has counted a dozen cases severe enough to warrant hospitalization this year, cases in which artificial intelligence “played a significant role in their psychotic episodes.” As this situation unfolds, a catchier definition has taken off in the headlines: “AI psychosis.”

Some patients insist the bots are sentient or spin new grand theories of physics. Other physicians tell of patients locked in days of back-and-forth with the tools, arriving at the hospital with thousands upon thousands of pages of trannoscripts detailing how the bots had supported or reinforced obviously problematic thoughts.
Reports like this are piling up, and the consequences are brutal. Distressed users and family and friends have described spirals that led to lost jobs, ruptured relationships, involuntary hospital admissions, jail time, and even death. Yet clinicians tell WIRED the medical community is split. Is this a distinct phenomenon that deserves its own label, or a familiar problem with a modern trigger?

AI psychosis is not a recognized clinical label. Still, the phrase has spread in news reports and on social media as a catchall denoscriptor for some kind of mental health crisis following prolonged chatbot conversations. Even industry leaders invoke it to discuss the many emerging mental health problems linked to AI. At Microsoft, Mustafa Suleyman, CEO of the tech giant’s AI division, warned in a blog post last month of the “psychosis risk.” Sakata says he is pragmatic and uses the phrase with people who already do. “It’s useful as shorthand for discussing a real phenomenon,” says the psychiatrist. However, he is quick to add that the term “can be misleading” and “risks oversimplifying complex psychiatric symptoms.”

It’s hard enough to know what Neanderthals looked like, let alone sounded like. However, there’s good reason to suspect our extinct hominin cousins were capable of complex language – and not just grunts and groans.

You may be one of the millions of people who have seen the viral clip from the BBC show Neanderthal: The Rebirth, which explores the so-called “high-pitched voice theory”.

To recreate the sound of Neanderthal voices, vocal coach Patsy Rodenburg worked with a 3D model of a Neanderthal vocal tract, alongside knowledge of their anatomy, such as a deep rib cage, heavy skull, and large nasal cavity. When she applied this understanding to a human singer, the result was an uncanny, high-pitched, nasal shriek. (Watch video here)

This is just one interpretation, but it’s a question that has been attacked from a more scientific angle.

The idea that Neanderthals are heavy-browed, brutish cavemen is an outdated one. An overwhelming amount of evidence now shows that, just like us, they were highly intelligent, culturally complex, and emotionally sensitive beings.

We know that Neanderthals had similar cognitive abilities to us, plus their brains were roughly the same size or bigger than ours, albeit with a different shape, so we can loosely assume they had the necessary neural hardware to compute language.

A look back at everyone on Earth #OTD in 1977.

NASA's Voyager 1 took the first image of the Earth and Moon in the same frame 48 years ago today. The spacecraft, which began its mission two weeks earlier, captured this scene from 7.25 million miles (11.7 million km) away.

From butterflies to grasshoppers, many delicate little things that run our world are in dire trouble. Not just in regions where human activity directly affects the landscape, but even in remote, human-free zones, a new study finds.

The loss of insects in key areas around the globe has been attributed in the past to the deliberate reduction of biodiverse habitats and changes in local climate. Now it's clear these forces stretch far beyond our immediate spheres of influence.

In areas relatively undisturbed by direct human activity, University of North Carolina biologist Keith Sockman recorded a dramatic drop of over 70 percent of flying insects in just 20 years.

NorthropGrumman's Cygnus XL cargo craft, packed with over 11,000 pounds of science and supplies, was captured by the Canadarm2 robotic arm operated by JonnyKimUSA as Zenanaut monitored at 7:24am ET today.

Neutrinos are one of the most enigmatic particles in the standard model. The main reason is that they're so hard to detect. Despite the fact that 400 trillion of them created in the Sun are passing through a person's body every second, they rarely interact with normal matter, making understanding anything about them difficult.

To help solve their mysteries, a new neutrino detector in China recently started collecting data, and hopes to provide insight on between 40 and 60 neutrinos a day for the next 10 years.

The detector, known as the Jiangmen Underground neutrino Observatory, or JUNO, is located in between two huge nuclear plants at Yangjian and Taishan. Both of those fission plants create their own artificial neutrinos in addition to the ones created by the Sun, meaning the general area should be awash with barely interacting particles.

That's despite the fact that, like most neutrino detectors, it's located underground. 700 meters underground in fact. The physical bulk of the Earth's crust is meant to block most other particles, like muons, from getting to it, and at other installations, like IceCube, it does a pretty good job.

In a major breakthrough for pregnancy research, scientists have 3D bioprinted “mini placentas”. The miniature organs, or “organoids”, closely resemble human placental tissue, providing an accurate model for studying the early placenta – something that has been sorely lacking until now.

The placenta plays a vital role in supporting fetal development, and its dysfunction is linked to numerous pregnancy complications, including miscarriage, preeclampsia, preterm birth, and stillbirth. These conditions can also have long-term health implications for both mother and baby, increasing the risk of future cardiovascular, endocrine, and neurological diseases. Yet, treatment options are limited, as is our understanding of placental biology.

Check out the view from up here! As the Curiosity rover continues to climb the flanks of Mt. Sharp, it has captured an extraordinarily clear panorama of Gale Crater's rim...and even the lands beyond. Details:
https://t.co/Zv9gQzybLO

The first living creatures intentionally sent into space were fruit flies! They travelled aboard a V2 rocket on 20 February 1947. NASA currently recognises the altitude of 100km as the point where space begins, making the fruit flies the first living creatures considered to have reached space.