Spotted: Falcon 9 [2nd stage] impact site

NASA’s Lunar Reconnaissance Orbiter (LRO) has captured images of a new crater on the Moon. The crater formed on Aug. 5, 2026, when a SpaceX Falcon 9 upper stage impacted the lunar surface. go.nasa.gov/4zw1I1g

The nature of supermassive black holes (SMBHs) is a hot topic in space science. It appears that all large galaxies like ours host an SMBH in their centers, but how these behemoths grew so massive is not clear. Neither are their origins, though most researchers agree that mergers certainly play a role.

A new discovery is bolstering the idea that mergers are behind the evolution and growth of SMBHs. Astronomers working with the JWST have found three actively accreting black holes in a single galaxy. No triple black hole galaxy has ever been seen before, though instances of three interacting galaxies with SMBHs have been observed.

The discovery is presented in new research in Astronomy and Astrophysics titled "BlackTHUNDER: Evidence of three massive black holes in a z ∼ 5 galaxy." The lead author is Hannah Übler from the Max Planck Institute for extraterrestrial Physics. The term BlackTHUNDER refers to a JWST observing program called "Black holes in the early Universe and their dense surroundings."

The galaxy in question is named J0148-4214 and it's more than 12.5 billion light years away. The JWST observed it as it appeared about 1.2 billion years after the Big Bang. BlackTHUNDER was aimed at observing black holes this old, and the JWST used its NIRSpec Integrated Field Unit to detect the three actively accreting black holes.

They have solar masses of 80 million, 0.6 million, and 2 million. The first two are closest together while the third is more distant. "“The JWST data allowed us not only to identify the three black holes, but also to estimate their masses, accretion rates, and the stellar mass of the galaxy,” said Dr. Giovanni Mazzolari, second author of the study and researcher at MPE. “We find a total stellar mass of about 1.3 billion suns, and the black holes represent a significant fraction of that.”

School was difficult for Makayla Caliendo when she was young. She remembers staring out the window while her classmates were paying attention in class. Some clothes, tags and socks were uncomfortable and left her unable to focus. She even had to repeat preschool.

“I left those early school years with less knowledge than my peers,” Caliendo wrote as a high-school senior in Attention magazine. That was back in 2021. Those experiences had “made me feel dumb.”

Scientists Say: Neurodivergent

But Caliendo wasn’t — and isn’t — dumb. Her brain just works differently than most of her peers.

Caliendo has a condition known as attention deficit/hyperactivity disorder, or ADHD. It affects how people think, learn and process information. Although sometimes called a “disorder,” ADHD is actually a neurotype. That means it’s a natural — and fairly common — variation in how the brain develops.

Some 22 million people in the United States have been diagnosed with ADHD. This includes more than 6 million children.

People with ADHD produce a lower amount of some of the chemicals that brain cells use to communicate. One is dopamine. It affects motivation and mood. Another is norepinephrine (Nor-ep-ih-NEF-rin). It plays a role in attention and excitement.

Explainer: What is dopamine?

These brain differences can make it harder for people with ADHD to focus on certain tasks when compared to folks without ADHD. People with ADHD may have trouble sitting still or controlling their emotions and actions.

Such traits are not always a problem, notes Jamal Williams at the University of Buffalo in New York. He studies the genetics behind ADHD. People have had traits linked to ADHD for a very long time. Recent research suggests ADHD-like behaviors, such as risk-taking and impulsivity, would have made people better foragers in ancient hunter-gatherer societies.

However, people today are expected to sit still in classes and offices. They may have to be quiet for hours at a time. This can make it difficult for people with ADHD to succeed in such environments.

Rice University researchers have shown that tiny wrinkles in graphene can change the material's electrical properties, providing evidence for flexoelectricity, a phenomenon in which a material generates an electric charge when it bends unevenly. The findings are published in Advanced Materials.

The discovery suggests scientists may be able to control electricity in atomically thin materials by changing their shape instead of adding new chemicals or materials. The approach could one day lead to more sensitive sensors and ultrathin electronic devices.

"Our work shows that even an ordinary wrinkle can become an extraordinary electronic feature when viewed at the atomic scale," said Pulickel Ajayan, the Benjamin M. and Mary Greenwood Anderson Professor of Engineering and co-corresponding author of the study. "By demonstrating that geometry alone can reshape electrical behavior in graphene, we open a new pathway for designing materials whose properties can be controlled through structure rather than chemistry."

A new AI framework, known as SHIC-XE, has been developed to detect signs of pain in horses from video analysis while providing stable, anatomically consistent explanations for its decisions. The framework was developed by an international team of researchers led by Dr. Marcelo Feighelstein, head of the Artificial Intelligence Systems Engineering Program at Tel-Hai University's new Cluster of Engineering and Advanced Computing.

For the first time, these AI-generated explanations can be quantitatively compared with expert assessments, representing a significant breakthrough in the field of explainable artificial intelligence and an important step toward AI systems that can be trusted in real-world clinical and health care environments.

The study, published in the International Journal of Computer Vision, addresses a longstanding challenge in both veterinary and human medicine: how to identify pain and distress in individuals who cannot communicate what they are experiencing.

Building a bridge between humans and animals
Animals cannot express pain in words, and even experienced veterinarians and caregivers can struggle to recognize signs of suffering. Feighelstein's research aims to bridge that communication gap by using artificial intelligence to interpret facial expressions, body language and movement patterns. Over the years, his team and collaborators have developed AI-based tools capable of recognizing pain and emotions in cats, dogs, rabbits, sheep, cattle and now horses.

"My motivation is to build a bridge between humans and animals," said Feighelstein. "We want to give a technological voice to those who have no words, allowing caregivers and professionals to better understand their condition, their emotions and their suffering."

A major challenge in AI-based video analysis is explainability. Existing methods often highlight different areas of an image from one video frame to the next, creating unstable and difficult-to-interpret explanations. To overcome this limitation, the researchers developed SHIC-XE, a novel framework that projects the model's attention onto a fixed three-dimensional representation of a horse's face. The result is a consistent anatomical explanation, even when the horse moves, changes its head position or is filmed from different angles.

Due to an ongoing commercial spacecraft attitude control issue, NASA and Katalyst Space announced Wednesday the LINK spacecraft will not capture or boost an agency satellite to a higher altitude to extend its science mission as planned. However, LINK still will attempt to conduct rendezvous and proximity operations with NASA’s Neil Gehrels Swift Observatory to demonstrate key capabilities for the future of space exploration.

NASA and Katalyst are working closely to assess next steps for rendezvous and gather as much data as possible to inform future satellite servicing operations.

Without intervention, NASA anticipates Swift is likely to re-enter Earth’s atmosphere later this year.

Read more:
go.nasa.gov/4ggh7ty

A poop transplant — otherwise known as a fecal microbiota transplant (FMT) — may help adults with long-term insomnia sleep better, a small study suggests.

After the transplant, study participants' gut bacteria became more diverse than an untreated group of people with insomnia. Their sleep efficiency, meaning the percentage of time they spent asleep while in bed, also increased, researchers reported July 22 in the Journal of Internal Medicine.

If future studies confirm that FMT is safe and effective for the treatment of insomnia, it could become another tool against the sleep disorder, joining sleeping pills and cognitive behavioral therapy (CBT).

For decades, both scientific theories and everyday thinking have treated decision-making as a separate step between perception and action. Under this traditional view, information moves through a linear sequence from sensing to thinking to acting, with each stage linked to a distinct brain function.

Many research methods, especially those used in model-based cognitive neuroscience, are built around this idea and reinforce it. It also matches our everyday experience. As James puts it, “Our actions feel like they are caused by decisions based on desires, beliefs, and intentions.”

However, James argues that this linear framework, often called the “sandwich model,” does not fit what scientists know about the brain. While researchers have identified neural systems for sensing and movement, there is no comparable brain mechanism that clearly serves as a dedicated decision-making stage.

From Decision-Making to Action Selection
Instead of proposing a separate decision-making center that directs behavior, James argues that actions emerge from the combined activity of sensory, sensorimotor, and motor processes. He prefers the term “action selection,” which reflects an ongoing interaction among the brain, body, and environment rather than a simple step-by-step process. He says this view also requires research methods that can better capture these dynamic interactions.

That does not mean decisions are unreal.

As James explains, “Of course they do. We use this language all the time, and it’s very helpful in terms of describing behavior. The leap, I think, is to say that the brain works by having decision-making or control processes. It produces behavior that is well described in that way. But it doesn’t need a process that does that to make it look that way.”

Bepi Colombo is approaching Mercury 🛰️🌑☀️ And we’re here for it! 🙌

On 3 September, the European Space Agency / JAXA en mission will carry out its first arrival manoeuvre, kicking off the Mercury Arrival Phase.

Want to watch live on 3 Sept?

Stay tuned here and on ESA Science for more details on 27 August 🗓️

This year's El Niño is on track to be the strongest on record, threatening to unleash extreme weather across the globe and make 2027 the hottest recorded year by far. It is expected to be the most "intense" El Niño in more than a century, the head of long-range forecasting at the U.K.'s Met Office warned Friday.

Stacked on top of long-term human-caused climate change, the temporary jolt will offer a glimpse of the heat the world can expect by the late 2030s, experts say.

Here's what to know.

Perspective is everything. 🌍 Gazing down at Earth from above offers an incredibly profound one, as does looking back at the Space Station from a window in the Russian segment. Our crewmate Anna Kikina captured this incredible footage during Anil Menon and my spacewalk on Aug. 6. Even suited up in our bulky spacesuits, we look like tiny little worker bees at the hive given the massive scale of the station.

I’ll just say it: What the hell is going on with AI “reasoning”?

Sorry for the air quotes. That punctuational side-eye was more common in 2024, when the specially trained cousins of LLMs now known as “large reasoning models,” or LRMs, were still new. Nowadays it may seem downright churlish, though, given that a “general-purpose reasoning model” from OpenAI solved a famous open mathematical research problem in one shot in May 2026. Still, I’m not sure how else to acknowledge my intellectual whiplash over the scientific interpretation of what these AI systems are actually doing.

Reasoning comes in many technically defined forms, but the basic procedure is easily recognizable: arriving at a sound conclusion by linking together intermediate steps that logically follow from each other. We do this with thoughts; LRMs use so-called chains of thought, a term of art for the streams of synthetic text that the models emit before arriving at an answer to a complex query. One minute, the idea that AI could reason via these chains was being prominently and credibly critiqued (by a team of researchers from Apple) as an “Illusion of Thinking” subject to “complete accuracy collapse” under surprisingly simple conditions. The next minute, LRMs were bagging gold medals at the International Mathematical Olympiad, a feat so challenging that “even very successful mathematicians and scientists may well highlight [it] on their CVs all their lives,” as the scientist and AI critic Gary Marcus and Ernest Davis wrote in 2025. If that’s not a sign of “real” reasoning, what is?

But wait — soon after, more research, from the Santa Fe Institute, showed that LRMs can crush even carefully designed benchmarks for reasoning (like a collection of analogy-like visual puzzles) using mere “surface-level ‘shortcuts.’”

What they were doing looked less like generalizable reasoning than just gaming the system. Then, as if on cue, another “hold my beer” moment: Google DeepMind and the mathematician Terence Tao (the GOAT!) used AI to rediscover or improve the solutions to 67 problems “spanning mathematical analysis, combinatorics, geometry, and number theory.” Deal with it, haters!

What about additional evidence that LRMs can’t reason reliably, even when they possess the necessary algorithm and computational budget to do so, and suffer from a list of scientifically documented failure states long enough to use as a Slip ’N Slide? Whatever — I guess that’s just “jagged intelligence” for you (AI-speak for “when it works, it works”).

And so it went from late 2025 into 2026. I’ve been a science journalist for 20 years and an AI journalist for half of that, so I know better than to expect tidy consistency out of rapidly advancing research. But even for me, this back-and-forth has been a bit much. To quote Al Pacino in The Insider, “I’m getting two things: pissed off, and curious.” I don’t believe there’s fraud to be found here. I just want to know which way is up. Can AI reasoning somehow be both BS and not at the same time? And if so, how on Earth does that work?

I knew just who to call first.