Toggle light / dark theme

Astronomer Calvin Leung was excited last summer to crunch data from a newly commissioned radio telescope to precisely pinpoint the origin of repeated bursts of intense radio waves—so-called fast radio bursts (FRBs)—emanating from somewhere in the northern constellation Ursa Minor.

Leung, a Miller Postdoctoral Fellowship recipient at the University of California, Berkeley, hopes eventually to understand the origins of these mysterious bursts and use them as probes to trace the large-scale structure of the universe, a key to its origin and evolution. He had written most of the computer code that allowed him and his colleagues to combine data from several telescopes to triangulate the position of a burst to within a hair’s width at arm’s length.

The excitement turned to perplexity when his collaborators on the Canadian Hydrogen Intensity Mapping Experiment (CHIME) turned optical telescopes on the spot and discovered that the source was in the distant outskirts of a long-dead elliptical galaxy that by all rights should not contain the kind of star thought to produce these bursts.

In today’s AI news, Galileo launched an Agent Leaderboard on Hugging Face, an open-source AI platform where users can build, train, access, and deploy AI models. The leaderboard is meant to help people learn how AI agents perform in real-world business applications and help teams determine which agent best fits their needs.

In other advancements, Bloomberg reported Friday that xAI is canvassing existing investors, including Sequoia Capital, Andreessen Horowitz, and Valor Equity Partners for the round, which would bring xAI’s total raised to $22.4 billion, according to Crunchbase. Bloomberg also noted that discussions are ongoing and that the terms of the fundraising round may change.

Ve done mobile app development will know how challenging it can be to deliver the right kind of experience on a smartphone. + And, while speaking with former U.K. Prime Minister Tony Blair at the World Governments Summit in Dubai on Wednesday, Oracle cofounder and executive chairman, Larry Ellison said that while government organizations collect massive amounts of data, it is highly fragmented, making it hard to feed it into an AI model.

In videos, the Imagination in Action video series from Davos 2025 is being uploaded and we’re featuring the sessions in today’s newsletter. First we dive into an in-depth panel discussion featuring AI visionaries Max Tegmark, Demis Hassabis, Yoshua Bengio, Dawn Song, and Ya-Qin Zhang. In this engaging conversation, the experts unpack the distinctions between narrow AI, AGI, and super intelligence …

And, an expert panel explores how regulation can drive innovation in AI, featuring perspectives from panelists: Robert Mahari, JD-PhD at Massachusetts Institute of Technology and Harvard Law School, Pablo Arredondo, Vice President of CoCounsel at Thomson Reuters and Founder of Casetext, Part of Thomson Reuters, Julia Apostle, Partner at Orrick, Herrington & Sutcliffe LLP, Gabriele Mazzini, Fellow at MIT Connection Science and Architect of EU AI Act.

This process, which cannot be understood satisfactorily by classical physics alone, occurs constantly in green plants and other photosynthetic organisms, such as photosynthetic bacteria. However, the exact mechanisms have still not been fully elucidated. Hauer and first author Erika Keil see their study as an important new basis in the effort to clarify how chlorophyll, the pigment in leaf green, works.

Applying these findings in the design of artificial photosynthesis units could help to utilize solar energy with unprecedented efficiency for electricity generation or photochemistry.

Most metals expand as their temperature rises. The Eiffel Tower, for example, stands about 10 to 15 centimeters taller in summer than in winter due to thermal expansion. However, this effect is highly undesirable for many technical applications. As a result, researchers have long sought materials that maintain a constant length regardless of temperature. One such material is Invar, an iron-nickel alloy known for its extremely low thermal expansion. The physical explanation for this property, however, remained unclear until recently.

Now, a collaboration between theoretical researchers at the Vienna University of Technology (TU Wien) and experimentalists at the University of Science and Technology Beijing has led to a significant breakthrough. Using complex computer simulations, they have unraveled the invar effect in detail and developed a so-called pyrochlore magnet—an alloy with even better thermal expansion properties than Invar. Over an exceptionally wide temperature range of more than 400 Kelvins, its length changes by only about one ten-thousandth of one percent per Kelvin.

Based on outstanding technical progress by research teams to date, DARPA has pivoted the third and final phase of its NOM4D (pronounced nō- mad) program[1] from planned further laboratory testing to conducting a pair of small-scale orbital demonstrations to evaluate novel materials and assembly processes in space.

As commercial space companies continue to expand access to orbit for U.S. economic and national security needs, a major roadblock for building large-scale structures in orbit remains: the size and weight limits imposed by a rocket’s cargo fairing. In 2022, DARPA introduced NOM4D to break this cargo-constraint mold by exploring a new paradigm. Instead of folding or compacting structures to fit them into a rocket fairing to be unfurled or deployed in space, DARPA proposed stowing novel lightweight raw materials in the rocket fairing that don’t need to be hardened for launch. The intent of this approach is to allow in-orbit construction of vastly larger and more mass-efficient structures than could ever fit in a rocket fairing. Additionally, this concept enables mass-efficient designs of structures that would sag under their own weight on Earth but are optimized for the low-gravity environment of space.

“Caltech [California Institute of Technology] and the University of Illinois Urbana-Champaign have demonstrated tremendous advances in the first two phases and have now partnered in Phase 3 with space-launch companies to conduct in-space testing of their novel assembly processes and materials,” said Andrew Detor, DARPA NOM4D program manager. “Originally, Phase 3 was going to be about making things more precisely in the lab than we did in Phase 2. But we said, ‘You know, the maturity is there, and there would be more impact if we took the capabilities we have now and actually go demonstrate them in space to show that it can be done.’ Pushing the performers to do a demo in space means they can’t just sweep challenges under the rug like they could in a lab. You better figure out how it’s going to survive in the space environment.”

🛟🧬🧫🩹⚙️


Wound infections are common combat injuries and can take otherwise able-bodied personnel out of operations and/or result in severe medical complications. Current standard of care relies on complicated and often time-consuming tests to identify the specific infection-inducing pathogens that caused the wound infection. Therapeutic treatments rely on broad-spectrum and high-dose antibiotics alongside surgical excision – which are not pathogen specific, drive antibiotic resistance, can have toxic side effects, require advanced medical training, and can result in high treatment costs and burden on patients. A game-changing approach to managing infection of combat wounds, particularly one that can be applied autonomously, would benefit warfighter readiness and resilience.

The BioElectronics to Sense and Treat (BEST) program seeks to meet this need by developing wearable, automated technologies that can predict and prevent a wound infection before it can occur, and to eliminate an infection if it has already taken hold. To achieve this, DARPA is seeking researchers to develop novel bioelectronic smart bandages comprised of wound infection sensor and treatment modules. The sensors should be high-resolution and provide real-time, continual monitoring of wounds based on, for example, the person’s immune state and the collection of bacteria that live in and around a wound. Data from these sensors will be used to predict if a wound will fail to heal due to infection, diagnose the infection, and regulate administration of targeted treatments – using closed-loop control to prevent or resolve infection for improved wound healing.

“Given that infection initiates at the time of injury and can take hold before aid arrives, particularly in austere environments, the earlier we can deploy these technologies, the bigger impact they will have,” noted Dr. Leonard Tender, BEST program manager. “Even if medivac occurs immediately, without the ability to prevent infection, the downstream care required to treat the surge of wound infections resulting from a large-scale combat operation could easily overwhelm care capacity.”

Summary: A new study suggests that ChatGPT’s responses in psychotherapy scenarios are often rated higher than those written by human therapists. Researchers found that participants struggled to distinguish between AI-generated and therapist-written responses in couple’s therapy vignettes. ChatGPT’s responses were generally longer and contained more nouns and adjectives, providing greater contextualization.

This additional detail may have contributed to higher ratings on core psychotherapy principles. The findings highlight AI’s potential role in therapeutic interventions while raising ethical and practical concerns about its integration into mental health care. Researchers emphasize the need for professionals to engage with AI developments to ensure responsible oversight.