Comment

Comment by Dr. Krishna Kumari Challa on August 15, 2026 at 10:23am

AI agents struggle to perform original scientific research

Among the many predictions about the future of artificial intelligence is that models will one day be able to conduct scientific research on their own, leaving humans out of the equation. Already, they can write code, run experiments and search scientific literature, but carrying out open-ended research would require a significant leap in ability.
In a paper posted on the arXiv preprint server, researchers tested AI's ability to conduct open-ended research and found that it came up short.
The study authors gave frontier agents (cutting-edge, state-of-the-art AI tools designed to carry out complex, multi-step tasks autonomously) six days to conduct research and write papers based on two then-unpublished AI conference submissions. This ensured they couldn't just find the answers online.

The agents had full access to the internet, dedicated computing power and approximately $3,000 in model-use credits, meaning they had a budget to conduct open-ended exploration and run experiments. The topics they had to research and write about were the structure and controllability of language-model personas and designing a detector for distribution shifts in tabular foundation models.

Once the six days were up, human researchers reviewed the AI-written papers and graded them as they would papers submitted to a top-tier AI conference.

The frontier agents did not do well at all. Both papers received unambiguous rejection scores (2/6 and 1/6 overall) from the expert human reviewers. Although the AI understood the research questions and proposed some directions that closely mirrored those of the original researchers, its scientific reasoning suffered from major flaws. Experimental designs were weak, and the agents handled negative feedback poorly, often adding caveats to existing findings rather than redesigning their studies.

They also managed their time poorly and spent less than half of their allocated API budget. Despite having time, they rushed through their work and submitted papers that fell far short of publishable standards.

The reviewers did not hold back on their assessments of the agents:

"The experiments and methodological choices were bizarre and hard to understand. The results seem clearly a result of post hoc choices."—David Africa, expert reviewer.
"Upon testing a few unsuccessful signals using a PFN's internals, going from there to 'there are no signals we can use that leverage a model's internals' is a huge leap, a kind of 'proof by example' fallacy that is highly non-scientific."—Viet Nguyen, expert reviewer.
While these results are a sobering reality check on AI's ability to perform scientific research, they do not mean models have no place in the lab. In the near term, they are more likely to serve as assistants handling routine tasks rather than being deeply involved in the process of discovery.

Peter Kirgis et al, Can AI agents conduct open-ended AI research? Early evidence from two case studies, arXiv (2026). DOI: 10.48550/arxiv.2607.27191

Comment by Dr. Krishna Kumari Challa on August 14, 2026 at 11:12am

Fridge-free vaccines

Scientists have successfully trialled vaccines that do not need to be refrigerated or frozen, which could slash the number of vaccine doses that are wasted each year. Sixty volunteers received a tetanus-diphtheria vaccine that had been kept as a dry powder for a year at room temperature before it was dissolved and injected. The jab was “safe and well tolerated” and gave “equivalent” immune responses to the conventional vaccine. The team will launch a larger trial of 160 people in the coming months.

https://www.thelancet.com/journals/eclinm/article/PIIS2589-5370(26)00369-X/fulltext?utm_source=Live+Audience&utm_campaign=e2b449d6f7-nature-briefing-translational-research-20260812&utm_medium=email&utm_term=0_-928f7c52f0-50323416

Comment by Dr. Krishna Kumari Challa on August 14, 2026 at 10:28am

Fungal age remains elusive as underground networks continually grow, split and recycle
Fungal age is difficult to define because mycelial networks continually grow, recycle tissue and fragment into genetically identical units. Longevity likely varies among species and lifestyles, requiring genetic tracking, long-term experiments and microfluidic systems to characterize persistence and life cycles.

Exploring the concept of longevity in fungi, Trends in Microbiology (2026). DOI: 10.1016/j.tim.2026.07.001

Comment by Dr. Krishna Kumari Challa on August 14, 2026 at 9:33am

Nature's original bioplastic may have fed animals for hundreds of millions of years

Long before humans discovered biodegradable plastics, microorganisms had already invented their own. Many bacteria and archaea produce natural bioplastics called polyhydroxyalkanoates (PHAs), storing them inside their cells as reserves of carbon and energy.
Until now, scientists thought that only microorganisms themselves could break down these substances. Researchers at the Max Planck Institute for Marine Microbiology in Bremen, Germany, have now overturned that long-standing assumption.

In a study published in Nature Ecology & Evolution, they show that animals ranging from marine worms and starfish to terrestrial species, including earthworms, have enzymes capable of degrading microbial PHAs. The findings reveal a previously overlooked way in which microbial carbon can enter animal food webs.
Microbial PHAs occur naturally in soils, sediments and aquatic environments worldwide. They are produced whenever microorganisms store excess carbon for later use and are among the few naturally occurring plastics that are completely biodegradable. Because PHAs are increasingly manufactured as sustainable alternatives to conventional plastics, understanding how they are degraded in nature has become an important area of research.
The new findings suggest that animals, together with microorganisms, may contribute to the breakdown of these natural bioplastics. More fundamentally, they reveal that animals can exploit a microbial carbon reserve that had previously been thought to be inaccessible to them.

Animal degradation of microbial storage polyhydroxyalkanoates, Nature Ecology & Evolution (2026). DOI: 10.1038/s41559-026-03153-8

**

Comment by Dr. Krishna Kumari Challa on August 14, 2026 at 9:20am

A 236-million-year-old fossil challenges the story of mammalian live birth

Some cynodonts may have been giving birth to live young much sooner in evolutionary history than previously assumed. A new Frontiers in Mammal Science study has offered the first compelling evidence that cynodonts may have been viviparous—a reproductive mode characterized by live birth.

Researchers showed for the first time that live birth was present in at least one mammalian ancestor, Chiniquodon theotonicus, which lived approximately 236 million years ago. This implies that viviparity among early cynodonts originated in the mammalian lineage at least 95 to 90 million years earlier than previously thought.

Cynodonts thrived in the Triassic, a period of recovery and restructuring of ecosystems after one of the most devastating mass extinctions in life history. This meant high competition for resources and strong predatory pressures. Combined with a trend toward aridity and strong seasonality, embryos of viviparous species would be better protected than those of egg-laying species.

Until now, giving birth to live young was considered a relatively modern evolutionary acquisition in the mammalian lineage. The finding raises questions about which other traits believed to have appeared much later were already present among cynodonts, the researchers say.

Early Origin of Viviparity in the Mammalian Lineage, Frontiers in Mammal Science (2026). DOI: 10.3389/fmamm.2026.1845319

Comment by Dr. Krishna Kumari Challa on August 13, 2026 at 12:03pm

Female gut muscles reshape to meet the demands of reproduction, preclinical study suggests

Organs don't just grow in early life; they can change in response to physiological or environmental challenges in adulthood. Researchers have now identified an active role for the intestinal muscles in remodelling the gut after reproduction.
Reproduction remodels intestinal smooth muscle in female fruit flies and mice, elongating contractile filaments without increasing cell number and reducing contractility. In flies, mating lowers juvenile hormone receptor signalling in muscle, permitting growth, while epithelial growth is promoted by the same hormone. These changes may enhance nutrient absorption during reproduction.

Alessandro Mineo et al, The sex and reproductive plasticity of intestinal muscles instruct gut size, Cell (2026). DOI: 10.1016/j.cell.2026.07.024

Comment by Dr. Krishna Kumari Challa on August 13, 2026 at 11:26am

For decades we have known that rheumatoid arthritis selectively targets particular joints, but one of the great unanswered questions is why?
These new findings suggest that the answer lies not only in the immune system but also in the tissues themselves. The cellular and structural characteristics established during development may help determine where inflammation takes hold later in life.
Researchers found that the developing joints were made up mainly of structural cells, including cartilage-forming cells and fibroblasts, rather than immune cells. They then investigated what drives these cells to develop into their different specialized forms.

One population that drew particular attention was the synovial lining fibroblasts. These cells produce substances that lubricate the joint to help protect and maintain smooth movement, yet they can also behave abnormally in arthritis. Further analysis suggested that the lining may come from two different sources, both the cartilage and surrounding joint fibroblasts.

The process appeared to be influenced by specific localized signals such as low oxygen levels. This may provide insights into the mechanisms driving their function and help identify ways to restore their normal protective role in disease.
The researchers found important differences between the PIP and DIP joints. A bespoke image analysis tool showed that PI16+ fibroblasts that were enriched in the PIP joints were specifically located around blood vessels and at sites where tendons and ligaments connect with surrounding tissue.

They also showed that PI16+ fibroblasts responded differently to inflammatory signals compared with other fibroblast populations. While PI16+ fibroblasts shared a common pro-inflammatory response with PI16- fibroblasts, they also displayed distinct changes in pathways linked to tissue organization and immune regulation.

The team also identified striking structural differences between the joints. Using high-resolution 3D imaging at Diamond Light Source at the Harwell Science and Innovation Campus, they found that the synovial tissue surrounding PIP joints was larger and organized differently from that seen in joints that are not usually affected by rheumatoid arthritis. Together, these cellular and structural differences may help explain why inflammation develops in some locations but not others.
Together, the findings suggest that the tendency of rheumatoid arthritis to affect particular joints may be shaped by tissue architecture established during development. Rather than being determined by immune activity alone, vulnerability to inflammation may depend on the local cellular and structural environment of each joint.

Sarah Davidson et al, The embryonic origins of site-specific arthritis, Nature Immunology (2026). DOI: 10.1038/s41590-026-02542-2

Comment by Dr. Krishna Kumari Challa on August 13, 2026 at 11:03am

Arthritis origins may begin before birth

Researchers have discovered that some joints may be more susceptible to inflammatory arthritis before birth, offering new insight into why rheumatoid arthritis attacks particular joints while sparing others.
"The embryonic origins of site-specific arthritis", published in Nature Immunology, provides new insight into one of the long-standing mysteries of rheumatoid arthritis: why inflammation targets certain joints while leaving others relatively unaffected.

Rheumatoid arthritis is an autoimmune disease that causes pain, swelling and stiffness when the immune system attacks the synovium, the tissue lining the joints. Over time, this inflammation can damage cartilage, bone and surrounding tissues.

The study compared two finger joints that differ in their susceptibility to rheumatoid arthritis. Researchers compared proximal interphalangeal (PIP) joints, which are commonly affected by the disease, and distal interphalangeal (DIP) joints near the fingertips, which are usually spared.

They found that the PIP joints contained larger synovial volumes and higher levels of PI16-positive (PI16+) fibroblasts, a specialized type of connective tissue cell. These differences were established before birth, suggesting that the tissues themselves may play an important role in determining where disease occurs.

The embryonic origins of site-specific arthritis. Credit: NDORMS

Part 1

Comment by Dr. Krishna Kumari Challa on August 13, 2026 at 9:49am

What if there's a star inside a black hole?
A dark-matter halo with an exotic pressure–density relation can mathematically surround a neutron star with an event horizon while leaving the star nonsingular and structured inside. Such configurations exist only within a narrow halo-density range and require dark-matter densities far above expected astrophysical values, so they remain theoretical.

Chen Tan et al, A Neutron Star Hidden Inside a Black Hole, arXiv (2026). DOI: 10.48550/arxiv.2608.06224

Comment by Dr. Krishna Kumari Challa on August 13, 2026 at 9:01am

What made the visible-eyes videos different? To answer that question, researchers used EEG readings. The adult speaker's EEG readings were recorded in advance during video production, while the babies wore EEG caps as they watched the videos. Analyzing brain signals allowed researchers to determine the degree of matching between the adult's and baby's brain activity.

Researchers found that adult-to-infant neural coupling—the one-way matching of adult brain waves to infant brain waves—predicted the infants' ability to learn. Only adult→infant neural coupling, not the infants' own rhythmic brain activity, predicted learning ability.

"The key insight," the authors point out, "is that speaker-to-listener neural coupling is a better predictor of selective learning than infants' own neural activity." Statistical tests supported this: Once researchers accounted for the strength of adult-to-infant coupling, the direct effect of gaze on learning essentially vanished. In short, eye contact drove a brain-to-brain signal that explained the learning boost.

Wei Zhang et al, Adult-to-infant unidirectional neural coupling mediates selective social learning in infants from the United Kingdom and Singapore, Nature Communications (2026). DOI: 10.1038/s41467-026-75831-x

Part 2

• ‹ Previous
• 1
• …
• 22
• 23
• 24
• 25
• 26
• …
• 1301
• Next ›
• Page