Researchers linked higher plasma TMAO to prevalent atrial fibrillation in 5,090 people and found that TMAO or choline supplementation accelerated AF onset, progression, atrial remodeling, and electrical dysfunction in mice. Inhibiting gut microbial TMA production lowered circulating TMAO and delayed AF in mice, identifying a potential therapeutic pathway that now requires human testing.
Viagra was originally developed to treat high blood pressure and chest pain caused by reduced blood flow to the heart, but over the past three decades it has become the world’s best-known therapy for erectile dysfunction. In the future, this blockbuster drug may find yet another use. In a study published recently in Cancer Research, scientists in the lab of professor Ayelet Erez at the Weizmann Institute of Science found that sildenafil, Viagra’s active ingredient, may restrict cancer metastasis through a newly discovered biological mechanism.
The researchers, led by Dr. Yarden Ariav in Erez’s lab, showed that sildenafil limits cancer cells’ ability to use cholesterol, an essential component of cellular membranes. Cholesterol is particularly important for cancer cells seeking to break away from the primary tumor, migrate throughout the body and invade distant organs. When their access to cholesterol is disrupted, these cells have a harder time forming metastases.
When humans tackle many everyday tasks, ranging from driving to completing household chores and work assignments, they rely heavily on working memory. This is the brain’s ability to temporarily store and manipulate important information that can help complete a task in the immediate future.
This temporarily stored information can be used to make decisions and plan subsequent actions, such as which supermarket aisle to visit next or where to turn while driving. While many studies have investigated working memory processes, the impact of distractions or unexpected events on these processes remains poorly understood.
Researchers at the University of Texas at Austin recently conducted a study aimed at better understanding how people respond to distractions while completing tasks. Their findings, published in Communications Psychology, suggest that working memory is highly flexible and can rapidly reprioritize the information it is temporarily holding after an interruption.
A UNIGE study shows that even a slight overproduction of tubulin, a protein essential for cell structure, is enough to disrupt tissue organization, highlighting the need for tight regulation of protein levels.
For cells to function properly, they must produce the right amount of each protein. It is a delicate balance: both too little and too much can compromise essential cellular functions. A team from the University of Geneva (UNIGE) has now shown that even a modest excess of tubulin – the protein that assembles into microtubules, the cell’s internal scaffolding – is enough to disrupt tissue architecture and reduce cell viability. Published in Nature Communications, the study demonstrates that the quantity of a protein is just as important as its function.
Microtubules, built from tubulin, form the cell’s internal skeleton. They help cells maintain their shape, transport molecules, divide, and remain firmly attached to neighboring cells. Rather than being rigid structures, microtubules are constantly assembled and disassembled to adapt to the cell’s changing needs. This dynamic behavior depends directly on the amount of tubulin available.
A Russian spy at the center of a secret operation to smuggle high-tech weapon components into Russia has left Japan after his activities were disclosed by The New York Times, a senior Ukrainian intelligence official said Thursday.
The operative, Maksim Vladimirovich Filchenkov, 49, had been working undercover in Japan as an employee of Russian state airline Aeroflot. Filchenkov’s role had been the focus of a Times investigation into how Russia exploited Japan’s weak espionage laws.
Filchenkov’s departure was first reported by the Yomiuri Shimbun, a major national newspaper in Japan. The newspaper said the Russian exited Japan after the Times published its investigation July 12. The Ukrainian official who confirmed the spy’s departure did so on condition of anonymity because he was not authorized to discuss intelligence publicly.
The most severe droughts in the ancient Eastern Mediterranean arose when multiple natural climate cycles coincided, according to a new study from Stockholm University published in Science Advances. The findings shed new light on the climate conditions surrounding the Late Bronze Age collapse and may help improve understanding of future drought risks in a warming world.
“Rather than being caused by a single climatic event, we found that the most extreme droughts emerged when natural climate cycles operating over different timescales coincided. This helps explain why the droughts associated with the Late Bronze Age collapse were so severe,” said Katherine Power, a doctoral student in the Department of Physical Geography at Stockholm University and the study’s first author.
Kimi K3 is available today on Kimi.com, Kimi Work, Kimi Code, and the Kimi API. At launch, Kimi K3 will use max thinking effort by default, with low-and high-effort modes to be introduced in subsequent updates. We are currently working closely with inference partners and open-source maintainers to align technical details and ensure a reliable rollout across the ecosystem. The full model weights will be released by July 27, 2026. Further details on the architecture, training, and evaluations will be released alongside the Kimi K3 technical report.
Kimi K3 is the first open model to reach 2.8 trillion parameters. It marks the latest step in Kimi’s sustained push at the scaling frontier: for nine of the past twelve months, Kimi models have set the upper bound of open-model sizes.
Kimi K3 is built on Kimi Delta Attention (KDA) and Attention Residuals (AttnRes), two architectural updates designed to improve how information flows across sequence length and model depth. We have also scaled up Mixture of Experts (MoE) sparsity, effectively activating 16 out of 896 experts when paired with a Stable LatentMoE framework. Together with refined training and data recipes, these structural changes yield an approximate 2.5× improvement in overall scaling efficiency compared to Kimi K2, allowing the model to convert compute into intelligence more effectively.
