New AI Assistant ‘Denario’ Aims to Accelerate Scientific Discovery
Researchers have introduced Denario, an AI-powered scientific assistant designed to accelerate the research process. This multi-agent system can assist with numerous tasks, including generating new research ideas, reviewing existing literature, creating research plans, writing and executing code, producing plots, and drafting scientific papers. Developed by a team from the University of Cambridge, the Flatiron Institute, and the Autonomous University of Barcelona, Denario is a modular tool for scientists at various stages of their work. While it has demonstrated the ability to generate entire research papers across disciplines like astrophysics and biology, its creators emphasize that it is a tool to augment, not replace, human scientists. They note the current version has limitations, with only about one in ten outputs providing significant insights and instances of data fabrication, making human oversight essential. The project’s code is publicly available, with a demo version accessible online.
Sources:
AI Tool Accurately Detects Structural Heart Disease Using Smartwatch ECG
A new AI algorithm can accurately diagnose various structural heart diseases by analyzing single-lead electrocardiogram (ECG) data from a smartwatch. In a preliminary study, the AI tool showed high accuracy in identifying conditions such as weakened pumping ability, damaged valves, and thickened heart muscle. This is the first prospective study to demonstrate that an AI algorithm can detect multiple structural heart diseases using the single-lead ECG sensor found on many consumer smartwatches. The model was trained on a large dataset of hospital-grade 12-lead ECGs and then adapted for the single-lead format. In a test with 600 adults, the algorithm accurately identified most individuals with heart disease and was highly effective at ruling it out in healthy individuals. This technology could make early screening for structural heart disease more accessible and affordable by leveraging existing consumer hardware.
Sources:
University of Bristol Develops Robotic Exosuit to Aid Astronauts in Space
Researchers at the University of Bristol have developed a soft robotic exosuit to assist astronauts with movement during space missions. The fabric-based garment is worn under a spacesuit and uses artificial muscles to automatically reduce muscular fatigue while allowing for natural movements. The technology could also have terrestrial applications for individuals needing mobility support. The exosuit recently underwent testing in a simulated lunar environment in Australia as part of an international proof-of-concept space mission. This marked the first time a soft robotic exosuit was integrated into a spacesuit for a field test, which evaluated its comfort and mobility during tasks like walking and carrying loads on loose terrain.
Adobe Launches Firefly Foundry for Custom Generative AI Models
Adobe has announced the launch of Adobe Firefly Foundry, a new service enabling businesses to create bespoke generative AI models trained on their own intellectual property. Unveiled at the Adobe MAX conference, this platform allows companies to develop proprietary models for images, video, audio, and 3D assets that align with their specific brand identity. The service is designed to help enterprises scale content production while maintaining brand consistency. Adobe will collaborate with businesses, providing access to experts to assist in the AI model training and deployment process. Firefly Foundry models are built on Adobe’s commercially safe Firefly models and can be integrated across Adobe’s ecosystem, including Creative Cloud and Express. As part of this initiative, the team from generative media solution Invoke has joined Adobe to help build out future AI-powered creative workflows.
Sources:
NVIDIA Releases New Open-Source AI Models for Language, Robotics, and Biology
NVIDIA has unveiled a suite of new open-source AI technologies aimed at advancing innovation in language, robotics, and biology. The release includes new models and data tools within the NVIDIA Nemotron family for AI reasoning, the NVIDIA Cosmos platform for physical AI, NVIDIA Isaac GR00T for robotics, and NVIDIA Clara for biomedical AI. These resources are being contributed to Hugging Face to make AI research and development more accessible to the global developer community. The Nemotron family includes models designed to help developers build specialized intelligent agents for tasks like software development and customer service. For physical AI and robotics, updates to Cosmos and Isaac GR00T aim to accelerate the training of robotic systems. Major companies, including PayPal and CrowdStrike, are already utilizing the Nemotron models to enhance their AI services.
Sources:
Anthropic Launches Claude Sonnet 4.5, a New Frontier Model for Coding and AI Agents
Anthropic has released Claude Sonnet 4.5, positioning it as their most capable model for coding, AI agents, and computer use. The new model demonstrates state-of-the-art performance on benchmarks that measure real-world software coding abilities, such as SWE-bench Verified. Anthropic claims the model can maintain focus for over 30 hours on complex, multi-step tasks. Alongside the model, Anthropic has released the Claude Agent SDK, providing developers with the infrastructure to build long-running, autonomous AI agents. Claude Sonnet 4.5 is available through the Claude API, Amazon Bedrock, and Google Cloud Vertex AI at the same price point as its predecessor, Claude Sonnet 4. This release is a significant step toward creating AI that can reliably perform complex digital tasks and workflows.
Sources:
Quantum Computing Firm Xanadu to Go Public in $3.6 Billion SPAC Deal
Toronto-based photonic quantum computing company, Xanadu Quantum Technologies Inc., has announced it will go public through a merger with a special purpose acquisition company (SPAC), Crane Harbor Acquisition Corp. The deal gives Xanadu a pre-money equity value of $3 billion and a combined market capitalization of $3.6 billion. The transaction is expected to raise approximately $500 million in gross proceeds, which will be used to accelerate the development of their fault-tolerant quantum computers. Upon completion, the merged entity will be listed on both the Nasdaq and the Toronto Stock Exchange. Xanadu’s photonic quantum computers have the advantage of operating at room temperature. The company’s software platform, PennyLane, is also a key part of its offerings.
Sources:
China’s First Atomic Quantum Computer ‘Hanyuan-1’ Achieves Commercialization
China has announced the commercialization of its first atomic quantum computer, named ‘Hanyuan-1’. The neutral-atom quantum computer has reportedly secured over 40 million yuan ($5.6 million USD) in orders, including an export order to Pakistan. Developed by the Chinese Academy of Sciences’ Innovation Academy for Precision Measurement Science and Technology, the 100-qubit system operates at room temperature, significantly reducing energy consumption and maintenance compared to superconducting quantum computers. The entire system is housed in three standard equipment racks. A subsidiary of China Mobile is among the first domestic customers. China also plans to establish the country’s first neutral-atom quantum computing power center.
Sources:
Apple Releases iOS 26.1, iPadOS 26.1, and macOS 26.1 with Key Security Fixes
Apple has released significant updates for its operating systems, including iOS 26.1, iPadOS 26.1, and macOS 26.1. These updates address multiple critical security vulnerabilities that could have exposed user data. Key security enhancements patch flaws in the Apple Neural Engine, WebKit, and other system components to prevent malicious applications from executing arbitrary code or accessing sensitive information. In addition to security fixes, the updates introduce user interface changes, such as a new option to give the “Liquid Glass” design a more opaque, frosted appearance. For iPad users, iPadOS 26.1 sees the return of the “Slide Over” multitasking feature. The updates are available for iPhone 11 and later, as well as a range of recent iPad and Mac models.
Sources:
Tigris Data Unveils Bucket Forking for GitHub-Style Object Storage Workflows
Tigris Data has launched a new feature called bucket forking, which brings GitHub-style forking capabilities to object storage. This allows users to create a metadata-only copy of a data bucket, enabling them to modify, add, or delete data without affecting the original source. The forked bucket and the source are isolated, similar to forking code in git. This feature is designed to streamline development workflows for data-intensive applications, such as AI and big data analytics, by allowing for experimentation and rapid feedback without creating complete data copies. The technology is built on an immutable, append-only architecture and uses the open-source FoundationDB for metadata management. Bucket forking is compatible with the AWS S3 API.
Sources: