New AI System Enhances At-Home Balance Training for Physical Therapy
Researchers at the University of Michigan have developed a new machine learning model to support home-based balance training for physical therapy patients. The system utilizes data from four wearable sensors to predict how a physical therapist would assess a patient’s performance on balance exercises. This AI-driven feedback can help patients make faster progress and maintain their abilities after their formal therapy sessions have concluded. The model was developed by training convolutional neural networks with sensor data and evaluations from physical therapists, achieving nearly 90% accuracy in predicting patient balance ratings. The technology aims to make balance training more accessible, particularly for those in rural areas, and to assist physical therapists in making healthcare decisions.
Flexiv Robotics Demonstrates Advanced Adaptive Robots at iREX 2025 in Tokyo
Flexiv Robotics, a developer of general-purpose robots, made its debut at the International Robot Exhibition (iREX) in Tokyo, a significant move in its global expansion. The company demonstrated the capabilities of its Rizon series robots through five live demonstrations, including automated car bodywork polishing and precision server component insertion. A notable feature was a massage station where the Rizon 4s adapted to each visitor’s body, showcasing the precision and responsiveness of its adaptive robotic technology. Flexiv views the Japanese market as a priority due to the high demand for flexible and intelligent automation in the country’s industries.
Sources:
Waymo Issues Software Recall for Self-Driving Cars Illegally Passing School Buses
Autonomous ride-hailing service Waymo is planning a voluntary software recall following reports of its self-driving vehicles illegally passing stopped school buses. The National Highway Traffic Safety Administration (NHTSA) initiated an investigation in October after a media report and has received documentation of multiple instances of this issue. Waymo has stated that no injuries have occurred due to this problem and that a software update is expected to address the issue. Despite this, the company emphasizes its overall safety record, citing significantly fewer crashes with serious injuries compared to human-driven vehicles.
Sources:
War Department Launches GenAI.mil Platform Powered by Google’s Gemini AI
The War Department has launched GenAI.mil, a new artificial intelligence platform aimed at fostering an ‘AI-first’ workforce and enhancing operational efficiency. The inaugural AI capability featured on the platform is Google Cloud’s Gemini for Government. This initiative will make advanced AI models available to all military personnel, civilians, and contractors, in line with the White House’s AI Action Plan. The platform is certified for Controlled Unclassified Information (CUI) and Impact Level 5 (IL5), ensuring security for operational use. The Under Secretary of War for Research and Engineering, Emil Michael, stated that the rapid deployment of powerful AI capabilities is crucial for maintaining America’s dominance in the global AI race.
Sources:
NVIDIA Expands Open-Source AI with New Models and Tools at NeurIPS
At the NeurIPS conference, NVIDIA expanded its suite of open AI models, datasets, and tools to support research in both digital and physical AI. The releases include Alpamayo-R1, described as the first industry-scale open reasoning vision language action (VLA) model for autonomous driving research. In the realm of digital AI, NVIDIA introduced new multi-speaker speech AI models, including MultiTalker Parakeet for automatic speech recognition and Sortformer for real-time speaker diarization. These initiatives underscore NVIDIA’s commitment to open-source development in the AI community.
Sources:
Kubernetes 1.35 Release Preview: New Alpha Features for Enhanced Flexibility and Security
The Kubernetes 1.35 release is scheduled for December 17, 2025, and will introduce a variety of experimental alpha features aimed at improving infrastructure flexibility and security. This release will include 15 new alpha features. Some of the notable additions include watch-based route controller reconciliation, Gang Scheduling for AI/ML workloads, and a new secrets field for passing Service Account tokens. Other experimental features focus on mutable volume attach limits and proxying API server requests to address version skew. The release cycle for version 1.35 began on September 15, 2025, with the code and test freeze occurring in early November.
Sources:
01 Quantum and Bullfrog Power Launch Quantum-Safe Environmental Tokens
01 Quantum Inc. has partnered with Bullfrog Power to integrate post-quantum cryptography (PQC) into Bullfrog’s Environmental Token, creating a secure digital record for businesses to verify their environmental claims long-term. This collaboration aims to protect environmental impact records from the future threat of quantum computers, ensuring the integrity and verifiability of sustainability claims for decades. Staples Canada is the first company to pilot the adoption of these quantum-safe tokens. The integration of NIST-endorsed PQC technology is intended to set a new standard for secure and transparent sustainability reporting.
Sources:
Quobly and STMicroelectronics Advance European Quantum Processor Supply Chain
Quobly, in collaboration with Soitec and STMicroelectronics, has achieved a significant milestone by processing the first 28Si FD-SOI wafers in ST’s 300 mm fab. This marks a critical step in establishing a European industrial supply chain for quantum processors. The use of 28Si-enriched FD-SOI substrates is designed to reduce isotopic noise, which is crucial for developing reliable and scalable silicon-based quantum processors. This advancement is a key part of Quobly’s roadmap towards million-qubit technology and accelerates the industrialization of quantum processor units. The first performance metrics from prototype devices are anticipated in the first quarter of 2026.
Sources:
Gartner Names Capgemini a Leader in 2025 Magic Quadrant for Custom Software Development
Capgemini has been named a Leader in the 2025 Gartner Magic Quadrant for Custom Software Development Services. This recognition highlights the company’s capabilities in designing, building, and modernizing custom software applications to meet unique client needs. Pierre-Yves Glever, a global head of applications at Capgemini, stated that the company is delighted with the acknowledgment, viewing it as a testament to their commitment to delivering business value through custom software. The company’s expertise spans AI-powered software engineering, composable and intelligent software, software modernization, and product engineering.
Sources:
Amazon CTO Werner Vogels Defines the ‘Renaissance Developer’ for the AI Era
In his final keynote at AWS re:Invent 2025, Amazon CTO Werner Vogels addressed the impact of AI on software development, introducing the concept of the ‘renaissance developer’. Vogels suggested that while AI will handle more coding tasks, it will not make developers obsolete if they evolve. He argued that the AI era requires developers to be modern polymaths who understand systems, business context, and the human element behind the technology. According to Vogels, the ‘renaissance developer’ is curious, a systems thinker, a skilled communicator, and takes ownership of the outcomes, not just the code. He emphasized that human judgment in the code review process is more critical than ever in an AI-augmented world.
Sources:
Introducing Whamm: A New Open-Source Framework for WebAssembly Instrumentation
A new open-source project called Whamm has been introduced for instrumenting WebAssembly (Wasm) applications. Developed by Elizabeth Gilbert, a doctoral candidate at Carnegie Mellon University, Whamm is a framework for Wasm application monitoring and manipulation. It allows users to instrument their Wasm applications using a programming language or code directly within the modules for tasks like debugging and monitoring. Whamm is designed to be a versatile tool for a range of use cases, from performance tuning to security audits, and supports dynamic analysis and bytecode rewriting. The project aims to provide a language-agnostic tooling approach, given that many languages can compile to WebAssembly.
Sources: