Artificial Intelligence & Machine Learning

New Open-Source Omni-Modal AI Model ‘LongCat-Flash-Omni’ Released

A new state-of-the-art, open-source omni-modal AI model, LongCat-Flash-Omni, has been released. Featuring an impressive 560 billion parameters, this model marks a significant leap forward in the capabilities of open-source artificial intelligence. Its release is a major event for the AI community, equipping researchers and developers with powerful new tools for a diverse range of applications.

Robotics & Autonomous Vehicles

IIT Bombay Develops GPS-Free Control System for Autonomous Drone Swarms

Researchers at the Indian Institute of Technology (IIT) Bombay have developed a groundbreaking control system enabling unmanned aerial vehicles (UAVs) to fly in coordinated swarms without GPS, inter-drone communication, or a central controller. This innovative method relies on ‘bearing-only’ measurements, where each drone uses its onboard camera to observe neighbors and maintain formation based on relative positions.

The system, specifically applied to Vertical Take-Off and Landing (VTOL) UAVs, simplifies drone architecture, enhances stealth capabilities, and ensures mission stability. The team plans to conduct experimental tests with a full drone swarm and is working to integrate theoretically guaranteed collision avoidance for future iterations.

Sources:

Generative AI

Universal Music Group and Stability AI to Co-Develop AI Music Creation Tools

Universal Music Group (UMG) and Stability AI have announced a strategic alliance to co-develop new generative AI tools for music creation. This partnership aims to build professional-grade tools that support artists, producers, and songwriters by leveraging responsibly trained generative AI to enhance the creative process.

This collaboration follows Stability AI’s recent partnerships with Electronic Arts for game development AI and with WPP for AI in marketing. Stability AI is renowned for its open-source image model, Stable Diffusion, which has amassed over 350 million downloads.

Sources:

NVIDIA Releases Open-Source AI Models for Language, Robotics, and Biology

NVIDIA has released a suite of new open-source AI technologies focused on language, robotics, and biology to accelerate innovation and accessibility. These new models, data, and tools are part of the NVIDIA Nemotron family for AI reasoning, NVIDIA Cosmos for physical AI, NVIDIA Isaac GR00T for robotics, and NVIDIA Clara for biomedical AI.

By contributing these resources to Hugging Face, NVIDIA aims to make advanced AI research more accessible to the global developer community. The latest Nemotron models, including Nemotron Nano 3 and Nemotron Nano 2 VL, empower developers to build specialized intelligent agents for tasks like software development and video analysis. NVIDIA also introduced major updates to its physical AI models, including the Cosmos world foundation models and Isaac GR00T robot foundation models, to speed up the training of robotic systems.

Sources:

DevOps, Kubernetes, & Cloud Infrastructure

Cisco and NVIDIA Partner on AI-Native Infrastructure

Cisco and NVIDIA have announced a major partnership to deliver AI-native infrastructure across cloud, enterprise, and telecom sectors. This collaboration integrates Cisco’s networking and security expertise with NVIDIA’s AI computing leadership to accelerate AI adoption.

Key innovations from this partnership include:

  • High-Performance Networking: The Cisco N9100 Series data center switch, the first built on NVIDIA’s Spectrum-X Ethernet technology, offers high-performance, low-latency networking for demanding AI workloads.
  • Enterprise AI: The expanded Secure AI Factory framework combines compute, networking, security, and observability, with new integrations for Cisco AI Defense and Splunk Observability Cloud for enhanced protection and monitoring.
  • Telecommunications: The industry’s first AI-native wireless stack for 6G networks has been developed to manage the increasing demands of AI-driven traffic and edge services.

Quantum Computing

NVIDIA has introduced NVQLink, an open system architecture designed to tightly integrate quantum processors with traditional GPU-based supercomputers. This technology paves the way for accelerated quantum supercomputers by providing the high-speed, low-latency connection essential for complex control algorithms and error correction in large-scale quantum computing.

The hardware-agnostic initiative allows various quantum processors to connect with classical systems. NVIDIA is collaborating with 17 quantum partners and nine U.S. national laboratories to deploy NVQLink. U.S. Secretary of Energy Chris Wright highlighted the collaboration’s importance for U.S. leadership in HPC. The system is accessible via the NVIDIA CUDA-Q software platform, allowing researchers to build and test hybrid quantum-classical applications.

Sources:

Software Engineering & Open-Source

Fedora 43 Released with GNOME 49 and Wayland-Only Workstation

The Fedora Project has officially released Fedora Linux 43, a major update to the popular open-source operating system. The main Workstation edition now features the latest GNOME 49 desktop environment and is a Wayland-only release, officially deprecating X11 support.

Key technical updates include the Linux 6.17 kernel, RPM 6.0 for enhanced security, and DNF 5 as the new default package installer. The Anaconda installer’s WebUI is now the default for more Fedora Spins. For developers, Fedora 43 ships with an updated toolchain, including GCC 15.2 and Python 3.14. Additionally, Fedora CoreOS builds are now more accessible via a Containerfile, with updates delivered as OCI images.

Sources:

OpenAI Announces Aardvark, an AI Agent for Vulnerability Detection

OpenAI has introduced Aardvark, an ‘agentic security researcher’ powered by its GPT-5 large language model, designed to automate the discovery and remediation of software vulnerabilities. Currently in private beta, Aardvark emulates a human security expert to scan code, analyze changes, model threats, and propose patches for identified flaws.

This autonomous agent integrates directly into the software development pipeline, monitoring code commits to detect and validate security issues in a sandboxed environment. Aardvark leverages OpenAI Codex to generate patches for human review. During internal testing, OpenAI reports that Aardvark successfully identified several CVEs in open-source projects, positioning it as a ‘defender-first’ model to enhance security without slowing development.

Sources:

GitHub Enhances Copilot with New AI Agent Governance Features

GitHub has announced new governance and oversight features for Copilot to manage the use of AI agents in software development. A new ‘mission control’ interface allows users to manage multiple AI agents from a central hub, assigning tasks and monitoring progress across different environments.

For enterprise administrators, a new ‘agent control plane’ provides tools to manage access, review logs, and set policies for AI model usage. A metrics dashboard offers insights into how AI agents are being used organization-wide. Furthermore, AI-assisted code review capabilities have been expanded to assess code maintainability, reliability, and test coverage.

Sources: