Open-Source

Uber and Rivian Announce Major Partnership for Autonomous Robotaxi Fleet Uber Technologies and Rivian Automotive have forged a strategic partnership to deploy up to 50,000 fully autonomous R2 robotaxis. As part of this landmark deal, Uber is set to invest up to $1.25 billion in the electric vehicle manufacturer, with the investment contingent on Rivian meeting specific autonomous performance milestones. An initial commitment of $300 million has been made, pending regulatory approval. The first deployment phase will introduce 10,000 vehicles, with initial rollouts planned for San Francisco and Miami in 2028, expanding to 25 cities by 2031. These fully autonomous Rivian R2 robotaxis will be exclusively bookable through the Uber platform. ...

March 2026 Tech News: AI Agents, Kubernetes 1.36, and Quantum Leaps AI Industry Shifts Focus to Autonomous Agents and Advanced Hardware The artificial intelligence sector is undergoing a significant transition, moving from generative assistance models to autonomous agentic systems capable of executing complex, multi-step workflows. This evolution towards more sophisticated ‘digital coworkers’ is exemplified by the development of models like GPT-5.4. Concurrently, the industry is preparing for a major leap in hardware with Nvidia expected to unveil its new ‘Rubin’ platform, which promises a generational advancement in computing power for AI. These developments are occurring as the financial and physical sustainability of the current AI infrastructure boom faces increased scrutiny, sparking debates on the required capital investment and growing local opposition to new data centers due to environmental concerns. ...

Meta Unveils Four New Generations of Custom AI Chips to Reduce Nvidia Reliance Meta has announced a significant expansion of its custom silicon efforts with four new generations of its in-house Meta Training and Inference Accelerator (MTIA) chips. The new chips—MTIA 300, 400, 450, and 500—are engineered to manage the massive AI workloads for Meta’s services, including ranking, recommendation systems, and generative AI. This strategic move aims to reduce the company’s reliance on external chip suppliers like Nvidia while improving performance and efficiency for its specific needs. The MTIA 300 is already in production, with subsequent generations planned for a rapid six-month release cadence. The more advanced MTIA 450 and 500 models are specifically optimized for generative AI inference and are slated for mass deployment in 2027. ...

Agentic AI Integration Accelerates in Desktop and Creative Software The trend of agentic AI is accelerating as major tech companies embed these autonomous AI agents into everyday software. Anthropic has launched ‘Claude in Chrome,’ allowing the AI to connect to a user’s session to perform actions across the web. The company has also introduced Claude integrations for Excel and PowerPoint. Similarly, OpenAI has released a standalone Codex application for Mac and ‘ChatGPT for Excel,’ both powered by its latest models. These desktop-native AI tools are experiencing rapid growth, with OpenAI’s Codex reaching two million weekly active users in early March. This shift signifies a move for AI from a content generator to a powerful workflow automator capable of executing complex, multi-step tasks. ...

Samsung Enters AI Smart Glasses Market with Project HAEAN Samsung has officially announced its entry into the AI smart glasses market with ‘Project HAEAN’, revealed at MWC 2026. The device will feature an eye-level camera and connect to a smartphone for processing, a design choice aimed at keeping the glasses lightweight. Positioned as a direct competitor to Meta’s Ray-Ban glasses, which currently dominate the market, Samsung’s offering is slated for a 2026 launch. The smart glasses are being developed in partnership with Qualcomm and Google, leveraging their respective expertise in chips and AI. This move from Samsung is expected to significantly intensify competition in the nascent AI smart glasses sector. ...

Artificial Intelligence & Machine Learning AI Industry Sees Major Developments in Early March 2026 The artificial intelligence sector has experienced a flurry of significant activity in early March 2026, marked by major infrastructure adjustments and new AI model advancements. In a notable development, Oracle and OpenAI have reportedly canceled plans to expand a major AI data center in Texas. This facility was a key part of the ‘Stargate’ infrastructure project, but disagreements over financing and evolving demands from OpenAI led to its cancellation. Concurrently, the AI model landscape continues its rapid evolution with OpenAI’s launch of GPT-5.4, a new model boasting enhanced reasoning and coding capabilities. Adding to the industry’s momentum, the Allen Institute for AI (AI2) has released Olmo Hybrid, a new open-source 7B-parameter model demonstrating improved data efficiency. These events highlight the dynamic nature of the AI industry, with simultaneous shifts in both large-scale infrastructure and the continuous innovation of AI models. ...

Northwestern’s Evolving Robots Showcase ‘Athletic Intelligence’ and Self-Repair Researchers at Northwestern Engineering have pioneered a new class of modular robots, dubbed “legged metamachines,” that exhibit athletic intelligence and can dynamically adapt to their environment. These robots are built from autonomous, LEGO-like modules, allowing for assembly into numerous configurations. A standout feature is their remarkable ability to recover from damage; for example, if a leg breaks, the remaining parts can continue to function and reconfigure for mobility. The designs were optimized using an AI-driven evolutionary algorithm that simulated natural selection to identify the most efficient and versatile body types for movement. According to the researchers, these are the first robots to be evolved within a computer and then successfully deployed in a real-world outdoor environment. ...

IIT Kharagpur Launches New School for Applied AI to Boost Industry Innovation The Indian Institute of Technology (IIT) Kharagpur has announced the launch of its new School of Digital Engineering, Applied Artificial Intelligence, and Machine Learning. This new school is designed to be a premier hub for research and innovation, with a strong focus on industry integration to promote the real-world application of AI technologies. According to the institution’s director, Suman Chakraborty, this initiative represents a significant shift in institutionalizing applied AI. The ultimate goal is to position IIT Kharagpur as a leading global center for developing responsible and sovereign AI for key sectors like healthcare and manufacturing. ...

AI News: Anthropic Alleges Data Theft & Samsung Unveils Gemini-Powered S26 The AI sector is buzzing with significant developments, led by a serious accusation from Anthropic. The company has identified and taken action against coordinated campaigns by three Chinese AI labs to improperly access and learn from its Claude AI. The labs allegedly involved are: DeepSeek Moonshot AI MiniMax These labs reportedly created over 24,000 fraudulent accounts to generate more than 16 million interactions with Claude. The efforts were aimed at ‘model distillation,’ a technique used to transfer the capabilities of a more advanced model to a new one. Anthropic noted the campaigns specifically targeted Claude’s key strengths in agentic reasoning, coding, and tool-use. ...

Critical Security Flaws in Anthropic’s Claude Code Expose Users to Remote Code Execution Cybersecurity researchers have discovered critical security vulnerabilities in Anthropic’s AI-powered coding assistant, Claude Code. The identified flaws could potentially allow for remote code execution (RCE) and the theft of sensitive API credentials. These vulnerabilities reside in configuration mechanisms like Hooks and Model Context Protocol (MCP) servers. If exploited, an attacker could execute arbitrary shell commands and exfiltrate Anthropic API keys when a developer clones and opens an untrusted repository. One significant issue is a code injection vulnerability that bypasses user consent when initializing Claude Code in a new project directory. ...