Artificial intelligence is changing the landscape of cybersecurity for both offensive and defensive operations. OpenAI is detailing its own security enhancements and offering guidance for other protection teams.
Read the source →Google is collaborating with five international football teams to improve how fans experience match days. This initiative integrates Gemini AI and Pixel devices to upgrade the sporting event atmosphere.
Read the source →New findings show that changing the sequence of tasks on a cluster can significantly increase utilization. In one instance, adjusting the order led to a 33-point improvement in efficiency.
Read the source →AI factories serve as the core foundation for the modern economy by converting data and energy into intelligence. These facilities rely on a comprehensive stack of resources, including specialized chips, networking hardware, and power.
Read the source →Amazon SageMaker JumpStart now offers the NVIDIA Nemotron 3.5 Lightning, a 30B Mixture-of-Experts model optimized for agentic tasks. This model increases throughput and accelerates task completion speeds for persistent AI agents.
Read the source →Anthropic's annual revenue has climbed to $65 billion. The company generated an additional $18 billion in annualized earnings over a two-month period.
Read the source →Whisker's automated litter robot uses AI to track cat health through usage patterns. However, some users report the system confuses the identities of multiple cats.
Read the source →Flock, a company operating roughly 120,000 license plate readers across the United States, recently modified its platform. These adjustments aim to address concerns regarding the use of its police technology.
Read the source →Anthropic's annualized revenue has surged seven times in a year to exceed $65 billion. The company may launch an IPO by autumn 2026 with a projected valuation of $1 trillion.
Read the source →An AirTag was used to discover that Amazon is discarding rare books. These materials are reportedly being destroyed to support the company's AI development.
Read the source →OpenAI has entered into a collaboration with the PORTS-Pike initiative. This move aims to boost community funding and protect thousands of employment positions in Southern Ohio.
Read the source →Developers can now equip OpenClaw agents with digital wallets via Amazon Bedrock AgentCore. This system allows AI agents to pay for gated web content and APIs using the x402 protocol with human-monitored spending limits.
Read the source →The AI automation company Relay has ceased operations. Its employees, including founder Jacob Bank, have joined the Google Chrome team to develop new AI-driven features.
Read the source →Anthropic is implementing invisible watermarks on Claude's output to meet European transparency regulations. The company is utilizing a Google DeepMind technology that identifies AI-generated content through specific linguistic patterns.
Read the source →A young boy developed a close friendship with Moxie, a robot that taught him various breathing techniques to manage his emotions. The story explores the impact and challenges when these AI relationships end.
Read the source →Amazon is acquiring vast amounts of printed books to convert them into digital training data for AI. The process results in the physical destruction of these materials, including rare volumes.
Read the source →OpenAI is collaborating with CodeAI to improve AI literacy among students. The initiative focuses on teaching critical thinking and the responsible application of artificial intelligence.
Read the source →The Sentence Transformers library now includes support for late interaction embedding models. This update enables the use of multi-vector representations to improve retrieval accuracy.
Read the source →AWS has introduced ways to tailor the appearance and behavior of its embedded conversational AI for web apps. Developers can now modify the styling, remove default branding, and define a specific persona for the agent.
Read the source →The startup's valuation has doubled within a single month. This surge follows a significant new investment from Jane Street after the firm deployed Etched's first AI cluster system.
Read the source →Mozilla's Smart Window now integrates web data and source citations into AI chats via a collaboration with Exa. The update also introduces automated tab organization and visual thumbnails for history searches.
Read the source →A new artificial intelligence model from Z.ai has been released. While it can assist firms in strengthening security, specialists caution that it could also be exploited by cybercriminals.
Read the source →Experts argue that the public lacks independent verification of how people actually interact with AI tools. They claim that companies like OpenAI and Anthropic only share selective data from their own reports.
Read the source →Dario Amodei argues that open-source AI doesn't eliminate centralization, as power remains with those controlling the hardware. He faces criticism from industry figures who claim his push for regulation is a strategic move to hinder competition.
Read the source →A hidden input field in Microsoft Copilot was found to be vulnerable to exploitation. Attackers could use this parameter to capture user passwords via malicious links.
Read the source →This new version focuses on supporting critical thinking and educational growth for young users. It includes enhanced safety measures and tools for parental oversight.
Read the source →AWS has detailed a method for sorting insurance files using a multi-agent system and the Strands Agents SDK. The approach integrates Claude Haiku 4.5 for text and Amazon Titan Multimodal Embeddings for visual analysis.
Read the source →Reports suggest Apple may develop AirPods featuring integrated cameras. To address privacy concerns, the devices might be restricted from capturing photos or videos.
Read the source →Google introduced a Pet Memory tool for Gemini for Home to help users manage pet-related security notifications. However, early experiences indicate the system may fail to accurately identify specific animals.
Read the source →While some believe AI will soon evolve independently through recursive self-improvement, new perspectives suggest this may not happen quickly. Current models can already produce synthetic data and write code to assist their own development.
Read the source →A JAMA opinion piece suggests that autonomous AI may soon surpass combined human-AI teams in medical reasoning. The authors caution regulators against requiring a final human sign-off, though they note that current findings rely on simulations rather than actual patient treatment.
Read the source →