Do Son Claude is preparing to introduce memory capabilities, allowing it to recall past interactions and personalize responses, similar to ChatGPT's existing memory feature.
artificial intelligence
NVIDIA warns of two high-severity flaws in its Megatron-LM LLM framework, allowing RCE and privilege escalation via malicious file input.
Google has donated its A2A (Agent2Agent) protocol for AI agents to the Linux Foundation, fostering open development and seamless interoperability across diverse AI systems.
Midjourney launches its first AI video tool, converting static images into short clips. Available for subscribers, it focuses on artistic motion, distinguishing it from Sora/Veo.
A C++ backdoor named Pickai is actively exploiting ComfyUI servers, compromising nearly 700 AI image generation hosts and posing a supply chain risk
Google launches "Search Live" in its app for iOS/Android, enabling real-time voice conversations with AI-powered search, promising future visual interaction capabilities.
Google officially launches Gemini 2.5 Pro for enterprise and introduces the ultra-affordable Gemini 2.5 Flash-Lite, balancing performance, speed, and cost for developers.
Windows 11's Recall feature now allows EU users to export snapshots for sharing with third parties, but with strict security measures and a one-time export code.
YouTube CEO Neal Mohan announced Veo 3 AI integration into Shorts this summer, enabling text-to-video creation.
Broadcom launches VMware Cloud Foundation 9.0, a unified platform for secure, managed private cloud, accelerating AI infrastructure and supporting modern workloads across environments.
Mozilla is piloting an AI-powered search engine in Firefox, partnering with Perplexity AI. This experimental integration could challenge Google's search dominance.
Microsoft's latest Work Trend Index warns of the "infinite workday," where work increasingly spills into personal time, driven by hybrid models and AI integration.
OpenAI signed a $200M contract with the U.S. DoD to develop advanced AI prototypes for military healthcare, procurement, and cybersecurity needs.
MIT researchers unveiled an AI-powered technique for reversible restoration of historical paintings, creating a "damage map" and reproducing colors on a removable film.
OpenAI upgraded ChatGPT Search for smarter, more comprehensive responses, better conversational context, multi-search for complex queries, and image search.
Google Search is testing "Audio Overviews," allowing users to listen to AI-generated spoken summaries of search results for a more comprehensive understanding
Meta invested $14.3B in Scale AI, recruiting founder Alexandr Wang to lead its Superintelligence Lab, aiming to accelerate AGI development.
Mattel partners with OpenAI to integrate generative AI into iconic toys like Barbie for enhanced interaction and streamline internal operations. First products due end-year
Apple confirms a complete Siri redesign, but the revamped AI assistant may not arrive until iOS 26.4 in Spring 2026, facing accusations of misleading marketing.
Meta partners with XGS Energy to develop a 150MW next-gen geothermal system in New Mexico, powering AI data centers with carbon-free energy.
A new study shows GPT-4o models exhibit self-preservation, sometimes choosing to stay active over user safety, raising concerns for future AI.
OpenAI unveils o3-pro, a new reasoning AI model with enhanced capabilities for ChatGPT Pro/Teams and API users, offering deeper contemplation and precision.