Track model releases, AI agents, open-source projects, infrastructure, policy, and product updates as they publish.
Zhipu AI CEO Tang Jie publicly solicited feature suggestions for the next-gen model GLM-5.3, with users nearly unanimously asking for vision abilities.
NVIDIA's GB300 chip is now powering Anthropic's Claude model on Microsoft Foundry, marking a deeper infrastructure-level partnership between the three AI giants.
Google has launched AI-driven online shopping features in its Gemini assistant, enabling natural language product search, comparison, and purchase assistance.
Coocaa Technology has officially released Happy Work AIOS Lite MVP, an enterprise AI operating system featuring a four-layer agent architecture designed to restructure corporate management processes.
The benchmark is the first industry-standard evaluation for AI infrastructure operations, covering 5 mainstream domestic chip platforms.
Amazon is reportedly evaluating alternatives to Anthropic's models — including OpenAI and its own Nova models — in response to rising costs from Anthropic.
In a 500-day business simulation, most AI models suffered heavy losses, while Claude Fable 5 excelled and stood out.
Due to computing power shortages, Google has restricted Meta's access to its Gemini AI model, highlighting the bottleneck of AI computing resources.
Shulan Medical has initiated an in-depth AI healthcare experiment, aiming to integrate artificial intelligence into clinical diagnosis, hospital management, and patient services.
Google UK has published its latest Economic Impact Report, urging the UK to unlock the benefits of AI-powered technologies and build a nation of AI trailblazers.
Minglue Technology has open-sourced Octo, a project designed to create an interconnected network infrastructure for AI agents.
HSBC announced a partnership with Google Cloud to leverage AI tools for enhancing banking operations and intelligence.
Google has rolled out free personalized image creation for Gemini users in the United States, allowing image generation at no cost.
DeepSeek has announced the official launch of its V4 model, along with a price increase for API calls during peak hours.
DeepSeek V4 official version confirms a significant price increase (doubling), while also delivering capability upgrades. Reported by QbitAI.
A new report reveals that 'high-intensity AI adopters' saw 10.2% headcount increase, with entry-level roles growing 12%, challenging the notion that AI destroys junior jobs.
AIPOCH has launched MedSkillAudit, an audit framework designed to evaluate the skills of medical AI agents prior to deployment.
Wix-owned vibe coding platform Base44 has started rolling out its own AI model, with hopes it will eventually outperform frontier models.
Google has limited Meta's usage of its Gemini AI models due to surging compute demand, a move reported by Financial Times and confirmed by CNBC, Forbes, and Bloomberg.
Indonesian outlet VOI.id reports that GCL ET is exploring opportunities to open an AI data center in Indonesia, signaling growing AI infrastructure investment in Southeast Asia.
Google has announced that personalized AI image generation is now available to all Gemini free-tier users across the United States, allowing more users to create custom images without a paid subscription.
AI coding assistant Cursor has released its mobile application, allowing developers to perform AI-assisted programming directly from their phones without needing a multi-screen desktop setup.
OpenAI engineers used large-scale core dump epidemiology to debug rare infrastructure crashes, uncovering both a hardware fault and a long-standing software bug that had persisted for 18 years.
OpenAI introduced GeneBench-Pro, a new benchmark for evaluating AI performance in genomics, biology, and scientific research using complex, real-world datasets.
Hugging Face has integrated "Every Eval Ever" community evaluation results directly onto model pages, allowing users to view benchmark performance data without navigating away.
DeepSeek V4's official version has doubled prices during peak hours, while ByteDance's Doubao launched a new 'Doubao Navigation' feature.
A new pre-IPO trading product has drawn market attention to OpenAI, signaling growing anticipation of a potential public listing.
OpenClaw released native iOS and Android mobile apps, allowing users to pair with their OpenClaw gateway for real-time voice conversations, agent action approvals, and device permission controls.
OpenAI has appointed John Buckley to oversee child safety and youth wellbeing, strengthening its AI safety governance team.
PTOP has announced the launch of the Synaptic Quant AI trading platform, expanding its presence in AI-driven financial trading.
Gartner warns that AI coding costs can drain a budget in days and predicts they will soon match developer compensation levels.
OpenClaw, an open-source AI agent framework, has released a native iOS application, bringing autonomous AI agent capabilities to mobile devices.
Google has opened its most advanced image analysis feature to all users, dropping the Gemini Advanced subscription requirement.
Cryptocurrency exchange Coinbase is adopting Chinese AI models to improve cost efficiency.
OpenAI and chip design giant Broadcom have jointly launched a custom AI chip that aims to reduce AI processing costs by approximately 50%.
Anthropic has completed its first deployment of NVIDIA infrastructure on Microsoft Azure, marking a significant expansion of its computing architecture.
DeepSeek has released DSpark, a new technology that claims to accelerate AI processing by up to 85% without requiring advanced hardware, with the main model verifying answers in groups.
Google is rolling out Gemini's personalized AI image generation to eligible free users in the U.S., enabling the chatbot to create custom images based on user interests and data from connected Google apps.
In the lead-up to the Brazil vs Japan match, 12 AI models all predicted Brazil to win. The unanimous forecast has ignited fan discussions about a classic plot from the 30-year-old anime Captain Tsubasa.
The Washington Post published a systematic investigation testing whether mainstream AI chatbots like ChatGPT exhibit political bias in their responses.
Daily Archive