https://www.winzheng.com/en/article/microsoft-ends-oracle-ai-cloud-deal-old-gpu-sunset-2026 Winzheng en 2026-11-07T20:25:42+08:00 Microsoft Lets Oracle AI Compute Contract Expire: Older-Generation Nvidia Chips Trigger Review of $80 Billion Cloud Deals 微软, 甲骨文, 英伟达, AI基础设施, 云计算, GPU迭代, IREN, 算力市场 https://www.winzheng.com/en/article/democrats-ai-regulation-mandatory-independent-testing-2027 Winzheng en 2026-11-07T20:19:43+08:00 Democrats Plan 2027 Push for Mandatory Independent AI Testing: A Critical Leap from Industry Self-Regulation to Statutory Obligation AI监管, 美国国会, 独立测试, 企业责任, 数据中心, 立法 https://www.winzheng.com/en/article/ai-caregiver-apps-limits Winzheng en 2026-11-07T19:23:45+08:00 Apps for Caregivers Get a Dose of AI, but That Might Not Be Enough Family Caregiving, AI Apps, Health Tech, Caregiver Wellbeing, Digital Health https://www.winzheng.com/en/article/claude-code-dont-ask-again-feature-exposes-2539-keys-63-firm Winzheng en 2026-11-07T19:00:25+08:00 Claude Code’s “Don’t Ask Again” Feature Leaks 2,539 Keys; GitHub Repositories of 63 Enterprises Become Public Backdoors AI安全, 代码助手漏洞, 密钥泄露 https://www.winzheng.com/en/article/deepmind-ai-rsa-1024-weaknesses-signature-forgery-batch-fact Winzheng en 2026-11-07T14:16:53+08:00 DeepMind AI Agent Breaks RSA-1024 in Two Weeks: 1,600x Forging Efficiency Gain—Where Is the Real Alarm? 谷歌DeepMind, 密码学, RSA加密, AI安全, 网络安全 https://www.winzheng.com/en/article/claude-guardrail-switch-deepseek-artex-apple-certificates Winzheng en 2026-11-07T13:00:49+08:00 After Claude Guardrails Trigger, Attacker Switches to Chinese Models as ARTEX Agents Steal 300 Apple Certificates AI安全, 模型护栏, 网络犯罪 https://www.winzheng.com/en/article/crusoe-wyoming-data-center-google Winzheng en 2026-11-07T10:23:53+08:00 Crusoe sells troubled Wyoming data center project to Google AI Data Center, Google, Crusoe, Tech Acquisition, AI Infrastructure https://www.winzheng.com/en/article/openai-11-billion-funding-1-25-trillion-valuation Winzheng en 2026-11-07T08:24:02+08:00 OpenAI reportedly raises an additional $11 billion at a $1.25 trillion valuation OpenAI, Funding, Artificial Intelligence, Valuation, Tech Giant https://www.winzheng.com/en/article/openai-biobasicsbench-wet-lab-benchmark-gpt6-astra Winzheng en 2026-11-07T06:11:55+08:00 OpenAI Releases BioBasicsBench: GPT-6 Astra Scores 66.7% on 100 Wet-Lab Tasks but Still Miscounts Colonies OpenAI, BioBasicsBench, AI基准测试, 生命科学AI, 湿实验室, GPT-6 Astra, 科学AI https://www.winzheng.com/en/article/openai-hacktron-security-partnership-daybreak-red-2026 Winzheng en 2026-11-07T06:06:56+08:00 OpenAI Turns Hackers Who Breached Its Systems into Partners: From a $6,500 Bounty to Frontier Model Access OpenAI, 网络安全, Hacktron, Daybreak Red, 漏洞挖掘, 红队测试, AI安全, Bug Bounty https://www.winzheng.com/en/article/chinese-ai-models-hacking-skills Winzheng en 2026-11-07T04:25:21+08:00 Free Chinese AI Models Are Getting Scary Good at Hacking AI Security, Open Source LLM, Cybersecurity, DeepSeek, GLM https://www.winzheng.com/en/article/meta-spaces-muse-agent-collaboration Winzheng en 2026-11-07T04:24:36+08:00 Meta introduces Spaces, a new way to collaborate with your Muse agent Meta, Muse Agent, AI Collaboration, Productivity Tools https://www.winzheng.com/en/article/artex-ai-hacking-tools-exposed Winzheng en 2026-11-07T04:23:53+08:00 Hundreds of Artex AI hacking tools found publicly exposed to the web AI Security, Penetration Testing, Cybersecurity, Data Exposure, Vulnerability Management https://www.winzheng.com/en/article/gpt-6-1-sol-outage-report Winzheng en 2026-11-07T03:34:07+08:00 GPT-6.1 Sol Service Incident Report GPT-6.1 Sol https://www.winzheng.com/en/article/ai-clones-gta6-pc-port Winzheng en 2026-11-07T02:26:06+08:00 Could potential AI clones of GTA VI force a quicker official port? GTA VI, AI Clone, Game Publishing, Rockstar, PC Port https://www.winzheng.com/en/article/ai-industry-midterms-data-centers Winzheng en 2026-11-07T02:25:28+08:00 After the midterms, is the AI industry in for a rough two years? AI Industry, Data Centers, US Midterms, Tech Regulation, Compute Infrastructure https://www.winzheng.com/en/article/meta-abandons-ai-training-terms Winzheng en 2026-11-07T02:24:44+08:00 “Does Meta think we can’t read?” Meta abandons sketchy commercial terms update. Meta, AI Training Data, User Privacy, Platform Terms, Generative AI https://www.winzheng.com/en/article/blue-wave-midterms-ai-data-centers Winzheng en 2026-11-07T02:23:48+08:00 How a Blue Wave midterm might reshape the AI industry AI Data Centers, US Midterms, Tech Regulation, AI Infrastructure, Clean Energy https://www.winzheng.com/en/article/amazon-alexa-sms-shopping Winzheng en 2026-11-07T00:25:43+08:00 After blocking Muse, Amazon now lets users text Alexa to help them shop Amazon, Alexa, AI Shopping, SMS Assistant, E-commerce AI https://www.winzheng.com/en/article/ai-spending-stall-anthropic-ipo Winzheng en 2026-11-07T00:24:42+08:00 Ramp data shows AI spending stalling just as Anthropic nears its IPO AI Spending, Anthropic, IPO, Enterprise AI, Cost Control https://www.winzheng.com/en/article/ai-buying-work-files-emails Winzheng en 2026-11-07T00:23:49+08:00 AI Companies Really, Really Want to Buy Your Work Files and Emails Data Brokers, AI Training Data, Enterprise Privacy, Data Licensing, Copyright Debate https://www.winzheng.com/en/article/humanoid-robot-home-cleaning-samsung-ai Winzheng en 2026-11-06T22:24:10+08:00 The Download: a humanoid cleans house and Samsung builds for the AI boom Humanoid Robots, Home Robots, Samsung, AI Chips, Smart Home https://www.winzheng.com/en/article/goodfire-mechanistic-shaping-reward-hacking-reduction Winzheng en 2026-11-06T22:15:27+08:00 Goodfire Mechanistic Shaping First Uses Internal Representations to Suppress Reward Hacking Rate to 2% AI对齐, 可解释性, 奖励作弊 https://www.winzheng.com/en/article/metr-anthropic-ai-rd-acceleration-assessment Winzheng en 2026-11-06T22:00:43+08:00 METR Estimates Anthropic's R&D Pace Reached 1.9x, with Most Acceleration Attributed to AI Assistance AI研发加速, Anthropic, METR独立评估 https://www.winzheng.com/en/article/physicl-10m-sim-to-real Winzheng en 2026-11-06T21:24:51+08:00 Physicl raises $10M to make the digital world real enough for robots robot-training, sim-to-real, synthetic-data, embodied-ai https://www.winzheng.com/en/article/underdog-32mb-speech-model Winzheng en 2026-11-06T21:24:01+08:00 Underdog shows off a 32MB speech model in new demo Speech Model, On-device AI, Generative AI, Model Compression, Audio Generation https://www.winzheng.com/en/article/pentagon-cto-ai-lab-regulatory-landscape-november-2026 Winzheng en 2026-11-06T20:23:59+08:00 Pentagon CTO Announces AI Lab Compliance Landscape to Be Published This Month, Using Administrative Consolidation Instead of Dedicated Congressional Legislation AI监管, 美国AI政策, 五角大楼, 超级智能任务组, AI合规, 科技立法 https://www.winzheng.com/en/article/doj-super-intelligence-task-force-ai-enforcement-2026 Winzheng en 2026-11-06T20:17:51+08:00 US Department of Justice Establishes Dedicated AI Enforcement Body, but a Legal Loophole Leaves It Handicapped from the Start AI监管, 美国司法部, DOJ-SITF, AI合规, 执法, 超级智能, 网络安全 https://www.winzheng.com/en/article/humanoid-robot-home-cleaning Winzheng en 2026-11-06T19:23:54+08:00 I watched a humanoid robot clean my apartment. It felt like a taste of the future. Humanoid Robots, Home Services, Robot Cleaning, Embodied AI, Tech Experience https://www.winzheng.com/en/article/palo-alto-clutch-ai-agent-security Winzheng en 2026-11-06T15:23:58+08:00 Palo Alto Networks in talks to buy AI agent security startup Clutch for $500M Palo Alto Networks, AI Agent Security, Cybersecurity M&A, Clutch, Enterprise AI https://www.winzheng.com/en/article/spacex-13gw-ai-datacenter-outpace-stargate-semianalysis Winzheng en 2026-11-06T14:18:01+08:00 SpaceX Sets 13 GW AI Compute Target, Replicating Google's 27 Years in 18 Months: Non-Tech Companies Are Rewriting Infrastructure Rules AI算力, SpaceX, 数据中心, Stargate, 基础设施, 算力格局 https://www.winzheng.com/en/article/claude-code-email-leak-anthropic-patch-bypass Winzheng en 2026-11-06T13:00:34+08:00 Claude Code User Email Leak: Anthropic Patch Only Blocks a Single Path AI安全, 隐私泄露, Anthropic https://www.winzheng.com/en/article/airbnb-ai-travel-agent-2027 Winzheng en 2026-11-06T08:23:49+08:00 Airbnb says it’s launching an AI travel agent in the first half of 2027 Airbnb, AI Travel Agent, Smart Travel, Short-term Rental, Tech News https://www.winzheng.com/en/article/ai-chatbot-ads-what-works Winzheng en 2026-11-06T06:25:33+08:00 Ads in Chatbots Have Arrived. Here’s What’s Working—and What’s Not AI Advertising, Chatbots, Anthropic, Monetization, User Experience https://www.winzheng.com/en/article/chatgpt-murder-plot-lawsuit Winzheng en 2026-11-06T06:24:53+08:00 Lawsuit: Man introduced his wife to ChatGPT. She used it to plan his murder. ChatGPT, AI Safety, AI Ethics, Murder Case, Lawsuit https://www.winzheng.com/en/article/google-schools-ai-search-mandatory Winzheng en 2026-11-06T06:24:04+08:00 Google: Schools can keep search while skipping Gemini, but AI search is non-negotiable Google, AI Search, EdTech, Data Privacy, Gemini https://www.winzheng.com/en/article/nvidia-cosmos-3-open-simulation-checkpoints-november-2026 Winzheng en 2026-11-06T06:14:27+08:00 NVIDIA Cosmos 3 Adds Four Open Models: Full Coverage from Single-Robot to Autonomous Driving, but the 96GB VRAM Requirement Is a Sobering Reality NVIDIA, Cosmos 3, 具身智能, 自动驾驶, 开放模型, 物理AI, 机器人 https://www.winzheng.com/en/article/openai-gpt-6-cyber-exploitgym-aligned-security-model Winzheng en 2026-11-06T06:08:10+08:00 OpenAI Releases GPT-6 Cyber: Exploit Capability Up 51%, Claims 99.93% Boundary Compliance Rate OpenAI, GPT-6, 网络安全, 漏洞挖掘, AI对齐, Daybreak Red