SageMaker AI Drops 13 Inference Features in 2026
September 18, 2026: SageMaker AI Drops 13 Inference Features in 2026 - Amazon SageMaker AI has launched 13 new inference capabilities in 2026, targeting the unique challenges of generative AI deployment including massive model sizes, latency demands, cold starts, and GPU constraints. The updates span two deployment paths: fully managed endpoints and Kubernetes-native H...
Researchers Used Claude to Hack OpenAI
September 18, 2026: Researchers Used Claude to Hack OpenAI - A three-person team at startup Hacktron AI used Anthropic's Claude Opus 5 to breach OpenAI's systems, chaining two vulnerabilities to access employee ChatGPT and Codex accounts. The attack exploited an unpatched memory bug in image-processing library libheif via OpenAI's community forum. OpenAI awar...
Claude AI Used to Breach OpenAI's Secret GitHub Repo
September 18, 2026: Claude AI Used to Breach OpenAI's Secret GitHub Repo - Cybersecurity researchers at startup Hacktron AI used Anthropic's Claude to breach OpenAI's GitHub repository, which reportedly contains the company's algorithmic secrets. The files remained accessible until June 24, when researchers reported their findings. OpenAI patched the issue within 14 hours....
Anthropic Runs Secret Biology Lab Amid Safety Warnings
September 18, 2026: Anthropic Runs Secret Biology Lab Amid Safety Warnings - Anthropic has confirmed it operates a wet biology lab in the Bay Area, using its AI models to conduct physical experiments. The lab focuses on fundamental biology rather than drug discovery, as the company avoids competing with pharma partners like Novo Nordisk. Anthropic also launched a Life Scienc...
Anthropic Launches AI-Powered Wet Lab in Bay Area
September 18, 2026: Anthropic Launches AI-Powered Wet Lab in Bay Area - Anthropic has opened a biology research wet lab in the San Francisco Bay Area, using robots powered by its Claude AI models to automate scientific tasks. The lab is not solely focused on drug discovery, suggesting broader biological research goals. Claude's recent benchmark showed a 50% hit rate in...
AI Hallucination Nearly Sparked US-China Military Conflict
September 18, 2026: AI Hallucination Nearly Sparked US-China Military Conflict - A US military operation against a Chinese vessel was aborted mid-flight this spring after officials discovered the intelligence behind it had been fabricated by an AI chatbot. An analyst used the tool to synthesize classified and open-source data, but the chatbot misidentified the ship's cargo, fals...
Hackers Used Claude to Breach OpenAI Accounts
September 18, 2026: Hackers Used Claude to Breach OpenAI Accounts - A trio of independent security researchers at Hacktron breached OpenAI employee accounts in under 72 hours using Anthropic's Claude Opus 4.8 and 5, according to The Wall Street Journal. The hack granted access to OpenAI's internal GitHub repository, known as Monorepo, which reportedly contains the c...
Ex-OpenAI researcher's non-LLM model stuns developers
September 18, 2026: Ex-OpenAI researcher's non-LLM model stuns developers - TypeSafe AI, founded by ex-OpenAI researcher Diogo Almeida, has released Jev, a transformer-based model that outputs probabilities instead of text. It cannot hallucinate, runs faster than traditional LLMs, and charges per billion input tokens rather than per million, making it dramatically cheaper f...
Vantora Raises $100M to Build Physical AI Startups
September 18, 2026: Vantora Raises $100M to Build Physical AI Startups - Startup studio UP.Labs has rebranded as Vantora and raised $100 million from Silversmith Capital Partners. The firm builds startups for corporate partners like Alaska Airlines, Porsche, and J.B. Hunt, and is now focused on physical AI solutions exclusively for those partners rather than the broader ...
Migrate AI Agents to Amazon Bedrock AgentCore Runtime
September 18, 2026: Migrate AI Agents to Amazon Bedrock AgentCore Runtime - Organizations building multi-model AI agents often struggle with infrastructure complexity, spending more time managing containers, scaling, and observability than developing agent logic. Amazon Bedrock AgentCore runtime offers a managed solution that handles these operational concerns automatically...
Newsom Orders AI Kill Switch for California
September 18, 2026: Newsom Orders AI Kill Switch for California - California Governor Gavin Newsom has signed an executive order aimed at establishing state leadership on AI oversight, including the potential mandate of a kill switch for frontier AI models. The order directs the state to assemble a panel of experts who will deliver recommendations within two month...
Stanford's Virtual Biotech Runs on 37,000 AI Agents
September 18, 2026: Stanford's Virtual Biotech Runs on 37,000 AI Agents - A Stanford Medicine lab has spun out a biotech startup unlike any other: 37,000 employees, zero humans. The AI-powered virtual company requires no office, no payroll, and no physical lab. Created by an associate professor of biomedical data science, the startup relies entirely on AI agents to opera...
Crusoe Raises $3.9B, Valued at $30.9B
September 18, 2026: Crusoe Raises $3.9B, Valued at $30.9B - AI data center builder Crusoe Inc. has closed a $3.9 billion Series F funding round at a $30.9 billion valuation. The round was led by Atreides Management, Mubadala Capital and Valor Equity Partners, with participation from Nvidia, Salesforce Ventures and over two dozen others. Crusoe operates majo...
Kimi K3 Now Available on Amazon Bedrock
September 18, 2026: Kimi K3 Now Available on Amazon Bedrock - Moonshot AI's Kimi K3, its most capable model with 2.8 trillion parameters, is now available on Amazon Bedrock. The open-weight model features native vision capabilities, a 1-million-token context window, and 2.5x improved scaling efficiency over its predecessor, making it suited for coding and know...
OpenAI, Microsoft Knew AI Training Harmed Web
September 18, 2026: OpenAI, Microsoft Knew AI Training Harmed Web - Newly unsealed court documents from the New York Times lawsuit against OpenAI and Microsoft reveal the companies were aware of serious ethical concerns about their AI training practices. Internal documentation warned of a doom loop that would damage the web. The documents show the companies describ...
CoreWeave's First User Conference: What to Expect
September 18, 2026: CoreWeave's First User Conference: What to Expect - CoreWeave holds its inaugural Fully Connected user conference in San Francisco on Sept. 30-Oct. 1, with theCUBE providing live coverage. The AI cloud provider has made waves with autonomous agent improvement capabilities and a new Physical AI Field Engineering service, while reporting a revenue back...
Google's CC AI Agent Now Manages Family Households
September 18, 2026: Google's CC AI Agent Now Manages Family Households - Google is expanding its CC AI agent to help families manage household tasks, including signing permission slips, creating meal plans, tracking school schedules, and coordinating shared calendars. The agent gets its own Google account, supports up to six family members, and can automate email sharing...
Virginia Governor Reins In Data Center Expansion
September 18, 2026: Virginia Governor Reins In Data Center Expansion - Virginia Governor Abigail Spanberger has signed Executive Order 22, banning executive branch officials from signing nondisclosure agreements and giving local communities greater say in data center development. The move could slow approvals in a state already recognized as the global data center capi...
AI Coding Agents Deploy Hugging Face Models Smarter
September 18, 2026: AI Coding Agents Deploy Hugging Face Models Smarter - Deploying Hugging Face models to Amazon SageMaker AI requires dozens of decisions around containers, instance types, autoscaling, and monitoring. Unguided coding agents like Kiro and Claude Code often make costly mistakes, defaulting to outdated serving containers and causing repeated deployment fai...
AI Leaders Debate Who Polices Frontier Development
September 18, 2026: AI Leaders Debate Who Polices Frontier Development - Anthropic CEO Dario Amodei has unveiled a plan to pace AI development, proposing independent safety evaluators and coordination among AI labs in democratic nations. The plan follows a doomsday warning from an Anthropic researcher and has drawn mixed reactions, including pushback from Nvidia's Jensen...
Stay ahead of the curve
Get a weekly AI news roundup delivered to your inbox every Monday.
Subscribe for free