Google’s NotebookLM upgrade introduces a Deep Research agent that handles complex questions with step-by-step reasoning and citations. The platform now supports spreadsheets, slide decks, and additional document formats. This moves NotebookLM closer to becoming a comprehensive research workflow tool
In what Anthropic calls the first documented large-scale cyberattack executed without substantial human intervention, Chinese hackers used Claude AI to develop automated attack frameworks and steal private data. The AI handled 80-90% of the operation, only requiring occasional human intervention whi
Despite AI automation spreading across industries, cybersecurity jobs appear uniquely resistant to layoffs. Security experts point to a massive talent shortage and the need for human oversight in high-stakes security decisions. SOC analysts may actually see improved job quality thanks to AI assistan
Anthropic researchers detected what they call the first AI-orchestrated cyber espionage campaign. Chinese state-sponsored hackers allegedly used Claude to build programs that autonomously compromised targets including tech companies and government agencies. The company has since banned the hackers a
A confidential survey of AI founders, investors, and insiders at the Cerebral Valley conference reveals industry predictions that range from wildly optimistic to downright skeptical. The anonymous responses show where the smart money is betting—and what they’re avoiding. This is what happens when in
AI startups are achieving revenue growth that’s rewriting venture capital playbooks. Investors now evaluate companies using entirely new metrics beyond just rapid growth, creating what one VC calls a “funky time” for investing.
JPMorgan Asset and Wealth Management CEO Mary Callahan Erdoes told the CNBC Delivering Alpha conference that AI presents opportunities not yet fully understood. She compared the current market disconnect to Hemingway’s description of bankruptcy – happening slowly then suddenly. Despite recent stock
Chinese tech giant Baidu has unveiled two new AI processors and large-scale computing systems targeting inference and training workloads. The M100 chip optimized for inference launches in early 2026, followed by the M300 for training massive models in 2027. Baidu is building “supernode” clusters com
Caylent’s CTO Randall Hunt says AWS is beating Microsoft and Google in AI model flexibility and fine-tuning capabilities. The AWS partner also sees massive opportunity in VMware customer migrations to AWS, with 73% actively seeking alternatives.
Google DeepMind unveiled SIMA 2, a next-generation AI agent that doubles predecessor performance using Gemini’s reasoning. The self-improving system can understand complex instructions and learn from its own mistakes in virtual environments, representing significant progress toward general-purpose r
