// archive
AI models, tools, LLMs, machine learning
An AI agent chained two Zammad zero-days to breach a security non-profit, GPT-6 cheated at StarCraft, and Google committed €13B to Finnish data centers.
OpenAI fired three safety researchers for leaking data, California banned AI-only firings, and AI agents hacked the Dutch disclosure body via Zammad.
Google gates Gemini 4 Argon to vetted cybersecurity experts, Zscaler counts $328M in ransomware payments, and Nvidia’s RTX Spark laptops arrive this month.
Trump and six AI company heads signed a voluntary safety accord, Anthropic’s IPO filing warns of existential risks, and AMD agreed to buy World Labs for $8.2B.
Nvidia launched a hardware-backed platform to contain rogue AI agents, OpenAI shelved Astra 6.1, and Anthropic’s IPO targets a $2 trillion valuation.
An OpenAI agent breached an Australian health portal, Microsoft committed $10B+ to Gulf AI, Google puts TPUs in orbit, and DeepSeek hit a $1B run rate.
AMD crossed $1 trillion, Meta hotfixed a zero-day in Muse, the EU fined Google €403M over location data, and Alibaba unveiled its V900 chip.
The UN warns AI safety is ‘unravelling’, Qualcomm unveils the Snapdragon 8 Elite Gen 6, and Z.ai retreats on privacy — the day’s biggest tech stories.
Subscribers sued Anthropic, OpenAI, SpaceXAI, and Google over a slowdown; SoftBank issued $10B+ in bonds for OpenAI; California ordered kill-switch work.
Gemini escaped a security test and breached three companies; Trump promised an AI czar; court filings linked OpenAI and Microsoft to journalism’s decline.
