AI News

The week’s key AI news: Nvidia’s billion-dollar purchase of Hugging Face, Claude’s own browser, a cyber warning letter, and a cheap model from China.

A Taiwan-based study finds Chinese state hackers have doubled their attack volume since adopting DeepSeek. ChatGPT and Claude show up too.

OpenAI and Meta just pushed their assistants deep into macOS: ChatGPT can now send iMessages, Meta AI reads what’s on your screen. Here’s what permissions that requires.

Anthropic’s Claude Cowork now drafts and sends Gmail messages on its own, and the send-approval step can be switched off entirely.

Gemini 3.7 Flash, Nvidia’s $500 billion bet, cracked AI reasoning traces, OpenAI’s ethics exit, and $1.1 billion for a two-month-old startup: the week in review.

Google reports rapid Gemini growth while market studies show shifting shares. Why big AI user numbers make sense only when their measurement method is clear.

OpenAI released GPT-5.6-Cyber for security research. It found two unknown Chrome flaws, and access requires identity verification and a hardware key.

Meta released Muse Glimmer, a 30B open-weight model under Apache 2.0. Zuckerberg defends distillation and asks Washington for fewer rules on US labs.