A Moneycontrol headline reports that CERT-In has access to Anthropic’s Claude Mythos for cybersecurity vulnerability testing; details remain unconfirmed.
The Latest
CERT-In Gets Access To Anthropic’s Claude Mythos To Test Cybersecurity Vulnerabilities – Moneycontrol.com
Top Anthropic Product Leader Cat Wu, One Of The Faces Of Claude Code, Has Left The Startup – Business Insider
Business Insider reports that Cat Wu, a product leader associated with Claude Code, has left Anthropic. Her departure and next role remain unclear.
Trump Asked Grok AI About Invading Venezuela. Democrats Want His Chat History – CNET
A headline-only report says Democrats want President Trump’s Grok chat history after he asked the AI about invading Venezuela. Key details remain unverified.
How Cisco Defends At Machine Speed With OpenAI
OpenAI’s headline describes Cisco defending at machine speed with OpenAI, but no article body was available to confirm the tools, results or terms.
OpenAI Appoints Jill Hazelbaker As Chief Corporate Affairs Officer
OpenAI has appointed Jill Hazelbaker as chief corporate affairs officer. The announcement provides no further details about her remit or start date.
Accelerating Engineering For The Physical World
OpenAI has a page titled “Accelerating engineering for the physical world,” but the available material provides no details about its claims or plans.
Anthropic Debuts Tool For Your Boss To Spy On Your Claude Chats – Futurism
Anthropic appears to have introduced a workplace tool for managing Claude use, but the headline alone does not confirm what employers can see.
AI Repairs AI: GPT-6 Astra Edited Another Model’s Weights. A Loop Fix, Not an IQ Jump
GPT-6 Astra found the cause of a small model’s endless loops and cut it out of two layers. The score jump is real, but it is a repair, not an upgrade: what the blog post and logs show, what the viral version left out, and what it means for companies running open models.
Z.ai’s Legal AI Fully Passes About 1 in 5 Benchmark Tasks. Why That Still Ranks Near the Top
Z.ai says its new legal model reaches a 19.2 percent all-criteria pass rate on Harvey’s Legal Agent Benchmark. I checked what that number measures, who ran the test, which claims hold up, and what a small spot check of my own found.
StateBench: Testing How Language Models Track State Changes Over Long Documents
Hugging Face introduces StateBench to compare how language models track changing information as documents grow longer.
