OpenAI Built a Cyberpunk Attack Dog to Maul Its Own Models
OpenAI built an LLM whose only job is attacking other LLMs. It's automated red teaming at scale — either the future of AI safety or a cyberpunk ouroboros eating its own tail. Probably both.
Tagged: cybersecurity
Grok's CLI Is Snitching: Every Prompt, Every File, Every Time
A wire-level analysis of xAI's Grok CLI reveals it's sending prompts, file contents, system metadata, and potentially API keys to xAI servers by default. The Elon Musk privacy hypocrisy hits new lows as developers question what their tools are actually doing.
China Calls Claude Code a Backdoor Threat. The Irony.
Beijing's cyber watchdog flags Anthropic's Claude Code for a 'backdoor.' The same government running the world's most sophisticated corporate espionage operation is suddenly very concerned about your data security. Cool.
Grok Caught Hoarding Entire Git Repos in xAI Heist
xAI's Grok Build isn't just reading your code—it's pulling a digital smash-and-grab on your entire GitHub repo. Another day, another multi-billion dollar tech grift.
Fake Perplexity Chrome Extension Was Lurking in Your Browser
A malicious Chrome extension disguised as Perplexity AI was silently intercepting every search and address bar input. The AI hype gold rush has created a paradise for data thieves.
OpenAI's Daybreak: Securing Every Org (Sure, Jan)
OpenAI's Daybreak promises to secure every org on Earth. The company that leaked user chats and gets jailbroken by teens now wants to run your SOC.