Keep pulling the thread on Project Glasswing.
Anthropic's Project Glasswing, the initiative for the limited rollout of its Mythos model, has identified over 10,000 high or critical severity software vulnerabilities.
Mozilla reported finding and fixing 271 vulnerabilities using Anthropic's Mythos model, which is more than 10 times the amount they found using Opus 4.6.
A financial firm used Anthropic's Mythos model to detect and prevent a fraudulent $1.5 million wire transfer.
Anthropic stated that no company has developed safeguards strong enough to prevent Mythos-class models from being misused and potentially causing severe harm, which is why they have not been publicly released.
The White House has approved a secret $9 billion budget request for the CIA and NSA to build their own inference cluster.
The $9 billion in funding for US intelligence agencies will go towards purchasing Nvidia Blackwell chips and supporting infrastructure.
White House Chief of Staff Susie Wiles has approved the use of Anthropic models at the NSA, despite the Pentagon designating Anthropic as a supply chain risk.
Anthropic and the US government are finalizing a contract to allow the NSA to use Anthropic models in classified settings.
The Pentagon disputed with Anthropic over limitations on using AI for autonomous weaponry and domestic surveillance, demanding contracts allow for "any lawful use."
Deepseek is proceeding with a $10 billion funding round that will reportedly value the company at $45 billion.
The Chinese government's AI industry investment fund, Tencent, and JD.com are expected to participate in Deepseek's upcoming $10 billion funding round.
Elon Musk announced that xAI has finished training a 1.5 trillion parameter foundation model called Grok V9 Medium.