Claude’s Mythos Too Powerful, DeepSeek, OpenAI Misses Targets, Anthropic Hits $1 Trillion, and Taylor Swift TMs Her Voice

News and Updates:

• Claude Mythos Finds Thousands of Flaws: Anthropic's unreleased Claude Mythos AI autonomously discovered thousands of zero-day vulnerabilities, including a 27-year-old OpenBSD flaw and critical Linux kernel exploits, without human assistance.
• DeepSeek V4 Goes Public: Chinese AI firm DeepSeek released V4, an open-source model matching top closed-source competitors at a fraction of the cost, with a 1-million-token context window and dramatic memory efficiency gains.
• OpenAI Misses Growth Targets: OpenAI reportedly fell short of internal ChatGPT user and revenue goals, rattling investors and sending Nvidia, AMD, Oracle, and CoreWeave shares lower in pre-market trading.
• Anthropic Surpasses $1 Trillion Valuation: Secondary market demand for Anthropic shares has pushed its valuation to $1 trillion on Forge Global, now trading above OpenAI despite OpenAI's larger official valuation of $852 billion.
• Taylor Swift Trademarks Her Voice and Likeness: Swift filed three U.S. trademark applications covering specific spoken phrases and a stage performance image, aiming to establish legal protection against unauthorized AI-generated deepfakes. Legal experts note Swift's approach of trademarking her voice is unprecedented in court, but could set a new legal standard for how public figures protect their identities in the AI era.