Anthropic's Mythos Model Poses Security Risks, OpenAI Raises Record $122B
This week's top AI news: Anthropic's powerful Mythos model discovered thousands of zero-day vulnerabilities, leading the company to restrict its release. Meanwhile, OpenAI raised a record $122B in funding and released GPT-5.4 with new capabilities. Meta also launched its first closed-source flagship AI model, Muse Spark.
Why it matters
These developments highlight the growing power and influence of AI companies, as well as the complex challenges around AI safety, governance, and competition.
Key Points
- 1Anthropic's Mythos model found thousands of zero-day vulnerabilities, prompting the company to limit access to 11 partners
- 2Anthropic's annual revenue surpassed $30B, overtaking OpenAI as the highest-earning AI company
- 3OpenAI raised a record $122B in funding, with Amazon as the largest investor, signaling a shift from its exclusive partnership with Microsoft
- 4Meta launched its first closed-source flagship AI model, Muse Spark, marking a strategic shift from its open-source advocacy
Details
This week saw major developments in the AI industry, with Anthropic and OpenAI taking center stage. Anthropic's powerful Mythos model was found to have discovered thousands of previously unknown zero-day vulnerabilities across major operating systems and browsers. Citing significant cybersecurity risks, Anthropic has restricted access to the model, allowing only 11 partner organizations to use it in a controlled environment. This decision has sparked debates around who should determine which models are
No comments yet
Be the first to comment