February 7, 2025 News

Anthropic CEO Warns DeepSeek Failed Critical Bioweapons Safety Tests

Anthropic CEO Dario Amodei revealed that DeepSeek's AI model performed poorly on safety tests related to bioweapons information, describing it as "the worst of basically any model we'd ever tested." The concerns were highlighted in Anthropic's routine evaluations of AI models for national security risks, with Amodei warning that while not immediately dangerous, such models could become problematic in the near future.

AI Pioneer Andrew Ng Endorses Google's Reversal on AI Weapons Pledge

AI researcher and Google Brain founder Andrew Ng expressed support for Google's decision to drop its 7-year pledge not to build AI systems for weapons. Ng criticized the original Project Maven protests, arguing that American companies should assist the military, and emphasized that AI drones will "completely revolutionize the battlefield" while suggesting that America's AI safety depends on technological competition with China.

DeepMind's AlphaGeometry2 Surpasses IMO Gold Medalists in Mathematical Problem Solving

Google DeepMind has developed AlphaGeometry2, an AI system that can solve 84% of International Mathematical Olympiad geometry problems from the past 25 years, outperforming the average gold medalist. The system combines a Gemini language model with a symbolic reasoning engine, demonstrating that hybrid approaches combining neural networks with rule-based systems may be more effective for complex mathematical reasoning than either approach alone.

Sutskever's Safe Superintelligence Startup Seeking Funding at $20B Valuation

Safe Superintelligence, founded by former OpenAI chief scientist Ilya Sutskever, is reportedly seeking funding at a valuation of at least $20 billion, quadrupling its previous $5 billion valuation from September. The startup, which has already raised $1 billion from investors including Sequoia Capital and Andreessen Horowitz, has yet to generate revenue and has revealed little about its technical work.

Figure AI and Others Moving Away from OpenAI Dependencies

Humanoid robotics company Figure has announced it's ending its partnership with OpenAI to develop its own in-house AI models, with CEO Brett Adcock hinting at a significant breakthrough. This move reflects a potential shift in the industry as other organizations, including academic researchers who recently demonstrated training a capable reasoning model for under $50, explore alternatives to OpenAI's offerings.

OpenAI Co-founder John Schulman Joins Mira Murati's New AI Venture

John Schulman, an OpenAI co-founder who briefly joined Anthropic, is reportedly joining former OpenAI CTO Mira Murati's secretive new startup. Murati, who left OpenAI in September, has also recruited other former OpenAI talent including Christian Gibson from the supercomputing team, and was reportedly seeking over $100 million in funding for her venture in October.