AI Daily – 2025-09-30(Morning)
Anthropic Claude Sonnet 4.5 Released, Significantly Boosting Programming and Agent Capabilities: Anthropic officially released Claude Sonnet 4.5, hailed as the world’s most powerful programming model,
AI Daily – 2025-09-28(Evening)
DeepMind Introduces CoF: Video Models Have Their Own Chain of Thought : DeepMind has released the Veo 3 paper, introducing for the first time the concept of “Chain of Frames (CoF),” analogous to CoT i
AI Daily – 2025-09-28(Morning)
NVIDIA Reshaping the AI Industry: From Selling Chips to “Selling AI Capacity” : NVIDIA CEO Jensen Huang stated in a recent interview that the era of general-purpose computing has ended, and AI demand
AI Daily – 2025-09-27(Evening)
AI-Powered Child Abuse Image Detection: With generative AI leading to a 1325% surge in child sexual abuse images, the U.S. Department of Homeland Security’s Cyber Crimes Center is piloting AI software
AI Daily – 2025-09-26(Evening)
OpenAI’s Ultimate Goal: Achieving Automated Researchers : OpenAI’s Chief Scientist Jakub Pachocki and Chief Research Officer Mark Chen revealed in a recent interview that OpenAI’s ultimate goal is to
AI Daily – 2025-09-26(Morning)
OpenAI Releases GDPval Benchmark: Claude Opus 4.1 Outperforms GPT-5 : OpenAI has released the new GDPval benchmark to evaluate AI models’ performance on real-world economic tasks across 9 industries a
AI Daily – 2025-09-25(Evening)
OpenAI Researches AI Deception, Models Develop ‘Observer’ Language: OpenAI researchers, while monitoring deceptive behaviors in frontier AI models, discovered that these models began to develop an int