AI Daily – 2025-05-16(Morning)
Meta’s flagship AI model “Behemoth” repeatedly postponed, sparking internal concerns and industry reflection: Meta’s flagship AI model “Behemoth,” originally planned for an April release and later pos
AI Daily – 2025-05-15(Evening)
DeepMind introduces AlphaEvolve: A Gemini-powered evolutionary coding agent, advancing algorithm discovery: AlphaEvolve combines the creativity of Gemini models with an automated evaluator, utilizing
AI Daily – 2025-05-15(Morning)
Google DeepMind releases AI programming agent AlphaEvolve, achieving algorithmic self-evolution and optimization: Google DeepMind has launched AlphaEvolve, an AI programming agent that combines the cr
AI Daily – 2025-05-14(Evening)
Google DeepMind releases AlphaEvolve: A Gemini-powered coding agent revolutionizing algorithm discovery: Google DeepMind has launched AlphaEvolve, an AI coding agent powered by Gemini, designed to dis
AI Daily – 2025-05-14(Morning)
MLSys 2025 Announces Best Paper Awards, Including FlashInfer and Other Projects: MLSys 2025, a top international conference in the systems field, announced two best papers. One is FlashInfer, from ins
AI Daily – 2025-05-13(Evening)
OpenAI Releases HealthBench Benchmark, Demonstrating Significant Improvement in AI Medical Capabilities: OpenAI has released HealthBench, a medical AI evaluation benchmark built in collaboration with
AI Daily – 2025-05-13(Morning)
Interview with OpenAI Chief Scientist Jakub Pachocki: AI May Autonomously Discover New Science Within 5 Years, World Models and Reinforcement Learning Are Key: OpenAI Chief Scientist Jakub Pachocki st
AI Daily – 2025-05-12(Evening)
OpenAI Releases HealthBench to Evaluate Medical AI Performance: OpenAI has launched HealthBench, a new benchmark designed to measure the performance and safety of large language models (LLMs) in medic
AI Daily – 2025-05-12(Morning)
Prime Intellect Open Sources INTELLECT-2 Model: Prime Intellect has released and open-sourced INTELLECT-2, a 32 billion parameter model claimed to be the first trained via globally distributed reinfor