2026-09-23 科技日报

扫描 574 篇候选内容 · 覆盖 650 个信息源 · 纳入 558 篇

📌 今日要点

• Qualcomm launches two new smartphone chips with emphasis on AI
• Meta admits Muse’s likeness to OpenClaw isn’t a coincidence
• OpenAI launches GPT-6 Sol and Luna, boasting lower cost and fewer mistakes
• Anthropic releases Opus 5.5 with lower prices and Fable-level performance
• AstroForge is putting AI in command of its next spacecraft
• Break Your Own RAG Pipeline Before Users Do

📝 科技简讯

• Qualcomm launches two new smartphone chips with emphasis on AI — TechCrunch AI
• Meta admits Muse’s likeness to OpenClaw isn’t a coincidence — TechCrunch AI
• OpenAI launches GPT-6 Sol and Luna, boasting lower cost and fewer mistakes — TechCrunch AI
• Anthropic releases Opus 5.5 with lower prices and Fable-level performance — TechCrunch AI
• AstroForge is putting AI in command of its next spacecraft — TechCrunch AI
• Break Your Own RAG Pipeline Before Users Do — Towards Data Science
• Build a Speaker-Recognition App with Claude Code — Towards Data Science
• 4 Ways to Use AI on a PhD Thesis — Towards Data Science
• An Introduction to Jev — Towards Data Science
• Better prompt caching for GPT-6 — OpenAI Blog
• Introducing GPT-6 Sol and Luna — OpenAI Blog
• Parallel cut research time and cost in half with GPT‑6 Astra — OpenAI Blog
• Debating RSI, the US-China Gap, and Jaggedness with JS Denain of Epoch AI — Interconnects (Nathan Lambert)
• OpenAI’s GPT-6 Sol and Luna cut prices in half but barely move the needle on performance — The Decoder
• Claude Opus 5.5 matches Fable 5.1 performance at lower cost and promises less “Claudish” writing — The Decoder
• OpenAI calls for international standards on AI that could improve itself — The Decoder
• A tiny software layer from lab-grown neurons promises faster, cheaper AI video — The Decoder
• Xiaomi’s affordable flagship AI leads the open models, and Anthropic says Claude helped get it there — The Decoder
• OpenAI says its internal model solved over 100 long-standing math problems after just a month of training — The Decoder
• Anthropic is setting up a biology lab where Claude guides robots through drug experiments — The Decoder
• Recognition, Simulation, and Refusal: A Contamination-Aware Study of Classic Psychological Effects in LLM Agents — ArXiv CL (cs.CL)
• Memory That Looks Forward: A Zero-Inference Prospective Term for Personal Memory Retrieval — ArXiv CL (cs.CL)
• Summarize, Judge, Refine: Decoupled Content Understanding and Policy Learning for Multimodal Content Moderation — ArXiv CL (cs.CL)
• AI-inferred expressed well-being and collective-action discourse in climate-change campaigns on X — ArXiv CL (cs.CL)
• Token Signatures of Code: Comparing Coding Behaviors Across Large Language Models — ArXiv CL (cs.CL)
• CHOREO: Every Humanoid Skill as a Trajectory — ArXiv RO (cs.RO)
• D3DWA: Adaptive Weight and Prediction-Horizon for Dynamic Window Approach via Dueling Double Deep Q-Network — ArXiv RO (cs.RO)
• DeViGrasp: Robust Visual Mobile Grasping for Quadruped Manipulators under Degraded Perception — ArXiv RO (cs.RO)
• ORDER: A Fictitious-World Benchmark for Domain-Adaptive Embodied AI — ArXiv RO (cs.RO)
• OJOx: Specification-Conditioned Demonstrations for Embodied AI in Construction — ArXiv RO (cs.RO)
• Did You Steal My Shot? Pioneering Camera Motion Plagiarism Detection in Generative Videos — ArXiv CV (cs.CV)
• Enabling Vision and Cross-Modal Learning for Multimodal Stroke Recurrence Prediction: An Interpretable Two-Step Framework — ArXiv CV (cs.CV)
• Moonworks Lunara: Modeling Artistic Intelligence — ArXiv CV (cs.CV)
• Performance vs Consistency: Evaluating a Foundation Model in Lung-RADS Screening — ArXiv CV (cs.CV)
• Brain-to-Image Generation: Reconstructing Visual Stimuli from EEG using Generative Adversarial Networks — ArXiv CV (cs.CV)
• 北京第一，杭州深圳进入前三！中国 AI 行业大洗牌了 — 新智元
• ChatGPT 版 Grok Bot 代码曝光！OpenAI 也要为你造不下班的 AI 同事 — 新智元
• Grok 4.7 上线，全球第三！马斯克：4.8 也训完了 — 新智元
• OpenAI 内部曝光：AI 开始自己造 AI！奥特曼急发全球暂停令 — 新智元
• 刚刚，中国 AI 占领全球前四！ — 新智元

📊 数据概览