2026-09-01 科技日报

扫描 514 篇候选内容 · 覆盖 650 个信息源 · 纳入 502 篇

📌 今日要点

• Apple shares ‘shocking evidence’ against former employee accused of stealing company data for OpenAI
• The Pentagon now has its own version of ChatGPT and Grok
• Instagram puts new limits on undisclosed AI profiles
• Harvard Law dropout raises $6M for Blue Voice to build a ‘Harvey for police officers’
• Clipto uses AI to search terabytes of video and is now valued at $250M
• Nvidia’s $3.5B MediaTek bet reveals its plan for tackling Big Tech’s AI chip buildout

📝 科技简讯

• Apple shares ‘shocking evidence’ against former employee accused of stealing company data for OpenAI — TechCrunch AI
• The Pentagon now has its own version of ChatGPT and Grok — TechCrunch AI
• Instagram puts new limits on undisclosed AI profiles — TechCrunch AI
• Harvard Law dropout raises $6M for Blue Voice to build a ‘Harvey for police officers’ — TechCrunch AI
• Clipto uses AI to search terabytes of video and is now valued at $250M — TechCrunch AI
• Nvidia’s $3.5B MediaTek bet reveals its plan for tackling Big Tech’s AI chip buildout — TechCrunch AI
• Meeting note-taker Circleback adds a free tier to attract more customers — TechCrunch AI
• GigaPath-Flash and GigaTIME-Flash: Toward population-scale discovery with efficient pathology foundation models — Microsoft Research
• Your LLM Can Return Perfect JSON and Still Be Wrong — Towards Data Science
• FAQ as RAG: When You Get to Design the Corpus — Towards Data Science
• Why RAG Complexity Should Be Earned — Towards Data Science
• AgentOps Is Not MLOps: What Breaks in Your Monitoring Stack When Agents Go to Production — Towards Data Science
• VC 疯了！200 万现金冠军奖，又花 4000 万造了一座 AI「创业乌托邦」 — 量子位
• 滴滴自动驾驶新一代车型开启载客测试服务 — 量子位
• 范式与华为达成重磅算力战略合作，成为首批拥抱国产最高端算力底座的 AI 企业 — 量子位
• 「GPT-6」灰测 demo 刷屏！周四发布在即 — 量子位
• OpenAI 买几万台 Mac 搞强化训练！英伟达的活被苹果抢了 — 量子位
• 全国第三，公司第二，“初创黑马”灵犀智涌用 ROSS Harness 把机器人送进工业具身智能第一梯队 — 量子位
• AQuA：让量化研究 Agent 持续进化，也让回测结果经得起检验 — 量子位
• Beyond Relative Geometry: Metric-Aware Geometry Perception for Robotics — ArXiv RO (cs.RO)
• Remote Human and Robot Interaction for Greenhouse Gardening Using Virtual Reality — ArXiv RO (cs.RO)
• Beyond Data Scaling: Representation-Centric Continued Pre-training for Vision-Language-Action Models — ArXiv RO (cs.RO)
• PHR-VLA: Planning Horizon Reasoning for Vision-Language-Action Models — ArXiv RO (cs.RO)
• One year in a forest: Analyzing the challenges of autonomous navigation in subarctic environments — ArXiv RO (cs.RO)
• Polimill builds Japan’s next-generation public AI infrastructure — OpenAI Blog
• Marginal Coverage Credit Reduces Redundant Exploration in Parallel State-Entropy Optimization — ArXiv ML (cs.LG)
• Quantization-Triggered Backdoors in Language Models: Cross-Quantizer Transferability and the Validation—Deployment Gap — ArXiv ML (cs.LG)
• DAMP: Decay-Aware Mixed-Precision Recurrent-State Quantization — ArXiv ML (cs.LG)
• A Deeper Analysis of Block-Sparse Featurizers — ArXiv ML (cs.LG)
• When Muon Meets Task Interference: A Spectral Perspective on Continual Learning and Model Merging — ArXiv ML (cs.LG)
• Dandelion: A Spherical Flower for Neural Simulation of Planetary Dynamics — ArXiv ML (cs.LG)
• Self-Explainable Multi-Label Graph Neural Network for Correlated Evidence Attribution — ArXiv ML (cs.LG)
• Speed Up LLM Inference with DSpark Speculative Decoding — KDnuggets
• 7 Python Mistakes Beginners Make (And What to Do Instead) — KDnuggets
• Connect an AgentCore Runtime hosted MCP server to Amazon Quick — AWS Machine Learning
• AWS recognized as a Leader in The Forrester Wave: AI Infrastructure Solutions, Q4 2025 — AWS Machine Learning
• Manage agents, tools and skills at scale with AWS Agent Registry — AWS Machine Learning
• Build observable enterprise agentic retrieval using Managed Amazon Bedrock Knowledge Base with AWS CloudFormation — AWS Machine Learning
• Build multi-tenant agentic chat applications on enterprise data with Amazon Bedrock Managed Knowledge Base — AWS Machine Learning
• Open Thread Autumn 2026 — LessWrong

📊 数据概览