2026-10-06 科技日报

扫描 452 篇候选内容 · 覆盖 650 个信息源 · 纳入 437 篇

📌 今日要点

• Introducing GLM 5.3 on Amazon Bedrock
• Supercharge regulated workloads with Claude Code and Amazon Bedrock
• New agent skill: Amazon SageMaker optimized generative AI inference for your coding agent
• Making Amazon Quick enterprise-ready: Automated, auditable cross-account resource promotion
• Agentic retrieval with LangChain and Amazon Bedrock Knowledge Bases
• Downgrading user roles in Amazon Quick

📝 科技简讯

• Introducing GLM 5.3 on Amazon Bedrock — AWS Machine Learning
• Supercharge regulated workloads with Claude Code and Amazon Bedrock — AWS Machine Learning
• New agent skill: Amazon SageMaker optimized generative AI inference for your coding agent — AWS Machine Learning
• Making Amazon Quick enterprise-ready: Automated, auditable cross-account resource promotion — AWS Machine Learning
• Agentic retrieval with LangChain and Amazon Bedrock Knowledge Bases — AWS Machine Learning
• Downgrading user roles in Amazon Quick — AWS Machine Learning
• Evaluating multi-agent systems for explainability and helpfulness with Amazon Bedrock AgentCore — AWS Machine Learning
• Meta Muse Explained: What It Is, How It Works, and What It Can Do — KDnuggets
• 3 Statsmodels Tricks for Time Series Analysis & Forecasting — KDnuggets
• Hybrid Machine Learning-Assisted Raman Spectroscopy with Generative Feature Augmentation for Pharmaceutical Identification — ArXiv ML (cs.LG)
• The Price of Greenwashing: Algorithmic Verification and Market Discipline using Conformal Machine Learning — ArXiv ML (cs.LG)
• State-Space Unlearning for Non-Stationary Bias in Land Surface Forecasting — ArXiv ML (cs.LG)
• Nearest-neighbour baselines for fingerprint prediction from MS/MS spectra under different assumptions — ArXiv ML (cs.LG)
• Rank-Aware Speculative Sampling for Diffusion Draft Trees — ArXiv ML (cs.LG)
• Counterfactual Predictions in Scientific Emulators Without Controlled Experiments — ArXiv ML (cs.LG)
• Approximation Property of Dropout Neural Networks: Sobolev Rates and Confidence Bounds — ArXiv ML (cs.LG)
• Our approach to EU text provenance rules — OpenAI Blog
• Building advertising for the way people use AI — OpenAI Blog
• Takes One to Know One - Training a model to grade reward hacks causes it to reward hack less itself — LessWrong
• You can graft SDF changes from base models onto post-trained models — LessWrong
• Kurzgesagt’s video on the Hugging Face incident: “AI Just Crossed the Terrifying Line - Now What?” — LessWrong
• Three types of AI risk: why (some) leftists are against (some) AI regulations — LessWrong
• A closer look at what ozone and nuclear arms control tell us about regulating AI — LessWrong
• OpenAI will start watermarking ChatGPT’s text in the EU — TechCrunch AI
• Reflection debuts Beam, an open-weight AI model to rival Chinese models at lower compute cost — TechCrunch AI
• NEEDLEWORK: Offline Rewriting of Robot Data with Verified Local Stitches — ArXiv RO (cs.RO)
• Filter-Aware Fine-Tuning for Safe Humanoid Whole-Body Tracking — ArXiv RO (cs.RO)
• Meta and Microsoft pull back from Claude as Anthropic transforms from partner into competitor — The Decoder
• Reka AI’s omni-model Rho-1 handles text, images, video, and robot control in a single model — The Decoder
• OpenAI will watermark ChatGPT text in the EU but makes it optional for API users worldwide — The Decoder
• Most Americans want AI development to slow down or stop entirely, new poll finds — The Decoder
• Anthropic is quietly becoming America’s biggest corporate donor ahead of its mega IPO — The Decoder
• Aleph Alpha releases Kolibri, an open-weight model that makes the case for European AI sovereignty — The Decoder
• ChatGPT’s new ad format fills the image generation loading screen with product carousels — The Decoder
• 刚刚，诺贝尔奖颁给光遗传学！ — 量子位
• 刚刚，Hinton 发了首篇 RSI 论文 — 量子位
• 限时 28 天！OpenAI 承诺没新功能就重置，网友：只想要 Opus — 量子位
• MintFlow: Minimal Trajectory Intervention for Constrained Flow Matching — arXiv AI (cs.AI)
• Fast Models, Slow Evidence: A Paired and Self-Audited Evaluation of System-1 Decision Models for LLM Agent Harnesses — arXiv AI (cs.AI)
• The AI Risk Observatory: What Can We Learn from AI Disclosures in Annual Reports About Societal Resilience? — arXiv AI (cs.AI)

📊 数据概览