2026-09-26 科技日报
4 min
扫描 515 篇候选内容 · 覆盖 650 个信息源 · 纳入 498 篇
📌 今日要点
- Ahead of US IPO, British AI neocloud Nscale secures $3.36B in convertible financing
- Meta’s Muse just stole the AI spotlight from OpenAI and Anthropic
- Proaction boosts sales 60% and saves 75+ hours with Codex
- 10 Things I’m Learning Beyond AI to Become More Technologically Fluent
- Your Model’s MSE Is Lying to You: Part II
- RAG Isn’t an Agent — I Built the Layer Between Retrieval and Action
📝 科技简讯
- Ahead of US IPO, British AI neocloud Nscale secures $3.36B in convertible financing — TechCrunch AI
- Meta’s Muse just stole the AI spotlight from OpenAI and Anthropic — TechCrunch AI
- Proaction boosts sales 60% and saves 75+ hours with Codex — OpenAI Blog
- 10 Things I’m Learning Beyond AI to Become More Technologically Fluent — Towards Data Science
- Your Model’s MSE Is Lying to You: Part II — Towards Data Science
- RAG Isn’t an Agent — I Built the Layer Between Retrieval and Action — Towards Data Science
- Jev vs. LLMs: When AI Moves from Generation to Decision-Making — Towards Data Science
- “AlphaGo”杀进足球场!自我对弈 140 年,机器人成“梅西终结者” — 量子位
- OpenAI 闯大祸!GPT 竟黑进医保系统,黄仁勋:管不住就关掉 — 量子位
- 亮出“中国最强 AI 芯片”还不够,平头哥又甩出一手开源 — 量子位
- 别人忙着卷 Code,Kimi 抽身反打浏览器插件:网页操作一秒变 Skill — 量子位
- 华为大模型双子星联手创业,要找物理世界的 Scaling Law — 量子位
- 呃…小扎「自研 Manus」刚成 Meta 太子,就塌房了 — 量子位
- 给机器人当老师,还能赚外快?“中国版 Index”觅蜂派来了 — 量子位
- Pentagon was right to slap Anthropic with a security supply chain risk label, federal court says — The Decoder
- Another Google Deepmind researcher quits, says building superintelligent AI soon is “inherently irresponsible” — The Decoder
- Microsoft gives Copilot another makeover, adding an Autopilot agent and usage-based billing — The Decoder
- Meta’s Muse agent gives every user a full cloud computer running Ubuntu Linux — The Decoder
- Intelligence doesn’t come cheap as AI drives up costs for the NSA, hospitals, and insurers — The Decoder
- Google’s “Call for Me” lets Gemini phone businesses for you — The Decoder
- Anthropic signs 500 billion in under a year — The Decoder
- Framing by Wording, Framing by Selection: A Large-Scale Two-Dimensional Audit of French News Headlines, 2022-2025 — ArXiv CL (cs.CL)
- Reward Hacking Challenges Oversight of Autonomous Research Agents — ArXiv CL (cs.CL)
- Benchmarking Argumentative Behaviour of LLMs: A Study of Defences Against Character Attacks — ArXiv CL (cs.CL)
- An Explainable DistilBERT-BiLSTM-Attention Framework for Binary and Multi-Class Hate Speech Detection — ArXiv CL (cs.CL)
- Batching by Length Instead of Looping Item by Item for SLM Optimization — KDnuggets
- 7 Advanced Python Tricks to Level Up Your Coding Skills — KDnuggets
- Stable and Faithful Explanations for Knowledge Tracing — ArXiv ML (cs.LG)
- SMILESGNN: Interpretable Clinical Toxicity Prediction via SMILES-Graph Cross-Attention Fusion — ArXiv ML (cs.LG)
- CFD Correction of Open Tip Clearance Flow in a Compressor Cascade Using VAE Latent Space Adaptation — ArXiv ML (cs.LG)
- CARE: Condition-Aware Representation Regularization for Diffusion Models — ArXiv ML (cs.LG)
- SpaFactor: Lightweight Spatial Context-Aware Gene Program Modeling for Histology-to-Transcriptomics Inference — ArXiv ML (cs.LG)
- When Explanations Cannot Be Read: Measuring and Correcting SHAP and LIME Rendering for Right-to-Left Languages — ArXiv ML (cs.LG)
- Leakage-Safe Machine Learning for Hydrogen Embrittlement Detection in 316L Stainless Steel: A Region-Held-Out Evaluation of Texture and Deep Features in SEM Micrographs — ArXiv ML (cs.LG)
- Scaling MoE reinforcement learning on Amazon EKS with EFA and DeepEP with 40% more throughput — AWS Machine Learning
- Accelerate multimodal RL training with SkyRL on Amazon SageMaker HyperPod — AWS Machine Learning
- NarrateAI: production-ready LLM quality assurance on Amazon Bedrock — AWS Machine Learning
- Deploying real-time personalized speech with Qwen3-TTS on Amazon SageMaker AI — AWS Machine Learning
- How Datacor built self-service rental analytics with Amazon Quick Sight — AWS Machine Learning
- Multi-Region training with Amazon SageMaker HyperPod and Qumulo — AWS Machine Learning
📊 数据概览
| 指标 | 数值 |
|---|---|
| 候选内容 | 515 |
| 去重后 | 498 |
| 纳入日报 | 498 |
| 主题分组 | 0 |
| 独立条目 | 40 |
| 信息源数量 | 650 |