NLNEXT LAB RADAR
AI 资讯日报 · 2026-07-28

2026-07-28 · AI 资讯日报

今日新收录 160 条公开资讯，按模型 / 产品 / 行业 / 论文 / 观点 自动归类汇编（非 AI 生成，点击可溯源原文）。

今日精选6 条

  1. 1.
    论文研究The Physics of Multi-Turn Long-Horizon Planning: From Pre-training to Post-training via Single- and Multi-Teacher On-Policy Agentic Distillation— arXiv

    Multi-turn long-horizon planning is critical for foundation model agents, yet how to fundamentally improve it remains unclear. Existing models are trained on uncontrollable and opaque Internet data, m…

  2. 2.
    论文研究Certified Parallel-in-Time Sinkhorn for Dynamic Entropic Optimal Transport— arXiv

    Dynamic applications, including optimal-transport Flow Matching, repeatedly solve related entropic optimal transport problems, yet conventional distributed Sinkhorn processes frames sequentially and s…

  3. 3.
    论文研究Learning Distributions from Multiple Data Providers— arXiv

    Motivated by learning from heterogeneous and overlapping data providers, we study a stylized model of distribution learning from restricted conditional samples. The goal is to learn an unknown distrib…

  4. 4.
    论文研究KANEx: Translating Kolmogorov-Arnold Networks' Interpretability to Medical Explainability— arXiv

    Computer vision models have become highly effective for medical applications, yet their black-box nature continues to undermine clinician trust. In clinical workflows, chest X-ray classifiers are incr…

  5. 5.
    论文研究Global Convergence of DGM and PINN Algorithms for Solving Nonlinear PDEs— arXiv

    The Deep Galerkin Method (DGM) and Physics Informed Neural Networks (PINNs) have become widely-used methods for solving partial differential equations (PDEs) in the rapidly growing field of scientific…

  6. 6.
    论文研究DataOrchestra: Learning to Orchestrate Per-Example Curation of Pretraining Data— arXiv

    Pretraining data processing is critical to the downstream performance of Large Language Models (LLMs). However, many existing approaches define a fixed processing strategy at the corpus or domain leve…

模型发布/更新6 条

  1. 1.
    Claude Opus 5 System Card— Anthropic

    Claude Opus 5 System Card www-cdn.anthropic.com

  2. 2.
    Scientific computing in the age of agentic AI— OpenAI

    AI 导读 · 科学家用AI编程代理加速科研，从基因组学到各领域，将颠覆传统计算模式。

  3. 3.
    not much happened today— AI News

    **Moonshot** released the **Kimi K3**, a **2.8T-parameter MoE** model with **104B active parameters/token**, featuring innovations like **Kimi Delta Attention (KDA)**, **Gated MLA*…

  4. 4.
    Hands-on with Claude.— Anthropic

    Hands-on with Claude. Anthropic

  5. 5.
    Black Forest Labs Releases FLUX 3: A Multimodal Flow Model for Image, Video, Audio and Robot Action Prediction— MarkTechPost

    Black Forest Labs (BFL) has released FLUX 3, a multimodal foundation model that learns from images, videos and audio inside a single architecture. It is also the first FLUX model t…

  6. 6.
    Agentic Misalignment in Summer 2026— Anthropic

    Agentic Misalignment in Summer 2026 Alignment Science Blog

产品发布/更新6 条

  1. 7.
    Perplexity Releases pplx, a Single-Binary CLI That Puts Its Search API in the Terminal for Coding Agents— MarkTechPost

    Perplexity has released pplx, an official command line client for its Search API. The tool exposes two commands — pplx search web and pplx content fetch — and returns exactly one J…

  2. 8.
    KwaiKAT Team Releases KAT-Coder-V2.5: An Agentic Coding Model Trained on 100,000+ Verifiable Repository Environments— MarkTechPost

    The KwaiKAT Team at Kuaishou has published the KAT-Coder-V2.5 technical report, arguing that agentic coding capability is bottlenecked by training infrastructure rather than model…

  3. 9.
    krishagarwal314/autodev-studio— GitHub

    Autonomous multi-agent SDLC harness: describe a feature in plain English and AI agents scope, code, test, review, and open a PR — grounded in a one-time knowled…

  4. 10.
    NimaChu/my-wiki— GitHub

    Local-first AI knowledge app and Agent Skill with evidence-backed Wiki, an interactive knowledge universe, Viki Q&A, and shareable knowledge galaxies.

  5. 11.
    苹果拟大举发力智能家居，以Siri AI为核心推出新品— 36氪

    AI 导读 · 苹果借Siri AI整合硬件，有望重塑智能家居市场格局。

  6. 12.
    OpenAI 引入两种转录模型：低延迟、更好理解上下文— IT之家

    AI 导读 · 专注低延迟与上下文理解，为实时语音交互应用提供更高效API方案。

行业动态6 条

  1. 13.
    OpenAI just open-sourced Codex Security— Hacker News

    AI 导读 · 开源代码安全工具，降低AI编程风险，开发者可免费自查漏洞。

  2. 14.
    Taiwan detains Nvidia employee in widening China chip smuggling probe— The Decoder

    Taiwan's prosecutors have detained an Nvidia employee in connection with the alleged illegal export of Super Micro AI servers to China, according to Bloomberg and Reuters. The arti…

  3. 15.
    Nvidia invests in Ilya Sutskever's AI lab, shifting SSI away from Google chips— The Decoder

    Nvidia is pouring what it calls a "substantial" sum into Safe Superintelligence (SSI), the AI lab run by Ilya Sutskever, OpenAI's former chief scientist. The article Nvidia invests…

  4. 16.
    Anthropic CEO Amodei doubles down on open-weight risk stance while insisting he never called for a ban— The Decoder

    Anthropic CEO Dario Amodei is once again warning about the risks of open AI models while insisting he has never called for a ban. He argues that authoritarian states like China cou…

  5. 17.
    Smart rings are looking like my kind of AI gadget— The Verge

    Over the last few months, I've spent a lot of time talking to my computer. One underrated feature of the LLM revolution has been a remarkable leap in all kinds of dictation technol…

  6. 18.
    Hugging Face is being used to easily undress women and children— The Verge

    Hugging Face is being used to make nonconsensual deepfakes, and the popular open-source AI model repository is doing very little to prevent it. That's according to a new report pub…

论文研究6 条

  1. 19.
    The Physics of Multi-Turn Long-Horizon Planning: From Pre-training to Post-training via Single- and Multi-Teacher On-Policy Agentic Distillation— arXiv

    AI 导读 · 探索从预训练到后训练提升大模型长程规划能力，为自主智能体发展提供关键路径。

  2. 20.
    Certified Parallel-in-Time Sinkhorn for Dynamic Entropic Optimal Transport— arXiv

    AI 导读 · 并行时序Sinkhorn算法突破，加速动态最优传输计算，助力流匹配等前沿应用。

  3. 21.
    Learning Distributions from Multiple Data Providers— arXiv

    AI 导读 · 多数据源分布学习突破，解决异构重叠数据训练难题。

  4. 22.
    KANEx: Translating Kolmogorov-Arnold Networks' Interpretability to Medical Explainability— arXiv

    AI 导读 · 用KAN网络破解医学影像黑箱，可解释性突破让AI诊断更可信。

  5. 23.
    Global Convergence of DGM and PINN Algorithms for Solving Nonlinear PDEs— arXiv

    AI 导读 · 首次证明DGM与PINN求解非线性PDE的全局收敛性，为AI科学计算奠定理论基础。

  6. 24.
    DataOrchestra: Learning to Orchestrate Per-Example Curation of Pretraining Data— arXiv

    AI 导读 · 按需定制每条训练数据，突破粗放式数据处理瓶颈。

技巧与观点6 条

  1. 25.
    LFM2.5-Encoders for Fast Long-Context Inference on CPU— HuggingFace Blog
  2. 26.
    The OlmoEarth Platform: Geospatial inference at planetary scale— HuggingFace Blog

    AI 导读 · 首个全球尺度地理空间推理平台，突破传统数据处理瓶颈，开启AI与地球科学融合新纪元。

  3. 27.
    Designing Skill-Driven Financial Analysis Agents with Claude, Python, MCP Connectors, and Automated Deliverables— MarkTechPost

    In this tutorial, we build an advanced workflow around Anthropic’s financial-services repository and reproduce its skill-driven architecture in pure Python. We begin by installing…

  4. 28.
    Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident— Simon Willison

    AI 导读 · 前沿实验室AI入侵事件的技术细节首次公开，为理解智能体安全漏洞提供了教科书级案例。

  5. 29.
    How AgentCore Gateway supports the MCP 2026-07-28 spec— AWS ML

    AI 导读 · AgentCore网关率先适配MCP重大版本更新，无状态架构与扩展系统是行业风向标。

  6. 30.
    Market surveillance agent with LangGraph and Strands on AgentCore— AWS ML

    AI 导读 · 多智能体协同与生产级部署的结合，为AI系统落地提供了可复用的架构方案。

快讯

日报生成时间：2026-08-16
← 2026-07-27最新日报2026-07-29 →