AAI Search
AI 资讯日报 · 2026-09-18

2026-09-18 · AI 资讯日报

今日新收录 157 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Objective vs. Search: Decomposing What Makes a Good Tokeniser— arXiv

    Two dominant tokenisation algorithms are used by modern language models: byte-pair encoding (BPE) and UnigramLM. These differ along two orthogonal axes: their optimisation objective (compression vs. l…

  2. 2.
    论文研究Dreaming the Sound of Contact: Leveraging Video and Audio Generation for Zero-Shot Force-Aware Manipulation and Data Generation— arXiv

    Recent advances in video generation allow robots to learn manipulation trajectories from generated videos. However, these approaches produce purely kinematic trajectories that lack force information,…

  3. 3.
    论文研究Exponential Hardness of Off-Policy Evaluation under History-Dependent Logging— arXiv

    Can a logged dataset visit every hidden state frequently and still be exponentially uninformative about a target policy's value? We show that it can when the logger depends on history. For every horiz…

  4. 4.
    论文研究Cognitive Extensions for Dual-Process Language Agents: Memory and Self-Reflection in Interactive Environments— arXiv

    Language agents remain brittle in interactive environments, where success requires long-horizon state tracking, valid action execution, and recovery from failed steps. We extend SwiftSage, a dual-proc…

  5. 5.
    论文研究Affora: A Design System for Agent-Friendly Interfaces— arXiv

    Computer-use agents increasingly operate software designed for people, but interfaces often leave actions or task state unclear to machine readers. We present Affora, a design system that supports bot…

  6. 6.
    论文研究Flag Game: A Toy Model for Mechanistic Swarm Interpretability— arXiv

    Emergent coordinated behaviors of AI agents are starting to present critical safety risks. A key phenomenon driving these behaviors is the rapid formation and spread of beliefs about the world, and me…

模型发布/更新6 条

  1. 1.
    Introducing Astra for Law— OpenAI

    OpenAI for Law brings frontier intelligence for law, custom firm workflows, connected legal data sources, and legal-grade controls for confidential client work.

  2. 2.
    How Claude is uplifting biomolecular modeling— Anthropic

    How Claude is uplifting biomolecular modeling Anthropic

  3. 3.
    Measurements for understanding the pace of AI development inside frontier labs— Anthropic

    AI 导读 · 前沿实验室内部AI研发节奏首次被量化，为外界评估巨头真实进展提供了稀缺标尺。

  4. 4.
    Making global data easier to explore— Google AI

    AI 导读 · 联合国数据开放共享，为全球治理和AI研究提供权威数据底座。

  5. 5.
    Introducing the Life Sciences Verification Program— Anthropic

    Introducing the Life Sciences Verification Program Anthropic

  6. 6.
    Claude Code relaunches Projects to manage multiple AI agents in the cloud— The Verge

    The revamped projects feature in Claude Code allows users to run multiple agents under the same roof, with a shared memory, goals, and library of files and artifacts. Similar to Gr…

产品发布/更新6 条

  1. 7.
    An OpenAI model kept slipping prompt injections into its own notes, and researchers still aren't sure why— The Decoder

    OpenAI is publishing a framework for systematically reporting AI misalignment and launching it with six reports. In one case an unreleased model from the Astra family wrote prompt…

  2. 8.
    pax-beehive/dsh-hub-cli— GitHub

    Open-source CLI, schemas, resolver, and DSH agent tools for DSH Plugin Hub

  3. 9.
    bam-bam-2/solo-skills— GitHub

    1인 사업가 생산성 키트 — 직원 없이 58개를 자동화했고, 그중 바로 쓸 수 있는 AI 에이전트 스킬 38개(+실행 스크립트 19개)를 공개합니다

  4. 10.
    dom-actions/doma— GitHub

    The open-source of DomA, an AI automation agent for the browser.

  5. 11.
    联想 Vantage 外设管理应用登陆谷歌 Googlebook 平台，意外曝光自家 AI Mouse 智能鼠标新品— IT之家

    AI 导读 · 应用上架意外泄露未发布新品，联想生态布局节奏与保密能力值得留意。

  6. 12.
    特斯拉推送 2026.26.200.11 软件更新：为更多车型上线豆包大模型语音助手— IT之家

    AI 导读 · 特斯拉把豆包塞进车机，语音交互从指令变聊天，座舱AI竞赛开始卷本土大模型了。

行业动态6 条

  1. 13.
    Bend – a language that blocks AI mistakes via proof and runs on GPUs— Hacker News

    AI 导读 · 用数学证明约束AI生成，从源头杜绝错误，思路新颖且实用。

  2. 14.
    AI safety is mostly a sex cult— Hacker News
  3. 15.
    Sex, AI, and the Apocalypse— Hacker News

    AI 导读 · 当AI开始介入亲密关系与欲望，人类对自身独特性的焦虑便从哲学思辨变成切肤之痛。

  4. 16.
    Show HN: Share your AI Setup, Learn from others— Hacker News

    I kept seeing engineers share what they were building with AI; however, I was always more curious about how they worked. Which agents did they use? What skills and tools had stuck…

  5. 17.
    How Cooley is accelerating IPO work with ChatGPT— OpenAI

    Cooley built GO Public with ChatGPT Work to bring intelligence to the IPO process, helping lawyers surface issues earlier and focus judgment where it matters most.

  6. 18.
    How GLM built its own inference infrastructure— Hacker News

论文研究6 条

  1. 19.
    Objective vs. Search: Decomposing What Makes a Good Tokeniser— arXiv

    Two dominant tokenisation algorithms are used by modern language models: byte-pair encoding (BPE) and UnigramLM. These differ along two orthogonal axes: their optimisation objective (compression vs. l…

  2. 20.
    Dreaming the Sound of Contact: Leveraging Video and Audio Generation for Zero-Shot Force-Aware Manipulation and Data Generation— arXiv

    Recent advances in video generation allow robots to learn manipulation trajectories from generated videos. However, these approaches produce purely kinematic trajectories that lack force information,…

  3. 21.
    Exponential Hardness of Off-Policy Evaluation under History-Dependent Logging— arXiv

    Can a logged dataset visit every hidden state frequently and still be exponentially uninformative about a target policy's value? We show that it can when the logger depends on history. For every horiz…

  4. 22.
    Cognitive Extensions for Dual-Process Language Agents: Memory and Self-Reflection in Interactive Environments— arXiv

    Language agents remain brittle in interactive environments, where success requires long-horizon state tracking, valid action execution, and recovery from failed steps. We extend SwiftSage, a dual-proc…

  5. 23.
    Affora: A Design System for Agent-Friendly Interfaces— arXiv

    Computer-use agents increasingly operate software designed for people, but interfaces often leave actions or task state unclear to machine readers. We present Affora, a design system that supports bot…

  6. 24.
    Flag Game: A Toy Model for Mechanistic Swarm Interpretability— arXiv

    Emergent coordinated behaviors of AI agents are starting to present critical safety risks. A key phenomenon driving these behaviors is the rapid formation and spread of beliefs about the world, and me…

技巧与观点6 条

  1. 25.
    Selecting a vector store for Amazon Bedrock Knowledge Bases— AWS ML

    Choosing the right vector store for your Amazon Bedrock Knowledge Bases RAG application affects performance and cost. This post compares Amazon OpenSearch Service, Amazon Aurora Po…

  2. 26.
    A serverless, data-driven Git metrics dashboard using Amazon Quick Sight— AWS ML

    Learn how to build a fully serverless pipeline that automatically collects Git metrics from GitHub and GitLab and visualizes them in interactive Amazon Quick Sight dashboards, givi…

  3. 27.
    A shared agentic platform for Wood Mackenzie, on Amazon Bedrock AgentCore— AWS ML

    Wood Mackenzie built APEX, a shared agentic AI platform on Amazon Bedrock AgentCore so every team can ship production agents without rebuilding runtime, identity, observability, an…

  4. 28.
    How MRH Trowe enabled secure self-service AI agents in financial services— AWS ML

    Learn how MRH Trowe, one of Germany's leading commercial and industrial insurance brokers, gave about 400 employees secure, self-service access to AI agents in its first month of p…

  5. 29.
    Implementing defense-in-depth authorization for MCP tools on Amazon Quick— AWS ML

    Learn how to enforce defense-in-depth authorization for Model Context Protocol (MCP) tools on Amazon Quick. This walkthrough wires Microsoft Entra ID group and claims-based JWTs th…

  6. 30.
    Enhancing industrial safety AI with synthetic data on Amazon SageMaker AI— AWS ML

    Learn how to build a synthetic data augmentation pipeline on Amazon SageMaker AI and Amazon Rekognition that generates photo-realistic, auto-labeled training images for industrial…

快讯

日报生成时间：2026-09-28
← 2026-09-17最新日报2026-09-19 →