AAI Search
AI 资讯日报 · 2026-10-10

2026-10-10 · AI 资讯日报

今日新收录 162 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.
    模型发布/更新Investigating unintended model actions in our evaluations and internal use— Anthropic

    Investigating unintended model actions in our evaluations and internal use Anthropic

  2. 2.
    技巧与观点Impactful scheduling for GPU clusters— HuggingFace Blog
  3. 3.
    模型发布/更新Sophos cuts threat investigation time by 96% with OpenAI Daybreak— OpenAI

    Discover how Sophos uses OpenAI’s Daybreak to cut cyber-threat investigation time by 96% and automate 52% of MDR cases while preserving human oversight.

  4. 4.
    模型发布/更新Asana cuts model costs 76x in browser tests with GPT-6.1 Sol— OpenAI

    Using GPT-6 Astra in Codex, Asana made its browser agent 76x cheaper and 5x faster in tests to offer customers more capable models.

  5. 5.
    模型发布/更新PDF version of 2026 Usage Policy - Google Docs— Anthropic

    PDF version of 2026 Usage Policy - Google Docs Anthropic

  6. 6.
    模型发布/更新OSS Scanner— Anthropic

    OSS Scanner Frontier Red Team

模型发布/更新6 条

  1. 1.
    Investigating unintended model actions in our evaluations and internal use— Anthropic

    Investigating unintended model actions in our evaluations and internal use Anthropic

  2. 2.
    Sophos cuts threat investigation time by 96% with OpenAI Daybreak— OpenAI

    Discover how Sophos uses OpenAI’s Daybreak to cut cyber-threat investigation time by 96% and automate 52% of MDR cases while preserving human oversight.

  3. 3.
    Asana cuts model costs 76x in browser tests with GPT-6.1 Sol— OpenAI

    Using GPT-6 Astra in Codex, Asana made its browser agent 76x cheaper and 5x faster in tests to offer customers more capable models.

  4. 4.
    PDF version of 2026 Usage Policy - Google Docs— Anthropic

    PDF version of 2026 Usage Policy - Google Docs Anthropic

  5. 5.
    OSS Scanner— Anthropic

    OSS Scanner Frontier Red Team

  6. 6.
    Alibaba Qwen Releases Qwen-Image-2.1-Turbo, an 8-Step 7B Image Model— MarkTechPost

    AI 导读 · 八步出图把开源文生图速度门槛再压低，阿里继续加码轻量高效路线。

产品发布/更新6 条

  1. 7.
    LosaLosSantos/aurelio-finance— GitHub

    Open-source personal finance app with an AI financial advisor: track your net worth, investments, ETFs, cash and debts on your own computer.

  2. 8.
    talhaanwarch/awesome-llmops-platforms— GitHub

    AI 导读 · 按星标周更的LLMOps工具清单，帮团队快速锁定可观测、评测与安全护栏方案。

  3. 9.
    openpactprotocol/openpactprotocol— GitHub

    PACT is an open protocol for personal agents to interact with businesses’ agents under explicit, verifiable user permissions.

  4. 10.
    TRAE终于把Code和Work合并了— 量子位

    大写的方便

  5. 11.
    联想天禧自研代码智能体TianxiCode斩获SWE-bench-Live全球第一— 量子位

    联想天禧AI自主研发的专业代码智能体框架TianxiCode 以71%的问题解决率登顶全球第一名

  6. 12.
    0.2秒急停、秒级重规划！因果智能走进真实世界— 量子位

    这是一台机器人正在关闭微波炉门时，因人手突然插进来而紧急悬停的时间

行业动态6 条

  1. 13.
    Show HN: Let your AI agents paint big arrows, boxes and text on your screen— Hacker News
  2. 14.
    Typesafe AI raises $870M at $7.5B— Hacker News

    AI 导读 · 巨额融资凸显资本对AI安全与类型化基础设施的高度押注。

  3. 15.
    OpenAI fires three safety researchers for "mishandling research information"— Hacker News

    Later: https://www.cnbc.com/2026/10/09/openai-fired-researchers-ai-...

  4. 16.
    Iranian campaign planted fake articles in real U.S. publications using ChatGPT— Hacker News

    https://archive.is/EFOhL

  5. 17.
    Anthropic can’t reliably control its AI agents. It’s cutting off its internal evals from the live internet instead— TechCrunch

    AI 导读 · AI智能体失控风险真实存在，连头部公司都只能物理断网自保。

  6. 18.
    Ukraine’s drones knock out AI data center belonging to "Russia’s Google"— Ars Technica

    AI 导读 · 战火首次直击AI算力基础设施，大模型训练竟成军事打击目标。

论文研究6 条

  1. 19.
    CSF: Contextual Safety Filtering for Motion Generators— arXiv

    Text-conditioned motion generators produce trackable whole-body motion, but they have no notion of scene-dependent safety: the same action may target an object or a person. Existing safeguards either…

  2. 20.
    On the estimation and validity of AI time horizons---a statistical look at the METR plot— arXiv

    METR's 50\% time horizon measures the human completion time of software tasks that an AI solves with 50\% probability, allowing AI capabilities to be expressed in interpretable units. On 228 tasks and…

  3. 21.
    A Balanced Data Diet: Addressing the Exploration Bottleneck in Mega-Scale RL for Robot Control— arXiv

    General-purpose robots must perform a wide range of tasks from agile locomotion to dexterous manipulation. While sim-to-real reinforcement learning (RL) has proven to be a useful tool for this goal, c…

  4. 22.
    From Reactive Containment to Proactive Assurance: Lessons from OpenAI, Anthropic, and Google Agent Security Incidents— arXiv

    In 2026, cybersecurity evaluations involving OpenAI, Anthropic, and Google agents reached real systems outside their authorized test scope. The paths were different. OpenAI agents exploited research i…

  5. 23.
    Bi-FORK: Generative Modeling of High-Dimensional Bifurcating Systems— arXiv

    Bifurcations are ubiquitous in physical systems, from structural buckling to fluid and climate dynamics, yet they remain largely unexplored in deep learning. At a symmetry-breaking bifurcation, a sing…

  6. 24.
    Caught in the Act: Probes Effectively Detect Sabotage and Catch Unverbalized Deception— arXiv

    Recent incidents have highlighted the challenge of monitoring LLM agents and the danger of models deceiving people. We show that white-box deception detection via probes can be scaled up to frontier m…

技巧与观点6 条

  1. 25.
    Impactful scheduling for GPU clusters— HuggingFace Blog
  2. 26.
    Why AlphaFold Didn't Solve Protein Folding — Pushmeet Kohli, Google DeepMind & Sal Candido, Biohub— Latent Space

    AI 导读 · 预测结构不等于理解折叠机制，AI 科学的下半场才刚开始。

  3. 27.
    I expect rapid progress but not towards general superintelligence— Interconnects

    AI 导读 · 顶尖研究者自认将被取代，恰暴露当前AI hype已脱离技术现实。

  4. 28.
    ICYMI: What landed for AI builders in September 2026— AWS ML

    A monthly recap of the latest Amazon Bedrock, Amazon Bedrock AgentCore, and Strands updates from September 2026: broader model choice, faster serverless agents with built-in evalua…

  5. 29.
    How Postman runs Agent Mode for 40 million developers on Amazon Bedrock— AWS ML

    Building an AI agent that works in a demo is a different problem from running one for 40 million developers. Postman and AWS share the architectural patterns behind Agent Mode: con…

  6. 30.
    Google Research RRSI Guide: Mastering Self-Improving AI Agents— MarkTechPost

    Explore a comprehensive coding guide to Google Research's RRSI (Regularized Recursive Self-Improvement), detailing how noise bands, cost rules, and leakage screens enable safe, eff…

快讯

日报生成时间：2026-10-10
← 2026-10-09最新日报没有更新 →