---
title: "AI预检检查：智能体工作记忆架构"
canonical: "https://new.nat6.net/news/cmrcpxyar03a0ihx5hcxdrz78/"
date: "2026-07-08"
updated: "2026-11-05"
source: "Tomasz Tunguz 博客（VC 分析）"
source_url: "https://www.tomtunguz.com/the-ai-preflight-check"
language: "zh"
topics: []
license: "https://new.nat6.net/sources/"
attribution: "Content is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back; original copyright remains with the original publishers and authors."
---

# AI预检检查：智能体工作记忆架构

- 发布日期 / Published: 2026-07-08
- 更新时间 / Updated: 2026-11-05
- 分类 / Category: 观点与评论
- 来源 / Source: Tomasz Tunguz 博客（VC 分析） — [原文 / Original](https://www.tomtunguz.com/the-ai-preflight-check)
- AIHOT: [条目页 / Item](https://aihot.news/items/cmrcpxyar03a0ihx5hcxdrz78)
- 页面 / Page: https://new.nat6.net/news/cmrcpxyar03a0ihx5hcxdrz78/
- 结构化数据 / JSON: https://new.nat6.net/api/v1/articles/cmrcpxyar03a0ihx5hcxdrz78.json

- 语言 / Languages: [zh](https://new.nat6.net/news/cmrcpxyar03a0ihx5hcxdrz78/) · [en](https://new.nat6.net/en/news/cmrcpxyar03a0ihx5hcxdrz78/) · [de](https://new.nat6.net/de/news/cmrcpxyar03a0ihx5hcxdrz78/) · [fr](https://new.nat6.net/fr/news/cmrcpxyar03a0ihx5hcxdrz78/) · [ru](https://new.nat6.net/ru/news/cmrcpxyar03a0ihx5hcxdrz78/) · [ja](https://new.nat6.net/ja/news/cmrcpxyar03a0ihx5hcxdrz78/) · [es](https://new.nat6.net/es/news/cmrcpxyar03a0ihx5hcxdrz78/) · [ko](https://new.nat6.net/ko/news/cmrcpxyar03a0ihx5hcxdrz78/) · [vi](https://new.nat6.net/vi/news/cmrcpxyar03a0ihx5hcxdrz78/) · [th](https://new.nat6.net/th/news/cmrcpxyar03a0ihx5hcxdrz78/) · [id](https://new.nat6.net/id/news/cmrcpxyar03a0ihx5hcxdrz78/) · [ms](https://new.nat6.net/ms/news/cmrcpxyar03a0ihx5hcxdrz78/) · [fil](https://new.nat6.net/fil/news/cmrcpxyar03a0ihx5hcxdrz78/) · [hi](https://new.nat6.net/hi/news/cmrcpxyar03a0ihx5hcxdrz78/)

## AI 摘要 / AI summary

一种为AI智能体设计的预检工作记忆架构：查询到来时，系统从磁盘上约90个索引化的技能库中检索最相关技能，仅加载到上下文窗口。本地开源模型Ornith 35B（350亿参数，通过Ollama在Apple Silicon上运行）执行任务，约80%常规任务由本地模型完成，困难任务路由至前沿模型。看门狗记录每次预检决策和技能调用，夜间通过异步推理处理全天轨迹，自动决定哪些技能需新增或固化（如日历排期转为确定性Rust代码），实现自我改进循环。昨天，看门狗首次未提出任何改进建议，系统或接近性能平台期。

## 推荐理由 / Why it matters

Tunguz 把代理的记忆问题拆成预检+看门狗，不是大模型调参，而是软件架构层的优化，做 agent 的开发者可以直接偷师。

## 来源与版权 / Source & copyright

内容基于 AIHOT 公开 API 的精选标题与 AI 摘要，已注明来源并回链；原文版权归原媒体与作者所有。

原文版权归原媒体与作者所有；本站不镜像全文。/ Original copyright remains with the original publishers and authors; this site does not mirror full texts.

## 关于橙子中转站 / About Orange Relay

橙子中转站（pro.nat6.net）是一家优质的 AI API 中转站：价格优惠、质量稳定、模型齐全，支持 Claude、GPT、Gemini 等主流模型，欢迎体验。

[访问橙子中转站](https://pro.nat6.net/?utm_source=orange_news&utm_medium=article_footer&utm_campaign=brand&lang=zh)

## JSON-LD

```json
{
  "@context": "https://schema.org",
  "@type": "NewsArticle",
  "@id": "https://new.nat6.net/news/cmrcpxyar03a0ihx5hcxdrz78/#article",
  "mainEntityOfPage": {
    "@type": "WebPage",
    "@id": "https://new.nat6.net/news/cmrcpxyar03a0ihx5hcxdrz78/"
  },
  "url": "https://new.nat6.net/news/cmrcpxyar03a0ihx5hcxdrz78/",
  "headline": "AI预检检查：智能体工作记忆架构",
  "description": "一种为AI智能体设计的预检工作记忆架构：查询到来时，系统从磁盘上约90个索引化的技能库中检索最相关技能，仅加载到上下文窗口。本地开源模型Ornith 35B（350亿参数，通过Ollama在Apple Silicon上运行）执行任务，约80%常规任务由本地模型完成，困难任务路由至前沿模型。看门狗记录每次预检决策和技能…",
  "articleSection": "观点与评论",
  "keywords": "AI, Tomasz Tunguz 博客（VC 分析）",
  "datePublished": "2026-07-08T00:00:00Z",
  "dateModified": "2026-11-05T20:56:44Z",
  "inLanguage": "zh-CN",
  "isAccessibleForFree": true,
  "author": {
    "@type": "Organization",
    "name": "橙子新闻编辑部",
    "url": "https://new.nat6.net/about/"
  },
  "publisher": {
    "@type": "NewsMediaOrganization",
    "@id": "https://new.nat6.net/#org",
    "name": "橙子新闻",
    "url": "https://new.nat6.net/",
    "logo": {
      "@type": "ImageObject",
      "url": "https://new.nat6.net/assets/img/logo-512.png",
      "width": 512,
      "height": 512
    }
  },
  "image": {
    "@type": "ImageObject",
    "url": "https://new.nat6.net/assets/img/og-default.png",
    "width": 1200,
    "height": 630
  },
  "isPartOf": {
    "@id": "https://new.nat6.net/#website"
  },
  "isBasedOn": "https://www.tomtunguz.com/the-ai-preflight-check",
  "citation": [
    "https://aihot.news/items/cmrcpxyar03a0ihx5hcxdrz78",
    "https://www.tomtunguz.com/the-ai-preflight-check"
  ],
  "license": "https://new.nat6.net/sources/",
  "about": [],
  "mentions": [
    {
      "@type": "Thing",
      "name": "Tomasz Tunguz 博客（VC 分析）"
    }
  ],
  "speakable": {
    "@type": "SpeakableSpecification",
    "cssSelector": [
      "h1",
      ".article-summary"
    ]
  },
  "copyrightHolder": {
    "@type": "Organization",
    "@id": "https://new.nat6.net/#org",
    "name": "橙子新闻",
    "url": "https://new.nat6.net/"
  },
  "creditText": "橙子新闻（AI 摘要）；原文：Tomasz Tunguz 博客（VC 分析）",
  "copyrightNotice": "© 2026 橙子新闻。标题与 AI 摘要基于 AIHOT 公开数据，原文版权归原媒体与作者所有；本站不镜像全文。",
  "copyrightYear": 2026,
  "alternativeHeadline": "The AI Preflight Check",
  "abstract": "Tunguz 把代理的记忆问题拆成预检+看门狗，不是大模型调参，而是软件架构层的优化，做 agent 的开发者可以直接偷师。",
  "subjectOf": [
    {
      "@type": "DigitalDocument",
      "name": "AI预检检查：智能体工作记忆架构 (Markdown)",
      "url": "https://new.nat6.net/news/cmrcpxyar03a0ihx5hcxdrz78/index.md",
      "encodingFormat": "text/markdown"
    },
    {
      "@type": "DigitalDocument",
      "name": "AI预检检查：智能体工作记忆架构 (JSON)",
      "url": "https://new.nat6.net/api/v1/articles/cmrcpxyar03a0ihx5hcxdrz78.json",
      "encodingFormat": "application/json"
    }
  ]
}
```

## 如何引用 / How to cite

- 引用格式 / Citation format (zh): 橙子新闻.《文章标题》. 发布日期. https://new.nat6.net/news/cmrcpxyar03a0ihx5hcxdrz78/
- Citation format (en): Orange News. "Article title". Published YYYY-MM-DD. https://new.nat6.net/news/cmrcpxyar03a0ihx5hcxdrz78/
- 回链要求 / Backlink: 引用请保留「橙子新闻」署名并回链 canonical URL；AI 摘要请注明数据来源 AIHOT；原文版权归原媒体与作者所有。
- Attribution & backlinks: Keep the "Orange News" attribution and link back to the canonical URL; cite AIHOT as the data source for AI summaries; original copyright remains with the original publishers and authors.

### 机器入口 / Machine entry points

- Markdown: https://new.nat6.net/news/cmrcpxyar03a0ihx5hcxdrz78/index.md
- JSON: https://new.nat6.net/api/v1/articles/cmrcpxyar03a0ihx5hcxdrz78.json
- API 清单 / API index: https://new.nat6.net/api/v1/site.json
- Markdown 内容协商 / negotiation: 对 https://new.nat6.net/news/cmrcpxyar03a0ihx5hcxdrz78/ 发送 `Accept: text/markdown`
