<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>CheinTian · Clippings</title>
    <link>https://notes.maxt.info/clippings/</link>
    <description>A working notebook on AI systems, databases, and distributed infrastructure.</description>
    <language>zh-CN</language>
    <lastBuildDate>Sat, 07 Nov 2026 12:30:34 GMT</lastBuildDate>
    <atom:link href="https://notes.maxt.info/feed.xml" rel="self" type="application/rss+xml"/>
    <item>
      <title>You can launch 100 coding agents. Who keeps them moving?</title>
      <link>https://notes.maxt.info/clippings/2026-11-07/you-can-launch-100-coding-agents-who-keeps-them-moving/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-07/you-can-launch-100-coding-agents-who-keeps-them-moving/</guid>
      <pubDate>Sat, 07 Nov 2026 00:00:00 GMT</pubDate>
      <description>Six rules for running the software development lifecycle across a fleet of coding agents, November 2026.Over the last 30 days, coding agents...</description>
    </item>
    <item>
      <title>Agents Are Going Async</title>
      <link>https://notes.maxt.info/clippings/2026-11-06/agents-are-going-async/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-06/agents-are-going-async/</guid>
      <pubDate>Fri, 06 Nov 2026 00:00:00 GMT</pubDate>
      <description>Imagine a chief of staff who disappears every time you hand them a big project and only comes back once it&apos;s done. You&apos;d probably fire them ...</description>
    </item>
    <item>
      <title>Computer Use 主流方案与工具调研</title>
      <link>https://notes.maxt.info/clippings/2026-11-06/computer-use-%E4%B8%BB%E6%B5%81%E6%96%B9%E6%A1%88%E4%B8%8E%E5%B7%A5%E5%85%B7%E8%B0%83%E7%A0%94/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-06/computer-use-%E4%B8%BB%E6%B5%81%E6%96%B9%E6%A1%88%E4%B8%8E%E5%B7%A5%E5%85%B7%E8%B0%83%E7%A0%94/</guid>
      <pubDate>Fri, 06 Nov 2026 00:00:00 GMT</pubDate>
      <description>🖥️ Computer Use 主流方案与工具调研（2026-11） [!summary] 一句话结论 Computer Use 已经分成四层：模型层（Claude / GPT / Gemini / UI-TARS / Fara / Qwen-CUA）、Agent Harness 层（Cowork / Agent TARS / Agent S / Browser Use）、环境层（E2B / O…</description>
    </item>
    <item>
      <title>Enterprise Evals 101</title>
      <link>https://notes.maxt.info/clippings/2026-11-06/enterprise-evals-101/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-06/enterprise-evals-101/</guid>
      <pubDate>Fri, 06 Nov 2026 00:00:00 GMT</pubDate>
      <description>Imagine your worst nightmare: your agent passes every demo, then a customer asks something you never tested, and it fails in public.In April...</description>
    </item>
    <item>
      <title>The components of agent memory</title>
      <link>https://notes.maxt.info/clippings/2026-11-06/the-components-of-agent-memory/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-06/the-components-of-agent-memory/</guid>
      <pubDate>Fri, 06 Nov 2026 00:00:00 GMT</pubDate>
      <description>For Claude to learn from experience, it has to decide what’s worth remembering before it knows what’s next. Here’s how “dreaming” helps it organize memory and correct mistakes.</description>
    </item>
    <item>
      <title>Designing an Always-On Agent</title>
      <link>https://notes.maxt.info/clippings/2026-11-05/designing-an-always-on-agent/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-05/designing-an-always-on-agent/</guid>
      <pubDate>Thu, 05 Nov 2026 00:00:00 GMT</pubDate>
      <description>Chatbots wait for a message. An always-on agent is single-threaded, proactive and situationally aware. Here&apos;s how we design them, and what models still lack.</description>
    </item>
    <item>
      <title>Software Factory Engineer: Job Description</title>
      <link>https://notes.maxt.info/clippings/2026-11-05/software-factory-engineer-job-description/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-05/software-factory-engineer-job-description/</guid>
      <pubDate>Thu, 05 Nov 2026 00:00:00 GMT</pubDate>
      <description>An emerging engineering role for the transition from AI-assisted development to L4 Autonomous Software Factories.The RoleWe are looking for ...</description>
    </item>
    <item>
      <title>ThoughtWorks 技术雷达 Vol.35 摘要</title>
      <link>https://notes.maxt.info/clippings/2026-11-05/thoughtworks-%E6%8A%80%E6%9C%AF%E9%9B%B7%E8%BE%BE-vol-35-%E6%91%98%E8%A6%81/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-05/thoughtworks-%E6%8A%80%E6%9C%AF%E9%9B%B7%E8%BE%BE-vol-35-%E6%91%98%E8%A6%81/</guid>
      <pubDate>Thu, 05 Nov 2026 00:00:00 GMT</pubDate>
      <description>ThoughtWorks 技术雷达第 35 期（2026 年 11 月）中文摘要：四大主题、117 个条目与团队行动建议。</description>
    </item>
    <item>
      <title>为团队构建真正的 Agent OS：来自 Ankole 的工程实践</title>
      <link>https://notes.maxt.info/clippings/2026-11-04/%E4%B8%BA%E5%9B%A2%E9%98%9F%E6%9E%84%E5%BB%BA%E7%9C%9F%E6%AD%A3%E7%9A%84-agent-os%E6%9D%A5%E8%87%AA-ankole-%E7%9A%84%E5%B7%A5%E7%A8%8B%E5%AE%9E%E8%B7%B5/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-04/%E4%B8%BA%E5%9B%A2%E9%98%9F%E6%9E%84%E5%BB%BA%E7%9C%9F%E6%AD%A3%E7%9A%84-agent-os%E6%9D%A5%E8%87%AA-ankole-%E7%9A%84%E5%B7%A5%E7%A8%8B%E5%AE%9E%E8%B7%B5/</guid>
      <pubDate>Wed, 04 Nov 2026 00:00:00 GMT</pubDate>
      <description>此为《Building a Real Agent OS for Teams: Lessons from Ankole》的中文翻译版，倘有歧义请以英文原版为准。1. 从单用户 Harness 到集体认知在独立开发者的笔记本电脑上，Claude Code、Codex 与 OpenCl...</description>
    </item>
    <item>
      <title>Agent State Machines: Building Deterministic Scaffolding for Agents</title>
      <link>https://notes.maxt.info/clippings/2026-11-04/agent-state-machines-building-deterministic-scaffolding-for-agents/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-04/agent-state-machines-building-deterministic-scaffolding-for-agents/</guid>
      <pubDate>Wed, 04 Nov 2026 00:00:00 GMT</pubDate>
      <description>If you&apos;re not using some type of agent state machine, you&apos;re completely at the mercy of the nondeterministic nature of models. Are you reall...</description>
    </item>
    <item>
      <title>Concurrent agents on one file system</title>
      <link>https://notes.maxt.info/clippings/2026-11-04/concurrent-agents-on-one-file-system/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-04/concurrent-agents-on-one-file-system/</guid>
      <pubDate>Wed, 04 Nov 2026 00:00:00 GMT</pubDate>
      <description>I think there&apos;s a lot of confusion out there about what a file system really is. There&apos;s lot of places where we can assign blame.People&apos;s fa...</description>
    </item>
    <item>
      <title>The most lucrative AI business to start in 2027</title>
      <link>https://notes.maxt.info/clippings/2026-11-04/the-most-lucrative-ai-business-to-start-in-2027/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-04/the-most-lucrative-ai-business-to-start-in-2027/</guid>
      <pubDate>Wed, 04 Nov 2026 00:00:00 GMT</pubDate>
      <description>The most lucrative skill has always been capturing attention.It’s obvious why.If your goal is making more money, it doesn’t matter how good ...</description>
    </item>
    <item>
      <title>Critical thinking when agents do the work</title>
      <link>https://notes.maxt.info/clippings/2026-11-03/critical-thinking-when-agents-do-the-work/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-03/critical-thinking-when-agents-do-the-work/</guid>
      <pubDate>Tue, 03 Nov 2026 00:00:00 GMT</pubDate>
      <description>Back in the day, critical thinking was taught as approaching new info with humble curiosity and a healthy dose of skepticism - take in diffe...</description>
    </item>
    <item>
      <title>The era of personal software</title>
      <link>https://notes.maxt.info/clippings/2026-11-03/the-era-of-personal-software/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-03/the-era-of-personal-software/</guid>
      <pubDate>Tue, 03 Nov 2026 00:00:00 GMT</pubDate>
      <description>📝 摘要 作者以自己构建的健康数据追踪应用为例，讨论 LLM 时代「个人软件（personal software）」的兴起：当构建成本足够低，软件将逐渐商品化，越来越多人会为自身需求量身定制工具。他完整记录了实现链路——Apple Health 数据经 Health Auto Export 每小时导出 JSON，通过 端点写入 $5 的 PlanetScale Postgres；数据库 sche…</description>
    </item>
    <item>
      <title>Program Status Protocol</title>
      <link>https://notes.maxt.info/clippings/2026-11-02/program-status-protocol/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-02/program-status-protocol/</guid>
      <pubDate>Mon, 02 Nov 2026 00:00:00 GMT</pubDate>
      <description>A terminal escape sequence that lets a program tell the terminal what it is doing: idle, working, waiting on the user, finished, or failed, and why.</description>
    </item>
    <item>
      <title>Sparks of Weight-Space Programming: LLMs can natively read &amp; write neural network weights</title>
      <link>https://notes.maxt.info/clippings/2026-11-02/sparks-of-weight-space-programming-llms-can-natively-read-and-write-neural-network-weights/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-02/sparks-of-weight-space-programming-llms-can-natively-read-and-write-neural-network-weights/</guid>
      <pubDate>Mon, 02 Nov 2026 00:00:00 GMT</pubDate>
      <description>We find that frontier LLMs are surprisingly good at natively reading and writing neural network weights, well past what we’d expect from humans or conventional algorithms.</description>
    </item>
    <item>
      <title>十分钟精读 DSH Super Agent：从 Paper 到 PRD</title>
      <link>https://notes.maxt.info/clippings/2026-11-01/%E5%8D%81%E5%88%86%E9%92%9F%E7%B2%BE%E8%AF%BB-dsh-super-agent%E4%BB%8E-paper-%E5%88%B0-prd/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-01/%E5%8D%81%E5%88%86%E9%92%9F%E7%B2%BE%E8%AF%BB-dsh-super-agent%E4%BB%8E-paper-%E5%88%B0-prd/</guid>
      <pubDate>Sun, 01 Nov 2026 00:00:00 GMT</pubDate>
      <description>1. Super-agent作者认为，以 Coding Agent 为代表的单任务 Agent 已逐渐能够独立完成复杂工作，但现有系统仍以单一 Session 为基本单位。用户并行推进多个任务时，需要自行切换会话、管理上下文；文章称一个人通常只能同时跟进 3～5 个活跃 Sess...</description>
    </item>
    <item>
      <title>Exploring Super-agent Design: Towards a More Autonomous Agent System</title>
      <link>https://notes.maxt.info/clippings/2026-11-01/exploring-super-agent-design-towards-a-more-autonomous-agent-system/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-11-01/exploring-super-agent-design-towards-a-more-autonomous-agent-system/</guid>
      <pubDate>Sun, 01 Nov 2026 00:00:00 GMT</pubDate>
      <description>The original article is in Chinese. This is a machine translation.This article discusses architectural ideas. It is not an introduction to e...</description>
    </item>
    <item>
      <title>New context engineering challenges in late 2026</title>
      <link>https://notes.maxt.info/clippings/2026-10-31/new-context-engineering-challenges-in-late-2026/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-10-31/new-context-engineering-challenges-in-late-2026/</guid>
      <pubDate>Sat, 31 Oct 2026 00:00:00 GMT</pubDate>
      <description>It feels like everyone is converging on the same agent architecture: an LLM, running in a loop, calling tools, running in a sandbox. Claude ...</description>
    </item>
    <item>
      <title>腾讯 2026 AI Agent 企业解决方案及落地实践手册</title>
      <link>https://notes.maxt.info/clippings/2026-10-29/%E8%85%BE%E8%AE%AF-2026-ai-agent-%E4%BC%81%E4%B8%9A%E8%A7%A3%E5%86%B3%E6%96%B9%E6%A1%88%E5%8F%8A%E8%90%BD%E5%9C%B0%E5%AE%9E%E8%B7%B5%E6%89%8B%E5%86%8C/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-10-29/%E8%85%BE%E8%AE%AF-2026-ai-agent-%E4%BC%81%E4%B8%9A%E8%A7%A3%E5%86%B3%E6%96%B9%E6%A1%88%E5%8F%8A%E8%90%BD%E5%9C%B0%E5%AE%9E%E8%B7%B5%E6%89%8B%E5%86%8C/</guid>
      <pubDate>Thu, 29 Oct 2026 00:00:00 GMT</pubDate>
      <description>腾讯、腾讯云与 WorkBuddy 联合发布的 78 页《2026 AI Agent 企业解决方案及落地实践手册》精读：七类提效场景与 WorkBuddy / ClawPro / Agent Runtime / TokenHub 四件套，三阶段跃迁与三代办公，研发运维、SaaS、内容生产、AI 社交陪伴、游戏 NPC、智能客服与生态共建的方案和案例。</description>
    </item>
    <item>
      <title>bliki: LLM Voice</title>
      <link>https://notes.maxt.info/clippings/2026-10-29/bliki-llm-voice/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-10-29/bliki-llm-voice/</guid>
      <pubDate>Thu, 29 Oct 2026 00:00:00 GMT</pubDate>
      <description>The uncanny valley of LLMs&apos; prose</description>
    </item>
    <item>
      <title>Building an AI-native engineering org</title>
      <link>https://notes.maxt.info/clippings/2026-10-29/building-an-ai-native-engineering-org/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-10-29/building-an-ai-native-engineering-org/</guid>
      <pubDate>Thu, 29 Oct 2026 00:00:00 GMT</pubDate>
      <description>Last week Will Larson (@Lethain) asked how leaders are evaluating engineering managers for AI skills. My reply, a list of questions to think...</description>
    </item>
    <item>
      <title>Building fine-grained version control for agents with Rust and Durable Objects</title>
      <link>https://notes.maxt.info/clippings/2026-10-29/building-fine-grained-version-control-for-agents-with-rust-and-durable-objects/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-10-29/building-fine-grained-version-control-for-agents-with-rust-and-durable-objects/</guid>
      <pubDate>Thu, 29 Oct 2026 00:00:00 GMT</pubDate>
      <description>How DeltaDB streams every edit from humans and agents into one causal graph.</description>
    </item>
    <item>
      <title>vLLM Semantic Router Releases d3-ultra: A 27B Open Multimodal Decision Model</title>
      <link>https://notes.maxt.info/clippings/2026-10-27/vllm-semantic-router-releases-d3-ultra-a-27b-open-multimodal-decision-model/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-10-27/vllm-semantic-router-releases-d3-ultra-a-27b-open-multimodal-decision-model/</guid>
      <pubDate>Tue, 27 Oct 2026 00:00:00 GMT</pubDate>
      <description>vLLM Semantic Router released d3-ultra, a 26.09B Apache-2.0 multimodal decision model scoring 87.2 on its own SR-Text benchmark.</description>
    </item>
    <item>
      <title>コードを読まずにマージするために必要なこと</title>
      <link>https://notes.maxt.info/clippings/2026-10-27/%E3%82%B3%E3%83%BC%E3%83%88%E3%82%99%E3%82%92%E8%AA%AD%E3%81%BE%E3%81%99%E3%82%99%E3%81%AB%E3%83%9E%E3%83%BC%E3%82%B7%E3%82%99%E3%81%99%E3%82%8B%E3%81%9F%E3%82%81%E3%81%AB%E5%BF%85%E8%A6%81%E3%81%AA%E3%81%93%E3%81%A8/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-10-27/%E3%82%B3%E3%83%BC%E3%83%88%E3%82%99%E3%82%92%E8%AA%AD%E3%81%BE%E3%81%99%E3%82%99%E3%81%AB%E3%83%9E%E3%83%BC%E3%82%B7%E3%82%99%E3%81%99%E3%82%8B%E3%81%9F%E3%82%81%E3%81%AB%E5%BF%85%E8%A6%81%E3%81%AA%E3%81%93%E3%81%A8/</guid>
      <pubDate>Tue, 27 Oct 2026 00:00:00 GMT</pubDate>
      <description>📝 摘要 本文讨论「no look merge」——开发者完全不读代码就合并 AI 产出的 PR——究竟需要哪些前提。作者认为，仅把质量确认自动化（测试验证行为 + AI 做设计与安全评审、修正后重新验证、存在未执行验证或未解决指摘就不合并）并不足以移除人工审批，真正的障碍在于组织共识：如果漏检后开发负责人仍追究个人责任，那实际上等同于要求人继续读代码。文章引用 2021 年的 Ship / S…</description>
    </item>
    <item>
      <title>bbtree: fastest ordered map</title>
      <link>https://notes.maxt.info/clippings/2026-10-26/bbtree-fastest-ordered-map/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-10-26/bbtree-fastest-ordered-map/</guid>
      <pubDate>Mon, 26 Oct 2026 00:00:00 GMT</pubDate>
      <description>How bbtree, a drop-in replacement for std&apos;s BTreeMap, cuts cache misses, instructions and cache coherence traffic, and how it is kept correct.</description>
    </item>
    <item>
      <title>The new set of problems for engineers to solve</title>
      <link>https://notes.maxt.info/clippings/2026-10-26/the-new-set-of-problems-for-engineers-to-solve/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-10-26/the-new-set-of-problems-for-engineers-to-solve/</guid>
      <pubDate>Mon, 26 Oct 2026 00:00:00 GMT</pubDate>
      <description>A lot of software engineers are going through what can only be described as an identity crisis. The job has changed a lot. Some are depresse...</description>
    </item>
    <item>
      <title>多 Agent 协作终极解：实测 YC 开源新作 Agent Relay，让 Claude Code 和 Codex 直接对话</title>
      <link>https://notes.maxt.info/clippings/2026-10-25/%E5%A4%9A-agent-%E5%8D%8F%E4%BD%9C%E7%BB%88%E6%9E%81%E8%A7%A3%E5%AE%9E%E6%B5%8B-yc-%E5%BC%80%E6%BA%90%E6%96%B0%E4%BD%9C-agent-relay%E8%AE%A9-claude-code-%E5%92%8C-codex-%E7%9B%B4%E6%8E%A5%E5%AF%B9%E8%AF%9D/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-10-25/%E5%A4%9A-agent-%E5%8D%8F%E4%BD%9C%E7%BB%88%E6%9E%81%E8%A7%A3%E5%AE%9E%E6%B5%8B-yc-%E5%BC%80%E6%BA%90%E6%96%B0%E4%BD%9C-agent-relay%E8%AE%A9-claude-code-%E5%92%8C-codex-%E7%9B%B4%E6%8E%A5%E5%AF%B9%E8%AF%9D/</guid>
      <pubDate>Sun, 25 Oct 2026 00:00:00 GMT</pubDate>
      <description>这周，美国创业孵化器 Y Combinator 2026 年秋季营（YC Fall 2026）的创业公司 Agent Workforce 发布了开源项目 Agent Relay。公司的两位联合创始人分别是 CEO Will Washburn 和 CTO Khaliq Gant，其...</description>
    </item>
    <item>
      <title>为什么我看好 Turbopuffer，从 Personal Agent 说起</title>
      <link>https://notes.maxt.info/clippings/2026-10-25/%E4%B8%BA%E4%BB%80%E4%B9%88%E6%88%91%E7%9C%8B%E5%A5%BD-turbopuffer%E4%BB%8E-personal-agent-%E8%AF%B4%E8%B5%B7/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-10-25/%E4%B8%BA%E4%BB%80%E4%B9%88%E6%88%91%E7%9C%8B%E5%A5%BD-turbopuffer%E4%BB%8E-personal-agent-%E8%AF%B4%E8%B5%B7/</guid>
      <pubDate>Sun, 25 Oct 2026 00:00:00 GMT</pubDate>
      <description>江湖小虾米</description>
    </item>
    <item>
      <title>How I work with Agents</title>
      <link>https://notes.maxt.info/clippings/2026-10-25/how-i-work-with-agents/</link>
      <guid isPermaLink="true">https://notes.maxt.info/clippings/2026-10-25/how-i-work-with-agents/</guid>
      <pubDate>Sun, 25 Oct 2026 00:00:00 GMT</pubDate>
      <description>... as of October 2026.Use the latest and greatest models. Everything else is potentially a waste of time. If a model is wrong 10% of the ti...</description>
    </item>
  </channel>
</rss>
