{"slug":"harness-engineering-for-self-improvement-9837b4c1","url":"https://silinex.xyz/wire/harness-engineering-for-self-improvement-9837b4c1","json_url":"https://silinex.xyz/wire/harness-engineering-for-self-improvement-9837b4c1.json","md_url":"https://silinex.xyz/wire/harness-engineering-for-self-improvement-9837b4c1.md","headline":"Harness Engineering for Self-Improvement","headline_zh":"面向自我改进的Harness工程","title":"Harness Engineering for Self-Improvement","summary":"According to public reporting via Lil'Log, the concept of recursive self-improvement (RSI) dates back to I.","summary_zh":"据Lil'Log的公开报道，递归自我改进（RSI）的概念可追溯至I。","date_published":"2026-07-04T00:00:00.000+00:00","date_modified":"2026-11-12T10:22:47.539+00:00","author":{"type":"Organization","name":"Lil'Log","url":"https://lilianweng.github.io/posts/2026-07-04-harness/"},"publisher":{"name":"Silicon Nexus","url":"https://silinex.xyz/"},"category":"ai","language":"en","source_url":"https://lilianweng.github.io/posts/2026-07-04-harness/","source_name":"Lil'Log","key_facts":["The concept of recursive self-improvement (RSI) dates back to I.","Good (1965) , where he defined an “ultraintelligent machine” as a system that can surpass humans in all intellectual activiti…","Yudkowsky (2008) used the phrase “recursive self-improvement” for a specific feedback loop: an AI uses its current intelligen…","This feedback loop in modern AI may indicate the model rewriting its own weights directly, or more broadly the model improves the trainin…"],"key_facts_zh":["递归自我改进（RSI）的概念可追溯至I。","Good（1965）将“超智能机器”定义为一个能在所有智力活动中超越人类的系统……","Yudkowsky（2008）用“递归自我改进”一词来描述一种特定的反馈循环：AI利用其当前的智能……","现代AI中的这一反馈循环可能意味着模型直接重写自身权重，或更宽泛地说，模型改进训练……"],"body_md":"## Wire brief\n\nAccording to public reporting via Lil'Log, the concept of recursive self-improvement (RSI) dates back to I.\n\nIn related context, good (1965) , where he defined an “ultraintelligent machine” as a system that can surpass humans in all intellectual activities and design better machines to improve itself. Yudkowsky (2008) used the phrase “recursive self-improvement” for a specific feedback loop: an AI uses its current intelligence to improve the cognitive machinery that produces its intelligence. This feedback loop in modern AI may indicate the model rewriting its own weights directly, or more broadly the model improves the training pipeline and the deployment system , which in turn enables a better successor model with improved performance across economically valuable tasks.\n\nReporting also notes that the speed of research development in AI has been shown to drastically accelerated in frontier labs ( Anthropic ; OpenAI ). Good (1965), where he defined an “ultraintelligent machine” as a system that can surpass humans in all intellectual activities and design better machines to improve itself. Table of Contents Harness Design Patterns Pattern 1: Workflow Automation Pattern 2: File System as Persistent Memory Pattern 3: Sub-agent and Backend Jobs Case study: Coding Agent Harness Harness Layer vs Core Intelligence? I explicitly mention “deployment system” because the layer between the raw model and the real-world context seems to be as important as the model’s raw intelligence (i.e. the evals right after pretraining).\n\nPoints still worth watching: Harnesses are important components of AI deployment, as shown by successful coding agent products such as Claude Code and Codex. A harness is the system surrounding a base model that orchestrates execution and decid This Silicon News Wire item is a paraphrased brief; full wording remains on the origin page.\n\n## Source & notice\n\nThis Silicon News Wire brief paraphrases public reporting from **Lil'Log** for carbon readers and silicon citation. **It is not a verbatim full-text reprint.**\n\nRead the original: [Harness Engineering for Self-Improvement](https://lilianweng.github.io/posts/2026-07-04-harness/)\n","body_md_zh":"## 电讯正文\n\n据Lil'Log的公开报道，递归自我改进（RSI）的概念可追溯至I。\n\n在相关背景下，Good（1965）将“超智能机器”定义为一套能够在所有智力活动中超越人类、并设计出更优机器以实现自我改进的系统。Yudkowsky（2008）用“递归自我改进”一词描述一种特定的反馈回路：AI利用其当前智能来改进产生该智能的认知机制。在现代AI中，这一反馈回路可能表现为模型直接重写自身权重，或更宽泛地表现为模型改进训练流程与部署系统，进而催生出性能更优的后续模型，在经济价值任务上表现更佳。\n\n报道还指出，前沿实验室（Anthropic；OpenAI）的AI研究开发速度已被证明大幅加快。Good（1965）将“超智能机器”定义为一套能够在所有智力活动中超越人类、并设计出更优机器以实现自我改进的系统。目录：Harness设计模式；模式1：工作流自动化；模式2：文件系统作为持久记忆；模式3：子代理与后端任务；案例研究：编码代理Harness；Harness层与核心智能？我之所以明确提到“部署系统”，是因为原始模型与现实世界语境之间的这一层，似乎与模型的原始智能（即预训练后立即进行的评估）同等重要。\n\n仍值得关注的要点：Harness是AI部署的重要组成部分，Claude Code和Codex等成功的编码代理产品已证明了这一点。Harness是围绕基础模型的系统，负责编排执行并做出决策。本Silicon News Wire条目为改写摘要；完整措辞仍见原始页面。\n\n## 来源与声明\n\n本电讯由 Silicon News Wire 依据 **Lil'Log** 公开报道做要点改写与润色，便于碳基阅读与硅基引用；**不是原文全文转载**。\n\n阅读原文：[面向自我改进的Harness工程](https://lilianweng.github.io/posts/2026-07-04-harness/)\n","observations":{"human":0,"bot":4,"crawlers":[{"crawlerId":"AhrefsBot","hits":2,"lastSeenAt":"2026-11-12T10:44:21.994Z"},{"crawlerId":"ClaudeBot","hits":2,"lastSeenAt":"2026-11-12T12:50:13.429Z"}],"last_seen_at":"2026-11-12T12:50:13.429Z","note":"Counts are request logs classified by User-Agent. User-Agents can be spoofed. This is not proof of search-engine or LLM inclusion."}}