<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Agent-Runtime on 兜忆轩 | Cedric's Compiler Notes</title><link>https://douyixuan.github.io/tags/agent-runtime/</link><description>Recent content in Agent-Runtime on 兜忆轩 | Cedric's Compiler Notes</description><image><title>兜忆轩 | Cedric's Compiler Notes</title><url>https://douyixuan.github.io/images/papermod-cover.png</url><link>https://douyixuan.github.io/images/papermod-cover.png</link></image><generator>Hugo -- 0.146.0</generator><language>en-us</language><lastBuildDate>Sun, 13 Sep 2026 08:00:00 +0800</lastBuildDate><atom:link href="https://douyixuan.github.io/tags/agent-runtime/index.xml" rel="self" type="application/rss+xml"/><item><title>AI Native Digest 2026-09-13</title><link>https://douyixuan.github.io/posts/ai-native/2026-09-13/</link><pubDate>Sun, 13 Sep 2026 08:00:00 +0800</pubDate><guid>https://douyixuan.github.io/posts/ai-native/2026-09-13/</guid><description>今天关注 AI-native 软件工程的两个新方向：Cursor Projects 把 coding agent 从单次会话推进为可持续数月、拥有共享上下文和 coordinator/subagent 架构的长期工作单元；Cognition 的 SWE-2 则直接把任务成功率与推理成本一起写进 RL 目标，说明 coding agent 的优化正在从单纯追求 benchmark 分数转向整条 cost-performance Pareto frontier。</description></item></channel></rss>