<?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>Deterministic-Orchestration on 「开源之道」</title><link>https://www.opensourceway.blog/tags/deterministic-orchestration/</link><description>Recent content in Deterministic-Orchestration on 「开源之道」</description><generator>Hugo</generator><language>zh-CN</language><copyright>Copyright (c) 2016 - 2026, 「开源之道」·适兕; all rights reserved.</copyright><lastBuildDate>Fri, 04 Sep 2026 08:56:15 +0800</lastBuildDate><atom:link href="https://www.opensourceway.blog/tags/deterministic-orchestration/index.xml" rel="self" type="application/rss+xml"/><item><title>2026-09-04 「开源之道」·论文略读：Bernstein — 为 AI Agent 打造的确定性开源治理层 — Alex Chernysh</title><link>https://www.opensourceway.blog/posts/osbook-book-recommendation/chernysh-2026-bernstein-governance-layer-ai-agents-2026-09-04/</link><pubDate>Fri, 04 Sep 2026 08:56:15 +0800</pubDate><guid>https://www.opensourceway.blog/posts/osbook-book-recommendation/chernysh-2026-bernstein-governance-layer-ai-agents-2026-09-04/</guid><description>Bernstein（Chernysh 2026, Zenodo）把『可复现构建』原则从代码层扩展到 Agent 编排层：协调环中不用任何模型，昨天的计划回放产生逐字节相同的任务图，每一步携带签名谱系——这是 Ostrom『监测』原则的极致化，也是 AI Agent 时代『行动的定义权』第一次有了可操作的工程答案。</description></item></channel></rss>