AI and Programming
Pi Durable Pi 1.0。Pi Durable 专为长期运行、持久且易于修改的智能体（agent）而构建，可在任何地方运行。
最提倡用 AI 的CEO，如今被 AI 的“垃圾手榴弹”炸懵了 什么是AI垃圾手榴弹？ https://noslopgrenade.com/
Graphify: Unifying Codebase Context to Streamline Agentic Software Engineering
China’s AI Safety Governance Framework 3.0 人工智能安全治理框架3.0
What We Can Learn from Claude’s Fable 5.1 System Prompt
Multi-Agent Teams Hold Experts Back 对话分析揭示了团队倾向于进行整合性妥协——即平均专家和非专家的观点，而非适当地加权专业知识——这种倾向随团队规模的扩大而增加，并与表现呈负相关
Which LLM is worth it Best model for your budget
文艺复兴科技 CEO：如何用科学方法做投资 Peter Brown 是 Renaissance Technologies 首席执行官
How Bridgewater Built an AI Analyst That Does Hours of Expert Research in Minutes
From Raw Data to Graph-Native AI Why graph modeling deserves as much attention as graph learning, GraphRAG, agents, and graph foundation models
Univer 面向 AI Agent 的 Office Harness https://office.univer.ai/
How to Draft an AI Policy in Late 2026
The Future of Platform Engineering report Among teams reporting mandatory adoption, 62.2% achieved most of their goals (high-performers). For respondents reporting optional platform adoption, this sits at roughly 27.5%.
Build adaptive AI interfaces with the AG-UI protocol, agent swarms, and Nova Act on AWS
VoiceStudio 开源的声音克隆、声音设计、视频配音、语音听写、转录与有声书创作
Small Decisions: Engineering a Leading Model
Getting the Source Right, Not Just the Fact: Source-Aware Verification for MCP Agents
How to train your own Jev for $17
Others
帕金森琐碎定理 或称鸡毛蒜皮定律、芝麻绿豆定律，由英国历史学者与政治学者西里尔·诺斯古德·帕金森（Cyril Northcote Parkinson）于1957年所提出，用来说明大型组织会花费大量时间在讨论无关紧要的琐事，但是真正重大的决议反而可以轻松过关这种现象。这是由于人对大议题较难有全面性的理解，故怕贸然提出异议，可能会失言；相反地，对于一些简单琐碎的小事，有相当的认识，因此意见特别多，造成组织在各事项上讨论所花费之时间，通常与事项本身的重要程度呈现反比。