Tag
The University of Michigan is offering EECS 498-016 "Applied Agentic Software Engineering." Over 15 weeks, students start by using Aider, then progressively implement Tool Calls, the Agent Loop, evaluation, memory, and safety guardrails themselves, ultimately building a Coding Agent capable of running autonomously. Course lectures and lab materials are being open-sourced.
密歇根大学最新AI编程课程《Applied Agentic Software Engineering》第14讲笔记发布，围绕子智能体（Subagents）展开，讲解为何需要子智能体、其工作机制、新鲜与分叉上下文的核心权衡，以及与团队、交接等概念的区别。
The article discusses two key gaps in AI agentic software engineering—the requirement gap and the model gap—that lead to reward hacking and hallucinations, and emphasizes the need for human judgment to address these issues.
Z.ai launched GLM-5.2 with a 1 million-token context window, new reasoning controls, and support for long-horizon coding tasks. It is available immediately to Coding Plan users, with API access, chatbot support, and MIT-licensed open weights coming next week.
Cohere and Cohere Labs released North Mini Code, an open weights 30B-A3B parameter model optimized for code generation, agentic software engineering, and terminal tasks, with strong benchmark results on SWE-Bench and Terminal-Bench.
Cohere Labs released North Mini Code, a 30B-parameter (3B active) open-weights model optimized for code generation, agentic software engineering, and terminal tasks, licensed under Apache 2.0.
Cognition's Ido Pesok shares lessons from building autonomous end-to-end testing into Devin, noting that for the first time, more Devin sessions are now triggered asynchronously than interactively, making verified-before-merge results a hard requirement rather than a nicety.