1
Windows 任务管理器之父打造：Win11 版 TMOG 发布，107 页文档让 AI 编码
IT之家 8 月 26 日消息，微软 Windows 退休工程师、任务管理器之父戴夫・普拉默（Dave W. Plummer）于 8 月 24 日在 X 平台发布推文， 宣布面向 Windows 11、Linux 系统发布原生任务管理器 TMOG，最新版本号为 0.1.1。 普拉默在推文中表示，19…
2
Activation-Weighted Seeded Residual Coding for Low-Bit LLM Weight Repair
一种结合激活加权与残差编码的低比特LLM权重修复新法，为压缩模型性能恢复提供创新思路。
arXiv:2608.23144v1 Announce Type: new Abstract: Low-bit weight quantization saves storage but leaves errors that degrade language-model quality. We in…
3
Architecture as Capability Equalizer for Coding Agents
揭秘架构如何成为编码智能体的能力均衡器，为提升代码生成质量提供新视角。
arXiv:2608.21747v1 Announce Type: cross Abstract: LLM-based coding agents generate complete software systems from high-level descriptions, yet little …
4
CyrillicQA: The Influence of Phonetically Encoded Secret Language on LLM Performance
用西里尔字母谐音伪装提问，检验大模型能否破解这层“密语”并保持问答性能，脑洞与严谨兼备
arXiv:2608.21462v1 Announce Type: cross Abstract: Due to the selection of their training data, large language models (LLMs) perform best on standard-l…
5
PromptResponse: Optimizing Prompts for LLM Coding Tasks
格式与提示词微观调整如何影响大模型编码表现？这项对照实验给出可复现的优化路径。
arXiv:2608.21074v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used in research workflows and software development pi…
6
PrimeAgentOrchestrator: Memory-Primed Agent Spawning for Personal AI Infrastructure
编码代理失忆难题有解了：为新会话预载记忆，让AI助手接续上下文
arXiv:2608.20342v1 Announce Type: new Abstract: Large language model (LLM) coding agents start each session with an empty context window, discarding a…
7
Why Your AI Coding Agent Should Never See Your .env
AI编码助手可能一键泄露你的.env密钥，这份安全指南教你如何防住。
Why Your AI Coding Agent Should Never See Your .env Your AI agent uses your API keys. It NEVER sees them. Not in context. Not in logs. Not in chat. No…
8
Slack wants to drag AI coding out of the terminal and into the group chat
Slack把AI编程代理拖进群聊，边聊边写代码，协作效率拉满
Slack wants to drag AI coding out of the terminal and into the group chat. The Salesforce-owned messaging platform today announced Slack Code , a new …
9
Show HN: Epho – run Claude Code with a curl
一条curl命令即可在云端沙箱运行Claude Code等编码代理，免去配置环境的麻烦。
Hey folks, Burak here. Epho is an API that allows running Claude Code, Codex or Opencode in a sandbox in the cloud. It abstracts away sandboxes, and a…
10
科技爱好者周刊（第 409 期）：程序员的职业未来
AI大模型与编码代理的双重冲击下，程序员如何应对未来？本期周刊还教你一招动态开放SSH端口的妙计。
这里记录每周值得分享的科技内容，周五发布。...
11
Show HN: CrewCode – Open-source Mission Control for AI coding agents
开源多智能体编程控制台，跨平台管理多个AI编码代理，Git工作树协作不失控，适合团队提效。
I've been experimenting heavily with coding agents and found myself managing agents more than writing code. I build CrewCode to solve that. Its an ope…
12
DeepSeek Harness (dsh) คืออะไร? เจาะลึกคู่แข่งโค้ดโอเพ่นซอร์สของ Claude
DeepSeek Harness 开源工具，让模型化身编码代理，被视作 Claude Code 的开源劲敌，值得一试。
DeepSeek เปิดตัว DeepSeek Harness ( dsh ) เมื่อวันที่ 13 สิงหาคม 2026 ซึ่งไม่ใช่โมเดล แต่เป็นเครื่องมือสำหรับรันโมเดลให้ทำงานเป็น coding agent ได้จริง…
13
Inside the Tokenizer: Why the Same Prompt Costs Different Amounts on Every Model
揭秘不同大模型计费差异的根源，读懂Tokenizer机制帮你节省API调用成本
If you build with LLMs, you pay by the token. Not by the word, not by the character — the token. And yet most of us treat the tokenizer as a black box…
14
MoRFI: Monotonic Sparse Autoencoder Feature Identification
揭秘大模型幻觉根源：MoRFI方法用单调稀疏自编码器精准锁定知识特征，为可解释性带来新突破。
arXiv:2604.26866v2 Announce Type: replace-cross Abstract: Large language models (LLMs) acquire most of their factual knowledge during the pre-training…
15
Programmatic memory for long-horizon LLM agents
长时任务中让智能体“记性好”的新方案：PRO-LONG 用程序化记忆显著提升编码表现，费用还省数倍。
Article URL: https://github.com/alexisfox7/PRO-LONG Comments URL: https://news.ycombinator.com/item?id=49329010 Points: 1 # Comments: 0
16
Engineering Signals of Human-AI Collaboration in the Agentic Coding Era: A Longitudinal Analysis of 33,228 Pull Requests from vLLM and SGLang with Implications for Biomedical AI Agents and Bioinformatics Pipeline Developmen
分析3.3万个PR，揭示智能体编程时代人机协作信号，为生物医学AI与生信开发提供实证参考。
arXiv:2608.13884v1 Announce Type: cross Abstract: The rapid adoption of AI coding assistants and autonomous agentic development systems has coincided …
17
Show HN: Rungraph – See your AI coding-agent runs as a graph
把AI编程代理的每次运行变成可视化图谱，快速定位问题，提升调试效率。
Article URL: https://github.com/fayzan123/rungraph Comments URL: https://news.ycombinator.com/item?id=49321388 Points: 1 # Comments: 0
18
ProofRun – a local verification receipt for AI coding agents
给你的AI编码代理加上加密验证收据，让“测试通过”不再只凭一句空话。
Article URL: https://github.com/yebiguo/ProofRun Comments URL: https://news.ycombinator.com/item?id=49316605 Points: 2 # Comments: 0
19
Spec-Driven Development in the Age of AI: OpenSpec vs. GitHub Spec Kit
AI编程会话一长就丢失需求？用规范驱动开发把决策固化进Git，对比OpenSpec与GitHub Spec Kit，告别千行diff的噩梦。
Originally published on labitcode.com . In the early wave of generative AI, the software industry embraced "vibe coding" — prompting an LLM in an open…
20
QuoteBench: How Matched Scores Can Hide Command-Path Failures
匹配分数会掩盖命令路径失败？QuoteBench用最终状态验证揭示LLM编码代理的真实边界。
arXiv:2608.13547v1 Announce Type: new Abstract: LLM coding agents issue Bash commands through interfaces that may serialize, wrap, and reparse model o…