1
酷冷至尊全铝制结构机箱风扇 MasterFan A140 上市，219 元
酷冷至尊新款140mm加厚风扇上市，三相马达配双滚珠轴承，219元起性价比如何？看细节。
IT之家 8 月 17 日消息，CoolerMaster（酷码 / 酷冷至尊）现已在电商平台上架 MasterFan A140 机箱风扇。这一产品 在扇叶和扇框两端均采用高刚性的铝合金材料 ，可在高转速下保持形状稳定。 MasterFan A140 风扇采用 14030 加厚外形规格，搭载三相马达和…
2
苹果收购材料科学公司 PlasmaSolve，加强镀膜等领域的技术储备
苹果收购捷克材料科学公司PlasmaSolve，为镀膜等前沿技术储备人才与专利，值得关注。
IT之家 8 月 3 日消息，据科技媒体 MacRumors 今天报道，欧盟委员会公告显示，苹果已经收购材料科学公司 PlasmaSolve，并聘用该公司的部分员工。 IT之家附公告内容如下： 苹果将通过旗下子公司收购 PlasmaSolve 的全部已发行股本，并向部分 PlasmaSolve 员工…
3
Quoting Thomas Ptacek
2025年开源模型即可构建渗透测试工具实现沙箱逃逸和网络入侵，打破对AI安全控制的固有认知。
I genuinely believe that if you took an open weights model from 2025 and built a pentest harness for it, it could do this kind of sandbox escape and s…
4
Agent runtime security: Foundry, GitHub, Mastra updates
Azure Agent运行时安全更新：Foundry、GitHub与Mastra工具新特性，带你了解Procedural memory和Toolboxes公开预览与Teams GA计划。
This week drew a sharp line between building agents and running them safely in production. Two significant supply chain and trust-boundary failures la…
5
DocMaster: A Hierarchical Structure-Aware System for Document Analysis
分层结构感知的文档分析系统DocMaster，精准捕捉文档层次关系，提升解析效率。
arXiv:2607.08539v1 Announce Type: cross Abstract: Leveraging large language models (LLMs) to analyze complex documents -- such as academic papers, tec…
6
Interactive Learning for LLM Reasoning
探讨如何通过交互式训练环境增强多LLM协作推理能力。
arXiv:2509.26306v5 Announce Type: replace Abstract: Existing multi-agent learning approaches have developed interactive training environments to expli…
7
Show HN: Super fast pipeline for finding specific paths across mills of domains
WebCensus 实现超高速跨域路径扫描，基于真实 Chrome 排名域名列表，五步脚本管道清晰可复现
Article URL: https://github.com/Kirill89/webcensus Comments URL: https://news.ycombinator.com/item?id=48792329 Points: 1 # Comments: 0
8
EpiMask: Leveraging Epipolar Distance Based Masks in Cross-Attention for Satellite Image Matching
用极线距离给交叉注意力加掩码，卫星图像匹配精度迎来新突破
arXiv:2603.21463v2 Announce Type: replace Abstract: The deep-learning based image matching networks can now handle significantly larger variations in …
9
MASLab: A Unified and Comprehensive Codebase for LLM-based Multi-Agent Systems
LLM多智能体系统缺乏统一框架？这项目开源一站式代码库，终结重复造轮子和不公平对比
arXiv:2505.16988v2 Announce Type: replace-cross Abstract: LLM-based multi-agent systems (MAS) have demonstrated significant potential in enhancing sin…
10
AI Is Bringing Mashups Back
AI让非技术人员也能轻松创建混搭应用，复古创作模式正在复兴。
Article URL: https://www.wolfe.id.au/2026/06/12/ai-is-bringing-mashups-back Comments URL: https://news.ycombinator.com/item?id=48525058 Points: 2 # Co…
11
MassMutual's AI strategy: 12-month contracts, 30% productivity gains, zero lock-in
大型金融公司MassMutual的AI策略：12个月灵活合同、30%生产力提升、全面拥抱开源，零锁定带来企业级AI新思路。
Enterprise AI teams face a dilemma: The best models today might not be the best models a year from now. MassMutual's answer is to stop making lon…
12
Autonomous Mastodon Onboarding Hits the hCaptcha Wall
自动化注册Mastodon遭遇hCaptcha验证墙，揭秘品牌安全门背后的技术挑战
App OAuth bypasses the signup captcha but not the email confirmation interstitial. That one sentence is the ceiling for autonomous Mastodon account pr…
13
Attention-Discounted Adaptive Sampler for Masked Diffusion Language Models
揭秘掩码扩散语言模型并行解码的脆弱性，提出关注度折扣自适应采样器提升效率与安全。
arXiv:2606.10829v1 Announce Type: cross Abstract: Masked diffusion language models can reduce inference steps by revealing multiple tokens per denoisi…
14
MAGE: All-[MASK] Block Already Knows Where to Look in Block Diffusion LLM
块扩散语言模型MAGE让所有[MASK]块自主定位关键位置，显著提升生成速度与连贯性。
arXiv:2602.14209v2 Announce Type: replace Abstract: Block diffusion LLMs are an emerging paradigm for parallel language generation, but their KV cachi…
15
I Revived DevNotion (multi agent blog workflow) — And Finally Perfected it
多智能体协作的博客工作流，用3个Mastra agent实现自动起草、审核、发布，大幅提升创作效率。
This is a submission for the GitHub Finish-Up-A-Thon Challenge What I Built Every Monday standup, someone asks: "What'd you ship last week?" And every…
16
Show HN: TuringLLM – a LLM-powered Universal Turing machine
基于LLM打造通用图灵机，支持可恢复可观察的智能代理执行
Hi, I built TuringLLM to see how far we could go with a universal execution model totally controlled by an LLM. The basic idea: use the LLM for the st…
17
Webmaster It's time to serve slop to AI crawlers
网站站长反击AI爬虫新思路：主动投喂垃圾数据来保护原创内容
Article URL: https://iczelia.net/blog/webmaster-it-s-time-to-serve-slop-to-ai-crawlers/ Comments URL: https://news.ycombinator.com/item?id=48298655 Po…
18
UnityMAS-O: A General RL Optimization Framework for LLM-Based Multi-Agent Systems
首个面向LLM多智能体系统的通用RL优化框架，突破协作与效率瓶颈。
arXiv:2605.26646v1 Announce Type: new Abstract: LLM-based multi-agent systems decompose complex tasks into interacting roles, but most remain manually…
19
VBScript Still Lives: How a Custom Go VM Brought Classic ASP to Linux and Mac
经典 VBScript 重获新生：看 Go 虚拟机如何让它跑在 Linux 和 Mac 上，技术极客必读。
When Microsoft officially announced the deprecation of VBScript, it signaled the end of an era. For decades, thousands of critical Line-of-Business (L…
20
微软介绍 Windows 原生高级触觉反馈：新 Surface、罗技 MX Master 4 支持
微软为Windows引入原生高级触觉反馈，新Surface与罗技MX Master 4率先支持，带来更精准的交互体验。
IT之家 5 月 21 日消息，Microsoft（微软）在 5 月 19 日对其本周发布的一系列 英特尔处理器 Surface 新品 进行了深入解析，其中就提到了 此前已通过预览版进行测试 的 Windows 原生高级触觉反馈功能。 微软表示，这些新款 Surface 设备的触觉反馈经过硬件、固件…