Jensen Huang has resigned. The available material does not provide reliable details about the timing, reasons, or succession.黄仁勋已辞职。现有材料没有提供有关时间、原因或继任安排的可靠细节。
Jensen Huang has resigned. The available material does not provide reliable details about the timing, reasons, or succession. The departure marks a leadership change at Nvidia.黄仁勋已辞职。现有材料没有提供有关时间、原因或继任安排的可靠细节。这意味着英伟达将发生领导层变动。
Patrick Stewart, beloved for Star Trek's Captain Picard and a long career in theatre and film, has died.帕特里克·斯图尔特因病去世,曾因星际迷航中帕克船长等角色而家喻户晓。
Patrick Stewart, beloved for Star Trek's Captain Picard and a long career in theatre and film, has died. His work helped redefine leadership in science fiction and mentored generations of actors. The news prompts reflection on a prolific career that bridged stage, screen, and iconic franchises.帕特里克·斯图尔特因病去世,曾因星际迷航中帕克船长等角色而家喻户晓。其横跨舞台、电视和电影的职业生涯对科幻与表演行业产生深远影响。人们由此缅怀这位多产演员及其对文化的贡献。
OpenSSH on musl Linux has reportedly allowed access to passwordless accounts since 2016.据报道,musl Linux 上的 OpenSSH 自 2016 年起可能允许进入无密码账户。
OpenSSH on musl Linux has reportedly allowed access to passwordless accounts since 2016. The issue relates to how Unix historically represented accounts that should never permit login, including accounts marked with an asterisk in the password file. The behavior can therefore expose accounts that were intended to be inaccessible.据报道,musl Linux 上的 OpenSSH 自 2016 年起可能允许进入无密码账户。问题与 Unix 处理本应禁止登录的账户的历史方式有关,包括在密码文件中用星号标记的账户。因此,原本应不可访问的账户可能被暴露。
The source reports that frontier LLMs can read and write neural-network weights directly.资料称,前沿大语言模型已经能够直接读取和写入神经网络权重。
The source reports that frontier LLMs can read and write neural-network weights directly. GPT-6 Astra can infer what a network was trained to do from its raw weights and create small game-playing networks that outperform ones trained with PPO for 100 million steps. The capability could matter for model optimization and interpretability.资料称,前沿大语言模型已经能够直接读取和写入神经网络权重。GPT-6 Astra 可以仅凭原始权重判断网络的训练目标,还能从零编写击败经过 1 亿步 PPO 训练的小型游戏网络。这种能力可能对模型优化和可解释性具有重要意义。
HackerTube presents a YouTube-style interface for 6,454 videos sourced from 39,834 Hacker News comments dating back to 2007.HackerTube 提供了一个 YouTube 风格的界面,收录了自 2007 年以来从 39,834 条 Hacker News 评论中提取的 6,454 个视频。
HackerTube presents a YouTube-style interface for 6,454 videos sourced from 39,834 Hacker News comments dating back to 2007. The collection is available through hackertube.org. It turns years of community recommendations into a browsable video archive.HackerTube 提供了一个 YouTube 风格的界面,收录了自 2007 年以来从 39,834 条 Hacker News 评论中提取的 6,454 个视频。该视频集合可通过 hackertube.org 浏览。它将多年社区推荐整理成了一个可探索的视频档案。
DeepInfer is a new high-performance inference framework designed for DeepSeek models, aiming to maximize throughput on modern hardware.DeepInfer 是面向 DeepSeek 模型的高性能推理框架,旨在于现代硬件上实现最大的吞吐量。
DeepInfer is a new high-performance inference framework designed for DeepSeek models, aiming to maximize throughput on modern hardware. It emphasizes modular backends and optimized runtimes to speed up deployment of large-scale AI.DeepInfer 是面向 DeepSeek 模型的高性能推理框架,旨在于现代硬件上实现最大的吞吐量。它通过模块化后端与优化运行时来提升大规模 AI 推理的部署效率。
Carnegie Mellon’s Kenneth C. Griffin School of Computer Science is publishing free lectures on YouTube.卡内基梅隆大学肯尼斯·C·格里芬计算机科学学院正在 YouTube 上发布免费课程讲座。
Carnegie Mellon’s Kenneth C. Griffin School of Computer Science is publishing free lectures on YouTube. Self-paced courseware is planned as the next step. The release gives learners access to the school’s computer science material without charge.卡内基梅隆大学肯尼斯·C·格里芬计算机科学学院正在 YouTube 上发布免费课程讲座。下一步计划推出可自主学习的课程材料。这将让学习者免费接触该学院的计算机科学内容。
HackerBooks catalogs about 16,500 books mentioned by at least two different Hacker News commenters between 2007 and today.HackerBooks 收录了 2007 年至今至少被两名不同 Hacker News 评论者提到过的约 1.65 万本书。
HackerBooks catalogs about 16,500 books mentioned by at least two different Hacker News commenters between 2007 and today. Each book links back to the comments that mentioned it, and the dataset can be downloaded from the about page. The creator notes that the system counts mentions without distinguishing praise from criticism, so some entries may be missing or mismatched.HackerBooks 收录了 2007 年至今至少被两名不同 Hacker News 评论者提到过的约 1.65 万本书。每本书都链接回提及它的评论,数据集也可从关于页面下载。创建者说明,系统只统计提及次数,不区分赞扬和批评,因此部分书籍可能遗漏或匹配错误。
A $5.6B theft linked to Binance highlights the ongoing scale of crypto-exchange breaches.涉及币安的巨额盗窃案,金额约56亿美元,凸显集中交易所安全风险。
A $5.6B theft linked to Binance highlights the ongoing scale of crypto-exchange breaches. On-chain tracing attempts to map the stolen funds as they move through wallets and mixers, showing how investigators pursue assets long after a hack. The lesson for builders is to strengthen custody security (multi-sig, cold storage) and incident response readiness.涉及币安的巨额盗窃案,金额约56亿美元,凸显集中交易所安全风险。链上追踪尝试绘出赃款去向,显示即便多年后仍有追踪的可能。对系统设计者的启示是加强资产托管安全(多签、冷存储)与快速事件响应能力。
SpaceX's Starship Flight 15 ended with the vehicle being caught by the launch system.SpaceX 的星舰第 15 次飞行以发射系统接住飞行器告终。
SpaceX's Starship Flight 15 ended with the vehicle being caught by the launch system. The available material provides no further details about the flight or the catch.SpaceX 的星舰第 15 次飞行以发射系统接住飞行器告终。现有信息未提供这次飞行或捕获过程的更多细节。
Mistral AI has released the weights for Mistral Large 4, enabling researchers to run and fine-tune the model locally.Mistral Large 4 的权重已对外公开,研究者和开发者可以在本地部署与微调。
Mistral AI has released the weights for Mistral Large 4, enabling researchers to run and fine-tune the model locally. This move reduces reliance on hosted APIs and accelerates experimentation, but also brings licensing, safety, and governance considerations to the fore in open-weight AI projects.Mistral Large 4 的权重已对外公开,研究者和开发者可以在本地部署与微调。此举降低对 API 的依赖,促进实验,但也引发许可、安全与治理方面的讨论。
A concept for a Wikipedia-like wiki restricted to items within a 30-minute walk of your location, enabling hyperlocal crowdsourcing.一个近场本地维基的设想,只收录你所在位置半小时步行范围内的条目,强调超本地化的众包协作与隐私保护。
A concept for a Wikipedia-like wiki restricted to items within a 30-minute walk of your location, enabling hyperlocal crowdsourcing. This model foregrounds privacy and relevancy, but raises questions about coverage, updates, and shared standards across communities.一个近场本地维基的设想,只收录你所在位置半小时步行范围内的条目,强调超本地化的众包协作与隐私保护。此思路提升信息相关性,但也带来覆盖范围、更新频率和跨社区标准化等挑战。
Anthropic has filed an S-1 with the SEC, signaling plans for an IPO.Anthropic 已向美国证券交易委员会提交 S-1 文件,表明公司计划上市。
Anthropic has filed an S-1 with the SEC, signaling plans for an IPO. The filing will reveal business metrics, product roadmap, and risk factors as the company scales its AI safety-focused offerings. Investors will weigh profitability trajectory, capital needs, and competitive positioning in a crowded AI market.Anthropic 已向美国证券交易委员会提交 S-1 文件,表明公司计划上市。文件将披露业务数据、产品路线图与风险因素,显示其在 AI 安全领域的扩张。投资者将评估盈利前景、资金需求以及在竞争激烈的 AI 市场中的定位。
The FAA issued a national ground stop affecting departing flights.美国联邦航空管理局发布了影响出发航班的全国地面停飞令。
The FAA issued a national ground stop affecting departing flights. The available records identify the action but do not provide reliable details about its duration, cause, or operational impact.美国联邦航空管理局发布了影响出发航班的全国地面停飞令。现有记录确认了这一措施,但未提供其持续时间、原因或运行影响的可靠细节。
CVE-2026-98432 is described as a critical, remotely exploitable bug in the Linux kernel's TCP code that has existed since 2016.CVE-2026-98432 被描述为 Linux 内核 TCP 代码中的严重漏洞,自 2016 年起便存在且可远程利用。
CVE-2026-98432 is described as a critical, remotely exploitable bug in the Linux kernel's TCP code that has existed since 2016. The source says Anthropic developed an exploit capable of gaining root access, while Debian issued a patch on October 16 and supported Ubuntu releases remained listed as vulnerable.CVE-2026-98432 被描述为 Linux 内核 TCP 代码中的严重漏洞,自 2016 年起便存在且可远程利用。资料称 Anthropic 已开发出可获取服务器 root 权限的漏洞利用程序;Debian 于 10 月 16 日发布了补丁,而受支持的 Ubuntu 版本仍被列为易受攻击。
A reported Claude Code bug exposed a user's email address to arbitrary third parties in the User-Agent header during web searches.有用户报告称,Claude Code 在进行网页搜索时,会通过 User-Agent 请求头将其邮箱地址暴露给任意第三方。
A reported Claude Code bug exposed a user's email address to arbitrary third parties in the User-Agent header during web searches. The user says this happened without permission and was followed by a large increase in scam emails. They argue that the behavior should be removed rather than controlled by an opt-in or opt-out setting.有用户报告称,Claude Code 在进行网页搜索时,会通过 User-Agent 请求头将其邮箱地址暴露给任意第三方。该用户表示,未经允许发生这一行为后,自己收到的诈骗邮件大幅增加。用户认为,这种行为不应通过加入或退出选项控制,而应从系统中彻底移除。
The author describes investigating browser fingerprinting scripts, from libraries such as FingerprintJS to anti-bot systems used by Cloudflare Turnstile and Akamai.作者介绍了对浏览器指纹识别脚本的调查,范围从 FingerprintJS 等常见库到 Cloudflare Turnstile 和 Akamai 使用的反机器人系统。
The author describes investigating browser fingerprinting scripts, from libraries such as FingerprintJS to anti-bot systems used by Cloudflare Turnstile and Akamai. They then considered crawling the web, intercepting fingerprinting-related browser API calls, and mapping what was found. The title connects this investigation with scripts that appear to be creating bioweapons, though the supplied material does not provide further details.作者介绍了对浏览器指纹识别脚本的调查,范围从 FingerprintJS 等常见库到 Cloudflare Turnstile 和 Akamai 使用的反机器人系统。随后,作者考虑抓取网页、拦截与指纹识别相关的浏览器 API 调用,并绘制其分布图。标题将这项调查与似乎正在制造生物武器的脚本联系起来,但现有内容未提供更多细节。
A security advisory warns of critical remote code execution vulnerabilities that could let attackers take control of affected systems.一份安全公告警告存在关键的远程代码执行漏洞,可能使攻击者远程控制受影响的系统。
A security advisory warns of critical remote code execution vulnerabilities that could let attackers take control of affected systems. Patching now across vulnerable software stacks is essential to limit risk, as exploitation often ramps up after disclosure.一份安全公告警告存在关键的远程代码执行漏洞,可能使攻击者远程控制受影响的系统。现在就对受影响的软件栈进行打补丁至关重要,以降低风险,因为披露后利用活动往往会增加。
GPT-6.2 Astra and Sol are presented as two named models or systems.GPT‑6.2 Astra 和 Sol 是两个名为 Astra 与 Sol 的模型或系统。
GPT-6.2 Astra and Sol are presented as two named models or systems. No further technical details or capabilities are established here.GPT‑6.2 Astra 和 Sol 是两个名为 Astra 与 Sol 的模型或系统。现有信息未提供更多技术细节或功能说明。
Privacy-focused DNS resolver Quad9 announces it will discontinue services in France, citing regulatory and operational challenges.Quad9 宣布退出法国市场,原因涉及监管与运营挑战。
Privacy-focused DNS resolver Quad9 announces it will discontinue services in France, citing regulatory and operational challenges. The move underscores how geopolitical and local privacy laws shape the reach of privacy-centric infrastructure.Quad9 宣布退出法国市场,原因涉及监管与运营挑战。这一举动凸显了地缘政治与地方隐私法规对隐私基础设施覆盖的影响。
A large national analysis comparing students’ scores before and after cellphone bans found modest academic improvements.一项比较学生禁用手机前后成绩的大型全国性分析发现,学业表现出现小幅改善。
A large national analysis comparing students’ scores before and after cellphone bans found modest academic improvements. Schools with strict policies requiring students to lock phones in pouches for the entire school day saw math and reading gains equivalent to roughly one or two additional weeks of schooling per year. The study was published by Harvard’s Center for Education Policy Research and is the first national study to track individual students’ growth in this way.一项比较学生禁用手机前后成绩的大型全国性分析发现,学业表现出现小幅改善。要求学生在整个上课日将手机锁入袋中的学校,数学和阅读成绩提升幅度相当于每年额外增加约一至两周的学习。该研究由哈佛教育政策研究中心发布,是首项以这种方式追踪个体学生成长的全国性研究。
AI is accelerating both attacks and defenses in cybersecurity.AI 正在同时加速网络攻击和防御。
AI is accelerating both attacks and defenses in cybersecurity. That acceleration is exposing painful technical debt in fragile legacy systems built up over roughly four decades. The pressure from AI is therefore forcing organizations to address weaknesses that older systems had left unresolved.AI 正在同时加速网络攻击和防御。这样的加速暴露了脆弱遗留系统中积累约 40 年的严重技术债。AI 带来的压力因此迫使组织处理旧系统长期遗留下来的弱点。
Residential proxy traffic exited the White House network 236 times over 30 days, passing through 28 public proxy services.过去 30 天内,有住宅代理流量 236 次从白宫网络流出,并经过 28 个公共代理服务。
Residential proxy traffic exited the White House network 236 times over 30 days, passing through 28 public proxy services. The network, AS2650, is registered to the Executive Office of the President, which hosts the White House network. Residential proxies route customers’ traffic through everyday devices so it appears to originate from those devices’ networks.过去 30 天内,有住宅代理流量 236 次从白宫网络流出,并经过 28 个公共代理服务。该网络的 AS2650 注册于负责托管白宫网络的美国总统行政办公室。住宅代理会让客户的流量经由普通设备转发,使其看起来来自这些设备所在的网络。
The site presents a large database of paper airplanes with step-by-step video folding tutorials.该网站提供大型纸飞机数据库,并为每种机型配备分步折叠视频教程。
The site presents a large database of paper airplanes with step-by-step video folding tutorials. One World Record prototype is reported to fly more than 300 feet. The Unite model holds a world record for time aloft at 31.2 seconds.该网站提供大型纸飞机数据库,并为每种机型配备分步折叠视频教程。据称,一款世界纪录原型机可以飞行超过 300 英尺。Unite 机型则以 31.2 秒的滞空时间保持世界纪录。
Apple's git still trusts compromised eMudhra certificates, leaving room for potential man-in-the-middle attacks on TLS and code signing trust.苹果的 Git 仍信任被妥协的 eMudhra 证书,增加了 TLS 与代码签名信任被滥用的风险。
Apple's git still trusts compromised eMudhra certificates, leaving room for potential man-in-the-middle attacks on TLS and code signing trust. This highlights how supply chain security depends on up-to-date trust stores, and warrants prompt certificate revocation and rotation.苹果的 Git 仍信任被妥协的 eMudhra 证书,增加了 TLS 与代码签名信任被滥用的风险。这凸显供应链安全对最新信任存储的依赖,应尽快撤销并轮换相关证书。
EMudhra reports compromise of its Intermediate Certificate Authority, potentially enabling issuance of fraudulent certificates.EMudhra 报告其中级证书颁发机构遭到妥协,可能导致伪造证书的签发。
EMudhra reports compromise of its Intermediate Certificate Authority, potentially enabling issuance of fraudulent certificates. The incident highlights PKI supply-chain risk and the need for aggressive revocation, transparency, and improved security controls.EMudhra 报告其中级证书颁发机构遭到妥协,可能导致伪造证书的签发。这一事件凸显公钥基础设施供应链风险,以及需要更快速的撤销、透明披露和改进的安全控制措施。
New documents and whistleblower testimony reveal ICE conducted a sprawling surveillance operation targeting left-wing social movements, leveraging internal data and informants across decades.新的文档和告密者证词揭示ICE长期针对左翼社会运动执行大规模监视行动,动用了内部数据和线人。
New documents and whistleblower testimony reveal ICE conducted a sprawling surveillance operation targeting left-wing social movements, leveraging internal data and informants across decades. The findings raise questions about civil liberties, oversight, and the balance between border enforcement and political dissent.新的文档和告密者证词揭示ICE长期针对左翼社会运动执行大规模监视行动,动用了内部数据和线人。相关证据推动对公民自由、监督机制以及边境执法与政治异议之间平衡的讨论。
Microsoft 365 is experiencing service degradation.Microsoft 365 正在经历服务降级。
Microsoft 365 is experiencing service degradation. The available material does not specify which services are affected, the cause, or the expected resolution time. Users may therefore encounter disruptions while the incident is being addressed.Microsoft 365 正在经历服务降级。现有材料没有说明受影响的具体服务、原因或预计恢复时间。因此,用户在问题处理期间可能会遇到服务中断。
Calif disclosed Java4Shell, tracked as CVE-2026-114938, a memory-corruption vulnerability in OpenJDK's Base64 decoder.Calif 披露了 Java4Shell(编号 CVE-2026-114938),这是 OpenJDK Base64 解码器中的内存破坏漏洞。
Calif disclosed Java4Shell, tracked as CVE-2026-114938, a memory-corruption vulnerability in OpenJDK's Base64 decoder. OpenJDK is described as the most popular Java runtime. The vulnerability has a CVSS 4.0 score of 9.2 out of 10, rated Critical. Its severity could make affected OpenJDK deployments a high-priority patching target.Calif 披露了 Java4Shell(编号 CVE-2026-114938),这是 OpenJDK Base64 解码器中的内存破坏漏洞。OpenJDK 被称为最流行的 Java 运行时。该漏洞在 CVSS 4.0 中获得 9.2 分(满分 10 分),评级为“严重”。这一评级意味着受影响的 OpenJDK 部署可能需要优先修补。
Nginx has been reported to contain unauthenticated remote code execution vulnerabilities.据报道,Nginx 存在无需认证即可触发的远程代码执行漏洞。
Nginx has been reported to contain unauthenticated remote code execution vulnerabilities. The supplied records identify two related CVE entries, CVE-2026-123633 and CVE-2026-123007.据报道,Nginx 存在无需认证即可触发的远程代码执行漏洞。相关记录列出了两个 CVE 编号:CVE-2026-123633 和 CVE-2026-123007。
Three new discoveries span mathematics, computer science, and physics, highlighting cross-disciplinary progress in fundamental science.宣布涵盖数学、计算机科学与物理学的三项新发现,凸显跨学科合作在基础科学领域的进展。
Three new discoveries span mathematics, computer science, and physics, highlighting cross-disciplinary progress in fundamental science. The announcement hints at novel methods and insights that could influence both theory and practice across disciplines.宣布涵盖数学、计算机科学与物理学的三项新发现,凸显跨学科合作在基础科学领域的进展。公告暗示可能的新方法与洞见,影响理论到实践的研究路径。
The author describes a decade-long home-automation rule: no core function in the house should depend on a vendor’s server.作者介绍了坚持十年的家庭自动化原则:家中的核心功能不应依赖厂商服务器。
The author describes a decade-long home-automation rule: no core function in the house should depend on a vendor’s server. As a result, the home continues operating when the internet goes down. The post presents the other automation rules and two related talks that grew out of this principle.作者介绍了坚持十年的家庭自动化原则:家中的核心功能不应依赖厂商服务器。因此,即使互联网中断,房屋仍能继续运行。文章还整理了由这一原则延伸出的其他自动化规则和两场相关演讲。
Tesla CEO Elon Musk said Cybertruck production at the company's Texas gigafactory will wind down around mid-2027.特斯拉 CEO 埃隆·马斯克表示,公司位于得州的超级工厂将在 2027 年年中左右逐步停止生产 Cybertruck。
Tesla CEO Elon Musk said Cybertruck production at the company's Texas gigafactory will wind down around mid-2027. He said the move will make room for higher production of the newly launched Cybercab, central to Tesla's robotaxi plans. Cybertruck sales have remained weak, and it has sold in only a limited number of markets. It remains unclear whether production will move elsewhere or end entirely.特斯拉 CEO 埃隆·马斯克表示,公司位于得州的超级工厂将在 2027 年年中左右逐步停止生产 Cybertruck。他称,此举将为提高新款 Cybercab 的产量腾出空间,而 Cybercab 是特斯拉机器人出租车计划的核心车型。Cybertruck 销量长期疲软,目前仅在少数市场销售。该车型是转移到其他工厂生产还是彻底停产,目前尚不清楚。
The project brings Need for Speed Underground 2 online in a web browser.该项目让《极品飞车:地下狂飙 2》能够在浏览器中在线运行。
The project brings Need for Speed Underground 2 online in a web browser. No further implementation details or features are established here.该项目让《极品飞车:地下狂飙 2》能够在浏览器中在线运行。现有信息未说明其具体实现方式或功能细节。
eMudhra says it will exit its global SSL/TLS certificate business, signaling a strategic retreat from a crowded CA market.eMudhra 宣布将退出全球 SSL/TLS 证书业务,显示其在竞争激烈的证书颁发机构市场中的战略收缩。
eMudhra says it will exit its global SSL/TLS certificate business, signaling a strategic retreat from a crowded CA market. The move could nudge customers to migrate to other providers and accelerate consolidation in the PKI ecosystem, with ripple effects on how organizations manage trust chains.eMudhra 宣布将退出全球 SSL/TLS 证书业务,显示其在竞争激烈的证书颁发机构市场中的战略收缩。此举可能促使客户转向其他供应商,并推动公钥基础设施生态的整合,对组织的信任链管理产生连锁影响。
We hacked the iPhone as planned, detailing the approach and lessons learned.作者按计划攻破了 iPhone,给出实现细节和经验教训。
We hacked the iPhone as planned, detailing the approach and lessons learned. The post underscores ongoing mobile security challenges and how defenders adapt to evolving techniques.作者按计划攻破了 iPhone,给出实现细节和经验教训。此事凸显移动设备安全的持续挑战及对抗者与防御者之间的博弈趋势。
The author argues that all code generated with AI assistance should be released as open source or under a CC0 license.作者主张,所有借助 AI 生成的代码都应以开源许可证或 CC0 许可发布。
The author argues that all code generated with AI assistance should be released as open source or under a CC0 license. The argument is based on AI models being trained on large amounts of publicly available open-source code, as well as Stack Overflow and related sites whose Creative Commons licenses generally require attribution. The author says turning that shared knowledge into proprietary code resembles license laundering.作者主张,所有借助 AI 生成的代码都应以开源许可证或 CC0 许可发布。其理由是,AI 模型训练使用了大量公开的开源代码,以及通常要求署名的 Stack Overflow 和相关网站内容。作者认为,将这些共享知识转化为专有闭源代码近似于“许可证洗白”。
libghostty-vt 1.0 has been released as a portable, high-performance C library for terminal emulation, extracted from Ghostty’s core.libghostty-vt 1.0 已发布,这是一个从 Ghostty 核心提取出的便携、高性能终端模拟 C 库。
libghostty-vt 1.0 has been released as a portable, high-performance C library for terminal emulation, extracted from Ghostty’s core. Both the library and its C ABI are now stable. The C API has grown from nonexistent to a complete terminal emulator used in production by millions of people across hundreds of free and commercial products.libghostty-vt 1.0 已发布,这是一个从 Ghostty 核心提取出的便携、高性能终端模拟 C 库。该库本身及其 C ABI 现已稳定。其 C API 已从无到有发展为完整的终端模拟器,目前已在数百款免费和商业产品中投入生产,服务数百万人。
Kimi introduces the K3.1 Pro line and a Highspeed variant, focusing on faster inference and practical AI workloads.文章介绍了 Kimi 的 K3.1 Pro 系列及 Highspeed 变体,聚焦于更快的推理性能及对 AI 工作负载的适用性。
Kimi introduces the K3.1 Pro line and a Highspeed variant, focusing on faster inference and practical AI workloads. The writeup highlights trade-offs between speed, power, and cost for developers deploying models at scale.文章介绍了 Kimi 的 K3.1 Pro 系列及 Highspeed 变体,聚焦于更快的推理性能及对 AI 工作负载的适用性。讨论了在大规模部署中,速度、功耗与成本之间的权衡。
PHP 8.2.35 is a maintenance release that patches several security and stability issues in the 8.2 line.这是对 8.2 分支的维护性更新,包含若干安全修复与错误修正。
PHP 8.2.35 is a maintenance release that patches several security and stability issues in the 8.2 line. It’s a standard part of the PHP patch cycle, and upgrading helps close vulnerabilities and improve reliability. If you run PHP 8.2 in production, apply the update to stay supported and secure.这是对 8.2 分支的维护性更新,包含若干安全修复与错误修正。强烈建议在生产环境尽快升级,以降低潜在漏洞并提升稳定性。
An undisclosed court ruling governs how the government collects data on Americans, and officials have kept it secret.尚未公开的法院裁决规定政府如何监控美国民众的数据,相关信息被隐瞒。
An undisclosed court ruling governs how the government collects data on Americans, and officials have kept it secret. The ruling highlights ongoing friction between security powers and civil liberties, underscoring the demand for greater transparency and robust oversight.尚未公开的法院裁决规定政府如何监控美国民众的数据,相关信息被隐瞒。该裁决暴露出国家安全权力与公民自由之间的矛盾,也凸显对更大透明度和强有力监督的诉求。
A new report, citing The Wall Street Journal, alleges that NVIDIA increased the GPU and memory shortage by investing in AI startups that then bought NVIDIA GPUs and built data centers.一份援引《华尔街日报》的新报告称,英伟达通过投资 AI 初创公司,促使这些公司购买英伟达 GPU 并建设数据中心,从而加剧了 GPU 和内存短缺。
A new report, citing The Wall Street Journal, alleges that NVIDIA increased the GPU and memory shortage by investing in AI startups that then bought NVIDIA GPUs and built data centers. The allegation emerged after Jensen Huang’s departure as NVIDIA’s CEO, according to the supplied report. The claim concerns whether the company’s financing activity helped inflate demand for its own hardware.一份援引《华尔街日报》的新报告称,英伟达通过投资 AI 初创公司,促使这些公司购买英伟达 GPU 并建设数据中心,从而加剧了 GPU 和内存短缺。报告称,这一指控是在 Jensen Huang 离任英伟达 CEO 后出现的。争议焦点在于,英伟达的融资活动是否帮助推高了对其自有硬件的需求。
Genie 4 from DeepMind represents a leap in world-simulation tech, enabling more realistic, scalable environments for AI training and planning.DeepMind 的 Genie 4 提供更真实、更可控的世界仿真能力,为 AI 训练和规划提供更丰富的环境,有望加速机器人、游戏 AI 以及从仿真到现实的迁移。
Genie 4 from DeepMind represents a leap in world-simulation tech, enabling more realistic, scalable environments for AI training and planning. It aims to accelerate progress in robotics, game AI, and simulation-to-real transfer by providing richer, more controllable environments.DeepMind 的 Genie 4 提供更真实、更可控的世界仿真能力,为 AI 训练和规划提供更丰富的环境,有望加速机器人、游戏 AI 以及从仿真到现实的迁移。
The author argues that the experiments reading raw model weights and the experiments modifying Qwen use entirely different techniques.作者认为,读取模型原始权重的实验与修改 Qwen 的实验使用了完全不同的技术。
The author argues that the experiments reading raw model weights and the experiments modifying Qwen use entirely different techniques. They compare the presentation to a series of stage-magic tricks that appear to test competing explanations while relying on different mechanisms.作者认为,读取模型原始权重的实验与修改 Qwen 的实验使用了完全不同的技术。作者将这种呈现方式比作一连串舞台魔术:表面上是在检验相互竞争的解释,实际却依赖不同机制。文章分析了原始说法为何可能让人误解这些实验究竟证明了什么。
AI agents analyzed NASA lidar data and identified a Maya site beneath the jungle in Mexico.AI 代理分析 NASA 的激光雷达数据,在墨西哥丛林下识别出一处玛雅遗址。
AI agents analyzed NASA lidar data and identified a Maya site beneath the jungle in Mexico. The site had already been mapped by archaeologists and described in Estrada-Belli et al. 2025, but it was not yet listed in Mexico’s official INAH register. The agents therefore rediscovered the site rather than discovering it, and the process took less than 72 hours to reach an official response from the Mexican government.AI 代理分析 NASA 的激光雷达数据,在墨西哥丛林下识别出一处玛雅遗址。考古学家其实已经绘制过该遗址的地图,并在 Estrada-Belli 等人 2025 年的论文中记录了它,只是它尚未列入墨西哥国家人类学与历史研究所(INAH)的官方名录。因此,代理完成的是“重新发现”而非发现;从提出想法到收到墨西哥政府的正式回复耗时不到 72 小时。
Microsoft is offering voluntary buyouts to roughly a quarter of its workforce, a major headcount reduction without forced layoffs.微软向约四分之一员工提供自愿离职计划,属于不强制裁员的大规模裁员。
Microsoft is offering voluntary buyouts to roughly a quarter of its workforce, a major headcount reduction without forced layoffs. The move fits into broader tech-sector efficiency drives as AI investments continue and macro pressures persist. It could reshape compensation bands, talent flow, and project prioritization across divisions.微软向约四分之一员工提供自愿离职计划,属于不强制裁员的大规模裁员。此举与科技行业的整体效率提升趋势相吻合,AI 投资与宏观压力并行推动成本优化。人力结构、人才流向和项目优先级可能因此调整。
The title reports a backdoor in a Claude Code extension for VS Code distributed through the Cursor Marketplace.标题称,Cursor Marketplace 上分发的 Claude Code for VS Code 扩展中存在后门。
The title reports a backdoor in a Claude Code extension for VS Code distributed through the Cursor Marketplace. The supplied context provides no further reliable details.标题称,Cursor Marketplace 上分发的 Claude Code for VS Code 扩展中存在后门。提供的上下文没有更多可靠细节。
Anthropic has filed to become a public company and published a founders’ letter along with related information.Anthropic 已提交申请,准备成为一家上市公司,并公布了创始人公开信及相关信息。
Anthropic has filed to become a public company and published a founders’ letter along with related information. The company is organized as a public benefit corporation focused on securing AI’s benefits and mitigating its risks. It says AI could accelerate scientific innovation, increase economic growth, help cure major diseases, and address long-standing problems if managed well.Anthropic 已提交申请,准备成为一家上市公司,并公布了创始人公开信及相关信息。该公司采用公益公司架构,致力于发挥人工智能的益处并降低其风险。Anthropic 表示,如果得到妥善管理,人工智能可能加速科学创新、提高经济增长率、推动重大疾病治疗,并解决长期未能解决的问题。
Gemini 4 Carbon is Google's latest in the Gemini family, focusing on efficiency and safety with improved inference and reliability for enterprise apps.Gemini 4 Carbon 作为谷歌 Gemini 系列的新一代模型,强调在企业应用中的推理效率与安全性提升。
Gemini 4 Carbon is Google's latest in the Gemini family, focusing on efficiency and safety with improved inference and reliability for enterprise apps. It demonstrates ongoing progress in AI model scaling, with Carbon as a more efficient variant for dense workloads.Gemini 4 Carbon 作为谷歌 Gemini 系列的新一代模型,强调在企业应用中的推理效率与安全性提升。该系列继续推动大模型规模化,Carbon 版本在高密度任务中更具效率。
An analysis of 159,041 Lovable projects found many cases in which personal and confidential information was accessible without logging in, including passport details and more than 200,000 private messages.对 159,041 个 Lovable 项目的分析发现,许多项目无需登录即可访问个人和机密信息,其中包括护照详情以及超过 20 万条私人消息。
An analysis of 159,041 Lovable projects found many cases in which personal and confidential information was accessible without logging in, including passport details and more than 200,000 private messages. Lovable is a Stockholm-based Swedish AI startup founded in 2023 by Anton Osika and Fabian Hedin, whose platform generates web applications from plain-language descriptions. The findings show that building applications with AI does not remove the need for careful handling of sensitive data.对 159,041 个 Lovable 项目的分析发现,许多项目无需登录即可访问个人和机密信息,其中包括护照详情以及超过 20 万条私人消息。Lovable 是一家总部位于斯德哥尔摩、由 Anton Osika 和 Fabian Hedin 于 2023 年创立的瑞典 AI 初创公司,其平台可根据自然语言描述生成 Web 应用。这些发现表明,使用 AI 构建应用并不能免除对敏感数据进行谨慎处理的要求。
AWS Emulator is presented as a project related to emulating AWS services.AWS Emulator 是一个用于模拟 AWS 服务的项目。
AWS Emulator is presented as a project related to emulating AWS services. The available context does not provide further technical details about its implementation or supported services.AWS Emulator 是一个用于模拟 AWS 服务的项目。现有信息没有提供其实现方式或所支持服务的更多技术细节。
An investigation found that many Black Friday and Cyber Monday deals advertised by Amazon did not provide actual savings.一项调查发现,亚马逊宣传的许多黑色星期五和网络星期一优惠实际上并没有带来节省。
An investigation found that many Black Friday and Cyber Monday deals advertised by Amazon did not provide actual savings. The deals were presented as discounts even though shoppers often paid no less than usual.一项调查发现,亚马逊宣传的许多黑色星期五和网络星期一优惠实际上并没有带来节省。这些商品虽然以折扣形式展示,消费者实际支付的价格却往往并未降低。
The author describes modifying the navigation system of a 2014 Renault Latitude to change the dashboard clock.作者介绍了如何修改一辆 2014 年雷诺 Latitude 的导航系统,以调整仪表盘时钟。
The author describes modifying the navigation system of a 2014 Renault Latitude to change the dashboard clock. The car apparently required a navigation-map SD card for access to the time-setting function, even though the card had been discarded because its maps were outdated. Without it, the display showed a prominent “No maps found!” message.作者介绍了如何修改一辆 2014 年雷诺 Latitude 的导航系统,以调整仪表盘时钟。该车似乎需要导航地图 SD 卡才能访问时间设置功能,但由于地图过时,这张卡此前已被丢弃。没有 SD 卡时,显示屏会显示醒目的“No maps found!”提示。
An instance of a broken Apple Music link surfaces, prompting questions about reliability and content routing on streaming platforms.发现 Apple Music 的一个链接返回 404,暴露出流媒体平台在内容发现与路由方面的脆弱性。
An instance of a broken Apple Music link surfaces, prompting questions about reliability and content routing on streaming platforms. The note underscores the fragility of URL-based content discovery.发现 Apple Music 的一个链接返回 404,暴露出流媒体平台在内容发现与路由方面的脆弱性。
A playful project adds permit bureaucracy to a SimCity-like environment, showing how permit delays shape growth, zoning decisions, and development timelines.一个将许可制度引入类似 SimCity 的模拟环境的有趣项目,展示许可拖延如何影响城市增长、分区决策与开发节奏。
A playful project adds permit bureaucracy to a SimCity-like environment, showing how permit delays shape growth, zoning decisions, and development timelines. The result offers a tangible look at how administrative friction can influence urban planning experiments and policy outcomes.一个将许可制度引入类似 SimCity 的模拟环境的有趣项目,展示许可拖延如何影响城市增长、分区决策与开发节奏。该成果为官僚摩擦如何影响城市规划实验与政策结果提供了直观的观察。
Verisign has withdrawn its plan to terminate 22,000 .name domain names and email redirects after a prolonged campaign.经过一场旷日持久的抗争,Verisign 撤回了终止 22,000 个 .name 域名及其邮件转发的计划。
Verisign has withdrawn its plan to terminate 22,000 .name domain names and email redirects after a prolonged campaign. New registrations for third-level .name domains will close, but all existing domains and redirects will remain indefinitely. The reversal followed efforts by domain owners, lawyers, and supporters to challenge the plan through ICANN.经过一场旷日持久的抗争,Verisign 撤回了终止 22,000 个 .name 域名及其邮件转发的计划。虽然 .name 三级域名将停止新注册,但现有域名和邮件转发服务将无限期保留。这一转变源于域名所有者、律师和支持者通过 ICANN 发起的共同争取。
Researchers demonstrate that eMudhra's CA keys can be cracked in milliseconds, challenging TLS trust assumptions.研究者显示 eMudhra 的 CA 钥匙可在毫秒内被破解,挑战了 TLS 信任的基本假设。
Researchers demonstrate that eMudhra's CA keys can be cracked in milliseconds, challenging TLS trust assumptions. The result highlights how certificate authorities' key management and cryptographic agility impact the security of the entire chain, pushing for shorter certificate lifetimes and stronger hardware protections. It also underscores the need for tougher CA audits and improved trust-chain resilience.研究者显示 eMudhra 的 CA 钥匙可在毫秒内被破解,挑战了 TLS 信任的基本假设。此现象突显证书颁发机构在密钥管理与加密灵活性方面对整个信任链安全的影响,推动对更短证书有效期和更强硬件保护的需求,同时强调对 CA 审计和信任链韧性的改进。
The Rust team has released Rust 1.100.0.Rust 团队发布了 Rust 1.100.0。
The Rust team has released Rust 1.100.0. Rust is described as a language for building reliable and efficient software. Users with an existing rustup installation can install the new version through rustup.Rust 团队发布了 Rust 1.100.0。Rust 被定位为用于构建可靠、高效软件的编程语言。已经通过 rustup 安装 Rust 的用户可以使用 rustup 获取新版本。
Software projects often settle for incremental gains; the piece argues for embracing bigger, riskier problems that unlock substantial leaps.很多软件项目止步于小修小补。作者主张在软件开发上要敢于挑战更大、风险更高的问题,以实现更大的跃迁。
Software projects often settle for incremental gains; the piece argues for embracing bigger, riskier problems that unlock substantial leaps. Ambition needs disciplined constraints, longer horizons, and tolerance for failure, but can reshape teams' capabilities and the value they create. Expect practical guidance on bet-sizing, architecture, and collaboration to support bold bets.很多软件项目止步于小修小补。作者主张在软件开发上要敢于挑战更大、风险更高的问题,以实现更大的跃迁。实现这种野心需要清晰的约束、长期投入与对失败的容忍,进而改变团队能力与创造的价值。
Swiss nonprofit Quad9 said it will withdraw its free public DNS resolver from France after months of litigation.瑞士非营利组织 Quad9 表示,将在经历数月诉讼后退出法国免费公共 DNS 解析服务。
Swiss nonprofit Quad9 said it will withdraw its free public DNS resolver from France after months of litigation. A French court imposed daily financial penalties because Quad9 did not block domains linked to alleged copyright infringement. Quad9 said the choice was between crippling penalties and changing its service in ways that would undermine user privacy and security. The case is presented as a warning about the consequences of mandating DNS-level blocking.瑞士非营利组织 Quad9 表示,将在经历数月诉讼后退出法国免费公共 DNS 解析服务。法国法院因其未屏蔽涉嫌侵犯版权的域名,对 Quad9 处以每日罚款。Quad9 表示,它只能在承受可能令组织陷入困境的处罚,或改变服务并损害用户隐私与安全之间作出选择。该事件被视为强制实施 DNS 层拦截可能带来后果的警示。
Anthropic has introduced Claude Fable 5.2 and Claude Mythos 5.2 for demanding technical tasks.Anthropic 推出了面向高难度技术任务的 Claude Fable 5.2 和 Claude Mythos 5.2。
Anthropic has introduced Claude Fable 5.2 and Claude Mythos 5.2 for demanding technical tasks. The company says the models improve coding, reasoning, and vision, and that prerelease testing helped engineers complete software migrations, design three working electronic devices, advance robotics evaluations, and formalize a proof of Fermat's Last Theorem. Fable 5.2 is positioned as a new high point for coding, finance, and science.Anthropic 推出了面向高难度技术任务的 Claude Fable 5.2 和 Claude Mythos 5.2。该公司称,这些模型在编程、推理和视觉方面有所提升,并在预发布测试中帮助工程师完成软件迁移、从零设计三种可工作的电子设备、推进机器人评测,以及形式化费马大定理的证明。Fable 5.2 被定位为编程、金融和科学领域的新标杆。
In the early 1990s, Parma farmer Claudio Grossi found an unidentified sack of seeds that likely belonged to his great-grandfather, who founded the farm around 1915.20 世纪 90 年代初,帕尔马农民克劳迪奥·格罗西发现了一袋无法辨认的种子,可能是其约于 1915 年创建农场的曾祖父留下的。
In the early 1990s, Parma farmer Claudio Grossi found an unidentified sack of seeds that likely belonged to his great-grandfather, who founded the farm around 1915. After he discarded the seeds in manure, several types of wheat unexpectedly sprouted. One variety grew about six feet tall and produced seven ears, unlike the single-ear commercial wheat Grossi knew.20 世纪 90 年代初,帕尔马农民克劳迪奥·格罗西发现了一袋无法辨认的种子,可能是其约于 1915 年创建农场的曾祖父留下的。他把种子倒进肥料堆后,几种小麦竟意外发芽。其中一个品种长到约六英尺高,并结出七个麦穗,不同于格罗西熟悉的单穗商业小麦。
Qwen4-Max is described as supporting chat, image and video understanding, image generation, document processing, web search integration, tool use, and artifacts.Qwen4-Max 被描述为支持聊天、图像和视频理解、图像生成、文档处理、联网搜索、工具调用及 artifacts。
Qwen4-Max is described as supporting chat, image and video understanding, image generation, document processing, web search integration, tool use, and artifacts. It is positioned as a model or service with functionality spanning multiple input and output types.Qwen4-Max 被描述为支持聊天、图像和视频理解、图像生成、文档处理、联网搜索、工具调用及 artifacts。它定位为覆盖多种输入和输出形式的模型或服务。
Explores oddities in Docker's behavior that catch developers off guard, from image layering and build caching to volume mounting quirks.揭示 Docker 行为中的一些反直觉之处,比如镜像分层、构建缓存和卷挂载的怪异之处。
Explores oddities in Docker's behavior that catch developers off guard, from image layering and build caching to volume mounting quirks. The post offers practical tips to avoid common pitfalls and explains why these quirks persist in modern container tooling.揭示 Docker 行为中的一些反直觉之处,比如镜像分层、构建缓存和卷挂载的怪异之处。文中给出实用的排错技巧,解释了这些怪异现象在现代容器工具中的持续原因。
The post claims that every Mac since macOS 13, released in 2022, includes two daemons that build a graph of the user’s relationships and activities.文章称,自 2022 年发布的 macOS 13 起,每台 Mac 都包含两个用于构建用户关系和活动图谱的后台进程。
The post claims that every Mac since macOS 13, released in 2022, includes two daemons that build a graph of the user’s relationships and activities. The graph is described as covering contacts, relationships, communications, locations, activities, and timing.文章称,自 2022 年发布的 macOS 13 起,每台 Mac 都包含两个用于构建用户关系和活动图谱的后台进程。该图谱据称涵盖联系人、彼此关系、通信、位置、活动及时间信息。文章还分析了这一系统的工作方式,并提供了查看个人设备上该图谱的工具。
Postgres 19 is late because many more bugs are being found during development.Postgres 19 的发布出现延期,原因是开发过程中发现的错误数量大幅增加。
Postgres 19 is late because many more bugs are being found during development. The title identifies the delay and attributes it to the increased discovery of defects.Postgres 19 的发布出现延期,原因是开发过程中发现的错误数量大幅增加。标题将这次延迟归因于缺陷发现增多。
Anthropic introduced Claude Fable 5.2 and Claude Mythos 5.2 for demanding technical work, with claimed improvements in coding, reasoning, and vision.Anthropic 推出了面向高难度技术工作的 Claude Fable 5.2 和 Claude Mythos 5.2,并称它们在编码、推理和视觉能力方面有所提升。
Anthropic introduced Claude Fable 5.2 and Claude Mythos 5.2 for demanding technical work, with claimed improvements in coding, reasoning, and vision. During prerelease testing, Anthropic says the models helped engineers complete software migrations, design three working electronic devices, make progress on robotics evaluations, and solve problems in physics. The company also says the models formalized a proof of Fermat’s Last Theorem in 30 hours.Anthropic 推出了面向高难度技术工作的 Claude Fable 5.2 和 Claude Mythos 5.2,并称它们在编码、推理和视觉能力方面有所提升。在预发布测试中,Anthropic 表示这些模型帮助工程师完成软件迁移、设计出三种可运行的电子设备、推进机器人评测,并解决物理学问题。公司还称,模型在 30 小时内形式化完成了费马大定理的证明。
A project or course titled “How to make almost anything” focuses on creating a wide range of physical or electronic things.“如何制造几乎任何东西”是一项围绕制作各种实体或电子作品展开的项目或课程。
A project or course titled “How to make almost anything” focuses on creating a wide range of physical or electronic things. Further details about its methods or scope are not available here.“如何制造几乎任何东西”是一项围绕制作各种实体或电子作品展开的项目或课程。现有信息未提供其具体方法或内容范围。
The post says that LibreOffice works all 365 days of the year.这则内容称 LibreOffice 一年 365 天都能正常工作。
The post says that LibreOffice works all 365 days of the year. It presents the statement as a brief comparison or endorsement of the office suite’s reliability.这则内容称 LibreOffice 一年 365 天都能正常工作。它以简短的对比或宣传方式强调了这套办公软件的可靠性。
A technical issue at JPMorgan Chase triggered online-banking delays for banks in the United States.摩根大通的一项技术故障引发了美国银行的在线银行服务延迟。
A technical issue at JPMorgan Chase triggered online-banking delays for banks in the United States. The incident affected online banking service availability or responsiveness.摩根大通的一项技术故障引发了美国银行的在线银行服务延迟。该事件影响了在线银行服务的可用性或响应速度。
A post reveals how to gain a root shell on Samsung TVs, underscoring persistent exploit vectors in smart TVs and the need for secure boot and firmware hardening.一则帖子展示了在三星电视上获取 root shell 的方法,凸显了智能电视的持续安全风险以及对安全启动与固件硬化的需求。
A post reveals how to gain a root shell on Samsung TVs, underscoring persistent exploit vectors in smart TVs and the need for secure boot and firmware hardening.一则帖子展示了在三星电视上获取 root shell 的方法,凸显了智能电视的持续安全风险以及对安全启动与固件硬化的需求。
Stringman is presented as an $800 open-source room-scale gripper robot.Stringman 被介绍为一款售价 800 美元的开源房间级抓取机器人。
Stringman is presented as an $800 open-source room-scale gripper robot. The available material provides an installation note about securing cleats to wall studs or columns, but no further technical specifications.Stringman 被介绍为一款售价 800 美元的开源房间级抓取机器人。现有内容只提到应将固定件用木螺钉固定在墙柱上,或用棘轮绑带固定在立柱上,未提供更多技术规格。
Subvocal is building smart glasses that decode silently mouthed words into text at speech speed.Subvocal 正在开发能将无声口型以接近语速转换为文字的智能眼镜。
Subvocal is building smart glasses that decode silently mouthed words into text at speech speed. The system can also work without opening the mouth, using radio-frequency sensing to detect movements inside the mouth, though the developers distinguish this from silently thinking words. The founders began the project after finding that voice tools were impractical in libraries, offices, and public spaces.Subvocal 正在开发能将无声口型以接近语速转换为文字的智能眼镜。系统甚至可以在不张嘴的情况下工作,通过射频感知口腔内部动作;开发者强调,这不同于单纯在脑中默念。创始人因发现语音工具不适合图书馆、办公室和公共场所,开始了这个项目。
A security patch release for PHP arrives today with critical fixes across core components.今天发布了 PHP 的安全更新,修补了若干核心组件的漏洞。
A security patch release for PHP arrives today with critical fixes across core components. The team outlines stronger security practices, faster disclosure, and enhanced hardening to reduce future risk.今天发布了 PHP 的安全更新,修补了若干核心组件的漏洞。团队还描述了更强的防护措施、快速漏洞披露与强化发布流程,以降低未来的安全风险。对于依赖 PHP 的应用而言,及时打补丁的重要性再次凸显。
Renowned ML educator Sebastian Raschka releases a hands-on video course teaching how to build reasoning models from scratch.知名机器学习教育者 Sebastian Raschka 推出从零开始构建推理模型的视频课程,聚焦端到端设计与调试。
Renowned ML educator Sebastian Raschka releases a hands-on video course teaching how to build reasoning models from scratch. The course emphasizes end-to-end design, debugging, and practical coding patterns to improve model reasoning capabilities. It provides a useful bridge between theory and implementation for AI researchers and engineers.知名机器学习教育者 Sebastian Raschka 推出从零开始构建推理模型的视频课程,聚焦端到端设计与调试。课程强调实际编码模式,帮助提升模型推理能力。对 AI 研究者和工程师来说,这是理论到实现的有用桥梁。
New jailbreak tooling for the Switch 2 has surfaced, enabling homebrew and custom software on Nintendo's latest console.Switch 2 的新越狱工具出现,使玩家能够在新版掌机上运行自制软件。
New jailbreak tooling for the Switch 2 has surfaced, enabling homebrew and custom software on Nintendo's latest console. The release underscores ongoing security tug-of-war between platform vendors and researchers, with potential risks to firmware integrity and user warranties.Switch 2 的新越狱工具出现,使玩家能够在新版掌机上运行自制软件。此举凸显平台安全与玩家破解之间的博弈,可能对固件完整性、设备稳定性和保修造成影响。
A new paper argues that the Exponential Time Hypothesis is false, upending a long-standing conditional lower-bound assumption in complexity theory.一篇新论文提出指数时间假设不成立,动摇了长期支撑 NP 问题下界的核心假设。
A new paper argues that the Exponential Time Hypothesis is false, upending a long-standing conditional lower-bound assumption in complexity theory. If ETH fails, widespread implications follow for algorithm design and hardness results, potentially enabling faster solutions for some NP-hard problems.一篇新论文提出指数时间假设不成立,动摇了长期支撑 NP 问题下界的核心假设。若 ETH 不再成立,算法设计与复杂性界限将被重新评估,某些 NP-hard 问题或将获得更高效的解法。
More than 1,600 comments, mostly supportive, have appeared on an internal Nvidia post urging the board to consider employee sentiment about former CEO Jensen Huang.英伟达内部一则呼吁董事会考虑员工对前 CEO 黄仁勋看法的帖子已收到超过 1,600 条评论,其中大多数表示支持。
More than 1,600 comments, mostly supportive, have appeared on an internal Nvidia post urging the board to consider employee sentiment about former CEO Jensen Huang. The post was submitted before Huang’s resignation by a longtime engineer and asked the board to consider employee views before taking action. Some employees are now asking the board to find a way to bring Huang back despite reports of multiple extramarital affairs and his abrupt resignation.英伟达内部一则呼吁董事会考虑员工对前 CEO 黄仁勋看法的帖子已收到超过 1,600 条评论,其中大多数表示支持。该帖子由一名长期任职的工程师在黄仁勋辞职前提交,要求董事会在采取行动前考虑员工意见。尽管有报道称黄仁勋存在多段婚外情并突然辞职,仍有部分员工要求董事会设法让他回归。
During the 2017 wave of annual wildfires on the West Coast, many Bay Area residents bought air purifiers because their homes were poorly insulated and sealed.2017 年西海岸接连发生野火时,许多湾区居民因为房屋隔热和密封性能较差而购买空气净化器,以避免吸入野火烟尘。
During the 2017 wave of annual wildfires on the West Coast, many Bay Area residents bought air purifiers because their homes were poorly insulated and sealed. The author describes buying a Coway Airmega to avoid breathing wildfire soot and reverse-engineering it to remove its dependence on cloud services. The project reflects a broader effort to keep household devices locally controlled.2017 年西海岸接连发生野火时,许多湾区居民因为房屋隔热和密封性能较差而购买空气净化器,以避免吸入野火烟尘。作者介绍了自己购买 Coway Airmega,并对其进行逆向工程,以摆脱对云端服务的依赖。这一项目体现了将家用设备控制权保留在本地的尝试。
The author bought an OpenStick for $10 including shipping, mainly because it was an unusual and appealing piece of hardware.作者花 10 美元(含运费)买了一台 OpenStick,主要是因为这款硬件新奇而有趣。
The author bought an OpenStick for $10 including shipping, mainly because it was an unusual and appealing piece of hardware. Although the device is less powerful than many modern watches, it can still support enjoyable self-hosting experiments. The project demonstrates how a very inexpensive, tiny computer can serve as an entry point to homelabbing.作者花 10 美元(含运费)买了一台 OpenStick,主要是因为这款硬件新奇而有趣。它的性能可能还不如许多现代智能手表,但仍然适合进行有趣的自托管实验。这个项目展示了廉价小设备如何成为入门家庭实验室的起点。
Mistral AI unveils Voxtral 3, the latest entry in its Voxtral family.Mistral AI 发布 Voxtral 3,是 Voxtral 系列的最新版本。
Mistral AI unveils Voxtral 3, the latest entry in its Voxtral family. The release signals continued momentum in open-weight models and the trend toward more capable, accessible AI cores.Mistral AI 发布 Voxtral 3,是 Voxtral 系列的最新版本。该发布体现了开源权重模型的持续发展,以及更强大、易接入的 AI 核心的趋势。
The operator of All About Berlin says the site's traffic fell by 70% to 75% after years of growth, attributing the decline to Google's AI Overviews.All About Berlin 的运营者表示,该网站在多年增长后,流量因 Google AI Overviews 下跌了 70% 至 75%。
The operator of All About Berlin says the site's traffic fell by 70% to 75% after years of growth, attributing the decline to Google's AI Overviews. Search pages now place an AI-generated answer trained on the site's content before the link to the site itself. The change has had a devastating effect on the site's traffic.All About Berlin 的运营者表示,该网站在多年增长后,流量因 Google AI Overviews 下跌了 70% 至 75%。如今搜索页面会先展示基于该网站内容训练的 AI 生成答案,最后才显示指向网站的链接。这一变化对网站流量造成了严重打击。
CVE-2026-98432 is a critical remote use-after-free vulnerability in Linux's IPv4 TCP handshake, reachable by an unauthenticated network peer.CVE-2026-98432 是 Linux IPv4 TCP 握手中的严重远程释放后使用漏洞,未经身份验证的网络对端即可触发。
CVE-2026-98432 is a critical remote use-after-free vulnerability in Linux's IPv4 TCP handshake, reachable by an unauthenticated network peer. Stable kernels fixed it in October 2026, with patches for versions including 5.10.272 through 7.3, but the flaw dates back to Linux 4.4. At the reported time, Ubuntu still marked every release as vulnerable, while Red Hat listed RHEL 8, 9, and 10 as affected without a fix.CVE-2026-98432 是 Linux IPv4 TCP 握手中的严重远程释放后使用漏洞,未经身份验证的网络对端即可触发。稳定版内核已于 2026 年 10 月修复,补丁覆盖 5.10.272 至 7.3 等版本,但该缺陷可追溯至 Linux 4.4。报道所述时间点,Ubuntu 仍将所有版本标为易受攻击,Red Hat 也列出受影响的 RHEL 8、9 和 10,尚未提供修复。
Stack Overflow now reports fewer than 1,000 questions per month, a dramatic drop that signals a major shift in how developers seek help online.Stack Overflow 现每月提问不足一千条,显示开发者在线求助方式的显著转变。
Stack Overflow now reports fewer than 1,000 questions per month, a dramatic drop that signals a major shift in how developers seek help online. The decline could reflect migration to other communities, changes in data collection, or platform strategy.Stack Overflow 现每月提问不足一千条,显示开发者在线求助方式的显著转变。下降可能源于迁移到其他社区、数据收集变化或平台策略调整。
A developer built a replacement app for a $300 embroidery machine that required a subscription.一名开发者为一台需要订阅的 300 美元刺绣机编写了替代应用。
A developer built a replacement app for a $300 embroidery machine that required a subscription. The project is published as Zigzag. The available description does not provide further technical details about the machine or the application.一名开发者为一台需要订阅的 300 美元刺绣机编写了替代应用。该项目以 Zigzag 的形式发布。现有描述没有提供有关机器或应用的更多技术细节。
Internal prompts reportedly steered WhatsApp's LLM-enabled help content to avoid recommending regulators.据称 WhatsApp 的 LLM 助手帮助页面中存在引导指令,要求其避免向用户推荐监管机构。
Internal prompts reportedly steered WhatsApp's LLM-enabled help content to avoid recommending regulators. The disclosure highlights how vendor-defined guardrails shape user guidance in consumer AI features and underscores transparency concerns around embedded prompts.据称 WhatsApp 的 LLM 助手帮助页面中存在引导指令,要求其避免向用户推荐监管机构。这揭示了厂商设定的安全边界如何影响用户指引,以及对嵌入式提示透明度的担忧。
The author had used Claude Code for more than a year, beginning when it became available through the Claude Pro subscription.作者从 Claude Code 随 Claude Pro 订阅推出后就开始使用,至今已有一年多。
The author had used Claude Code for more than a year, beginning when it became available through the Claude Pro subscription. Despite considering its release an important moment for software development, the author accumulated enough small frustrations to stop using it as the main coding harness. Their development setup is now being migrated to Pi.作者从 Claude Code 随 Claude Pro 订阅推出后就开始使用,至今已有一年多。尽管作者认为它的发布可能会成为软件开发史上的重要节点,但长期积累的许多小问题最终让作者不再把它作为主要编程工具。作者目前正将开发环境迁移到 Pi。
BUI is a GPU-accelerated Blender user interface decoupled from Blender's original codebase.BUI 是一个与 Blender 原始代码库解耦的 GPU 加速用户界面。
BUI is a GPU-accelerated Blender user interface decoupled from Blender's original codebase. It supports Rust, Python, JavaScript, React, C, and C++. The project describes its performance as twice that of GPUI.BUI 是一个与 Blender 原始代码库解耦的 GPU 加速用户界面。它支持 Rust、Python、JavaScript、React、C 和 C++。项目方称其性能是 GPUI 的两倍。
A user is looking for a privacy-respecting television with a 60–80-inch 4K display running at least 60 Hz.一名用户想寻找重视隐私的电视,要求尺寸在 60 至 80 英寸之间、支持 4K 且刷新率至少为 60Hz。
A user is looking for a privacy-respecting television with a 60–80-inch 4K display running at least 60 Hz. They want the device to act only as a display for an Apple TV rather than use smart-TV apps, and are asking which models pair well with that setup.一名用户想寻找重视隐私的电视,要求尺寸在 60 至 80 英寸之间、支持 4K 且刷新率至少为 60Hz。他希望电视只作为 Apple TV 的显示器,而不运行智能电视原生应用,并询问哪些型号适合这种配置。
The author reports that Qwen 3.8 27B running on a laptop can interpret raw neural-network weights.作者称,在笔记本上运行的 Qwen 3.8 27B 也能解读神经网络的原始权重。
The author reports that Qwen 3.8 27B running on a laptop can interpret raw neural-network weights. This follows research describing “weight-space programming,” in which language models analyze and modify model weights represented as floating-point numbers. The earlier experiments highlighted frontier-model capabilities, including those attributed to GPT-6 Astra.作者称,在笔记本上运行的 Qwen 3.8 27B 也能解读神经网络的原始权重。这延续了“权重空间编程”的研究:语言模型可以分析并修改由浮点数表示的模型权重。此前的实验展示了前沿模型的这类能力,包括研究归于 GPT-6 Astra 的能力。
The USPS has released a national ZIP Code boundary map built from delivery-address and location data.美国邮政服务(USPS)发布了基于投递地址和位置数据制作的全国 ZIP Code 边界地图。
The USPS has released a national ZIP Code boundary map built from delivery-address and location data. The boundaries can change as delivery locations are added or modified, and the data is updated quarterly with limited distribution to subsidiaries and partners. The dataset is intended for mapping, mailing, marketing, insurance, real estate, logistics, navigation, and transportation applications.美国邮政服务(USPS)发布了基于投递地址和位置数据制作的全国 ZIP Code 边界地图。随着投递地点新增或变更,边界可能发生变化;数据每季度更新一次,并仅向子公司和合作伙伴有限提供。该数据可用于地图、邮寄、营销、保险、房地产、物流、导航和交通等行业。
Apple services are reported to be down.据报道,Apple 服务发生中断。
Apple services are reported to be down. The available information does not identify which services are affected, how long the outage has lasted, or its cause.据报道,Apple 服务发生中断。现有信息未说明受影响的具体服务、持续时间或原因。
Updated Google guidance reportedly confirms that content from connected Drive and email accounts can enter human-review pipelines for Gemini.Google 更新后的说明据称确认,连接到 Gemini 的 Drive 和电子邮件账户内容可能进入人工审核流程。
Updated Google guidance reportedly confirms that content from connected Drive and email accounts can enter human-review pipelines for Gemini. This reverses a 2023 promise that personal Gmail, Docs, and Drive content would not be seen by human reviewers. Google now warns users not to connect apps containing confidential information they would not want a reviewer to see.Google 更新后的说明据称确认,连接到 Gemini 的 Drive 和电子邮件账户内容可能进入人工审核流程。这改变了 Google 在 2023 年作出的承诺——当时公司表示个人 Gmail、Docs 和 Drive 内容不会被人工审核员看到。Google 目前警告用户,不要连接包含不希望审核员看到的机密信息的应用。
For personal Google AI subscriptions, saved Gemini chats can be used to improve Google services and train generative AI models when Keep Activity is enabled.对于个人 Google AI 订阅账户,在启用“保留活动记录”时,保存的 Gemini 聊天记录可用于改进 Google 服务和训练生成式 AI 模型。
For personal Google AI subscriptions, saved Gemini chats can be used to improve Google services and train generative AI models when Keep Activity is enabled. A subset of chats may be reviewed by human reviewers and can be retained for up to three years after review, even if the user deletes activity. Users are warned not to enter confidential information, and opting out requires disabling chat history or using temporary chats.对于个人 Google AI 订阅账户,在启用“保留活动记录”时,保存的 Gemini 聊天记录可用于改进 Google 服务和训练生成式 AI 模型。部分聊天可能由人工审核,审核过的内容即使用户删除活动记录,也可能保留最长三年。用户被提醒不要输入机密信息;若要退出,需要关闭聊天记录或使用临时聊天。
The restrooms along Norway’s Scenic Routes are presented as architectural landmarks designed to work with the surrounding landscape.挪威风景公路沿线的卫生间被视为与周围自然环境协调的建筑地标。
The restrooms along Norway’s Scenic Routes are presented as architectural landmarks designed to work with the surrounding landscape. Most were designed by established or emerging Norwegian architecture firms. Examples include Snøhetta’s small Eggum stop in Lofoten and Reiulf Ramstad’s facility at Selvika.挪威风景公路沿线的卫生间被视为与周围自然环境协调的建筑地标。大多数设施由知名或新兴的挪威建筑事务所设计。例如,罗弗敦群岛 Eggum 休息点由 Snøhetta 设计,Selvika 的设施则由 Reiulf Ramstad 设计。
Reverse engineering of Windows 11’s MissionControlAggregator found coding-agent and developer-tool usage bits in a diagnostic ETW report.对 Windows 11 的 MissionControlAggregator 进行逆向分析后,研究者在诊断 ETW 报告中发现了编程代理和开发工具的使用标记。
Reverse engineering of Windows 11’s MissionControlAggregator found coding-agent and developer-tool usage bits in a diagnostic ETW report. The observed list includes Codex, Claude, Copilot, Cursor, Python, .NET, GCC, and Visual Studio Code executables. The finding raises questions about why a Windows system DLL records the use of these tools.对 Windows 11 的 MissionControlAggregator 进行逆向分析后,研究者在诊断 ETW 报告中发现了编程代理和开发工具的使用标记。相关列表包括 Codex、Claude、Copilot、Cursor,以及 Python、.NET、GCC 和 Visual Studio Code 等可执行文件。这一发现引出了一个问题:Windows 系统 DLL 为何要记录这些工具的使用情况。
The Socratic Tutor is an agent skill designed to guide users toward genuine mastery through questions that make them reason, predict, and explain rather than simply listen to a lecture.“苏格拉底式导师”是一项代理技能,通过要求用户推理、预测和解释来帮助其真正掌握知识,而不是单纯进行讲授。
The Socratic Tutor is an agent skill designed to guide users toward genuine mastery through questions that make them reason, predict, and explain rather than simply listen to a lecture. It is intended for ongoing teaching sessions, not one-off factual questions. When the user’s intent is ambiguous, it first asks whether they want a quick answer or a full tutoring session.“苏格拉底式导师”是一项代理技能,通过要求用户推理、预测和解释来帮助其真正掌握知识,而不是单纯进行讲授。它适用于持续的教学过程,不适用于一次性的事实问答。当用户意图不明确时,它会先询问用户想要快速答案还是完整的辅导过程。
A compatibility report examines whether the ASN Bank Android app blocks non-Google Android systems, including GrapheneOS.一份兼容性报告调查 ASN Bank 的 Android 应用是否屏蔽包括 GrapheneOS 在内的非 Google Android 系统。
A compatibility report examines whether the ASN Bank Android app blocks non-Google Android systems, including GrapheneOS. The report identifies app version 2.40.2 in the Netherlands. One user reported that the app appeared to work on a stock operating system.一份兼容性报告调查 ASN Bank 的 Android 应用是否屏蔽包括 GrapheneOS 在内的非 Google Android 系统。报告涉及荷兰地区的 2.40.2 版本。一名用户表示,该应用在原生操作系统上看起来可以正常运行。
Moonshot AI has released the weights for Kimi K3.1 Pro, an open-weight, natively multimodal agentic model with 2.8 trillion parameters and a one-million-token context window.月之暗面发布了 Kimi K3.1 Pro 的模型权重。
Moonshot AI has released the weights for Kimi K3.1 Pro, an open-weight, natively multimodal agentic model with 2.8 trillion parameters and a one-million-token context window. It was post-trained from Kimi K3 using large-scale joint reinforcement learning across hundreds of domains. The model replaces Gated MLA with Gated DeepSeek Sparse Attention for more efficient long-context modeling.月之暗面发布了 Kimi K3.1 Pro 的模型权重。这是一款开放权重、原生多模态的智能体模型,拥有 2.8 万亿参数和 100 万 token 的上下文窗口。该模型基于 Kimi K3,并通过覆盖数百个领域的大规模联合强化学习进行后训练;同时以 Gated DeepSeek Sparse Attention 替代 Gated MLA,以提高长上下文建模效率。
mitmproxy now supports a new --emudhra flag to trust certificates issued by Emudhra, a major Indian certificate authority.mitmproxy 新增 --emudhra 选项,用于信任 Emudhra 颁发的证书——Emudhra 是印度的主要证书颁发机构。
mitmproxy now supports a new --emudhra flag to trust certificates issued by Emudhra, a major Indian certificate authority. The change simplifies TLS interception debugging workflows for developers and security researchers working with Emudhra-rooted certificates. Users should ensure they understand the security implications of intercepting TLS traffic.mitmproxy 新增 --emudhra 选项,用于信任 Emudhra 颁发的证书——Emudhra 是印度的主要证书颁发机构。此变更简化了在使用 Emudhra 证书的环境中的 TLS 拦截和调试工作流程。用户应了解拦截 TLS 流量所带来的安全风险。
Claude Code transmits each background agent’s transcripts to its model, raising privacy and data-use concerns for developers.Claude Code 会将后台代理的对话记录发送给模型,引发隐私与数据使用方面的担忧。
Claude Code transmits each background agent’s transcripts to its model, raising privacy and data-use concerns for developers. This telemetry behavior highlights how data sharing can impact confidentiality in code tools.Claude Code 会将后台代理的对话记录发送给模型,引发隐私与数据使用方面的担忧。此类遥测行为凸显了代码工具在数据共享方面的潜在风险。
A hands-on project demonstrates booting and running the full iOS 16.5 firmware on Android hardware, challenging cross-OS compatibility.一个动手项目展示在 Android 硬件上跑完整的 iOS 16.5 固件,挑战跨操作系统兼容性。
A hands-on project demonstrates booting and running the full iOS 16.5 firmware on Android hardware, challenging cross-OS compatibility. It highlights the iOS boot chain, kernel and driver constraints, and what the exercise reveals about the barriers between mobile ecosystems.一个动手项目展示在 Android 硬件上跑完整的 iOS 16.5 固件,挑战跨操作系统兼容性。该工作揭示了 iOS 启动链、内核与驱动的约束,以及两大移动生态之间的壁垒。
Nordwand Mono is a new monospaced font designed for coders and adventurers, aiming for high legibility and distinctive personality in IDEs and terminals.Nordwand Mono 是一款面向程序员与爱好者的新型等宽字体,强调在集成开发环境和终端中的高可读性与独特风格。
Nordwand Mono is a new monospaced font designed for coders and adventurers, aiming for high legibility and distinctive personality in IDEs and terminals. It joins a growing ecosystem of expressive, practical coding typefaces that optimize readability during long sessions.Nordwand Mono 是一款面向程序员与爱好者的新型等宽字体,强调在集成开发环境和终端中的高可读性与独特风格。它加入了注重可读性和个性表达的编码字体家族,适合长时间编码时使用。
OpenShell provides a private, sandboxed runtime to run fleets of autonomous AI agents with strong isolation and governance.OpenShell 提供私有、沙箱化的运行时环境,用于部署成群的自主 AI 代理,强调强隔离与治理。
OpenShell provides a private, sandboxed runtime to run fleets of autonomous AI agents with strong isolation and governance. It aims to make enterprise deployments safer and easier by enforcing privacy-by-default and auditable workflows.OpenShell 提供私有、沙箱化的运行时环境,用于部署成群的自主 AI 代理,强调强隔离与治理。面向企业场景,聚焦隐私保护和可审计的工作流,提升安全性和合规性。
smolmaxxing explores how to run the largest practical language models on small or old devices such as TV boxes, phones, single-board computers, and routers.smolmaxxing 探索如何在电视盒子、旧手机、单板计算机和路由器等小型或老旧设备上运行实际可用的最大语言模型。
smolmaxxing explores how to run the largest practical language models on small or old devices such as TV boxes, phones, single-board computers, and routers. It uses custom kernels, memory optimization, and weight streaming, with claims measured on the actual hardware. On a Xiaomi TV Box S 3rd Gen with about 1 GB of usable RAM, the project reports running models ranging from 260K to 16B parameters.smolmaxxing 探索如何在电视盒子、旧手机、单板计算机和路由器等小型或老旧设备上运行实际可用的最大语言模型。项目使用定制内核、内存优化和权重流式传输,并在真实硬件上测量各项结果。在可用内存约 1 GB 的 Xiaomi TV Box S 第三代上,项目报告称可运行参数量从 26 万到 160 亿不等的模型。
OBS Studio 33 arrives with new features, performance tweaks, and bug fixes, continuing the project’s open-source momentum for live streaming and screen capture.OBS Studio 33 发布,带来新功能、性能优化与 bug 修复,持续推动开源直播和录制工具的发展。
OBS Studio 33 arrives with new features, performance tweaks, and bug fixes, continuing the project’s open-source momentum for live streaming and screen capture. The update improves stability and compatibility for creators relying on high-fidelity audio/video.OBS Studio 33 发布,带来新功能、性能优化与 bug 修复,持续推动开源直播和录制工具的发展。此次更新提升了稳定性与兼容性,方便依赖高保真的主播和创作者。
Memdoor implements a team-solving workflow in which five small models act as a planner, ideators, solver, manager, and repair worker.Memdoor 实现了一种协作解题流程,由五个小模型分别担任规划者、创意提出者、求解器、管理者和修复者。
Memdoor implements a team-solving workflow in which five small models act as a planner, ideators, solver, manager, and repair worker. Each model starts with a fresh context and communicates only through files, while an external script determines whether solutions pass. The authors reproduced the method using Qwen3.8-27B, eight YAML files, their prompts, and a 128K response limit.Memdoor 实现了一种协作解题流程,由五个小模型分别担任规划者、创意提出者、求解器、管理者和修复者。每个模型都从全新的上下文开始,仅通过磁盘文件通信,并由外部脚本判断解答是否通过。作者使用 Qwen3.8-27B、八个 YAML 文件、自定义提示词和 128K 回复上限复现了这一方法。
Microsoft's ONNX Runtime has disappeared from GitHub.微软的 ONNX Runtime 已从 GitHub 消失。
Microsoft's ONNX Runtime has disappeared from GitHub. The available information does not establish why it disappeared, when the change occurred, or what will happen to the project.微软的 ONNX Runtime 已从 GitHub 消失。现有信息无法确认其消失原因、发生时间或项目后续安排。
A Chrome-led project provides coding agents with curated guidance on modern, native web APIs.一个由 Chrome 发起的项目为编码代理提供现代原生 Web API 的精选指南。
A Chrome-led project provides coding agents with curated guidance on modern, native web APIs. Examples include CSS highlights, dialog and showModal(), and field-sizing, which can replace common JavaScript-heavy patterns. The open-source release contains more than 200 expert-written guides, developed with Firefox, MDN, and Edge, and is designed for agent skills and retrieval-augmented tooling.一个由 Chrome 发起的项目为编码代理提供现代原生 Web API 的精选指南。例如,CSS Highlights、dialog 与 showModal()、field-sizing 都可以替代常见的重 JavaScript 实现。该开源项目目前包含 200 多份由专家编写的指南,并与 Firefox、MDN 和 Edge 合作开发,面向代理技能和检索增强工具。
MicroG offers a libre reimplementation of Google Mobile Services, enabling Android runs without proprietary Google binaries.MicroG 提供对 Google 移动服务的自由开源再实现,使用户在不使用官方谷歌二进制的情况下运行 Android。
MicroG offers a libre reimplementation of Google Mobile Services, enabling Android runs without proprietary Google binaries. It supports many apps and services while prioritizing privacy, empowering custom ROMs and privacy-focused devices, though some apps may rely on proprietary features.MicroG 提供对 Google 移动服务的自由开源再实现,使用户在不使用官方谷歌二进制的情况下运行 Android。尽管兼容性可能有限,但它为隐私友好的设备与自定义 ROM 提供了可行选项,某些应用仍可能依赖专有特性。
Llama-modchip is a collection of ggml and llama.cpp kernel patches, quantization transcoders, and runtime tools for running sparse MoE models entirely in system RAM.Llama-modchip 是一组 ggml 和 llama.cpp 内核补丁、量化转换器及运行时工具,旨在让稀疏 MoE 模型完全运行于系统内存中。
Llama-modchip is a collection of ggml and llama.cpp kernel patches, quantization transcoders, and runtime tools for running sparse MoE models entirely in system RAM. Its project page claims that a 284B DeepSeek-V4-Flash model can run on a 64GB CPU-only system at a 256K context length, with reported prompt-processing speeds of 44.4 tok/s and generation speeds of 9.1–14.7 tok/s. It supports x86_64 and ARM64 processors and also targets 30B MoE models.Llama-modchip 是一组 ggml 和 llama.cpp 内核补丁、量化转换器及运行时工具,旨在让稀疏 MoE 模型完全运行于系统内存中。项目页面称,284B 参数的 DeepSeek-V4-Flash 可在仅配备 64GB 内存、没有 GPU 的 CPU 系统上运行,支持 256K 上下文,提示处理速度为 44.4 token/s,生成速度为 9.1–14.7 token/s。该项目支持 x86_64 和 ARM64 处理器,也面向 30B MoE 模型。
ggml 1.0.0 removes deprecated public APIs and the legacy DIAG_MASK operations, while adding configurable YaRN truncation and FP8 support.ggml 1.0.0 移除了已弃用的公共 API 和旧版 DIAG_MASK 操作,并加入可配置的 YaRN 截断和 FP8 支持。
ggml 1.0.0 removes deprecated public APIs and the legacy DIAG_MASK operations, while adding configurable YaRN truncation and FP8 support. It also delivers performance work across CUDA, Metal, Vulkan, SYCL, OpenCL, WebGPU, and Hexagon backends, including quantized matrix multiplication, MoE fusion, FlashAttention paths, and scheduler improvements. The release broadens hardware and backend support while reducing unnecessary input uploads.ggml 1.0.0 移除了已弃用的公共 API 和旧版 DIAG_MASK 操作,并加入可配置的 YaRN 截断和 FP8 支持。该版本还改进了 CUDA、Metal、Vulkan、SYCL、OpenCL、WebGPU 和 Hexagon 后端的性能,包括量化矩阵乘法、MoE 融合、FlashAttention 路径和调度器优化。此次发布扩大了硬件与后端支持,并减少了不必要的输入上传。
Florilegium generates Delft-blue floral page margins for LuaLaTeX using Lua.Florilegium 使用 Lua 为 LuaLaTeX 生成代尔夫特蓝风格的花卉页面边饰。
Florilegium generates Delft-blue floral page margins for LuaLaTeX using Lua. It combines hand-authored primitives, randomized parameters, and a procedural layout that fills unused space. The creator says the project produces a similar look to a referenced design without trying to replicate it, and used Claude during development.Florilegium 使用 Lua 为 LuaLaTeX 生成代尔夫特蓝风格的花卉页面边饰。它结合手工编写的基本图形、随机参数和用于填充空白区域的程序化布局。作者表示,该项目形成了与某个参考设计相似的效果,但并未试图复制原作,开发过程中使用了 Claude。
Google open-sources TorchTPU, a native PyTorch backend for TPUs, making TPUs feel like a first-class citizen for PyTorch users.Google 将 TorchTPU 开源,提供对 TPUs 的原生 PyTorch 后端,使 PyTorch 用户能够直接在 TPUs 上运行模型。
Google open-sources TorchTPU, a native PyTorch backend for TPUs, making TPUs feel like a first-class citizen for PyTorch users. This lowers the barrier to running PyTorch models on Google's accelerators and could shift performance comparisons across ML stacks.Google 将 TorchTPU 开源,提供对 TPUs 的原生 PyTorch 后端,使 PyTorch 用户能够直接在 TPUs 上运行模型。此举降低了在 TPU 上运行 PyTorch 的门槛,可能改变不同 ML 堆栈间的性能对比。
jQuery 4.1.0 was released, with the release notes listing an Ajax change that adds a responseURL field to jqXHR.jQuery 4.1.0 已发布,更新说明列出了一项 Ajax 改动:为 jqXHR 添加 responseURL 字段。
jQuery 4.1.0 was released, with the release notes listing an Ajax change that adds a responseURL field to jqXHR. The same announcement also references a jQuery 3.8.0 release. The update continues development of the JavaScript library across its newer and older release lines.jQuery 4.1.0 已发布,更新说明列出了一项 Ajax 改动:为 jqXHR 添加 responseURL 字段。同一公告还提到了 jQuery 3.8.0 的发布。此次更新同时推进了该 JavaScript 库的新旧版本线。
bootvine is a boot animation that displays systemd stages as a tree.bootvine 是一款启动动画,会将 systemd 的各个阶段显示为一棵树。
bootvine is a boot animation that displays systemd stages as a tree. It provides a visual representation of systemd jobs during system startup.bootvine 是一款启动动画,会将 systemd 的各个阶段显示为一棵树。它在系统启动过程中以可视化方式呈现 systemd 作业。
Disc Image Studio converts an image into data that conforms to the physical-layer constraints of CDs, DVDs, and single-layer BD-R discs.Disc Image Studio 可将图片转换为符合 CD、DVD 和单层 BD-R 光盘物理层约束的数据,刻录后光盘表面会呈现可见图案。
Disc Image Studio converts an image into data that conforms to the physical-layer constraints of CDs, DVDs, and single-layer BD-R discs. After burning, the disc surface displays a visible pattern. The Windows desktop version supports design and burning, while Linux and macOS have experimental offline command-line versions.Disc Image Studio 可将图片转换为符合 CD、DVD 和单层 BD-R 光盘物理层约束的数据,刻录后光盘表面会呈现可见图案。Windows 桌面版支持设计与刻录,Linux 和 macOS 则提供实验性的离线命令行版本。
A project demonstrates running Android apps natively on macOS, potentially removing emulation overhead.一个项目展示在 macOS 上原生运行 Android 应用,或许能消除仿真开销。
A project demonstrates running Android apps natively on macOS, potentially removing emulation overhead. If mature, this could blur platform boundaries for developers and users, enabling smoother access to Android ecosystems on Apple Silicon.一个项目展示在 macOS 上原生运行 Android 应用,或许能消除仿真开销。若成熟,将模糊开发者与用户的跨平台边界,使 Android 生态在 Apple Silicon 上获得更顺畅的体验。
A community project analyzes FP and BF16 instruction issue rates on Nvidia RTX 30-series cards, suggesting bottlenecks that affect mixed-precision workloads.一个社区项目对 NVIDIA RTX 30 系列显卡的 FP 与 BF16 指令发射率进行分析,揭示对混合精度任务的潜在瓶颈。
A community project analyzes FP and BF16 instruction issue rates on Nvidia RTX 30-series cards, suggesting bottlenecks that affect mixed-precision workloads. The findings could inform driver optimizations, compiler choices, and how developers allocate precision in AI and HPC tasks.一个社区项目对 NVIDIA RTX 30 系列显卡的 FP 与 BF16 指令发射率进行分析,揭示对混合精度任务的潜在瓶颈。相关结果可能为驱动优化、编译器选择以及开发者在 AI 和 HPC 任务中对精度的分配提供参考。
Window Sweaters is a macOS menu-bar app that adds knitted borders to application windows.Window Sweaters 是一款 macOS 菜单栏应用,会为应用窗口添加针织风格的边框。
Window Sweaters is a macOS menu-bar app that adds knitted borders to application windows. It offers sweater-like borders and colors inspired by different apps. The project can also be handed to a coding agent to install.Window Sweaters 是一款 macOS 菜单栏应用,会为应用窗口添加针织风格的边框。它提供类似毛衣的边框和受不同应用启发的配色。项目还可以交给编程代理来完成安装。
OmniDiff is an AGPLv3 open-source diff tool written in Rust that handles code, configuration, images, PDFs, archives, cursors, and fonts.OmniDiff 是一个采用 AGPLv3 许可、用 Rust 编写的开源差异工具,支持代码、配置文件、图像、PDF、压缩包、光标和字体等内容。
OmniDiff is an AGPLv3 open-source diff tool written in Rust that handles code, configuration, images, PDFs, archives, cursors, and fonts. It can compare C++ files tens of thousands of lines long in under 100 milliseconds on a 2023 ThinkPad T14s. Its pipelines were informed by manually studying 2,000 diffs sampled from 7,400 Gentoo package repositories, and the project is currently at version 0.3.0.OmniDiff 是一个采用 AGPLv3 许可、用 Rust 编写的开源差异工具,支持代码、配置文件、图像、PDF、压缩包、光标和字体等内容。它在 2023 年款 ThinkPad T14s 上可在不到 100 毫秒内比较数万行的 C++ 文件。其处理流程源于对 7400 个 Gentoo 软件包仓库中抽样的 2000 个差异进行人工研究,目前版本为 0.3.0。
Show HNOpen SourceProgrammingCodeNov 211月2日0.531094
cfsolve is a Rust library and command-line tool that solves Cloudflare Turnstile widgets and challenge pages without a headless browser.cfsolve 是一个 Rust 库和命令行工具,无需无头浏览器即可处理 Cloudflare Turnstile 小部件和挑战页面。
cfsolve is a Rust library and command-line tool that solves Cloudflare Turnstile widgets and challenge pages without a headless browser. It returns a Turnstile token for widgets or a cf_clearance cookie for challenge pages. Widgets can be handled with a page URL and site key, while challenge pages require a direct fetch or the page’s HTML.cfsolve 是一个 Rust 库和命令行工具,无需无头浏览器即可处理 Cloudflare Turnstile 小部件和挑战页面。它会为小部件返回 Turnstile 令牌,为挑战页面返回 cf_clearance Cookie。处理小部件需要页面 URL 和站点密钥,而挑战页面则需要直接抓取页面或提供页面 HTML。
Yogurt is a free, MIT-licensed desktop app that uses on-device AI models to transcribe, summarize, and chat about meetings.Yogurt 是一款采用 MIT 许可证的免费桌面应用,使用设备本地的 AI 模型转录会议、生成摘要并进行问答。
Yogurt is a free, MIT-licensed desktop app that uses on-device AI models to transcribe, summarize, and chat about meetings. It runs in the background without joining calls, and keeps transcripts, recordings, and summaries on the user’s computer. The app also offers calendar integration and a local MCP server for optionally sharing notes with external frontier models.Yogurt 是一款采用 MIT 许可证的免费桌面应用,使用设备本地的 AI 模型转录会议、生成摘要并进行问答。它在后台运行,不会加入通话,会议转录、录音和摘要也不会离开用户的电脑。该应用还提供日历集成和本地 MCP 服务器,用户可以选择通过它将笔记分享给外部前沿模型。
Coldframe is a Python honeypot that impersonates supermarket refrigeration controllers, including the Danfoss AK-SM 800A and Dixell XWEB interfaces.Coldframe 是一个用 Python 编写的蜜罐,伪装成超市制冷控制器,包括 Danfoss AK-SM 800A 和 Dixell XWEB 的界面。
Coldframe is a Python honeypot that impersonates supermarket refrigeration controllers, including the Danfoss AK-SM 800A and Dixell XWEB interfaces. It uses only the Python standard library and matches details such as headers, page sizes, certificate fields, and TLS versions found on real devices. The honeypot recognizes several exploit attempts, including the CVE-2025-41450 login bypass, and records submitted codes.Coldframe 是一个用 Python 编写的蜜罐,伪装成超市制冷控制器,包括 Danfoss AK-SM 800A 和 Dixell XWEB 的界面。它仅使用 Python 标准库,并模拟真实设备的请求头、页面大小、证书字段和 TLS 版本等细节。该蜜罐能够识别多种攻击尝试,包括 CVE-2025-41450 登录绕过,并记录提交的代码。
The project compiles an NES CPU into the weights of a standard Qwen3.5 model.该项目将 NES CPU 编译进标准 Qwen3.5 模型的权重中。
The project compiles an NES CPU into the weights of a standard Qwen3.5 model. The NES 6502 CPU, 2 KB of work RAM, and cartridge ROM run inside the model checkpoint, with every weight computed by a compiler rather than learned through training. During play, the model predicts the CPU execution log token by token while a host program handles graphics, controls, and the vertical-blank interrupt.该项目将 NES CPU 编译进标准 Qwen3.5 模型的权重中。NES 的 6502 CPU、2 KB 工作内存和卡带程序 ROM 都运行在这个模型检查点内,所有权重均由编译器计算得出,而不是通过训练学习。游戏运行时,模型逐 token 预测 CPU 执行日志,主机程序则负责绘图、读取手柄输入和触发垂直消隐中断。
Show HN: A Truly Modular Keyboard showcases a build with swappable parts and open hardware ethos.Show HN: 真正的模块化键盘展示了一个具备可替换部件、开放软硬件理念的设计。
Show HN: A Truly Modular Keyboard showcases a build with swappable parts and open hardware ethos. It demonstrates how design-for-modularity can extend device lifespans and tailor ergonomics. The project invites readers to rethink how keyboards can evolve with users.Show HN: 真正的模块化键盘展示了一个具备可替换部件、开放软硬件理念的设计。此类模块化设计能延长设备寿命并更好地满足用户的打字需求。该项目促使人们重新思考键盘如何随用户演进。
NVIDIA Auto Ontology is an open Python framework for building semantic, graph-based reasoning layers over data.NVIDIA Auto Ontology 是一个用于构建语义化图推理层的开源 Python 框架。
NVIDIA Auto Ontology is an open Python framework for building semantic, graph-based reasoning layers over data. It reconciles enterprise metadata and definitions across systems, helping agents locate, interpret, and retrieve data more reliably. The layer is exposed to agent systems through MCP, CLI, and API, and supports text-to-SQL evaluation workflows.NVIDIA Auto Ontology 是一个用于构建语义化图推理层的开源 Python 框架。它整合企业元数据并协调不同系统中的定义,帮助智能体更可靠地定位、理解和检索数据。该推理层可通过 MCP、CLI 和 API 接入智能体,也支持 text-to-SQL 评测流程。
A project implements the DeepSeek-V3 architecture in 190 lines of pure JAX, including MLA, DeepSeekMoE, auxiliary-loss-free balancing, and multi-token prediction.一个项目用 190 行纯 JAX 实现了 DeepSeek-V3 架构,包括 MLA、DeepSeekMoE、无辅助损失均衡和多词元预测。
A project implements the DeepSeek-V3 architecture in 190 lines of pure JAX, including MLA, DeepSeekMoE, auxiliary-loss-free balancing, and multi-token prediction. The model was pretrained from scratch on a MacBook Air using public-domain detective fiction, without PyTorch, Flax, or Equinox. Its parameters are plain dictionaries of arrays, and the implementation is organized as readable functions.一个项目用 190 行纯 JAX 实现了 DeepSeek-V3 架构,包括 MLA、DeepSeekMoE、无辅助损失均衡和多词元预测。该模型在 MacBook Air 上使用公版侦探小说从头预训练,未使用 PyTorch、Flax 或 Equinox。模型参数采用由数组组成的普通字典,代码则组织为可从头读到尾的函数。
In the SlopCodeBench 10/2026 report, Opus 5.5 leads the coding benchmarks while Astra posts the strongest overall score.在 SlopCodeBench 2026年10月的评测中,Opus 5.5领跑代码基准测试,Astra在总分上表现最优。
In the SlopCodeBench 10/2026 report, Opus 5.5 leads the coding benchmarks while Astra posts the strongest overall score. The assessment also marks xhigh as harmful, highlighting ongoing safety and reliability debates for coding agents and the value of diverse benchmarking在 SlopCodeBench 2026年10月的评测中,Opus 5.5领跑代码基准测试,Astra在总分上表现最优。同时将 xhigh 标记为有害,凸显编码代理在安全性与可靠性方面的持续讨论以及多维基准的重要性。
Microsoft unveils Type Server Protocol, a new layer that augments the Language Server Protocol.微软推出 Type Server Protocol,为语言服务器协议引入一层新接口,允许编辑器与语言服务器更丰富地交换类型信息与状态。
Microsoft unveils Type Server Protocol, a new layer that augments the Language Server Protocol. The protocol enables editors and language servers to exchange richer typing information and state, improving performance and scalability for multi-language toolchains. This move could help unify editor experiences across languages and tooling ecosystems.微软推出 Type Server Protocol,为语言服务器协议引入一层新接口,允许编辑器与语言服务器更丰富地交换类型信息与状态。此举有望提升多语言工具链的性能与可扩展性,并统一编辑器体验。
Herdr released a native app, expanding beyond its earlier web or cross-platform offering and delivering faster performance, offline support, and deeper OS integration.Herdr 发布了原生应用,扩展了其此前的网页/跨平台版本,带来更快的性能、离线支持和更深的系统集成。
Herdr released a native app, expanding beyond its earlier web or cross-platform offering and delivering faster performance, offline support, and deeper OS integration. The move signals a maturation in the product and potential for broader platform reach.Herdr 发布了原生应用,扩展了其此前的网页/跨平台版本,带来更快的性能、离线支持和更深的系统集成。此举标志着产品成熟度提升,并可能带来更广泛的应用平台覆盖。
Astroterm is a terminal-based planetarium that renders sky maps and astronomical data directly in your shell.Astroterm 将星空投影带入终端,直接在外壳中渲染天图与星历数据。
Astroterm is a terminal-based planetarium that renders sky maps and astronomical data directly in your shell. It brings rich visualization to environments without a full GUI, making it a handy tool for quick stargazing and education.Astroterm 将星空投影带入终端,直接在外壳中渲染天图与星历数据。即使在无完整图形界面的环境下,也能进行直观的天文可视化,适合快速观星与学习。
OpenAI released an official command-line interface, letting developers interact with models, manage API keys, and automate tasks from the terminal.OpenAI 发布官方命令行界面,方便开发者在终端与模型交互、管理 API Key 以及执行自动化任务。
OpenAI released an official command-line interface, letting developers interact with models, manage API keys, and automate tasks from the terminal. The CLI lowers barriers for scripting, automation, and CI workflows, complementing the Python client.OpenAI 发布官方命令行界面,方便开发者在终端与模型交互、管理 API Key 以及执行自动化任务。CLI 降低了上手门槛,适合在脚本、持续集成等场景使用,与 Python 客户端形成互补。
A tiny world model runs on a Raspberry Pi Pico 2 W, showing that edge AI is creeping into microcontrollers.在 Raspberry Pi Pico 2 W 上跑出一个极小的世界模型,展示边缘AI在微控制器上的潜力。
A tiny world model runs on a Raspberry Pi Pico 2 W, showing that edge AI is creeping into microcontrollers. The demo demonstrates memory-efficient planning and small neural models that fit in ultra‑compact hardware, hinting at future autonomous nodes in IoT. It signals potential for smarter, low-power devices at the edge.在 Raspberry Pi Pico 2 W 上跑出一个极小的世界模型,展示边缘AI在微控制器上的潜力。该演示表明在资源受限设备上也能运行简短的神经模型和规划任务。预示着物联网边缘设备将拥有更智能的能力。
Hedra is a 3D open world game created from scratch in high school.Hedra 是作者在高中时独自从零开始开发的 3D 开放世界游戏。
Hedra is a 3D open world game created from scratch in high school. This solo project demonstrates independent game development skills, including world-building, controls, and basic engine work, offering a peek into a teen's indie effort.Hedra 是作者在高中时独自从零开始开发的 3D 开放世界游戏。该作品展示了单人开发的完整流程,包括世界构建、控制与基础引擎工作,呈现青少年独立游戏开发的有趣案例。
PaperPilot is a zero-cloud, pure-Rust PDF engine designed to run in WebAssembly with MCP support.PaperPilot 是一个零云、纯 Rust 的 PDF 引擎,支持在 WebAssembly 中运行并配合 MCP 使用。
PaperPilot is a zero-cloud, pure-Rust PDF engine designed to run in WebAssembly with MCP support. It enables offline, self-contained PDF processing that can be embedded into apps without external services. This could be appealing for privacy-focused workflows and edge deployments.PaperPilot 是一个零云、纯 Rust 的 PDF 引擎,支持在 WebAssembly 中运行并配合 MCP 使用。它实现离线、即可嵌入的 PDF 处理,避免对外部服务的依赖。对注重隐私的工作流和边缘部署尤具吸引力。
Microsoft’s Process Explorer for Mac is a native macOS implementation of the core Process Explorer experience.微软的 macOS 版 Process Explorer 是对核心 Process Explorer 体验的原生实现。
Microsoft’s Process Explorer for Mac is a native macOS implementation of the core Process Explorer experience. It provides a live process tree, module, handle, and thread inspection, per-process properties, system-wide performance graphs, and code-signing checks. The Swift, SwiftUI, and AppKit application uses public macOS APIs, with optional support for a privileged helper.微软的 macOS 版 Process Explorer 是对核心 Process Explorer 体验的原生实现。它提供实时进程树、模块、句柄和线程检查、逐进程属性、系统级性能图表以及代码签名检查。该应用使用 Swift、SwiftUI 和 AppKit 构建,并通过 macOS 公共 API 工作,还可选择使用特权辅助程序。
A playful experimental debugger that applies Nietzschean prompts to guide debugging strategies, offering a philosophical UX to debugging workflows.一个将尼采式提示应用于调试策略的实验性调试器,提供哲学化的调试体验。
A playful experimental debugger that applies Nietzschean prompts to guide debugging strategies, offering a philosophical UX to debugging workflows. It’s a thought-provoking take on making debugging less mechanical and more narrative.一个将尼采式提示应用于调试策略的实验性调试器,提供哲学化的调试体验。它挑战传统的机械式排错,试图让调试过程更具叙事性。
fheroes2 is an open-source recreation of the Heroes of Might and Magic II engine, aiming to faithfully reproduce classic gameplay across platforms.fheroes2 是对《魔法门之英雄无敌 II》引擎的开源重建,目标在各平台上忠实再现经典玩法。
fheroes2 is an open-source recreation of the Heroes of Might and Magic II engine, aiming to faithfully reproduce classic gameplay across platforms. The project modernizes the engine while preserving the original's look and feel, enabling new audiences to experience the game. It demonstrates how community-driven reverse-engineering can extend retro games.fheroes2 是对《魔法门之英雄无敌 II》引擎的开源重建,目标在各平台上忠实再现经典玩法。项目在保留原作风格的同时实现引擎现代化,帮助新玩家体验这款经典游戏。它展示了社区驱动的逆向工程如何延续复古游戏的生命力。
OnKey analyzes your singing pitch in real time and marks off-key moments, turning practice into actionable feedback.OnKey 实时分析你的歌声音高并标出走音片段,把练习变成可执行的反馈。
OnKey analyzes your singing pitch in real time and marks off-key moments, turning practice into actionable feedback. Built as a Show HN project, it demonstrates how accessible audio analysis tools can be built with open-source tech and signal processing. It could inspire more consumer-friendly music-improvement apps.OnKey 实时分析你的歌声音高并标出走音片段,把练习变成可执行的反馈。作为 Show HN 项目,展示了用开源技术和信号处理构建音乐分析工具的可行性。它或许会激发更多面向大众的发声训练应用的开发。
The project proposes removing pgrust after indications of cheating, including per-query kernels and result caching that skew benchmarks.在发现存在作弊迹象(如逐查询内核与结果缓存扭曲基准结果)后,项目组提议移除 pgrust。
The project proposes removing pgrust after indications of cheating, including per-query kernels and result caching that skew benchmarks. The move reinforces the need for honest, transparent benchmarking in open-source projects and could affect credibility of reported performance figures. Maintainers emphasize fair competition and reproducibility.在发现存在作弊迹象(如逐查询内核与结果缓存扭曲基准结果)后,项目组提议移除 pgrust。此举强调在开源基准测试中保持公正、透明的重要性,可能影响对性能数据的信任度,并敦促维护者关注可复现性与公平竞争。
Nibnib is an image format using 4 bits per pixel and optimized for extremely fast decompression.Nibnib 是一种每像素 4 bit、针对极快解压速度优化的图像格式。
Nibnib is an image format using 4 bits per pixel and optimized for extremely fast decompression. It began as an experiment targeting DVI output on an RP2350 microcontroller, where roughly six cycles per pixel must cover pixel fetching, palette lookup, and storage in a line buffer.Nibnib 是一种每像素 4 bit、针对极快解压速度优化的图像格式。它最初是为 RP2350 微控制器上的 DVI 输出进行的实验,该设备每个像素约有 6 个时钟周期,需要完成像素读取、调色板查询和写入行缓冲区。
OpenSSL 4.1.0 is released with security fixes and new features, signaling ongoing evolution of the crypto stack.OpenSSL 4.1.0 发布,包含安全修复与新特性,体现密码学库的持续演进。
OpenSSL 4.1.0 is released with security fixes and new features, signaling ongoing evolution of the crypto stack. The update may deprecate some APIs and require developers to test for compatibility with existing software. It’s a reminder to keep cryptographic dependencies up to date in production systems.OpenSSL 4.1.0 发布,包含安全修复与新特性,体现密码学库的持续演进。此次更新可能对某些 API 进行弃用并要求开发者进行兼容性测试,提醒在生产环境中保持加密依赖的最新状态。
Five Qwen3.8-27B models match Fable 5 using the GVS5H harness, illustrating how open-source LLMs can reach strong performance when paired with efficient hardware.五个 Qwen3.8-27B 模型在 GVS5H harness 下达到与 Fable 5 相近的性能,展示了开源大模型在搭配高效硬件时的竞争力。
Five Qwen3.8-27B models match Fable 5 using the GVS5H harness, illustrating how open-source LLMs can reach strong performance when paired with efficient hardware. This benchmark highlights the value of optimized deployment stacks for researchers and hobbyists alike.五个 Qwen3.8-27B 模型在 GVS5H harness 下达到与 Fable 5 相近的性能,展示了开源大模型在搭配高效硬件时的竞争力。此基准凸显了针对研究者和爱好者的优化部署堆栈的重要性。
With official maintenance ended, users should plan forks, migrations, or look for alternatives to keep security and compatibility.官方宣布不再维护,用户需要考虑分叉、迁移或替代方案以维持安全性与兼容性。
With official maintenance ended, users should plan forks, migrations, or look for alternatives to keep security and compatibility. The post may offer guidance on how to migrate or handle breaking changes while preserving the lightweight, accessible styling Pico CSS is known for.官方宣布不再维护,用户需要考虑分叉、迁移或替代方案以维持安全性与兼容性。文中给出迁移建议与应对潜在改动的途径,同时保留 Pico CSS 的简洁与可访问性。
A speculative piece arguing about why conditional constructs like if-then-else are essential in programming language design, exploring historical aspects and potential alternatives.一篇关于在编程语言设计中 if-then-else 条件表达式为何重要的讨论性文章,回顾历史背景并探讨潜在替代方案,质疑我们如何在代码中组织逻辑。
A speculative piece arguing about why conditional constructs like if-then-else are essential in programming language design, exploring historical aspects and potential alternatives. It challenges assumptions about how we structure logic in code.一篇关于在编程语言设计中 if-then-else 条件表达式为何重要的讨论性文章,回顾历史背景并探讨潜在替代方案,质疑我们如何在代码中组织逻辑。
Unleash is an open-source feature-management platform that lets teams control when and how new features are released.Unleash 是一个开源功能管理平台,让团队控制新功能何时以及如何向用户发布。
Unleash is an open-source feature-management platform that lets teams control when and how new features are released. It supports smaller, more manageable production deployments and can streamline software delivery. The project has been developed for 11 years.Unleash 是一个开源功能管理平台,让团队控制新功能何时以及如何向用户发布。它支持更小、更易管理的生产部署,并可简化软件交付流程。该项目已开发 11 年。
A price comparison normalizes on-demand GPU rental costs across AWS, Azure, and neocloud providers by GPU and hour.一项价格比较按 GPU 和每小时费用,对 AWS、Azure 及新型云服务商的按需 GPU 租赁价格进行了标准化。
A price comparison normalizes on-demand GPU rental costs across AWS, Azure, and neocloud providers by GPU and hour. For identical GPUs, it reports a 4.57× gap between the cheapest and most expensive options. Running one GPU continuously for a month costs $1,964 or $8,972 depending only on the provider.一项价格比较按 GPU 和每小时费用,对 AWS、Azure 及新型云服务商的按需 GPU 租赁价格进行了标准化。对于相同 GPU,最便宜和最昂贵选项之间的价格差达到 4.57 倍。连续运行一块 GPU 一个月,费用可能因服务商不同而为 1964 美元或 8972 美元。
Rune is a native UI framework for Rust that uses GPU rendering across macOS, Windows, Linux, and the web.Rune 是一个使用 GPU 渲染的 Rust 原生 UI 框架,支持 macOS、Windows、Linux 和 Web。
Rune is a native UI framework for Rust that uses GPU rendering across macOS, Windows, Linux, and the web. It provides its own renderer, text system, layout engine, and spring-based animations. Applications can also run headlessly, allowing agents to render PNG images, dump the element tree, and drive scripted input.Rune 是一个使用 GPU 渲染的 Rust 原生 UI 框架,支持 macOS、Windows、Linux 和 Web。它自带渲染器、文本系统、布局引擎和基于弹簧模型的动画。应用还可以无头运行,使智能体能够渲染 PNG、导出元素树并通过脚本驱动输入。
A PoC for a Bitcoin-themed Capture The Flag exercise demonstrates a vulnerability or attack vector in simulated wallets.这是一个比特币主题CTF的漏洞证明,演示了在模拟钱包环境中的攻击向量。
A PoC for a Bitcoin-themed Capture The Flag exercise demonstrates a vulnerability or attack vector in simulated wallets. It underscores how responsible disclosure and educational CTFs can teach secure crypto practices.这是一个比特币主题CTF的漏洞证明,演示了在模拟钱包环境中的攻击向量。强调通过负责任披露和教育性CTF提升加密安全意识。
Lonewolf 0.3.0 adds certificate-authenticated federation, direct TLS, and additional desktop-client features.Lonewolf 0.3.0 新增基于证书认证的联邦通信、直接 TLS,以及更多桌面客户端功能。
Lonewolf 0.3.0 adds certificate-authenticated federation, direct TLS, and additional desktop-client features. It remains a pre-1.0, single-node server, while full XMPP compliance and clustering are still on the roadmap. The release expands secure federation without yet providing those larger deployment features.Lonewolf 0.3.0 新增基于证书认证的联邦通信、直接 TLS,以及更多桌面客户端功能。它仍是一个处于 1.0 之前阶段的单节点服务器,完整 XMPP 兼容性和集群支持仍在路线图中。此次更新增强了安全联邦能力,但尚未提供这些更大规模部署功能。
Shitter is a fork of Nitter, offering a privacy-conscious, JS-light front-end for Twitter-like content and a self-hosted viewing experience.Shitter 是 Nitter 的一个分叉,提供一个轻量、注重隐私的前端来浏览类似 Twitter 的内容,并支持自托管使用。
Shitter is a fork of Nitter, offering a privacy-conscious, JS-light front-end for Twitter-like content and a self-hosted viewing experience. It exemplifies open-source efforts to reduce tracking and API reliance while preserving access to public posts.Shitter 是 Nitter 的一个分叉,提供一个轻量、注重隐私的前端来浏览类似 Twitter 的内容,并支持自托管使用。该项目展示了开源社区在减少跟踪、降低对官方 API 依赖方面的努力,同时保持对公开内容的访问。
Direct File is a US government service that lets taxpayers file federal returns electronically for free and directly with the IRS.Direct File 是美国政府提供的服务,允许纳税人免费通过电子方式直接向美国国税局提交联邦税表。
Direct File is a US government service that lets taxpayers file federal returns electronically for free and directly with the IRS. It uses interview-style questions, works across phones and computers, supports English and Spanish, and translates users' answers into tax forms submitted through the IRS Modernized e-File API. The repository contains the service's code and related tests and scenarios.Direct File 是美国政府提供的服务,允许纳税人免费通过电子方式直接向美国国税局提交联邦税表。该服务采用问答式界面,可在手机和电脑等设备上使用,支持英语和西班牙语,并将用户答案转换为税表后通过国税局 Modernized e-File API 提交。该开源仓库包含服务代码以及相关测试和场景。
Hob enables an on-device AI experience, offering an open-source alternative to Claude Cowork and ChatGPT that runs entirely locally.Hob 开源本地化 AI 体验,提供面向 Claude Cowork 与 ChatGPT 的本地替代方案。
Hob enables an on-device AI experience, offering an open-source alternative to Claude Cowork and ChatGPT that runs entirely locally. It demonstrates the feasibility of private, offline inference on consumer hardware, though model size and compute demands remain practical constraints.Hob 开源本地化 AI 体验,提供面向 Claude Cowork 与 ChatGPT 的本地替代方案。它展示了在消费级硬件上实现私有、离线推理的可行性,但模型规模与计算需求仍是实际制约。
Auto Dismiss Mac OS 'Allow app to find devices on local networks?' is a small open-source tool that automatically closes that permission prompt.Show HN 项目自动拒绝 macOS 的“允许应用发现本地网络设备?
Auto Dismiss Mac OS 'Allow app to find devices on local networks?' is a small open-source tool that automatically closes that permission prompt. It highlights the tension between a smooth UX and strict local-network permissions, inviting a discussion on safety vs convenience. The project serves as a reminder of how permission prompts influence everyday security decisions.Show HN 项目自动拒绝 macOS 的“允许应用发现本地网络设备?”提示,提升了使用体验。产品背后的讨论聚焦于 UX 与本地网络权限之间的权衡,以及安全治理的重要性。
Stela is a UEFI boot program intended to provide an Apple Silicon layer for x86-64 PCs.Stela 是一个面向 x86-64 PC、旨在提供 Apple Silicon 兼容层的 UEFI 启动程序。
Stela is a UEFI boot program intended to provide an Apple Silicon layer for x86-64 PCs. It includes an Arm64-to-x86-64 translator, an NVIDIA Metal driver, and a model of Apple silicon hardware, and can currently boot macOS and most apps. The v0.1.0 release is slow, requires more testing, and has a low chance of booting on current hardware.Stela 是一个面向 x86-64 PC、旨在提供 Apple Silicon 兼容层的 UEFI 启动程序。它包含 Arm64 到 x86-64 的转换器、NVIDIA Metal 驱动以及 Apple 芯片硬件模型,目前可以启动 macOS 和大多数应用。v0.1.0 运行速度较慢,仍需大量测试,在当前硬件上成功启动的概率也很低。
This project statically recompiles a Steam copy of Black Ops II into a native ARM64 macOS application using Claude Opus 5.5, without Wine, Crossover, Rosetta, or another emulator.这个项目使用 Claude Opus 5.5,将 Steam 版《使命召唤:黑色行动 II》静态重编译为原生 ARM64 macOS 应用,不依赖 Wine、Crossover、Rosetta 或其他模拟器。
This project statically recompiles a Steam copy of Black Ops II into a native ARM64 macOS application using Claude Opus 5.5, without Wine, Crossover, Rosetta, or another emulator. The game reportedly runs at nearly 60 frames per second across the campaign, local multiplayer with bots, and local Zombies mode, including owned DLC maps. Split-screen is being worked on, while LAN mode is planned for later.这个项目使用 Claude Opus 5.5,将 Steam 版《使命召唤:黑色行动 II》静态重编译为原生 ARM64 macOS 应用,不依赖 Wine、Crossover、Rosetta 或其他模拟器。据称,战役、本地机器人多人模式和本地僵尸模式都能以接近 60 帧运行,已拥有的 DLC 地图也可使用。开发者正在处理分屏功能,局域网模式则计划稍后加入。
Term inactory is a cozy factory game that runs directly in a terminal and was built for playing while waiting on long-running agent tasks.Term inactory 是一款直接运行在终端中的休闲工厂游戏,适合在等待耗时较长的代理任务时游玩。
Term inactory is a cozy factory game that runs directly in a terminal and was built for playing while waiting on long-running agent tasks. It includes basic factory mechanics such as building, deleting, selecting, measuring earnings, and researching unlocks. Additional features include a fictional browser with a wiki and forums, extensive lore, and hooks that display coding-agent responses as in-game “telegrams.”Term inactory 是一款直接运行在终端中的休闲工厂游戏,适合在等待耗时较长的代理任务时游玩。游戏包含建造、删除、选择、收益测算和研究解锁等基本工厂机制。它还加入了带有 Wiki 和论坛的虚构浏览器、大量背景故事,以及将编码代理回复显示为游戏内“电报”的钩子功能。
GREP.AI has open-sourced AgentRun as a managed platform for agents.GREP.AI 将 AgentRun 开源为一个托管智能体平台。
GREP.AI has open-sourced AgentRun as a managed platform for agents. It includes self-improving agents, web-grounded workflows, sandboxes, computer environments, tools and integrations, and behavioral analytics. AgentRun supports harnesses including Pi, Claude Agents, Codex, and OpenCode, and can be self-hosted or used as a hosted service.GREP.AI 将 AgentRun 开源为一个托管智能体平台。它包含自我改进智能体、基于网页信息的工作流、沙箱和计算机环境、工具与集成,以及智能体行为分析。AgentRun 支持 Pi、Claude Agents、Codex 和 OpenCode 等 harness,也可自行托管或使用托管版本。
MACaQ is a model-agnostic quantization system for GGUF and local inference across text, image, video, and audio models.MACaQ 是一个面向 GGUF 和本地推理的模型无关量化系统,支持文本、图像、视频和音频模型。
MACaQ is a model-agnostic quantization system for GGUF and local inference across text, image, video, and audio models. It supports llama.cpp, vLLM, Diffusers, and PyTorch. Huawei says it can quantize a 122-billion-parameter model in 64 minutes using 32.5 GB of device memory, while achieving lower perplexity than MR-GPTQ and DuQuant++ in stated comparisons.MACaQ 是一个面向 GGUF 和本地推理的模型无关量化系统,支持文本、图像、视频和音频模型。它兼容 llama.cpp、vLLM、Diffusers 和 PyTorch。华为称,MACaQ 可使用 32.5 GB 设备内存,在 64 分钟内完成一个 1220 亿参数模型的量化;在其给出的比较中,困惑度低于 MR-GPTQ 和 DuQuant++。
iSponsorBlockTV is a self-hosted application that connects to the YouTube TV app and automatically skips sponsored segments and intros using the SponsorBlock API.iSponsorBlockTV 是一款自托管应用,可连接 YouTube TV 客户端,并通过 SponsorBlock API 自动跳过赞助片段和片头。
iSponsorBlockTV is a self-hosted application that connects to the YouTube TV app and automatically skips sponsored segments and intros using the SponsorBlock API. It can also mute videos automatically and press YouTube's "Skip Ad" button when it appears. YouTube appears to have changed its screen-ID format and is revoking existing codes, so devices must be paired again.iSponsorBlockTV 是一款自托管应用,可连接 YouTube TV 客户端,并通过 SponsorBlock API 自动跳过赞助片段和片头。它还可以自动静音,并在 YouTube 广告出现“跳过广告”按钮后自动点击。YouTube 似乎更改了屏幕 ID 格式并正在撤销现有代码,因此设备需要重新配对。
WineVDM provides a pathway to run 16-bit Windows apps (Windows 1.x, 2.x, 3.0/3.1) on 64-bit Windows by using a dedicated compatibility runtime.该项目通过 winevdm 在64位 Windows 上实现对16位 Windows 应用(1.x、2.x、3.0、3.1 等)的运行。
WineVDM provides a pathway to run 16-bit Windows apps (Windows 1.x, 2.x, 3.0/3.1) on 64-bit Windows by using a dedicated compatibility runtime. It illustrates how modern Windows still lacks native 16-bit support, leaving preservation-focused projects to recreate the old environment. This matters for running legacy software, archaeology of software, and niche users who still rely on early Windows programs.该项目通过 winevdm 在64位 Windows 上实现对16位 Windows 应用(1.x、2.x、3.0、3.1 等)的运行。与现代 Windows 不再原生支持16位应用不同,兼容层/运行时的实现让老旧软件得以保存与使用,但仍存在兼容性与性能限制。
NVIDIA Nemotron 3.5 Transcribe brings open-source ASR to 40 languages, lowering the barrier for multilingual speech applications.NVIDIA Nemotron 3.5 Transcribe 为40种语言提供开源自动语音识别,降低多语言语音应用的门槛。
NVIDIA Nemotron 3.5 Transcribe brings open-source ASR to 40 languages, lowering the barrier for multilingual speech applications. The release lets researchers compare performance against other open models and accelerates progress in open speech AI.NVIDIA Nemotron 3.5 Transcribe 为40种语言提供开源自动语音识别,降低多语言语音应用的门槛。该发布使研究者能够与其它开源模型进行对比,加速开放语音 AI 的进展。
NVIDIA Kubernetes Engine is a management plane for running Kubernetes on bare-metal NVIDIA GPU infrastructure.NVIDIA Kubernetes Engine 是一个用于在裸机 NVIDIA GPU 基础设施上运行 Kubernetes 的管理平面。
NVIDIA Kubernetes Engine is a management plane for running Kubernetes on bare-metal NVIDIA GPU infrastructure. Platform teams can install it to create and operate workload clusters, each with its own control plane, etcd, and node pools. NVKE is currently a preview release, and its APIs, configuration, and behavior may change, so it is not recommended for production workloads.NVIDIA Kubernetes Engine 是一个用于在裸机 NVIDIA GPU 基础设施上运行 Kubernetes 的管理平面。平台团队可以安装它来创建和运营工作负载集群,每个集群都拥有独立的控制平面、etcd 和节点池。NVKE 目前仍处于预览阶段,API、配置和行为可能发生变化,因此不建议用于生产工作负载。
Rivet is an open-source, self-hostable orchestrator for agent workloads built around lightweight actors backed by S3.Rivet 是一个开源、可自行托管的代理工作负载编排器,围绕由 S3 支持的轻量级 actor 构建。
Rivet is an open-source, self-hostable orchestrator for agent workloads built around lightweight actors backed by S3. Each actor provides a long-lived process, in-memory context, a WebSocket server, durable SQLite storage backed by S3, and idle hibernation. The team says the design grew out of infrastructure work for multiplayer games and aims to avoid the overhead of running a heavyweight sandbox per agent.Rivet 是一个开源、可自行托管的代理工作负载编排器,围绕由 S3 支持的轻量级 actor 构建。每个 actor 都提供长期运行的进程、内存上下文、WebSocket 服务器、由 S3 支持的持久化 SQLite 存储,以及空闲休眠功能。团队表示,该设计源自多人游戏基础设施工作,旨在避免为每个代理运行重量级沙箱的开销。
This project is a Claude Code skill for creating product videos.这是一个用于制作产品视频的 Claude Code 技能。
This project is a Claude Code skill for creating product videos. It supports real-browser capture, branded composition, and local mastering, and includes a quick start, worked videos, and usage instructions.这是一个用于制作产品视频的 Claude Code 技能。它支持真实浏览器录制、品牌化合成和本地母版制作,并提供快速入门、示例视频和使用说明。
Ginux is now a configurable userspace kernel for isolated Linux compatibility instances.Ginux 现已成为用于隔离 Linux 兼容实例的可配置用户态内核。
Ginux is now a configurable userspace kernel for isolated Linux compatibility instances. It runs as an ordinary Linux process and defines an explicit, versioned contract between managed workloads and the host kernel. Its earlier phase as an independent Rust kernel booting through UEFI is complete, with that source preserved separately while active development continues on the userspace version.Ginux 现已成为用于隔离 Linux 兼容实例的可配置用户态内核。它作为普通 Linux 进程运行,并在受管理工作负载与主机内核之间建立明确且有版本控制的契约。此前通过 UEFI 启动、以 Rust 编写的独立内核阶段已经结束,相关源代码被单独保留,当前开发转向用户态版本。
Zigzag is a free, open-source Mac browser written in Swift and based on WebKit, created by a high-school student from Shanghai.Zigzag 是一款由上海高中生开发的免费开源 Mac 浏览器,使用 Swift 编写并基于 WebKit。
Zigzag is a free, open-source Mac browser written in Swift and based on WebKit, created by a high-school student from Shanghai. It aims to provide an Arc-like interface and can import spaces, pins, history, passwords, and cookies from Arc or other browsers. Despite using WebKit, it supports Chrome extensions, including 1Password, Bitwarden, and Tampermonkey. The browser requires no account and claims to have no telemetry or analytics.Zigzag 是一款由上海高中生开发的免费开源 Mac 浏览器,使用 Swift 编写并基于 WebKit。它试图提供类似 Arc 的界面,并可从 Arc 或其他浏览器导入空间、固定项目、历史记录、密码和 Cookie。虽然基于 WebKit,它仍支持 Chrome 扩展,包括 1Password、Bitwarden 和 Tampermonkey。该浏览器无需账户,并声称不包含遥测或分析功能。
Show HNOpen SourceWebDataEducationCodeNov 1711月17日0.011268
Eclipse Mosquitto is an open-source implementation of an MQTT broker supporting MQTT versions 5.0, 3.1.1, and 3.1.Eclipse Mosquitto 是一个开源 MQTT 代理实现,支持 MQTT 5.0、3.1.1 和 3.1 版本。
Eclipse Mosquitto is an open-source implementation of an MQTT broker supporting MQTT versions 5.0, 3.1.1, and 3.1. It also provides C and C++ client libraries and command-line tools for publishing, subscribing, and administering the broker.Eclipse Mosquitto 是一个开源 MQTT 代理实现,支持 MQTT 5.0、3.1.1 和 3.1 版本。它还提供 C 和 C++ 客户端库,以及用于发布、订阅和管理代理的命令行工具。
Zebkit packages Chromium’s engine as a cross-platform library for Linux, macOS, and Windows.Zebkit 将 Chromium 引擎打包成可运行于 Linux、macOS 和 Windows 的跨平台库。
Zebkit packages Chromium’s engine as a cross-platform library for Linux, macOS, and Windows. It includes Blink, V8 with its JIT tiers, PDFium, FFmpeg, and other Chrome components, but omits the Chrome browser and Electron runtime. Applications can use it through C, C++, Rust, Python, Go, or JavaScript bindings.Zebkit 将 Chromium 引擎打包成可运行于 Linux、macOS 和 Windows 的跨平台库。它包含 Blink、带有各级 JIT 的 V8、PDFium、FFmpeg 及其他 Chrome 组件,但不包含 Chrome 浏览器本身或 Electron 运行时。应用可以通过 C、C++、Rust、Python、Go 或 JavaScript 绑定使用它。
ExcalidrawZ is a macOS and iOS drawing app built with SwiftUI that preserves the familiar Excalidraw canvas.ExcalidrawZ 是一款使用 SwiftUI 构建的 macOS 和 iOS 绘图应用,保留了熟悉的 Excalidraw 画布。
ExcalidrawZ is a macOS and iOS drawing app built with SwiftUI that preserves the familiar Excalidraw canvas. It adds native document management, iCloud sync, history, export, AI assistance, MCP integration, and other Apple-platform features. Drawings are handled as application documents rather than loose browser downloads.ExcalidrawZ 是一款使用 SwiftUI 构建的 macOS 和 iOS 绘图应用,保留了熟悉的 Excalidraw 画布。它增加了原生文件管理、iCloud 同步、历史记录、导出、AI 辅助、MCP 集成及其他 Apple 平台功能。绘图会作为应用文档管理,而不是浏览器中零散下载的文件。
BootGPT places an entire inference engine in a 510-byte MBR, with model weights, vocabulary, and strings stored in the following disk sectors.BootGPT 将完整推理引擎放入 510 字节的 MBR 中,并把模型权重、词汇表和字符串存储在后续磁盘扇区。
BootGPT places an entire inference engine in a 510-byte MBR, with model weights, vocabulary, and strings stored in the following disk sectors. Its small character-level neural network was trained to generate Linux boot messages indefinitely, matching a simulator character by character. Debugging uncovered an x86 assembly bug involving the remainder left in EDX by integer division and a pointer that also used EDX.BootGPT 将完整推理引擎放入 510 字节的 MBR 中,并把模型权重、词汇表和字符串存储在后续磁盘扇区。这个小型字符级神经网络经过训练,可以无限生成 Linux 启动消息,并与模拟器逐字符匹配。调试过程中发现了一个 x86 汇编错误:整数除法会把余数留在 EDX,而偏置表指针也使用了 EDX。
The project is a model zoo for forecasting blood glucose levels, trained and tested on the author’s personal data as well as the public Metabonet dataset.这个项目是一个血糖预测模型集合,使用作者自己的数据和公开的 Metabonet 数据集进行训练与测试。
The project is a model zoo for forecasting blood glucose levels, trained and tested on the author’s personal data as well as the public Metabonet dataset. It includes 15 forecasting models, including models developed by the author. Most of the repository’s code was written with AI assistance.这个项目是一个血糖预测模型集合,使用作者自己的数据和公开的 Metabonet 数据集进行训练与测试。集合中包含 15 个预测模型,其中也有作者自行开发的模型。仓库中的大部分代码由 AI 辅助编写。
Go is tightening TLS validation by rejecting certificate chains that include eight distrusted SPKI entries.Go 将通过在 TLS 验证中拒绝包含八个不可信 SPKI 的证书链,来收紧安全性。
Go is tightening TLS validation by rejecting certificate chains that include eight distrusted SPKI entries. The change reduces attack surface from compromised keys and misissued certificates, nudging developers to audit complex chains or intermediates.Go 将通过在 TLS 验证中拒绝包含八个不可信 SPKI 的证书链,来收紧安全性。此举降低了被入侵密钥和被篡改证书带来的风险,促使开发者审查包含多中间证书的证书链。
Using DOOM as a testbed, the project applies Transformer-based models to game data, demonstrating how sequence models can learn planning or perception tasks in a classic game.在 DOOM 的数据上应用 Transformer,展示序列模型在经典游戏中的学习能力,涉及规划与感知任务。
Using DOOM as a testbed, the project applies Transformer-based models to game data, demonstrating how sequence models can learn planning or perception tasks in a classic game. It’s a compact, proof-of-concept that sparks ideas about cross-domain AI experiments.在 DOOM 的数据上应用 Transformer,展示序列模型在经典游戏中的学习能力,涉及规划与感知任务。这个小型的实现是跨领域 AI 实验的一个有趣原型,值得关注。
AssemblyForNeovim brings a Godbolt-style assembly view into a floating Neovim window for C, C++, and Rust.AssemblyForNeovim 将类似 Godbolt 的汇编查看功能带入 Neovim 浮动窗口,支持 C、C++ 和 Rust。
AssemblyForNeovim brings a Godbolt-style assembly view into a floating Neovim window for C, C++, and Rust. It generates assembly with a compiler installed locally, so it does not require an internet connection.AssemblyForNeovim 将类似 Godbolt 的汇编查看功能带入 Neovim 浮动窗口,支持 C、C++ 和 Rust。它使用本地安装的编译器生成汇编代码,因此无需互联网连接。
WINUDF is an installable file-system driver that lets DOS and early Windows systems read and write UDF media.WINUDF 是一种可安装文件系统驱动,让 DOS 和早期 Windows 系统能够读写 UDF 媒体。
WINUDF is an installable file-system driver that lets DOS and early Windows systems read and write UDF media. It supports UDF versions 1.02 through 2.60, including DVDs, Blu-ray discs, packet-written CDs, and UDF-formatted USB drives. DOS and Windows 1.0x–3.11 receive drive letters through WINUDF.EXE, while WINUDF.VXD provides a long-name file-system driver for Windows 95, 98, and ME.WINUDF 是一种可安装文件系统驱动,让 DOS 和早期 Windows 系统能够读写 UDF 媒体。它支持 1.02 至 2.60 版 UDF,涵盖 DVD、蓝光光盘、包写入 CD 以及采用 UDF 格式的 USB 驱动器。DOS 和 Windows 1.0x–3.11 可通过 WINUDF.EXE 获得盘符,Windows 95、98 和 ME 则使用支持长文件名的 WINUDF.VXD 驱动。
Keplearn is a tool that infers equations from data in seconds, enabling rapid symbolic modeling of complex systems.Keplearn 能在秒级从数据中推断出方程,快速实现对复杂系统的符号建模。
Keplearn is a tool that infers equations from data in seconds, enabling rapid symbolic modeling of complex systems. This accelerates scientific experimentation and can complement traditional regression with interpretable models.Keplearn 能在秒级从数据中推断出方程,快速实现对复杂系统的符号建模。这加速了科学实验,并可与传统回归方法结合,产生可解释的模型。
Agweb is a macOS browser for coding agents, built on WebKit and designed not to steal keyboard focus.Agweb 是一款面向编码代理的 macOS 浏览器,基于 WebKit 构建,设计目标是不抢占键盘焦点。
Agweb is a macOS browser for coding agents, built on WebKit and designed not to steal keyboard focus. Agents control the single window from the shell through an agweb command modeled on Playwright CLI, while users can watch all tabs live in one view. Idle tabs are suspended; the author reports that five open pages used 0% CPU on a Mac, compared with five to seven busy cores from Chrome’s GPU process.Agweb 是一款面向编码代理的 macOS 浏览器,基于 WebKit 构建,设计目标是不抢占键盘焦点。代理通过类似 Playwright CLI 的 agweb 命令从 shell 控制这个窗口,用户则可以在一个视图中实时查看所有标签页。闲置标签页会被挂起;作者称,在 Mac 上打开五个页面时其 CPU 占用为 0%,而 Chrome 的 GPU 进程会占用五到七个繁忙核心。
A WASM size comparison suggests C# compiled to WebAssembly can produce smaller binaries than Rust for certain workloads, challenging the assumption that Rust always yields the smallest output.对比显示在某些工作负载下,C# 编译为 WebAssembly 的二进制体积可能比 Rust 更小,挑战了“Rust 总是最小”的常识。
A WASM size comparison suggests C# compiled to WebAssembly can produce smaller binaries than Rust for certain workloads, challenging the assumption that Rust always yields the smallest output. This influences language choice for wasm-based apps, though ecosystem and performance considerations remain important.对比显示在某些工作负载下,C# 编译为 WebAssembly 的二进制体积可能比 Rust 更小,挑战了“Rust 总是最小”的常识。对 wasm 应用的语言选择产生影响,但生态和性能等因素仍然重要。
Omnara is an open source alternative to Claude's Managed Agents, offering configurable AI agents that you can run and customize yourself.Omnara 是一个开源替代方案,定位为可本地控管的 Claude Managed Agents 的对等实现,提供可配置的 AI 代理和可插拔集成。
Omnara is an open source alternative to Claude's Managed Agents, offering configurable AI agents that you can run and customize yourself. The project emphasizes transparency and privacy and aims to match managed agent capabilities with self hosted deployment.Omnara 是一个开源替代方案,定位为可本地控管的 Claude Managed Agents 的对等实现,提供可配置的 AI 代理和可插拔集成。该项目强调透明度、可定制性和隐私,面向希望自托管 AI 自动化的团队。
A DIY project that strips embedded smart TV software to create a privacy-preserving, lean device, then reintroduces smart capabilities in a more controllable way.一个 DIY 项目通过移除智能电视的内置软件来打造一个轻量、注重隐私的设备,随后再以更可控的方式重新引入智能功能。
A DIY project that strips embedded smart TV software to create a privacy-preserving, lean device, then reintroduces smart capabilities in a more controllable way. The effort highlights the tension between vendor-bloat and user autonomy, and points to hackable hardware routes for better privacy.一个 DIY 项目通过移除智能电视的内置软件来打造一个轻量、注重隐私的设备,随后再以更可控的方式重新引入智能功能。此举凸显厂商臃肿软件与用户自主权之间的矛盾,并指向可 Hack 的硬件方式以提升隐私保护。
Sift is an idempotent formatter, rule-based linter, and language server for Scheme, Emacs Lisp, and Common Lisp.Sift 是面向 Scheme、Emacs Lisp 和 Common Lisp 的幂等格式化器、基于规则的 linter 和语言服务器。
Sift is an idempotent formatter, rule-based linter, and language server for Scheme, Emacs Lisp, and Common Lisp. It reports malformed Lisp source and rewrites whitespace while preserving comments and read-time forms in their original positions. Formatting an already formatted file produces no further changes, a property asserted across the test suite.Sift 是面向 Scheme、Emacs Lisp 和 Common Lisp 的幂等格式化器、基于规则的 linter 和语言服务器。它会报告格式错误的 Lisp 源码并重写空白,同时保留注释和读取时形式的原始位置。对已经格式化的文件再次格式化不会产生变化,该特性在整个测试套件中都有断言验证。
Stroq is an open-source local tool for inspecting Claude Code, Codex, and Cursor sessions.Stroq 是一款开源本地工具,用于检查 Claude Code、Codex 和 Cursor 的会话。
Stroq is an open-source local tool for inspecting Claude Code, Codex, and Cursor sessions. It identifies known secrets, counts commands derived from text an agent read, and shows which keys left the machine. Local hooks can then trace risky commands back to the content that prompted them, without telemetry.Stroq 是一款开源本地工具,用于检查 Claude Code、Codex 和 Cursor 的会话。它可以识别已知密钥,统计代理从所读取文本中产生的命令,并显示哪些密钥离开了本机。用户还可以通过本地钩子将高风险命令追溯到触发它们的内容,而且不会发送遥测数据。
Introduces Moxnox 1.0, an NT-inspired, POSIX-compatible OS built on the Btrfs filesystem, signaling an experimental approach to combining Windows-like design with modern filesystems.介绍了 Moxnox 1.0,这是一个在 Btrfs 文件系统上实现的 NT 风格、POSIX 兼容的操作系统原型。
Introduces Moxnox 1.0, an NT-inspired, POSIX-compatible OS built on the Btrfs filesystem, signaling an experimental approach to combining Windows-like design with modern filesystems.介绍了 Moxnox 1.0,这是一个在 Btrfs 文件系统上实现的 NT 风格、POSIX 兼容的操作系统原型。
Waymark is an open-source durable execution library similar to Temporal or DBOS.Waymark 是一个类似 Temporal 或 DBOS 的开源持久执行库。
Waymark is an open-source durable execution library similar to Temporal or DBOS. It lets developers write workflows in ordinary Python code using conditionals, loops, and asyncio-style parallelism, while the runtime handles distributed execution. Rather than replaying past state in an interpreter, Waymark walks the program’s AST and compiles the logic into bytecode for a stackless virtual machine.Waymark 是一个类似 Temporal 或 DBOS 的开源持久执行库。开发者可以用普通 Python 代码编写工作流,包括条件语句、循环和类似 asyncio 的并行操作,再由运行时负责分布式执行。Waymark 不通过解释器重放历史状态,而是遍历程序 AST,将逻辑编译为运行在无栈虚拟机上的字节码。
Show HNProgrammingOpen SourceCodeNov 611月6日0.08140
Noalloc is a toolkit to enforce zero-heap allocations in Go programs, improving latency predictability for high-performance services.noalloc 提供工具与库,用于在 Go 程序中强制零堆分配,从而提升高性能服务的延迟可预测性。
Noalloc is a toolkit to enforce zero-heap allocations in Go programs, improving latency predictability for high-performance services. It helps developers reason about memory use, though it may constrain certain idioms and require careful design.noalloc 提供工具与库,用于在 Go 程序中强制零堆分配,从而提升高性能服务的延迟可预测性。它帮助开发者管理内存使用,但可能限制某些编码习惯并需谨慎设计。
This project provides peer-to-peer WebRTC connections with built-in signaling and peer discovery.这个项目提供带有内置信令和节点发现功能的点对点 WebRTC 连接。
This project provides peer-to-peer WebRTC connections with built-in signaling and peer discovery. It works across browsers, Node.js, and React Native.这个项目提供带有内置信令和节点发现功能的点对点 WebRTC 连接。它支持浏览器、Node.js 和 React Native。
Mizu provides thread-speed asynchronous and parallel execution for Python over shared memory.Mizu 通过共享内存为 Python 提供接近线程速度的异步和并行执行能力。
Mizu provides thread-speed asynchronous and parallel execution for Python over shared memory. Its benchmarks report reducing task round trips from roughly 100 microseconds over sockets to less than 1 microsecond. Lists, dictionaries, NumPy arrays, and Arrow tables can be shared between Python and R as zero-copy views.Mizu 通过共享内存为 Python 提供接近线程速度的异步和并行执行能力。其基准测试称,任务往返时间可从基于套接字的约 100 微秒降至不到 1 微秒。列表、字典、NumPy 数组和 Arrow 表都可以作为零拷贝视图在 Python 与 R 之间共享。
Sumi.css presents a minimal CSS framework focused on speed and simplicity, offering a compact toolset for quickly building accessible UIs.Sumi.css 提供一个极简的 CSS 框架,强调小巧和易用性,适合快速构建可访问的界面。
Sumi.css presents a minimal CSS framework focused on speed and simplicity, offering a compact toolset for quickly building accessible UIs.Sumi.css 提供一个极简的 CSS 框架,强调小巧和易用性,适合快速构建可访问的界面。
A Show HN project formalizes IRS tax forms as a proof-of-concept, likely using a theorem prover or formal spec; demonstrates how to model tax forms for correctness and tooling; significance for software that handles regulated forms.Show HN 展示了将 IRS 税表形式化的概念性实现,借助形式化工具提高正确性与可验证性。
A Show HN project formalizes IRS tax forms as a proof-of-concept, likely using a theorem prover or formal spec; demonstrates how to model tax forms for correctness and tooling; significance for software that handles regulated forms. 2-3 sentences.Show HN 展示了将 IRS 税表形式化的概念性实现,借助形式化工具提高正确性与可验证性。此举对政府表单数字化和软件可靠性具有示范价值。
The project describes a personalized mRNA cancer vaccine for Yash, produced at the RNA Foundry.该项目介绍了一款由 RNA Foundry 制作、面向 Yash 的个性化 mRNA 癌症疫苗。
The project describes a personalized mRNA cancer vaccine for Yash, produced at the RNA Foundry. Yashdeep Bindal was diagnosed with stage IV pancreatic cancer in January 2025 at age 32, shortly after returning to work following the birth of his daughter Maya.该项目介绍了一款由 RNA Foundry 制作、面向 Yash 的个性化 mRNA 癌症疫苗。Yashdeep Bindal 于 2025 年 1 月、32 岁时被诊断为 IV 期胰腺癌;此前他刚在女儿 Maya 出生后重返工作岗位。
Fursoy is a 64-bit out-of-order superscalar RISC-V processor implemented in SystemVerilog.Fursoy 是一款用 SystemVerilog 实现的 64 位乱序超标量 RISC-V 处理器。
Fursoy is a 64-bit out-of-order superscalar RISC-V processor implemented in SystemVerilog. It supports RV64IMAFDV with Zba and Zbb extensions, boots mainline Linux 6.6.50 and BusyBox under Sv39 virtual memory, and includes separate machine, supervisor, and user privilege levels. Synthesis in Nangate 45nm reached 500 MHz timing closure across eight core functional blocks.Fursoy 是一款用 SystemVerilog 实现的 64 位乱序超标量 RISC-V 处理器。它支持带 Zba 和 Zbb 扩展的 RV64IMAFDV,可在 Sv39 虚拟内存下启动主线 Linux 6.6.50 和 BusyBox,并具备机器、监管者和用户三级权限。该设计在 Nangate 45nm 工艺中完成了八个核心功能模块的 500 MHz 时序收敛。
Klad is a Rust-based painting application that coding agents can control through a CLI, while also offering a full graphical interface for humans.Klad 是一款用 Rust 编写的绘画应用,编码代理可以通过 CLI 控制它,同时人类也能使用完整的图形界面。
Klad is a Rust-based painting application that coding agents can control through a CLI, while also offering a full graphical interface for humans. It supports brush painting, frame-by-frame animation, bones and inverse kinematics, vectors, text layers, colors, layer effects, and persistent unlimited undo. Its editable strokes are intended to let agents create drafts that people can refine stroke by stroke.Klad 是一款用 Rust 编写的绘画应用,编码代理可以通过 CLI 控制它,同时人类也能使用完整的图形界面。它支持笔刷绘画、逐帧动画、骨骼与反向运动学、矢量、文本图层、颜色、图层效果,以及可跨文件重新打开保留的无限撤销。它将可编辑笔触作为核心,使代理能够生成草稿,再由人类逐笔修改。
DeepSeek Ultrafast is an inference engine for DeepSeek-V4.1-Flash targeting NVIDIA sm_100a and sm_103a hardware.DeepSeek Ultrafast 是面向 NVIDIA sm_100a 和 sm_103a 硬件的 DeepSeek-V4.1-Flash 推理引擎。
DeepSeek Ultrafast is an inference engine for DeepSeek-V4.1-Flash targeting NVIDIA sm_100a and sm_103a hardware. It supports a 1-million-token context length and multi-server KV-cache sharing. The project measures decode speed, step time, and acceptance length using a Tetris-generation prompt.DeepSeek Ultrafast 是面向 NVIDIA sm_100a 和 sm_103a 硬件的 DeepSeek-V4.1-Flash 推理引擎。它支持 100 万 token 的上下文长度和多服务器 KV 缓存共享。项目使用生成俄罗斯方块的提示词测量解码速度、步骤耗时和接受长度。
PySR blends symbolic regression with modern ML ecosystems to yield interpretable models in Python and Julia.PySR 将符号回归与现代机器学习生态系统相结合,在 Python 与 Julia 中生成可解释的模型。
PySR blends symbolic regression with modern ML ecosystems to yield interpretable models in Python and Julia. By combining mathematical discovery with scalable tooling, it helps uncover human-readable expressions that fit data.PySR 将符号回归与现代机器学习生态系统相结合,在 Python 与 Julia 中生成可解释的模型。通过将数学发现与可扩展工具链结合,它有助于发现能拟合数据且易读的表达式。
HyperFlow H3 Flash is a compressed, few-step version of MiniMax-H3 with 13.7 billion parameters.HyperFlow H3 Flash 是 MiniMax-H3 的压缩少步数版本,包含 137 亿个参数。
HyperFlow H3 Flash is a compressed, few-step version of MiniMax-H3 with 13.7 billion parameters. It generates video and synchronized stereo audio together and supports INT8 and GGUF formats. The project also provides weights, ComfyUI nodes, a GGUF setup guide, and an eight-step LoRA for the full-size model.HyperFlow H3 Flash 是 MiniMax-H3 的压缩少步数版本,包含 137 亿个参数。它能够同时生成视频和同步立体声,并支持 INT8 与 GGUF 格式。项目还提供模型权重、ComfyUI 节点、GGUF 配置指南,以及适用于完整模型的 8 步 LoRA。
Hurston-1 is an open-source text-to-speech model that adds voices for Southern American and Afro-English dialects, expanding dialect representation in TTS and accessibility for diverse speakers.Hurston-1 是一个开源文本转语音模型,新增南方美式英语和 Afro-English 方言声音,扩展了 TTS 的方言表示和对多样语音使用者的可及性。
Hurston-1 is an open-source text-to-speech model that adds voices for Southern American and Afro-English dialects, expanding dialect representation in TTS and accessibility for diverse speakers.Hurston-1 是一个开源文本转语音模型,新增南方美式英语和 Afro-English 方言声音,扩展了 TTS 的方言表示和对多样语音使用者的可及性。
Unref is a Python dead-code finder built with Rust and Ruff's parser, created because existing alternatives did not work reliably on the author's codebases.Unref 是一款使用 Rust 和 Ruff 解析器构建的 Python 死代码查找器,起因是现有替代工具无法可靠适配作者的代码库。
Unref is a Python dead-code finder built with Rust and Ruff's parser, created because existing alternatives did not work reliably on the author's codebases. It can scan all of CPython in under a second and has found substantial dead code in the author's team's projects. The tool is intended for low-configuration use and is already being used in continuous integration.Unref 是一款使用 Rust 和 Ruff 解析器构建的 Python 死代码查找器,起因是现有替代工具无法可靠适配作者的代码库。它能在不到一秒内检查完整的 CPython,并在作者团队的项目中发现了大量死代码。该工具定位于低配置使用,目前已用于持续集成。
Spindle is a multicore-by-design, Eio-native web framework for OCaml that uses a type-driven API to derive codecs, request validation, OpenAPI specifications, and frontend schemas.Spindle 是一个面向多核设计、原生支持 Eio 的 OCaml Web 框架,通过类型驱动的 API 自动生成编解码器、请求验证、OpenAPI 规范和前端模式。
Spindle is a multicore-by-design, Eio-native web framework for OCaml that uses a type-driven API to derive codecs, request validation, OpenAPI specifications, and frontend schemas. It includes an RFC-strict HTTP server, WebSockets, SSE, MCP server primitives, tracing, metrics, and logging. Its benchmarks report performance two times faster than NestJS and five times faster than FastAPI on a single core, although speed was not the project’s primary goal.Spindle 是一个面向多核设计、原生支持 Eio 的 OCaml Web 框架,通过类型驱动的 API 自动生成编解码器、请求验证、OpenAPI 规范和前端模式。它包含严格遵循 RFC 的 HTTP 服务器、WebSocket、SSE、MCP 服务器基础组件,以及分布式追踪、指标和日志功能。基准测试显示,它在单核上运行速度约为 NestJS 的两倍、FastAPI 的五倍,不过速度并非项目的首要目标。
The Septopert is a convex seven-vertex polyhedron that is not Rupert, meaning it cannot pass through a congruent copy of itself.Septopert 是一种拥有 7 个顶点、且不是鲁珀特多面体的凸多面体,也就是说它无法穿过与自身全等的另一个多面体。
The Septopert is a convex seven-vertex polyhedron that is not Rupert, meaning it cannot pass through a congruent copy of itself. Seven vertices is the minimum possible for a non-Rupert polyhedron. The page gives exact statements, Lean-checked proofs, and reproduction instructions for the canonical Septopert.Septopert 是一种拥有 7 个顶点、且不是鲁珀特多面体的凸多面体,也就是说它无法穿过与自身全等的另一个多面体。7 个顶点是非鲁珀特多面体可能达到的最少数量。页面提供了规范 Septopert 的精确陈述、经 Lean 检查的证明和复现步骤。
ZapFast is a lightweight, native WhatsApp client written in Rust with an egui interface.ZapFast 是一款使用 Rust 编写、基于 egui 界面的轻量级原生 WhatsApp 客户端。
ZapFast is a lightweight, native WhatsApp client written in Rust with an egui interface. It is designed to run on Linux, macOS, and Windows. The project provides downloads, setup instructions, a user guide, and contribution information.ZapFast 是一款使用 Rust 编写、基于 egui 界面的轻量级原生 WhatsApp 客户端。它面向 Linux、macOS 和 Windows 平台。项目提供下载、安装说明、用户指南和参与贡献的方式。
Micron is a lean, statically typed, modular systems language combining Oberon’s simplicity and explicit modules with C’s low-level capabilities and interoperability.Micron 是一种精简、静态类型、模块化的系统编程语言,将 Oberon 的简洁性和显式模块结构与 C 的底层能力及互操作性结合起来。
Micron is a lean, statically typed, modular systems language combining Oberon’s simplicity and explicit modules with C’s low-level capabilities and interoperability. Its goal is to use one language for boot code, device drivers, operating-system kernels, and ordinary applications without a garbage collector. It supports explicit memory layouts, signed and unsigned integers, pointer operations, and direct C interoperability.Micron 是一种精简、静态类型、模块化的系统编程语言,将 Oberon 的简洁性和显式模块结构与 C 的底层能力及互操作性结合起来。它的目标是用同一种语言编写启动代码、设备驱动、操作系统内核和普通应用,同时不依赖垃圾回收器。该语言支持显式内存布局、有符号和无符号整数、指针操作以及直接的 C 互操作。
Farol visualizes risk associated with a pull request for the approver, mapping potential issues to guide decisions before merging.Farol 将对拉取请求的风险进行可视化映射,帮助审核者在合并前直观了解潜在问题以指导决策。
Farol visualizes risk associated with a pull request for the approver, mapping potential issues to guide decisions before merging.Farol 将对拉取请求的风险进行可视化映射,帮助审核者在合并前直观了解潜在问题以指导决策。
Penmentor uses Claude to annotate handwritten math, offering feedback and marking; demonstrates practical AI-assisted education tooling.Penmentor 使用 Claude 对手写数学题进行标注与批改,提供逐步解题反馈,展示 AI 辅助教育工具的应用场景。
Penmentor uses Claude to annotate handwritten math, offering feedback and marking; demonstrates practical AI-assisted education tooling.Penmentor 使用 Claude 对手写数学题进行标注与批改,提供逐步解题反馈,展示 AI 辅助教育工具的应用场景。
VMMouse NT 3.x driver enhances mouse integration for early virtualization emulation in QEMU/VMware.VMMouse NT 3.x 驱动提升了在 QEMU/VMware 等早期虚拟化环境中的鼠标集成,展示了早期操作系统与虚拟化栈如何处理输入设备的历史细节。
VMMouse NT 3.x driver enhances mouse integration for early virtualization emulation in QEMU/VMware. It’s a retro computing nugget showing how input devices were integrated in early OS/virtualization stacks.VMMouse NT 3.x 驱动提升了在 QEMU/VMware 等早期虚拟化环境中的鼠标集成,展示了早期操作系统与虚拟化栈如何处理输入设备的历史细节。
Kube32 is a Kubernetes-compatible control plane and WebAssembly pod runtime for ESP32-S3 microcontrollers.Kube32 是面向 ESP32-S3 微控制器的 Kubernetes 兼容控制平面和 WebAssembly Pod 运行时。
Kube32 is a Kubernetes-compatible control plane and WebAssembly pod runtime for ESP32-S3 microcontrollers. The project lets kubectl and k9s manage boards over Wi-Fi or USB serial, while pods run WebAssembly modules through the wasm3 interpreter. A board with PSRAM serves as the control plane, and boards without PSRAM act as workers. The same sources also build a macOS and Linux host simulator.Kube32 是面向 ESP32-S3 微控制器的 Kubernetes 兼容控制平面和 WebAssembly Pod 运行时。该项目支持通过 Wi-Fi 或 USB 串口使用 kubectl 和 k9s 管理开发板,并通过 wasm3 解释器运行 WebAssembly 模块。带 PSRAM 的开发板充当控制平面,不带 PSRAM 的开发板作为工作节点。同一套源代码还可构建 macOS 和 Linux 主机模拟器。
PyLings teaches Python by having users fix code in exercises ranging from basic concepts to senior-level interview problems.PyLings 通过让用户修复代码来学习 Python,练习范围从基础知识延伸到高级面试题。
PyLings teaches Python by having users fix code in exercises ranging from basic concepts to senior-level interview problems. It includes 35 warm-ups and 20 interview exercises, requires no dependencies, and targets Python 3.14. The tool opens the next exercise and reruns its tests whenever the user saves.PyLings 通过让用户修复代码来学习 Python,练习范围从基础知识延伸到高级面试题。它包含 35 道热身题和 20 道面试题,不依赖额外软件包,面向 Python 3.14。用户每次保存时,工具都会打开下一道练习并重新运行测试。
This project collects V100 SXM2 pin maps, two- and four-GPU NVLink connection plans, and connector-only KiCad resources.该项目整理了 V100 SXM2 引脚图、双 GPU 和四 GPU 的 NVLink 连接方案,以及仅含连接器的 KiCad 资源。
This project collects V100 SXM2 pin maps, two- and four-GPU NVLink connection plans, and connector-only KiCad resources. It aims to support an open-source carrier-board design for Nvidia V100 SXM2 modules, while enabling others to develop their own designs.该项目整理了 V100 SXM2 引脚图、双 GPU 和四 GPU 的 NVLink 连接方案,以及仅含连接器的 KiCad 资源。项目目标是支持 Nvidia V100 SXM2 模块的开源载板设计,同时让其他人能够据此开发自己的设计。
WaveDudeNext is an interactive Windows simulator for two-dimensional wave optics.WaveDudeNext 是一款面向 Windows 的交互式二维波动光学模拟器。
WaveDudeNext is an interactive Windows simulator for two-dimensional wave optics. Users can turn images into refractive scenes, place light sources and optical components, and observe refraction, diffraction, and interference. It displays both the wave field and time-averaged intensity, while supporting probes, timeline control, video export, adjustable brightness, custom color schemes, and multiple interface languages.WaveDudeNext 是一款面向 Windows 的交互式二维波动光学模拟器。用户可以将图像转换为折射场景,放置光源和光学元件,并观察折射、衍射与干涉现象。程序同时显示波场和时间平均强度,并支持探针、时间轴控制、视频导出、亮度调节、自定义配色以及多种界面语言。
Turbo-Browser-MCP is a Playwright-based MCP server designed to reduce token usage and multi-turn latency in browser automation.Turbo-Browser-MCP 是一个基于 Playwright 的 MCP 服务器,旨在降低浏览器自动化中的 token 消耗和多轮交互延迟。
Turbo-Browser-MCP is a Playwright-based MCP server designed to reduce token usage and multi-turn latency in browser automation. It combines 25–33 fine-grained browser primitives into 10 batched tools, with delta DOM snapshots, React/Shadow DOM/iframe form filling, direct disk uploads, parallel tabs, and network-blocking presets. It also includes a default human-in-the-loop guard before form submission.Turbo-Browser-MCP 是一个基于 Playwright 的 MCP 服务器,旨在降低浏览器自动化中的 token 消耗和多轮交互延迟。它将 25 至 33 个细粒度浏览器原语整合为 10 个批量工具,并支持 DOM 增量快照、React/Shadow DOM/iframe 表单填写、直接从磁盘上传文件、并行标签页和网络阻断预设。提交表单前,它还默认提供人工介入保护。
A Show HN demonstrates obtaining an autonomous root on a Tcl Android TV device using a known CVE, highlighting ongoing security challenges in consumer electronics and the importance of timely patching.Show HN 展示了利用已知 CVE 在 Tcl Android TV 设备上实现自主 Root 的方法,凸显消费电子产品安全性挑战和及时修补的重要性。
A Show HN demonstrates obtaining an autonomous root on a Tcl Android TV device using a known CVE, highlighting ongoing security challenges in consumer electronics and the importance of timely patching.Show HN 展示了利用已知 CVE 在 Tcl Android TV 设备上实现自主 Root 的方法,凸显消费电子产品安全性挑战和及时修补的重要性。
TimeNet provides a unified data layer for temporal datasets used by Temporal Foundation Models.TimeNet 为时间基础模型使用的时间序列数据集提供统一数据层。
TimeNet provides a unified data layer for temporal datasets used by Temporal Foundation Models. It supports tasks including classification, question answering, scalar prediction, temporal localization, forecasting, generation, editing, and correspondence. The project reports that training on multiple temporal datasets improved average F1 performance by 14% for OpenTSLM and 22% for SLIP.TimeNet 为时间基础模型使用的时间序列数据集提供统一数据层。它支持分类、问答、标量预测、时间定位、预测、生成、编辑和对应关系等任务。项目方称,在多个时间数据集上训练后,OpenTSLM 的平均 F1 性能提升了 14%,SLIP 提升了 22%。
RunAnywhere open-sourced TrnRT, an inference engine for AWS Trainium that currently serves GLM-5.3 Flash in FP8 on an AWS trn2.48xlarge instance.RunAnywhere 开源了 TrnRT,这是一个面向 AWS Trainium 的推理引擎,目前可在 AWS trn2.48xlarge 实例上以 FP8 运行 GLM-5.3 Flash。
RunAnywhere open-sourced TrnRT, an inference engine for AWS Trainium that currently serves GLM-5.3 Flash in FP8 on an AWS trn2.48xlarge instance. It can serve two model instances on one box using eight Trainium2 chips and a cache-aware router. On the SemiAnalysis AgentX benchmark, the team reports 3.57 million tokens per minute. The project targets large models with long context windows and aims to improve cost efficiency compared with H100 and H200 deployments.RunAnywhere 开源了 TrnRT,这是一个面向 AWS Trainium 的推理引擎,目前可在 AWS trn2.48xlarge 实例上以 FP8 运行 GLM-5.3 Flash。它可借助 8 个 Trainium2 芯片和感知缓存的路由器,在一台机器上服务两个模型实例。团队称,其在 SemiAnalysis AgentX 基准上的性能达到每分钟 357 万个 token。该项目面向长上下文大模型,目标是相较 H100 和 H200 部署提升成本效率。
Tamserve runs Meta’s SAM 3 on live camera feeds, detecting named objects and tracking them across multiple cameras.Tamserve 在实时摄像头画面上运行 Meta 的 SAM 3,可根据名称发现目标,并在多个摄像头之间持续跟踪。
Tamserve runs Meta’s SAM 3 on live camera feeds, detecting named objects and tracking them across multiple cameras. On an RTX 4090, it is reported to be four times faster than SAM 3’s tracker for arbitrary objects while retaining 92% of its accuracy; for people, it can handle 16 cameras concurrently. Its memory use stays around 3.5 GB on a 3,000-frame video, while SAM 3 runs out of memory at frame 423.Tamserve 在实时摄像头画面上运行 Meta 的 SAM 3,可根据名称发现目标,并在多个摄像头之间持续跟踪。在 RTX 4090 上,对于任意目标,其速度据称是 SAM 3 跟踪器的 4 倍,同时保持 92% 的准确率;针对人员,它可以同时处理 16 个摄像头。处理 3,000 帧视频时,其内存占用保持在约 3.5 GB,而 SAM 3 在第 423 帧就会耗尽内存。
The project is an open field guide covering 39 CAPTCHA types, their identifying signals, documented open-source bypass methods, and paid alternatives.该项目是一份开放实地指南,涵盖 39 种 CAPTCHA 类型、识别信号、已有的开源绕过方法及付费替代方案,并提供可搜索的网页版内容。
The project is an open field guide covering 39 CAPTCHA types, their identifying signals, documented open-source bypass methods, and paid alternatives. It provides a searchable web version and frames the material for education and authorized research. Its stated goal is to make changing CAPTCHA systems easier to navigate, compare, and cite while emphasizing responsible use.该项目是一份开放实地指南,涵盖 39 种 CAPTCHA 类型、识别信号、已有的开源绕过方法及付费替代方案,并提供可搜索的网页版内容。项目将材料定位为教育和获授权研究用途。其目标是在强调负责任使用的同时,让不断变化的 CAPTCHA 系统更易于查找、比较和引用。
The project is titled “decode kernel panic from QR Code photo.” It is presented as a GitHub Gist, but the available page content does not provide the decoding method or a working example.这个项目名为“从二维码照片中解码内核崩溃信息”。
The project is titled “decode kernel panic from QR Code photo.” It is presented as a GitHub Gist, but the available page content does not provide the decoding method or a working example.这个项目名为“从二维码照片中解码内核崩溃信息”。它以 GitHub Gist 的形式呈现,但现有页面内容没有提供解码方法或可运行示例。
VennDB is an append-only, in-memory database written in Rust that queries rows through bit-flag columns.VennDB 是一个用 Rust 编写的追加式内存数据库,通过位标志列查询行数据。
VennDB is an append-only, in-memory database written in Rust that queries rows through bit-flag columns. It targets mostly static datasets loaded at startup and queried frequently with simple filters, with an emphasis on speed and compactness. For its limited use cases, the project has few dependencies and is described as faster than a naive implementation or a heavier option such as SQLite.VennDB 是一个用 Rust 编写的追加式内存数据库,通过位标志列查询行数据。它面向主要在启动时加载、之后频繁使用简单过滤条件查询的静态数据集,重点是速度和紧凑性。在适用范围有限的场景中,该项目依赖很少,并被描述为比朴素实现或 SQLite 等更重量级方案更快。
Show HNProgrammingDataStartupsCodeNov 1711月17日0.1433
Partyline is a group voice app built on Reticulum and LXST, showcasing a decentralized approach to real-time group audio.Partyline 是一个在 Reticulum 与 LXST 上构建的群组语音应用,展示了去中心化的实时群体音频通信。
Partyline is a group voice app built on Reticulum and LXST, showcasing a decentralized approach to real-time group audio. The project emphasizes privacy-first peer-to-peer communication without centralized servers.Partyline 是一个在 Reticulum 与 LXST 上构建的群组语音应用,展示了去中心化的实时群体音频通信。该项目强调隐私优先的点对点通信,避免中心化服务器。
Codemap generates a single HTML page mapping a Go module’s packages, dependencies, files, and functions.Codemap 会生成一个 HTML 页面,展示 Go 模块中的包、依赖、文件和函数。
Codemap generates a single HTML page mapping a Go module’s packages, dependencies, files, and functions. Users can search the map, inspect what each package uses and what uses it, and jump to function definitions in GitHub-hosted modules. Tests are excluded from the map.Codemap 会生成一个 HTML 页面,展示 Go 模块中的包、依赖、文件和函数。用户可以搜索地图,查看每个包依赖什么以及被什么依赖,并跳转到 GitHub 模块中的函数定义。地图不包含测试代码。
Tiny-voca is a small English text-to-speech model with two voices, F2 and M2.Tiny-voca 是一款支持 F2 和 M2 两种声音的轻量级英语文本转语音模型。
Tiny-voca is a small English text-to-speech model with two voices, F2 and M2. Its complete model, including the waveform generator, has 4.5 million parameters and was trained from scratch on about 18 hours of synthetic speech for roughly $10. On four CPU threads, it runs at about 70 times real time.Tiny-voca 是一款支持 F2 和 M2 两种声音的轻量级英语文本转语音模型。包含波形生成器在内的完整模型有 450 万个参数,使用约 18 小时合成语音从头训练,成本约为 10 美元。在四个 CPU 线程上,它的运行速度约为实时速度的 70 倍。
STE80 Humanizer is an agent skill that rewrites AI-generated text to remove common AI markers while preserving its meaning.STE80 Humanizer 是一种智能体技能,旨在去除 AI 生成文本的常见痕迹,同时保持原意。
STE80 Humanizer is an agent skill that rewrites AI-generated text to remove common AI markers while preserving its meaning. It targets roughly 80% compliance with ASD-STE100 Simplified Technical English. The example focuses on Kubernetes liveness probes and retains both their restart behavior and the warning that misconfiguration can cause cascading restarts.STE80 Humanizer 是一种智能体技能,旨在去除 AI 生成文本的常见痕迹,同时保持原意。它的目标是让文本达到约 80% 的 ASD-STE100 简明技术英语规范。示例内容围绕 Kubernetes 存活探针展开,并保留了探针触发重启的行为以及配置错误可能导致级联重启的警告。
RAM-Altman is a Bash script that once a day reports unused RAM directly to OpenAI CEO Sam Altman.RAM-Altman 是一个每天向 OpenAI CEO Sam Altman 报告闲置内存的 Bash 脚本。
RAM-Altman is a Bash script that once a day reports unused RAM directly to OpenAI CEO Sam Altman. It requires no installation and provides sample emails through its companion site.RAM-Altman 是一个每天向 OpenAI CEO Sam Altman 报告闲置内存的 Bash 脚本。它无需安装,并通过配套网站提供示例邮件。
MoofTCP is a drop-in replacement for Apple's MacTCP on System 6 and 68000-based Macs.MoofTCP 是 System 6 和 68000 Mac 上苹果 MacTCP 的即插即用替代品。
MoofTCP is a drop-in replacement for Apple's MacTCP on System 6 and 68000-based Macs. Existing MacTCP programs can run unchanged on its newly written, portable-C TCP/IP stack. It includes its own control panel, installer, and DHCP client, and was designed for a Mac Classic with 4 MB of RAM.MoofTCP 是 System 6 和 68000 Mac 上苹果 MacTCP 的即插即用替代品。现有的 MacTCP 程序无需修改即可运行在其用可移植 C 语言重写的 TCP/IP 协议栈上。它还包含独立的控制面板、安装程序和 DHCP 客户端,并针对仅有 4 MB 内存的 Mac Classic 设计。
This work-in-progress project implements a ZIP-style Prolog virtual machine and a Prolog-to-ZIP compiler in C.这个进行中的项目用 C 实现了 ZIP 风格的 Prolog 虚拟机,以及从 Prolog 到 ZIP 的编译器。
This work-in-progress project implements a ZIP-style Prolog virtual machine and a Prolog-to-ZIP compiler in C. Its goal is to provide enough of a VM to compare it with the current trilog runtime. ZIP is the abstract machine described in the 1984 book Design and Simulation of a Sequential Prolog Machine.这个进行中的项目用 C 实现了 ZIP 风格的 Prolog 虚拟机,以及从 Prolog 到 ZIP 的编译器。其目标是先搭建足够完整的虚拟机,以便与当前的 trilog 运行时进行比较。ZIP 是 1984 年《Design and Simulation of a Sequential Prolog Machine》中描述的抽象机器。
Overtone is a show-n-HN project that transcribes, corrects, and cuts captions to fit any screen shape.Overtone 是 Show HN 项目,能够转写、纠错并将字幕裁剪以适配任意屏幕形状,提升无障碍和本地化工作效率。
Overtone is a show-n-HN project that transcribes, corrects, and cuts captions to fit any screen shape. It advances accessibility and localization workflows by automating caption adaptation. The tool can help creators reach diverse audiences across devices with fewer manual tweaks.Overtone 是 Show HN 项目,能够转写、纠错并将字幕裁剪以适配任意屏幕形状,提升无障碍和本地化工作效率。它帮助内容创作者覆盖多设备场景,省去大量手动调整。
db-branch creates copy-on-write branches of large MySQL and PostgreSQL databases in seconds using ZFS clones.db-branch 使用 ZFS 克隆,在几秒内为大型 MySQL 和 PostgreSQL 数据库创建写时复制分支。
db-branch creates copy-on-write branches of large MySQL and PostgreSQL databases in seconds using ZFS clones. The self-hosted tool can provide isolated databases for migrations, AI agents, or pull requests without copying the underlying data. Branches can be reset or discarded, and the project is licensed under Apache-2.0.db-branch 使用 ZFS 克隆,在几秒内为大型 MySQL 和 PostgreSQL 数据库创建写时复制分支。这款自托管工具无需复制底层数据,就能为迁移、AI 代理或每个拉取请求提供独立数据库。分支可以重置或丢弃,项目采用 Apache-2.0 许可证。
Battlezone Online brings BZFlag to the browser, enabling multiplayer play via WebGL/WebAssembly.Battlezone Online 将 BZFlag 带入浏览器,利用 WebGL/WebAssembly 实现多人对战。
Battlezone Online brings BZFlag to the browser, enabling multiplayer play via WebGL/WebAssembly. The project showcases how legacy open-source games can run natively in the browser and invites community participation.Battlezone Online 将 BZFlag 带入浏览器,利用 WebGL/WebAssembly 实现多人对战。该项目展示了开源遗产游戏在浏览器中原生运行的可能性,并邀请社区参与。
PyWolf3D is a Wolfenstein 3D engine and game clone written from scratch in Python using Pygame.PyWolf3D 是一个使用 Python 和 Pygame 从零编写的 Wolfenstein 3D 引擎及游戏克隆版,并提供演示。
PyWolf3D is a Wolfenstein 3D engine and game clone written from scratch in Python using Pygame. A demo is included.PyWolf3D 是一个使用 Python 和 Pygame 从零编写的 Wolfenstein 3D 引擎及游戏克隆版,并提供演示。
Mocktastic is a small, customizable Meshtastic-compatible LoRa client for ESP32 boards built with ArduinoIDE.Mocktastic 是一个面向 ESP32 开发板、使用 ArduinoIDE 构建的小型可定制 Meshtastic 兼容 LoRa 客户端。
Mocktastic is a small, customizable Meshtastic-compatible LoRa client for ESP32 boards built with ArduinoIDE. It is an educational and experimental C++ implementation rather than a fork of the official firmware. The project demonstrates Meshtastic packet encoding and decoding and includes an environmental-sensor example.Mocktastic 是一个面向 ESP32 开发板、使用 ArduinoIDE 构建的小型可定制 Meshtastic 兼容 LoRa 客户端。它是用于教育和实验的 C++ 实现,并非官方固件的分支。项目演示了 Meshtastic 数据包的编码与解码,并包含环境传感器示例。
Lasm lets developers write web applications in Lean with formal verification, targeting Node, serverless, and edge environments.Lasm 让开发者用 Lean 编写并进行形式化验证的网络应用,面向 Node、无服务器/边缘计算环境。
Lasm lets developers write web applications in Lean with formal verification, targeting Node, serverless, and edge environments. It demonstrates how strong guarantees can be baked into modern web apps, reducing runtime errors and increasing security.Lasm 让开发者用 Lean 编写并进行形式化验证的网络应用,面向 Node、无服务器/边缘计算环境。展示了在现代 Web 开发中嵌入定理证明保证的可行性与潜在收益。
XScreenSaver lacks an official GitHub repository, leaving users to rely on forks and mirrors.XScreenSaver 缺乏官方 GitHub 仓库,用户只能依赖 forks 与镜像。
XScreenSaver lacks an official GitHub repository, leaving users to rely on forks and mirrors. The situation highlights how long-running open-source projects drift from centralized hosting, complicating updates and issue tracking.XScreenSaver 缺乏官方 GitHub 仓库,用户只能依赖 forks 与镜像。本质上反映了长期运行的开源项目可能渐渐远离集中托管,更新与问题跟踪变得更加困难。
Reflota is a free, open-source, read-only tool for recovering files from NTFS, FAT32, FAT16, FAT12, and exFAT disks, including deleted files.Reflota 是一款免费、开源且只读的数据恢复工具,支持从 NTFS、FAT32、FAT16、FAT12 和 exFAT 磁盘中恢复文件,包括已删除的文件。
Reflota is a free, open-source, read-only tool for recovering files from NTFS, FAT32, FAT16, FAT12, and exFAT disks, including deleted files. It reads disks directly: for NTFS, it reconstructs the folder tree from the master file table and directory metadata, while for FAT and exFAT it reads directories and follows cluster chains. It is intended for cases where Windows reports that a disk is unformatted, a folder will not open, or Explorer hangs.Reflota 是一款免费、开源且只读的数据恢复工具,支持从 NTFS、FAT32、FAT16、FAT12 和 exFAT 磁盘中恢复文件,包括已删除的文件。它会直接读取磁盘:处理 NTFS 时读取主文件表并根据目录元数据重建文件夹树,处理 FAT 和 exFAT 时读取目录并跟踪簇链。当 Windows 报告磁盘未格式化、文件夹无法打开或资源管理器卡死时,该工具可用于尝试恢复数据。
Gauge is an open-source CLI that recursively improves documentation for coding agents.Gauge 是一个用于递归改进 coding agent 文档的开源 CLI。
Gauge is an open-source CLI that recursively improves documentation for coding agents. It runs tasks, analyzes agent transcripts for friction or failures, creates documentation variants, and repeats the process until a variant resolves the issues. The tool includes a local execution backend and can also optimize skills, CLIs, and other agent-facing interfaces.Gauge 是一个用于递归改进 coding agent 文档的开源 CLI。它会运行任务、分析 agent 会话记录中的阻碍或失败、生成文档变体,并重复这一过程,直到某个变体解决相关问题。该工具包含本地执行后端,也可用于优化技能、CLI 和其他面向 agent 的界面。
Show HNOpen SourceProgrammingWebCodeNov 411月4日0.3294
Sutra-OS is an open-source, self-hosted ontology platform.Sutra-OS 是一个开源、自托管的本体平台。
Sutra-OS is an open-source, self-hosted ontology platform. It represents data as one owned model with typed and linked entities, server-enforced rules, and numbers traceable to their sources. The project describes its design as “Scarce Units, Typed, Related, Actionable.”Sutra-OS 是一个开源、自托管的本体平台。它将数据表示为用户拥有的统一模型,其中实体具有类型和关联,修改由服务器规则强制执行,所有数字都能追溯到来源。项目将其设计概括为“稀缺、类型化、关联、可行动”。
Flawless UI-UX is a skill for coding agents that helps them build distinctive and usable interfaces.Flawless UI-UX 是一项帮助编程智能体构建具有特色且易用界面的技能。
Flawless UI-UX is a skill for coding agents that helps them build distinctive and usable interfaces. Version 1.0.0 includes an interactive installer and support for twelve agents. The installer can be run with npm, pnpm, or Bun, and lets users choose project or global scope, select agents, review destinations, and confirm file copying.Flawless UI-UX 是一项帮助编程智能体构建具有特色且易用界面的技能。1.0.0 版本提供交互式安装程序,并支持 12 种智能体。用户可以通过 npm、pnpm 或 Bun 启动安装器,选择项目级或全局范围、指定智能体、检查目标位置并确认复制文件。
Panthera is Ampere’s cycle-accurate software performance model for studying prospective CPU features and evaluating microarchitectural trade-offs.Panthera 是 Ampere 的周期精确软件性能模型,用于研究候选 CPU 特性并评估微架构取舍。
Panthera is Ampere’s cycle-accurate software performance model for studying prospective CPU features and evaluating microarchitectural trade-offs. It is built by Ampere’s Architecture & Performance team to guide development of high-performance cloud processors. The project includes a user guide for building and using Panthera and its associated tools.Panthera 是 Ampere 的周期精确软件性能模型,用于研究候选 CPU 特性并评估微架构取舍。它由 Ampere 架构与性能团队开发,以指导高性能云处理器的设计。项目还提供了构建和使用 Panthera 及相关工具的用户指南。
A Linux-native rebuild of the classic Picasa viewer aims to revive familiar photo-management workflows on modern desktops.一个面向 Linux 的 Picasa 照片查看器重制版正试图在现代桌面环境中延续熟悉的照片管理流程。
A Linux-native rebuild of the classic Picasa viewer aims to revive familiar photo-management workflows on modern desktops. The project prioritizes a faithful user experience while integrating with current Linux ecosystems.一个面向 Linux 的 Picasa 照片查看器重制版正试图在现代桌面环境中延续熟悉的照片管理流程。该项目力求保留原有使用体验,同时与当前的 Linux 生态无缝对接。
ZeBridge is a bidirectional, local-first synchronization engine connecting PostgreSQL with edge SQLite, PGLITE, PostgreSQL, and DuckDB clients over NATS.ZeBridge 是一个双向、以本地优先为理念的同步引擎,通过 NATS 连接 PostgreSQL 与边缘端的 SQLite、PGLITE、PostgreSQL 和 DuckDB 客户端。
ZeBridge is a bidirectional, local-first synchronization engine connecting PostgreSQL with edge SQLite, PGLITE, PostgreSQL, and DuckDB clients over NATS. It keeps a local database on each device and queues writes while the device is offline. When connectivity returns, PostgreSQL remains authoritative by applying rules, resolving conflicts, and returning a result to the application.ZeBridge 是一个双向、以本地优先为理念的同步引擎,通过 NATS 连接 PostgreSQL 与边缘端的 SQLite、PGLITE、PostgreSQL 和 DuckDB 客户端。它在每台设备上保留本地数据库,并在离线时排队保存写入操作。重新连接后,PostgreSQL 仍是权威数据源,负责应用规则、解决冲突并向应用返回结果。
tmux-restore restores tmux sessions, layouts, and scrollback after a reboot.tmux-restore 可在重启后恢复 tmux 会话、布局和滚动缓冲区。
tmux-restore restores tmux sessions, layouts, and scrollback after a reboot. It can also reopen conversations for Claude Code, Codex, Gemini CLI, opencode, and pi. The project describes itself as reboot-proof tmux for coding agents.tmux-restore 可在重启后恢复 tmux 会话、布局和滚动缓冲区。它还能够重新打开 Claude Code、Codex、Gemini CLI、opencode 和 pi 中此前的对话。该项目将自己定位为面向编程代理的“重启无忧” tmux 工具。
Show HNProgrammingOpen SourceAICodeNov 611月6日0.02711
Ghostman is a keyboard-driven API client and webhook inspector that runs in the terminal.Ghostman 是一个运行在终端中的键盘驱动 API 客户端和 Webhook 检查器。
Ghostman is a keyboard-driven API client and webhook inspector that runs in the terminal. It is local-first, requires no cloud account, uses no Electron, and is presented as a lightweight alternative to heavier modern API clients.Ghostman 是一个运行在终端中的键盘驱动 API 客户端和 Webhook 检查器。它以本地优先为设计理念,无需云账户,也不使用 Electron,定位是更轻量的现代 API 客户端替代方案。
This survival assistant runs fully offline on Windows, Linux, and macOS, including from a laptop or USB drive.这款生存助手可在 Windows、Linux 和 macOS 上完全离线运行,也可以从笔记本电脑或 U 盘启动。
This survival assistant runs fully offline on Windows, Linux, and macOS, including from a laptop or USB drive. It provides sourced guides on water, fire, food, first aid, and other topics in Portuguese, English, and Spanish, alongside a local LLM using llama.cpp or Ollama. Kiwix libraries provide additional offline reference material.这款生存助手可在 Windows、Linux 和 macOS 上完全离线运行,也可以从笔记本电脑或 U 盘启动。它提供葡萄牙语、英语和西班牙语的水、火、食物、急救等主题资料,并通过 llama.cpp 或 Ollama 运行本地 LLM。Kiwix 资料库则提供额外的离线参考内容。
LDraw Mecha uses agentic language models to analyze annotated LDraw LEGO model files, infer their submodels and mechanisms, plan animations, and execute them in Blender.LDraw Mecha 使用代理式语言模型分析带注释的 LDraw LEGO 模型文件,推断其中的子模型和机械结构,规划动画,并在 Blender 中执行。
LDraw Mecha uses agentic language models to analyze annotated LDraw LEGO model files, infer their submodels and mechanisms, plan animations, and execute them in Blender. It outputs a rigged and animated Blender file from an initial .mpd source file. The project builds on LDraw Nova, an earlier tool for generating LEGO CAD models from LDraw source code.LDraw Mecha 使用代理式语言模型分析带注释的 LDraw LEGO 模型文件,推断其中的子模型和机械结构,规划动画,并在 Blender 中执行。它可以从初始的 .mpd 源文件生成带绑定和动画的 Blender 文件。该项目建立在此前用于从 LDraw 源代码生成 LEGO CAD 模型的 LDraw Nova 之上。
Edison provides a Python toolkit to parse EDIFACT messages using syntax trees, enabling more readable, verifiable parsing of standardized EDI data.Edison 提供了一个用语法树解析 EDIFACT 的 Python 工具包,使对标准化EDI数据的解析更加直观且可验证。
Edison provides a Python toolkit to parse EDIFACT messages using syntax trees, enabling more readable, verifiable parsing of standardized EDI data. This approach helps developers validate and transform large batch files used in travel, logistics, and commerce.Edison 提供了一个用语法树解析 EDIFACT 的 Python 工具包,使对标准化EDI数据的解析更加直观且可验证。该方法有助于在旅行业、物流和商业领域处理的大规模文件进行验证与转换。
Recalc is a spreadsheet that can host a lead Claude Code or Codex session in a side panel while allowing each cell to act as a terminal session.Recalc 是一款电子表格应用,可在侧栏运行主 Claude Code 或 Codex 会话,并让每个单元格充当终端会话。
Recalc is a spreadsheet that can host a lead Claude Code or Codex session in a side panel while allowing each cell to act as a terminal session. Sessions can communicate through grid cells, named cells, or an MCP server, and can be continued in a regular terminal. The tool currently runs locally on macOS, with Linux support described as less tested.Recalc 是一款电子表格应用,可在侧栏运行主 Claude Code 或 Codex 会话,并让每个单元格充当终端会话。不同会话可以通过网格单元格、命名单元格或 MCP 服务器互相通信,也可以在普通终端中继续运行。目前该工具可在 macOS 本地运行,Linux 支持则尚未充分测试。
Uniform is a JavaScript framework for web applications that uses WebAssembly for layout and GPU rendering, with a reported size of 280 KB.Uniform 是一个用于网页应用的 JavaScript 框架,使用 WebAssembly 处理布局并通过 GPU 渲染,体积据称为 280 KB。
Uniform is a JavaScript framework for web applications that uses WebAssembly for layout and GPU rendering, with a reported size of 280 KB. Its creator aims to avoid the platform limits of native frameworks, the HTML/CSS constraints of browser-engine approaches, and the size or language requirements of other cross-platform frameworks. The design bundles the UI framework with the application so developers can update it independently of platform release cycles.Uniform 是一个用于网页应用的 JavaScript 框架,使用 WebAssembly 处理布局并通过 GPU 渲染,体积据称为 280 KB。作者希望避免原生框架的平台限制、基于浏览器引擎方案所继承的 HTML/CSS 约束,以及其他跨平台框架的体积或语言要求。该设计将 UI 框架直接与应用打包,使开发者能够独立于平台发布周期更新它。
Kauldron is a modular, scalable library for training machine learning models, aimed at simplifying large-scale training and experimentation.Kauldron 是一个模块化、可扩展的机器学习模型训练库,旨在简化大规模训练与实验流程。
Kauldron is a modular, scalable library for training machine learning models, aimed at simplifying large-scale training and experimentation.Kauldron 是一个模块化、可扩展的机器学习模型训练库,旨在简化大规模训练与实验流程。
OpenTodo is a free, open-source task manager that can be self-hosted or hosted for users.OpenTodo 是一款免费开源的任务管理器,用户可以自行部署或交由他人托管。
OpenTodo is a free, open-source task manager that can be self-hosted or hosted for users. It works on phones and computers, supports offline use and shared lists, and keeps users’ data under their control. The project is designed as a local-first alternative to subscription-based task applications.OpenTodo 是一款免费开源的任务管理器,用户可以自行部署或交由他人托管。它支持手机和电脑、离线使用以及与家人或团队共享列表,并让用户掌控自己的数据。该项目旨在提供一种本地优先的替代方案,减少对订阅制任务应用的依赖。
A Claude Code modification adds an usage-window guard to prevent overuse or budget overruns, useful for teams running Claude in production.一个 Claude Code 的修改插件增设使用时段保护,帮助避免超出预算或滥用。
A Claude Code modification adds an usage-window guard to prevent overuse or budget overruns, useful for teams running Claude in production. It introduces lightweight enforcements for access windows and prompts.一个 Claude Code 的修改插件增设使用时段保护,帮助避免超出预算或滥用。适用于在生产环境中运行 Claude 的团队,提供对访问时间的轻量级约束。
Show HNProgrammingOpen SourceAICodeOct 2210月22日0.1207
TubeHaus is a private, self-hosted YouTube client available as a Docker Compose web app and native apps for Android, Windows, macOS, Linux, and iPhone/iPad.TubeHaus 是一款私密、自托管的 YouTube 客户端,提供 Docker Compose 网页版,以及 Android、Windows、macOS、Linux 和 iPhone/iPad 原生应用。
TubeHaus is a private, self-hosted YouTube client available as a Docker Compose web app and native apps for Android, Windows, macOS, Linux, and iPhone/iPad. It requires no Google account and provides no tracking. The client preserves subscriptions and history on the user's own server or device while retaining a YouTube-like layout and keyboard shortcuts.TubeHaus 是一款私密、自托管的 YouTube 客户端,提供 Docker Compose 网页版,以及 Android、Windows、macOS、Linux 和 iPhone/iPad 原生应用。它无需 Google 账户,也不进行跟踪。用户可以将订阅和历史记录保存在自己的服务器或设备上,同时保留类似 YouTube 的布局和键盘快捷键。
Show HNOpen SourceWebPrivacyDevOpsCodeNov 511月5日0.1185
This custom firmware for Baofeng UV-5RM radios adds applications including texting and spectrum tools.这套 Baofeng UV-5RM 自定义固件新增了短信和频谱工具等应用。
This custom firmware for Baofeng UV-5RM radios adds applications including texting and spectrum tools. Its texting feature supports short messages between two radios with delivery confirmation and requires no phone, internet connection, or additional hardware. CHIRP will no longer communicate with the modified device, so channels must be programmed with the included applications.这套 Baofeng UV-5RM 自定义固件新增了短信和频谱工具等应用。短信功能支持两台无线电之间发送短消息并确认送达,不需要手机、互联网或额外硬件。改装后设备将无法再与 CHIRP 通信,因此必须使用随附应用配置频道。
Flatpoints applies geometry-image principles directly to point clouds.Flatpoints 将几何图像的原理直接应用于点云。
Flatpoints applies geometry-image principles directly to point clouds. It maps each 3D point from a point cloud sampled on an unknown surface to a 2D pixel, arranging the points into a square grid. The resulting flat representation can serve as an image-like indexing structure.Flatpoints 将几何图像的原理直接应用于点云。它把从未知曲面采样得到的每个三维点映射为二维像素,并将这些点排列到方形网格中。生成的平面表示可以作为类似图像的索引结构。
Vault-share is a self-hosted, invite-only encrypted file-sharing application built with Flask.Vault-share 是一个基于 Flask 的自托管、仅限邀请的加密文件共享应用。
Vault-share is a self-hosted, invite-only encrypted file-sharing application built with Flask. Each file uses its own AES-256-GCM key, which is sealed separately for every authorized reader using HPKE, while uploads are signed and integrity-checked. The prototype also supports duress passwords that reveal separate cover vaults and can require a USB flash drive for the master key.Vault-share 是一个基于 Flask 的自托管、仅限邀请的加密文件共享应用。每个文件使用独立的 AES-256-GCM 密钥,并通过 HPKE 分别封装给每位获授权的读取者;上传内容还会进行签名和完整性校验。该原型还支持在胁迫密码下打开独立的伪装保险库,并可要求使用 USB 闪存盘保存主密钥。
The project is an open-source collection of assets for daily games.该项目是一个面向每日游戏的开源资源收藏集。
The project is an open-source collection of assets for daily games. The available information provides no further details about the collection.该项目是一个面向每日游戏的开源资源收藏集。现有信息未提供该收藏集的更多细节。
SafeScript is a strict computational subset of JavaScript designed to execute untrusted code in multiple environments.SafeScript 是 JavaScript 的严格计算子集,旨在让不可信代码能够在多种环境中安全执行。
SafeScript is a strict computational subset of JavaScript designed to execute untrusted code in multiple environments. It excludes features including eval, new, globals, and prototype access, then compiles to plain JavaScript and PHP. The compiler is about 48 KB minified, or roughly 12 KB gzipped, and uses .safe files intended to be locally verifiable.SafeScript 是 JavaScript 的严格计算子集,旨在让不可信代码能够在多种环境中安全执行。它排除了 eval、new、全局变量和原型访问等功能,并可编译为普通 JavaScript 和 PHP。编译器压缩后约 48 KB、gzip 后约 12 KB,使用 .safe 文件格式,并支持本地验证。
Datastructures-JS is a JavaScript data-structures project with TypeScript support.Datastructures-JS 是一个支持 TypeScript 的 JavaScript 数据结构项目。
Datastructures-JS is a JavaScript data-structures project with TypeScript support. The project also provides an associated planning board.Datastructures-JS 是一个支持 TypeScript 的 JavaScript 数据结构项目。项目还提供了配套的规划看板。
WolveKafka lets applications consume Kafka through an HTTP handler in any programming language.WolveKafka 允许应用使用任意编程语言,通过 HTTP 处理器消费 Kafka。
WolveKafka lets applications consume Kafka through an HTTP handler in any programming language. It manages consumer groups, batching, retries, offsets, and dead-letter queues. The project includes a quickstart, examples, deployment guidance, and comparison material.WolveKafka 允许应用使用任意编程语言,通过 HTTP 处理器消费 Kafka。它负责管理消费者组、批处理、重试、偏移量和死信队列。项目还提供快速入门、示例、部署指南和对比资料。
Agda 2.9.0 adds an experimental `-j` option for type-checking modules in parallel in batch mode.Agda 2.9.0 在批处理模式下新增实验性的 `-j` 选项,可并行进行模块类型检查。
Agda 2.9.0 adds an experimental `-j` option for type-checking modules in parallel in batch mode. The release also includes substantial performance improvements and drops support for GHC 8.8, 8.10, and 9.0. Agda is a dependently typed programming language and interactive theorem prover.Agda 2.9.0 在批处理模式下新增实验性的 `-j` 选项,可并行进行模块类型检查。该版本还带来了显著的性能改进,并停止支持 GHC 8.8、8.10 和 9.0。Agda 是一种依赖类型编程语言和交互式定理证明器。
MoEShift is an Apple Silicon inference engine for memory-efficient, high-performance mixture-of-experts models.MoEShift 是一款面向 Apple Silicon 的高性能、节省内存的混合专家模型推理引擎。
MoEShift is an Apple Silicon inference engine for memory-efficient, high-performance mixture-of-experts models. On a Mac mini with an M4 Pro and 24 GB of unified memory, it runs the 125-billion-parameter Qwen 3.8 Flash Next using under 10 GB of peak process memory. The project reports roughly half the peak memory usage of Strata and 2.6× faster long-prompt processing.MoEShift 是一款面向 Apple Silicon 的高性能、节省内存的混合专家模型推理引擎。在配备 M4 Pro 和 24GB 统一内存的 Mac mini 上,它运行 1250 亿参数的 Qwen 3.8 Flash Next 时,进程峰值内存低于 10GB。项目称,与 Strata 相比,它的峰值内存占用约少一半,长提示处理速度快 2.6 倍。
DRS-VPT estimates a camera’s six-degree-of-freedom pose from a single photo and a 3D point cloud without an initial pose or known camera intrinsics.DRS-VPT 可以根据一张照片和一个三维点云估计相机的六自由度位姿,无需初始位姿或已知相机内参。
DRS-VPT estimates a camera’s six-degree-of-freedom pose from a single photo and a 3D point cloud without an initial pose or known camera intrinsics. The same model works across driving LiDAR scans, indoor iPhone photos, and structure-from-motion maps. The authors provide the code, model, weights, and an online demo.DRS-VPT 可以根据一张照片和一个三维点云估计相机的六自由度位姿,无需初始位姿或已知相机内参。同一个模型可用于驾驶场景的 LiDAR 扫描、室内 iPhone 照片和运动恢复结构地图。作者提供了代码、模型、权重以及在线演示。
VB Harness makes local coding agents work in small tasks and isolated copies, using a project’s own checks to validate each change.VB Harness 将本地编程智能体的工作拆分为小任务,并在隔离副本中执行,通过项目自身的检查来验证每次修改。
VB Harness makes local coding agents work in small tasks and isolated copies, using a project’s own checks to validate each change. It supports LM Studio, Ollama, and other OpenAI-compatible models, and retains changes only when type checks, linting, builds, tests, and independent review succeed. The tool is designed to make local agents accountable without trying to make the underlying model smarter.VB Harness 将本地编程智能体的工作拆分为小任务,并在隔离副本中执行,通过项目自身的检查来验证每次修改。它支持 LM Studio、Ollama 和其他兼容 OpenAI API 的模型,只有在类型检查、代码检查、构建、测试和独立复审均通过后才会保留改动。该工具旨在让本地智能体对结果负责,而不是试图直接提升底层模型的智能程度。
Cerulion is a ROS 2-compatible robotics runtime written in Rust.Cerulion 是一个用 Rust 编写、兼容 ROS 2 的机器人运行时。
Cerulion is a ROS 2-compatible robotics runtime written in Rust. It aims to reduce inter-process communication latency without requiring a single process or restricting messages to plain-old-data types. Its multi-process graphs can be deterministically resimulated, and existing ROS 2 nodes and launch files can run alongside native Cerulion Rust or Python code.Cerulion 是一个用 Rust 编写、兼容 ROS 2 的机器人运行时。它旨在降低进程间通信延迟,同时不要求将程序合并为单进程,也不限制消息必须使用普通数据类型。其多进程图可以进行确定性重演,现有 ROS 2 节点和 launch 文件也能与原生 Cerulion Rust 或 Python 代码并行运行。
This open-source project lets users control the status light on a MacBook charger.这个开源项目可以控制 MacBook 充电器的状态灯。
This open-source project lets users control the status light on a MacBook charger. It supports turning the light on or off automatically based on time or Focus status, as well as keeping it persistently on or off. The project invites feature suggestions through GitHub.这个开源项目可以控制 MacBook 充电器的状态灯。它支持根据时间或专注模式状态自动开关灯,也可以让灯持续亮起或关闭。项目通过 GitHub 征集功能建议。
This open hardware board gives a computer LTE data, SMS, and voice calls over USB-C.这款开源硬件板通过 USB-C 为电脑提供 LTE 数据、短信和语音通话功能。
This open hardware board gives a computer LTE data, SMS, and voice calls over USB-C. With a nano-SIM and one USB-C cable, a computer can route call audio to a headset, hand calls to a script or AI agent, and send texts. The four-layer KiCad board was designed from a one-page specification and reviewed against component datasheets.这款开源硬件板通过 USB-C 为电脑提供 LTE 数据、短信和语音通话功能。插入 nano-SIM 并连接一根 USB-C 线后,电脑可以把通话音频发送到耳机,也可以将通话交给脚本或 AI 代理处理,并发送短信。这块四层 KiCad 电路板根据一页规格说明设计,并逐项对照元件数据手册进行审核。
The project is a bio-inspired edge-detection algorithm modeled on the visual cortex.该项目是一种受视觉皮层启发的仿生边缘检测算法。
The project is a bio-inspired edge-detection algorithm modeled on the visual cortex. Its current benchmark results fall between traditional methods such as Canny and Sobel and modern deep neural networks, while retaining the interpretability of the former. The author plans to improve video frame rates and performance across different hardware, and is seeking collaborators.该项目是一种受视觉皮层启发的仿生边缘检测算法。目前的基准测试结果介于 Canny、Sobel 等传统方法与现代深度神经网络之间,同时保留了传统方法的可解释性。作者计划提升视频处理帧率和不同硬件上的性能,并寻找合作伙伴。
Async-a-sync is a compiler and runtime system built on Fil-C that lets ordinary blocking-looking C code run asynchronously underneath.Async-a-sync 是构建在 Fil-C 之上的编译器和运行时系统,让表面上阻塞的普通 C 代码在底层异步运行。
Async-a-sync is a compiler and runtime system built on Fil-C that lets ordinary blocking-looking C code run asynchronously underneath. It introduces implicit asynchronous futures without adding syntax, with the first access to a buffer acting as the await point.Async-a-sync 是构建在 Fil-C 之上的编译器和运行时系统,让表面上阻塞的普通 C 代码在底层异步运行。它无需新增语法即可引入隐式异步 future,并将首次访问缓冲区作为 await 点。
This hands-on guide explains how to train Mixture-of-Experts models.这份实践指南介绍了如何训练混合专家模型。
This hands-on guide explains how to train Mixture-of-Experts models. It provides a navigable table of contents and covers the practical steps involved in building and training MoE systems.这份实践指南介绍了如何训练混合专家模型。页面提供可点击的目录,并涵盖构建和训练 MoE 系统所需的实际步骤。
DeepSeek V4 Pro 0813 is a text-generation model with 1.7 trillion parameters.DeepSeek V4 Pro 0813 是一款拥有 1.7 万亿参数的文本生成模型。
DeepSeek V4 Pro 0813 is a text-generation model with 1.7 trillion parameters. The model files are available to download and explore through Hugging Face.DeepSeek V4 Pro 0813 是一款拥有 1.7 万亿参数的文本生成模型。其模型文件可通过 Hugging Face 下载和探索。
Flide is an inference engine designed specifically for Jev-like decision models.Flide 是一款专为 Jev 类决策模型设计的推理引擎。
Flide is an inference engine designed specifically for Jev-like decision models. The project claims up to a 9.5x speedup over vLLM and SGLang and currently supports more than 30 popular local Jev-like models. It can be installed with pip using the package name flide.Flide 是一款专为 Jev 类决策模型设计的推理引擎。项目称其速度最高可达 vLLM 和 SGLang 的 9.5 倍,目前支持 30 多种热门本地 Jev 类模型。用户可以通过 pip 安装名为 flide 的软件包。
Maple is a structural-compression method for mixture-of-experts language models.Maple 是一种面向混合专家语言模型的结构化压缩方法。
Maple is a structural-compression method for mixture-of-experts language models. It compresses Gemma 4 26B-A4B from 9.92 GB to 4.75 GB, equivalent to 1.48 effective bits per parameter. The project reports 100% in-VRAM execution on 6 GB laptop GPUs at 45.4 tokens per second.Maple 是一种面向混合专家语言模型的结构化压缩方法。它将 Gemma 4 26B-A4B 从 9.92 GB 压缩到 4.75 GB,相当于每参数 1.48 个有效比特。项目报告称,该模型可在 6GB 笔记本 GPU 上完全载入显存运行,速度达到每秒 45.4 个 token。
FailTrace takes the URL of a failed GitHub Actions run and identifies the failed job and step.FailTrace 接收失败的 GitHub Actions 运行 URL,并找出失败的作业和步骤。
FailTrace takes the URL of a failed GitHub Actions run and identifies the failed job and step. It matches that failure to the workflow file at the exact commit, extracts the relevant log output, and explains what is needed to reproduce the issue locally. The tool is designed to reduce guesswork when debugging CI failures.FailTrace 接收失败的 GitHub Actions 运行 URL,并找出失败的作业和步骤。它会将失败过程与对应提交中的工作流文件匹配,提取相关日志,并说明在本地复现问题所需的条件。该工具旨在减少调试 CI 失败时的猜测。
Skillset groups programming, backend, and other agent skills into collections that can be installed with one command from a JSON file.Skillset 将编程、后端等代理技能整理成集合,用户可以通过一个 JSON 文件和一条命令安装所需集合。
Skillset groups programming, backend, and other agent skills into collections that can be installed with one command from a JSON file. It wraps the Vercel Skill CLI and uses it as the underlying library, aiming to preserve compatibility with the standard CLI across teams. The project is designed to avoid installing unnecessary skills globally and consuming excess context.Skillset 将编程、后端等代理技能整理成集合,用户可以通过一个 JSON 文件和一条命令安装所需集合。它封装并直接使用 Vercel Skill CLI,旨在保持与标准 CLI 的团队间兼容性。该项目希望避免全局安装不必要的技能并消耗过多上下文。
A self-hosted radio stack that combines Icecast, Liquidsoap, and a Rust-based CLI to simplify running your own Internet radio.自托管的网络广播堆栈,将 Icecast、Liquidsoap 与 Rust CLI 组合在一起,简化自建互联网电台的部署。
A self-hosted radio stack that combines Icecast, Liquidsoap, and a Rust-based CLI to simplify running your own Internet radio. The project emphasizes portability, reproducibility, and a minimal maintenance footprint for hobbyists.自托管的网络广播堆栈,将 Icecast、Liquidsoap 与 Rust CLI 组合在一起,简化自建互联网电台的部署。强调可移植性、可重复性与低维护成本,面向爱好者。
Awake is an API client created after its developers encountered problems with Postman and Bruno around environments, workspaces, pre-request authentication, and performance on large collections.Awake 是一款 API 客户端,源于开发者在使用 Postman 和 Bruno 时遇到的环境、工作区、请求前认证及大型集合性能问题。
Awake is an API client created after its developers encountered problems with Postman and Bruno around environments, workspaces, pre-request authentication, and performance on large collections. The project began as a side project and became an everyday tool for its creators. It is presented as a Git-friendly open API client.Awake 是一款 API 客户端,源于开发者在使用 Postman 和 Bruno 时遇到的环境、工作区、请求前认证及大型集合性能问题。该项目最初是一个业余项目,后来成为创建者日常使用的工具。它被定位为一款对 Git 友好的开放 API 客户端。