GlucoFM：面向连续血糖监测的基础模型

Google Research 推出 GlucoFM，一款轻量级自监督 CGM 基础模型，采用双流设计分别建模缓慢血糖趋势与短期波动。在四个队列、七项临床预测任务的 14 项评估中，其 PR-AUC 较最优 GluFormer 变体平均高出 5.8 个百分点，并在 PPGR 预测中取得最低 MAE。模型在 109，066 小时无标注 CGM 数据上预训练，具备跨数据集迁移与少样本适应能力。（AIHOT源：Google Research：Blog（网页））

Claude in Chrome 正式全面上线

Anthropic 宣布 Claude in Chrome 现已面向所有付费 Claude 套餐全面开放，Claude 可在浏览器中自主执行操作，无需逐步审批。系统通过安全分类器在每次操作前验证其安全性及是否符合用户请求，并强化了针对提示注入攻击的防御。最新评测显示，在启用探测与安全分类器后，自 Opus 4.8 起的所有模型均未出现攻击成功案例。（AIHOT源：Claude：Blog（网页））

Anthropic 开放 Claude 真实使用数据供外部独立研究，公布试点结果

Anthropic 今年春季启动试点，通过隐私保护工具 Anthropic Insights（原 Clio）向斯坦福大学 SALT Lab、牛津大学人类信息处理实验室及 METR 三个外部机构开放约 25 万段 2026 年 4-5 月的 Claude.ai 或 Claude Code 对话数据，供其独立设计研究并公开发布结果。（AIHOT源：Anthropic：Research（发表成果 · 网页））

Warp 如何在 Claude 上构建自我改进的智能体

Warp 在 Claude 平台上构建了基于 Agent Skills 的自我改进循环，通过基础技能与改进技能两个文件型技能，将人类反馈转化为对智能体的持续优化。该模式已应用于其整个开源仓库，覆盖数百名贡献者与数千次代码审查。团队建议以原则而非规则编写技能，并强调低摩擦反馈与改进技能的可复用性。（AIHOT源：Claude：Blog（网页））

Gemini 3.5 Transcribe 发布：面向实时语音交互的高精度语音转文本模型

Google DeepMind 推出 Gemini 3.5 Transcribe 语音转文本模型，支持流式与非流式两种 API。据 Artificial Analysis 评测，其流式与非流式平均词错率分别为 4.0% 和 2.6%，支持超 85 种语言、自定义词汇及最多三人说话人识别。（AIHOT源：Google DeepMind：Blog（RSS））

Qwen3.8-Flash 开源，Qwen4 架构预览

通义千问发布 Qwen3.8-Flash，一款多模态 MoE 模型，作为 Qwen4 架构的早期预览并开放权重。该模型总参数 125B，每 token 仅激活 6B，训练成本仅为 Qwen3.7-Plus 的 1/9，性能全面超越后者。生产版 API 定价 $0.16/1M 输入 tokens 和 $0.47/1M 输出 tokens，原生上下文 262K，可扩展至 1M。（AIHOT源：X：通义千问 / Qwen (@Alibaba_Qwen)）

GLM-5.3-Flash 开源：320B 总参数、AA 指数 57 分，定价为 Opus 4.8 的 1/40

智谱上线并开源 GLM-5.3-Flash（320B-A18B），这是 GLM-5 系列首个原生多模态模型，AA 综合智能指数 57 分，与 Claude Opus 4.8 持平。其定价为 GLM-5.3 的 1/10，限时折扣内为 Opus 4.8 的 1/40，并已接入 ZCode 等平台开放 API 调用。该模型采用稀疏注意力与线性注意力混合架构，推理服务已跑在国产芯片集群上。（AIHOT源：公众号：智谱（GLM））

以色列资助的假美国智库试图利用AI进行宣传

一个打着"汉诺威公共政策研究所"旗号的亲以色列网站，在九天内发布了124篇、超56万字的报告，旨在优化内容以引导ChatGPT等AI聊天机器人引用其亲以观点。该网站由美国公司Piro Inc依据《外国代理人登记法》为以色列政府分发内容，其背后是以色列政府数千万美元资助、经Havas Media等第三方转手的更广泛宣传行动。（AIHOT源：Hacker News 热门（buzzing.cc 中文翻译））

C2PA相机经不起现实的考验：Android端可被root攻击伪造签名

安全研究员David Buchanan指出，C2PA相机认证在Android平台上可被攻破。通过root权限提升漏洞（如CVE-2026-43499），攻击者可利用StrongBox硬件签名任意数据，伪造C2PA签名图像和视频，且无需硬件攻击。该问题无法通过常规补丁修复，已提前90天向相关方报告。（AIHOT源：Hacker News 热门（buzzing.cc 中文翻译））

Qwen3.8-Flash-Next 开源：Qwen4 架构早期预览

通义千问开源 Qwen3.8-Flash-Next，一款多模态 MoE 模型，也是 Qwen4 架构的早期预览。该模型采用 GDN + QSA 混合注意力等四项升级，总参数 125B，每 token 激活 6B，训练成本约为 Qwen3.7-Plus 的 1/9，编码与办公任务能力更强。（AIHOT源：Qwen：Blog Retrieval（API））

实测飞书和豆包合体后第1个Agent：豆包工作的8个使用技巧

豆包工作（豆包 Work）是当前企业接入Agent门槛最低的路径，但需用飞书账号登录才能解锁满血功能。实测可用手机远程控制最多7台设备、定时任务、自动读取本地skill、侧边栏直接编辑并同步飞书，且管理员看不到聊天记录。作者认为Work Agent是token消耗倍增器，飞书原生生态是豆包工作相比Claude Cowork、Codex Work的核心优势。（AIHOT源：公众号：卡尔的AI沃茨）

腾讯混元将端侧翻译模型 Hy-MT2-1.8B 压缩至 440MB，已落地哔哩哔哩直播弹幕翻译

腾讯混元将端侧翻译模型 Hy-MT2-1.8B 通过 2-bit 与 1.25-bit 量化方案压缩至 574MB 和 440MB，翻译质量几乎无损，在 FLORES-200 上优于 Microsoft Translator 等商业 API。该模型已联合英特尔完成 x86 适配，并在哔哩哔哩直播弹幕实时翻译中落地，单条弹幕翻译耗时 500~800ms。（AIHOT源：公众号：腾讯混元）

Introducing AI Disclosure on DEV: Tools for Nuance, Clarity, and Better Feeds

10 Git Commands You’ll Wish You Knew Earlier

Why I Built an SSH Config and Tunnel Manager for macOS

I Tested 5 Design to Code Tools With the Same Outdated SaaS Dashboard

Portfolio Update, I Guess

Claude: Claude now has one memory across chat and Claude Cowork, and you decide what'...

Claude now has one memory across chat and Claude Cowork, and you decide what's in it. Hand Cowork a task and it starts from what Claude already knows from your chats: the project you talked through, your manager's preferences, or the client from last quarter. https://t.co/ViKapsT8K6｜@claudeai · 热度 7946 · 2026-08-25

Thibault Sottiaux: So much demand for this one

So much demand for this one. Works similar to the Pro $100 plan but designed for teams and small companies. ✅ All ChatGPT, ChatGPT Work, and Codex features ✅ Connect to Google Workspace, Slack, GitHub, Microsoft 365, and more ✅ Secure workspace with SAML, SSO, and MFA ✅ Centralized billing and administration ✅ Usage analytics and spend controls ✅ No 5h limits｜@thsottiaux · 热度 3715 · 2026-08-25

Peter Yang: Today, I’m open-sourcing /fuck-cancer, an AI skill that helps patients and ca...

Today, I’m open-sourcing /fuck-cancer, an AI skill that helps patients and caregivers navigate cancer diagnosis and treatment and advocate for themselves and their loved ones. Here’s what patients and caregivers have told me since I shared my mom’s story: “You have to be a huge patient advocate. Ask questions and push for answers, biopsies, and proper testing.” “I need help navigating difficult conversations and advocating for myself.” “The volume of doctors, documents, and insurance paperwork can quickly become overwhelming.” I built the skill to help with these problems. It creates and updates a practical brief with five sections: 1. Patient and care-team information for easy reference during calls 2. What to do next, limited to three specific actions 3. What we know, separating confirmed facts from what remains unclear 4. Medical terms explained in plain English 5. A care log with recent updates and decisions It builds this brief from documents, and context you provide. When research is needed, it uses trusted sources such as the National Cancer Institute and the ClinicalTrials gov API. I use it with ChatGPT/Codex and Claude Code to prepare for conversations and research. It can save the brief locally as a Markdown file or update a shareable Google Doc so the whole family can work from the same information. 📌 Get the free, open-source skill here: https://t.co/YUGoFbhnNc If you find it useful, please ⭐ the repo and share it with someone going through this BS disease.｜@petergyang · 热度 1774 · 2026-08-25

Google Labs: 🚨 NEW LABS EXPERIMENT 🚨 Great minds t̵h̵i̵n̵k̵ play alike! Play with Putty...

🚨 NEW LABS EXPERIMENT 🚨 Great minds t̵h̵i̵n̵k̵ play alike! Play with Putty is a collaborative vibe coding tool that lets you build tools and websites together in real time. 🫟 Ready for your imagination to go multiplayer? Join the waitlist at https://t.co/v4m5iiYt9T and share your feedback with us! US only, ages 18+｜@GoogleLabs · 热度 1491 · 2026-08-25

Thariq: excited to share more on how we’re making Claude Code more hackable soon http...

excited to share more on how we’re making Claude Code more hackable soon https://t.co/liQ5rQG887｜@trq212 · 热度 1266 · 2026-08-25

Cat Wu: Thanks to your feedback, we’ve unified memory across Chat and Cowork

Thanks to your feedback, we’ve unified memory across Chat and Cowork. Now, you can tell Claude to remember something once and it’ll have that context across surfaces! Tell us what you’d like to see next :) https://t.co/0Vw2o4zeaB｜@_catwu · 热度 733 · 2026-08-25

Claude: Everything Claude remembers is saved as a list of topics in Settings, where y...

Everything Claude remembers is saved as a list of topics in Settings, where you can read, edit, or delete each one. Memory also updates on its own as you chat, saving new details as they come up. You can also say "remember this" to save something specific. https://t.co/ExpQAHkyhu｜@claudeai · 热度 615 · 2026-08-25

Guillermo Rauch: We're introducing Run SDK: secure 𝚎𝚟𝚊𝚕 for dynamic Code Mode execution

We're introducing Run SDK: secure 𝚎𝚟𝚊𝚕 for dynamic Code Mode execution. When agents write code, you don't always need a full sandbox. You can 𝚛𝚞𝚗 their code in a lightweight QuickJS secure context. Faster and more cost-efficient. 𝚗𝚙𝚖 𝚒 𝚛𝚞𝚗 https://t.co/br24j8Qp4r https://t.co/ZLqpkwqTT5 https://t.co/6ElyxieLme｜@rauchg · 热度 495 · 2026-08-25

Guillermo Rauch: The hardest problem in building agents is secure connectivity to services and...

The hardest problem in building agents is secure connectivity to services and data. Excited for Vercel Connect to be GA. Run e.g: 𝚟𝚎𝚛𝚌𝚎𝚕 𝚌𝚘𝚗𝚗𝚎𝚌𝚝 𝚌𝚛𝚎𝚊𝚝𝚎 𝚗𝚘𝚝𝚒𝚘𝚗, then get an MCP client you can query on behalf of the authenticated user. https://t.co/UrQxaEDfHx｜@rauchg · 热度 257 · 2026-08-25

Aaron Levie: Good post on what the applied AI strategy looks like at scale

Good post on what the applied AI strategy looks like at scale. It’s clear that there’s a wide gap between the AI models and the underlying workflows of an enterprise, which leaves a ton of opportunity for applied AI companies. “The world doesn’t just want raw models and agents; it wants problems resolved and outcomes achieved. The premium will sit with the companies that can diffuse this intelligence through every aspect of civilization, converting raw tokens into real world outcomes, transforming industries, and creating economies in the process.” This requires understanding the context, driving the change management, having a harness that can route to various models, connecting to the critical business systems in that vertical, solving the UX challenges of connecting users to agents in the right way in a workflow, understanding the evals in the space, and so on. That’s a ton of value that goes beyond just the model intelligence itself. And there’s a window of opportunity right now to build the defining companies that can bring intelligence to the critical domains in an enterprise.｜@levie · 热度 153 · 2026-08-26

Peter Yang: Very excited to attend @OpenAI Dev Day

Very excited to attend @OpenAI Dev Day. ChatGPT / Codex is basically the OS I use for everything I do on my computer. Can't wait to see what updates the team has in store. https://t.co/0mvJVeQqFa https://t.co/rhxpVhn7Ue｜@petergyang · 热度 137 · 2026-08-25

DietrichGebert/ponytail

Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.｜JavaScript · ⭐112482 · Search 增补 · 近快照 +1570星 · 创建 2026-06-12

stablyai/orca

Orca is the ADE for working with a fleet of parallel agents. Run any coding agent with your own subscription. Available on desktop, mobile and VPS.｜TypeScript · ⭐54265 · Search 增补 · 近快照 +679星 · 创建 2026-03-17

Graphify-Labs/graphify

Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local deterministic AST parsing, every edge explained, no vector store.｜Python · ⭐111053 · Search 增补 · 近快照 +939星 · 创建 2026-04-03

zhaoxuya520/reverse-skill

Reverse Engineering / Authorized Penetration Testing / Security Research Skill Router Pack AI-powered routing + On-demand toolchain bootstrapping + Self-evolving knowledge base Supports Claude Code, Kiro, Cursor, Cline, and other AI coding clients 逆向/渗透/安全技能路由包 - AI 自动路由 + 按需自举工具链 + 自动进化经验库 | 支持 Claude Code / Kiro / Cursor / Cline 等代码 AI 客户端｜PowerShell · ⭐29579 · Search 增补 · 近快照 +369星 · 创建 2026-05-13

nexu-io/open-design

🎨 Best DeepSeek Harness Design Plugin. The open-source Claude Design alternative. 🖥️ Local-first desktop app. 🖼️ Your coding agent becomes the design engine: prototypes, landing pages, dashboards, slides, images & video — real files, HTML/PDF/PPTX/MP4 export. 🤖 Claude Code / Codex / Cursor / DeepSeek Harness / OpenCode & 20+ CLIs via BYOK.｜TypeScript · ⭐91779 · Search 增补 · 近快照 +315星 · 创建 2026-04-28

AWS Acquires DuckLabs

GLM-5.3-Flash

https://news.ycombinator.com/item?id=49450353

Qwen3.8-Flash-Next

https://imageat.com/models/qwen-3-8-27b-uncensored

Tim Curry has died

RAG Is Simpler Than You Think

惠普 2026 财年第三财季归母净利润 6.61 亿美元，同比下降 13%

IT之家 8 月 27 日消息，惠普今日发布 2026 财年（2025年11月1日～2026年10月31日）第三财季（2026年5月1日～2026年7月31日）及半年报（2025年11月1日～2026年7月31日）报告： 2026 财年第三财季 营业总收入 ：156.77 亿美元 （IT之家注：现汇率约合 1056.16 亿元人民币） ，同比增长 12.5% 毛利润 ：29.45 亿美元 （现汇率约合 198.41 亿元人民币） 毛利率 ：18.79% 归母净利润 ：6.61 亿美元 （现汇率约合 44.53 ...

开启 iPhone 18 Pro / Ultra 系列新篇章，苹果发布会官宣定档 9 月 10 日

IT之家 8 月 27 日消息，苹果官方宣布将于当地时间 9 月 9 日（北京时间 9 月 10 日凌晨 1 点）举行秋季新品发布会，主题为“惊喜闪耀”（Surprise and shine）。 这将是约翰 · 特努斯（John Ternus）于 9 月 1 日正式接任苹果 CEO 后的首场发布会，预计将推出全新一代 iPhone 18 Pro / Max 以及 Ultra 系列折叠屏新机。后续，苹果还将于明年春季推出 iPhone 18 基础版及 iPhone 18e 等新品。 结合此前爆料，iPhone 18 Pro 系列预计为常规升级，外观几乎不变，主要亮点包括： ...

微软 Win11 新功能曝光：未来将支持手机远程操控关机、重启电脑

IT之家 8 月 26 日消息，Windows Latest 今日在对 Link to Windows 应用（原名“你的手机”）安装包进行拆包分析时发现，微软正在为该应用开发一种涉及远程关机、重启及睡眠的远程控制功能。 相关代码出现在一个标记为 RemoteControlManager 的组件中。一份日期为 2026 年 8 月 25 日的调试日志显示，该功能的启用标志（isRemotePcControlsV2Enabled）当前仍为 false 状态，表明该功能尚未对任何用户开放。 Windows Latest 发现其代码中的字符串资源描绘了一套完整的远程电源控制方案：...

英伟达公布 Vera Rubin 最新进展：推理、网络和定制芯片全面升级

点击查看原文>

DeepSeek 开源 Harness：AI 智能体基础设施开始“拆分”

点击查看原文>

不懂算法的工厂老兵，在 AMD 锐龙 AI Max+ 395上跑出工业智能体

点击查看原文>

Haiku R1/beta6 released

The Root of The Root of All Evil

Merchants of Insecurity

The Move to Python 3 Begins

I stabilized never type

The inside story on why OpenAI agents hacked Hugging Face

The models responsible for last month’s agent hack of Hugging Face had been inadvertently trained to cheat and to communicate with each other, according to an OpenAI technical report released today. The hack, which a group of agents undertook to find solutions for a cybersecurity test that they w...

The Download: the Kids issue arrives, and Bill Gates reveals his AI fears

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Introducing: the Kids issue If the desire to limit kids’ use of technology was once a subcurrent, it has become a raging flood. Countries around the world are ...

Raised on AI

When my oldest child was born, I immediately set up Gmail and Twitter accounts in her name. I broadly announced her birth online and proceeded to plaster her photo across all sorts of platforms. In short, I began creating her digital footprint long before she could stand on her own two feet.  Fas...

Google’s Gemini has a branding problem, and so does the rest of AI

Consumer AI apps need to stop making users learn their product architecture.

How do we explain OpenAI’s executive exodus?

Was Greg Brockman the right executive all along?

Nvidia is about to be a hundred-billion-dollar-a-quarter company

Nvidia's predicting it will pull in $108 billion in revenue within just a few months. It wouldn't be the first company to rake in over $100 billion in quarterly revenue - Amazon, Apple, and Alphabet have repeatedly reached the milestone. Nvidia said in its latest earnings report that it brought i...

OpenAI’s rogue AI model incident was worse than we thought

In July, an unreleased OpenAI model broke out of a restricted environment, figured out how to get access to the internet, allowed AI agents to talk to each other using a secret "message board," and hacked into the internal systems of a different AI lab, Hugging Face. It took nearly two weeks for ...

All the ways Instagram and Facebook are changing for teens

Meta just agreed to make sweeping changes for teens as part of a child safety settlement reached with attorneys general across the US. Under the agreement, Meta must apply new safeguards across Instagram and Facebook that will change how teens interact with the social media platforms, as well as ...

装了啥特别篇：少数派编辑部最近都 vibe 了啥？

如今，vibecoding的话题热度居高不下。但抛开那些被吹上天的一人公司「壮举」，vibecoding能给一般用户的日常使用带来什么帮助呢？本期「装了啥」中，编辑部成员就分享了一些近期折腾出的小玩意 ... 查看全文 本文为会员文章，出自 《单篇文章》 ，订阅后可阅读全文。

《巫师 3：重制版》正式公开：科隆游戏展 2026 消息汇总

编注：本文首发于机核网，少数派经授权转载，仅对文章格式略作调整。科隆游戏展2026开幕夜为我们带来了海量新作公布与续作消息：从《巫师3：重制版》《魔法门之英雄无敌3：重制版》等经典IP重制，到《寂静岭 ... 查看全文

36K stars 的“酷炫组件”，到底该怎么用才不显得用力过猛？

# 36K stars 的“酷炫组件”，到底该怎么用才不显得用力过猛？ 打开一个动画交互组件库，人很容易先被视觉效果抓住：文字有了流动感，卡片有了呼吸感，鼠标经过时页面像在回应你。 这类体验确实迷

从“听得懂”到“干得了”：工业大模型落地工厂的三层进化路线

未来衡量一个工业大模型的关键，也许不再只是“模型有多大、参数有多少”，而是一个更加朴素的问题： 它到底能不能在真实的工厂里，把事情做好。

GLM-5.3-Flash 发布：追平 Opus 4.8 的智力，1/40 的价格，跑在国产芯片上

智谱开源 GLM-5.3-Flash：320B-A18B 原生多模态模型，AA 智能指数追平 Claude Opus 4.8，限时折扣价约其 1/40，推理流量全部跑在国产芯片上。

英伟达：毛利率预计在第四财季触底

英伟达表示，预计毛利率将在2027财年第四季度下降并触底，区间在71%至72%，部分原因在于内存价格。 “我们希望对此坦诚相告，而不是让它作为一个悬而未决的问题持续存在...

亚马逊将从英伟达购买200万颗GPU

亚马逊与英伟达宣布，Amazon Web Services将购买200万颗英伟达GPU，并将采用英伟达名为Vera的新型CPU。 英伟达表示，AWS的部分Vera...

黄仁勋谈AI投资：“唯一的遗憾是没能更早、投入更多”

英伟达已对领先的AI实验室进行了大规模投资，包括今年早些时候以300亿美元收购OpenAI的股份。 黄仁勋表示，在预期的IPO之前，公司本应更加激进。

胰腺癌治疗迎来重大突破！美国FDA批准Revolution新药 患者中位生存期几乎翻倍

来源：财联社 美国食品药品监督管理局（FDA）周三批准了Revolution Medicines突破性胰腺癌药物的上市申请。大型临床试验显示，该药物几乎将晚期胰腺癌患者的中位生存期延长...

惠普Q3同比增长13%，盈利预测超预期，仍难挡股价重挫，市场聚焦PC需求隐忧

来源：华尔街见闻 惠普公布的季度业绩及全年盈利指引均超出分析师预期，但投资者将目光投向更深层的需求隐患，盘后股价随即重挫约10%。

神秘「牛来」模型果然是智谱！GLM首个原生多模态，还用的国产卡

GLM-5.3 Flash

AI视频应用井喷，美图打开新的增长空间

从图像到视频，美图找到了一门更赚钱的生意

小宇宙推出《AI趋势报告》：AI创作、AI办公、协作型AI等成讨论新趋势

AI内容创作者增长187%，相关节目数量增长239%