←搜索结果

按匹配度 · 共 1528 条结果 清除搜索

星期五

  1. Pydantic AI 官方文章

    2026 年 Sentry 替代方案盘点:Pydantic Logfire、GlitchTip 等

    Pydantic 官方文章盘点了 2026 年 Sentry 的替代方案,首推 Pydantic Logfire:它把错误作为 trace 的一部分而非独立事件,不设单独的错误计量,免费计划每月含 1000 万条记录,Team 计划 49 美元。文章称同样每月 100 万条错误在 Sentry 公开档位下约需 225 美元。其余选项包括兼容 Sentry SDK 协议的自托管 GlitchTip、轻量单机 Bugsink、捆绑 uptime 与 cron 监控的 Honeybadger,以及含每月 10 万条异常免费额度的 PostHog。

    your actual volume: capacity planning is the price of the low rates. 7. SigNoz: best open-source full observability SigNoz is the open-source path to the same conclusion Logfire draws: errors belong next to traces, logs, and metrics, on OpenTelemetry. The correlation is there; the cost is that you operate the platform yourself, ClickHouse included, and query it in its query builder or ClickHouse SQL rather than PostgreSQL. Easier to run than self-hosted Sentry, but not free of ops. 8. Rollbar: b

    匹配字段正文

  2. NVIDIA Developer · Generative AI

    NVIDIA 如何用生成式推荐模型重塑大规模 RecSys

    NVIDIA 的 recsys-examples 仓库为 HSTU 和 Semantic ID 两类生成式推荐模型提供生产级训练与推理实现,并集成 DynamicEmb 动态嵌入表与 Megatron-Core 分布式并行。HSTU 推理结合 PyTorch AOTInductor、Triton Inference Server、nv-embedding-cache 和 FlexKV KV cache,在全 GPU 缓存命中场景下比 Python 后端最高提速 2.38x;Semantic ID 专用推理框架在 H100 上相比 SGLang beam search 离线延迟降低 2.14x 至 2.27x、在线吞吐提升 1.85x。

    ility distribution of next action or item conditioned on a sequence of user history: The shift toward more homogeneous, transformer-like architectures can better leverage scaling laws, potentially unify retrieval and ranking within a single model, and more naturally integrate with the rapidly evolving LLM ecosystem. The two prevalent approaches that implement this objective in recommenders are Hierarchical Sequential Transduction Units (HSTU) and Semantic IDs: HSTU, introduced by Meta in 2024, i

    匹配字段正文

星期四

  1. The Register

    Slack 推出 Slack Code,把编码智能体放进项目频道

    Slack 推出 Slack Code,为编码智能体开设专属项目频道,用户可在 Slack 对话中 @ 智能体,让它创建代码频道并跟进任务。在频道里,成员可以查看智能体的操作、审阅其提出的改动、查看实时 HTML 预览,并在需要时让它改变方向;任务完成后频道可自行归档,留下可搜索的记录。Anthropic、Cognition、GitHub、OpenAI 的 ChatGPT 和 Vercel 等正在为 Slack Code 开发集成,Claude、Devin、Copilot 和 ChatGPT 等智能体将进入代码频道;Slack 还推出 Agent DM、Agents 标签页和 Add to Slack 功能。

    tonking Stingray' beta swims out: Deeper Rust 'oxidization' plus LLM speech-to-text GNOME 51 is here, and Ubuntu interim releases are getting interesting again; Revive Raskin's 3 laws of software with humane work that considers a dev's duty of care All this, and angry NeoVim users; Firefox 157: Could bold new look be just what Mozilla needs? Yet another new version of Firefox, with a significant reskin; KDE turns 30 and someone's brought an AI-native desktop proposal Akademy talk imagines Plasma

    匹配字段正文

  2. The Rundown AI

    Anthropic Claude 将蛋白质设计纳入能力范围

    Anthropic 的 Claude 基本自主完成蛋白质设计流程,设计结果经实验室验证,在 15 个靶点中的 14 个产生有效分子。Mythos Preview 和 Opus 4.8 仅凭一份专家提示词、互联网与工具开展任务,分子结合成功率达 22%–35%,高于 Anthropic 所称的行业典型水平 10%–15%;实际合成和测量由 Twist Bioscience 与 Adaptyv Bio 完成。

    Codex to codify ambiguous work into scripts Pro tip: Turn this self-improvement process into a reusable skill, then run it whenever a workflow becomes inconsistent. The Rundown : AWS re:Invent brings 2,200+ sessions to Las Vegas, and 70% are hands-on. Join builders’ sessions, workshops, code talks and more to learn by doing, not just through sessions. While onsite, you can : Bring edge-case questions to engineers and partners building on AWS Test services against realistic workloads in sandbox

    匹配字段正文

  3. ZDNET · Artificial Intelligence

    2025 年最值得购买的 Android 手机推荐

    ZDNET 测试了三星 Galaxy Z Fold 8、Google Pixel 11 系列等最新 Android 手机,并将起售价 1050 美元的 Samsung Galaxy S26 Ultra 选为综合最佳,其搭载 Snapdragon 8 Elite Gen 5 芯片与 Privacy Display 隐私屏。亚军为 425 美元的 Google Pixel 10a,提供七年系统与安全更新。本次更新新增 Pixel 11 Pro 系列、Galaxy Z Fold 8 和 Motorola Razr Ultra(2026)。

    op the list for our benchmark tests. The latest model, the Galaxy S26 Ultra, delivers in almost every way, from general performance to display technology to camera quality, with flaws that you really have to look closely to uncover. The standout feature of the Galaxy S26 Ultra is its Privacy Display, which dims half the screen’s pixels, so it appears tinted when viewed at an off angle. This complex display tech can be customized to hide important information, such as 2FA message alerts and passw

    匹配字段正文

星期三

  1. MIT News · Artificial Intelligence精选

    MIT 研究发现生成式 AI 图像往往无法追溯至具体训练数据

    MIT CSAIL 研究发现,生成模型的训练数据越多,单个样本对特定输出的影响越小,这一现象被称为“归因衰减”。团队构建“扩散模型集成”架构,无需重新训练即可精确移除样本影响,并在来自七个公开数据集、规模从 256 张到超过 160,000 张图像的实验中观察到反幂律收缩趋势。该开放获取论文发表于 Nature Communications。

    Press Contact: Previous image Next image When an artificial intelligence image generator produces a portrait, whose work went into it? The question sits at the center of lawsuits, licensing deals, and proposed regulations worldwide. Artists want credit. Companies want clarity. Policymakers want a way to assign responsibility. New work from a team of researchers at MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) suggests that for models trained on large datasets, the questio

    匹配字段正文

星期二

  1. MIT News · Artificial Intelligence

    问答:MIT 教授 Eugene Fitzgerald 重新思考创新如何发生

    MIT 教授 Eugene Fitzgerald 提出,创新源于个人与企业构成的去中心化集体智慧,在不确定性中创造市场惊喜,而非可被简单控制的线性过程。他以应变硅从 Bell Labs、MIT、创业公司走向行业应用为例,强调创新需要市场应用、技术与实施相互融合。

    Merton C. Flemings SMA Professor in MIT’s Department of Materials Science and Engineering, in his latest book, “The Invisible Engine: Why Innovation Evades Control.” Fitzgerald draws on a decade of work leading international research programs, including the MIT and Masdar Institute Cooperative Program and the MIT-Singapore Alliance for Research and Technology, where he explored innovation as the integration of market applications, technology, and implementation. Written at a moment when artific

    匹配字段正文

  2. Replit Blog

    Replit 上线黑盒渗透测试扫描

    Replit 推出黑盒渗透测试扫描,以外部攻击者视角通过网络和浏览器测试应用,无需访问源码,与已有的白盒代码扫描互补。扫描在私有沙箱中针对应用副本运行,先以未登录访客身份探测,再以普通登录用户身份检查越权访问,最后给出可确认的问题清单供 Replit Agent 修复。Replit 表示两种扫描的发现很少重叠,例如白盒扫描发现权限撤销后仍可操作,黑盒扫描发现管理后台地址可猜且无需登录。该功能加入 Replit Auto-Protect,平台现有 3 级临时安全扫描,Level 3 同时运行白盒与黑盒扫描。

    Black-box pen tests on Replit — cover image

    ng a pen test from a vendor that would cost thousands, and weeks of back-and-forth. A penetration (”pen”) test is a safe, authorized mock cyberattack that helps you fix security flaws before bad hackers find them. As we have democratized software creation, we have also sought to democratize securing it. A new generation of creators needs to ensure their apps are hardened against attackers, who look at apps built with AI as an opportunity as a potential opportunity. Our existing scans have an age

    匹配字段正文

星期一

  1. The Register

    Mozilla 为 iOS 版 Firefox 加入广告拦截功能

    Mozilla 正通过渐进式推送为 iOS 版 Firefox 引入可选广告拦截功能,默认关闭,基于 EasyList 过滤列表在广告加载前拦截第三方广告网络、广告追踪器、网站自带广告以及弹窗和浮层。该功能不拦截搜索结果中的广告(包括 Google 和 Bing),也不拦截 Firefox 主页或新标签页上的“赞助内容”。

    tonking Stingray' beta swims out: Deeper Rust 'oxidization' plus LLM speech-to-text GNOME 51 is here, and Ubuntu interim releases are getting interesting again; Revive Raskin's 3 laws of software with humane work that considers a dev's duty of care All this, and angry NeoVim users; Firefox 157: Could bold new look be just what Mozilla needs? Yet another new version of Firefox, with a significant reskin; KDE turns 30 and someone's brought an AI-native desktop proposal Akademy talk imagines Plasma

    匹配字段正文

  2. The Register

    微软将于 9 月 14 日停用 Excel 的 COPILOT() 函数

    微软将于 2026 年 9 月 14 日停用 Excel 的 COPILOT() 函数,该函数 2025 年 8 月才在预览版亮相,原计划 2027 年正式可用。微软表示将改由 Copilot 侧边栏提供摘要文本、分类数据、生成内容和检索网络信息等同类 AI 能力,并称此举是为简化 Excel 内的 Copilot 体验。

    tonking Stingray' beta swims out: Deeper Rust 'oxidization' plus LLM speech-to-text GNOME 51 is here, and Ubuntu interim releases are getting interesting again; Revive Raskin's 3 laws of software with humane work that considers a dev's duty of care All this, and angry NeoVim users; Firefox 157: Could bold new look be just what Mozilla needs? Yet another new version of Firefox, with a significant reskin; KDE turns 30 and someone's brought an AI-native desktop proposal Akademy talk imagines Plasma

    匹配字段正文

  3. The Register

    微软称 AI 找 bug 拖慢 Exchange SE CU1 发布,无法给出日期

    微软 Exchange 团队确认,Exchange Server Subscription Edition 的 Cumulative Update 1(CU1)因 AI 工具发现大量漏洞而延期,目前无法给出发布日期。该更新原计划 2026 年上半年发布,后改为下半年;团队称需等到出现一个没有紧急安全补丁的月份才会发布 CU1。

    pt launchpads Redmond’s missive also points to Microsoft’s pledge to “prioritize security above all else” as a reason for delays. A reminder: Microsoft adopted that stance after flaws in Exchange led to an attack on Exchange by suspected Chinese operatives, earning it a tongue-lashing from the US government. The Exchange team says that while trying to stay on top of bugs, it is also working on CU1. “We are regularly rolling our monthly security payload into our internal CU1 build and plan to rel

    匹配字段正文

星期六

  1. Augment Code 博客

    Augment Code 重建 Auggie CLI 架构,任务成本比 Claude Code 低 53%

    Augment Code 宣布重建 Auggie CLI 的 harness,在 SWE-bench Pro 上相同通过率下,Auggie v2 完成任务成本为 $1.27,比 Claude Code 的 $2.70 低 53%。重建基于开源项目 Pi,通过精简工具面、用检索代替探索和主动压缩上下文实现效率提升,CLI 将于 9 月或 10 月推出。

    How we rebuilt the Auggie CLI harness to make tasks 53% cheaper than Claude Code

    inal, so every rendering bug became a debugging session through a React reconciler and a flexbox engine. And there were limits we never engineered around: the whole-transcript redraws flickered, and layout occasionally came out wrong. The AI SDK and Mastra prototypes moved to OpenTUI, which maintains a virtual terminal buffer and keeps it in sync with the real one. That fixes flicker, but you give up the native buffer along the way: the transcript is no longer plain text in your terminal's scrol

    匹配字段正文

星期五

  1. Hugging Face

    Hugging Face 发布 2026 夏季开源模型观察:Qwen 衍生模型达 15 万个

    Hugging Face 发布 2026 夏季开源模型观察报告,基于 Hub 前七个月数据指出点赞与下载记录的是两种不同行为:今年发布的模型无一进入下载量前 25,而 25 个中有 13 个来自 2022 年,all-MiniLM-L6-v2 七个月被下载 15.5 亿次。报告称 Qwen 衍生模型已达 151,448 个,是 Meta 总规模的 2.6 倍、Llama 仓库的 4.7 倍,前七个月以每天约 180 至 210 个新仓库增长。报告还提到 7 月 Claude Code 占智能体流量的 44.4%,Codex 从 10.4% 升至 20.8%,近四分之一智能体流量来自尚未登记的工具。

    rt, the geographical rebalancing of power continues to accelerate. While U.S. open source models continue to be competitive, the race between several Chinese frontier model labs draws strong attention. Many likes on these frontier models point to what excites the community the most, and growth opportunity for companies leveraging the attention for valuations. However, the AI race is not only sprints, but also a marathon; tools like llama.cpp helps deploying the big models locally, but a broad mo

    匹配字段正文

  2. Augment Code 博客

    Augment Code 如何优化 PR 到合并的完整循环

    Augment Code 将 Cosmos 系统从代码审查扩展到 PR 到合并的完整循环,由专用智能体处理机械工作、验证正确性并响应反馈,人类保留最终合并决策权。此前该系统已帮助其工程组织将代码产出提升 3 倍,同时缩短中位合并时间。

    e company’s AI Code Review agent, bringing research-grade program analysis and modern GenAI techniques together to automate one of the most time-consuming parts of the SDLC. Before Augment, Akshay spent several years at Uber advancing automated code review and repair systems, and conducted research across AWS’s Automated Reasoning Group, Google’s Android Static Analysis team, and UCLA. His work sits at the intersection of AI, software engineering, and programming-language theory. Related reading

    匹配字段正文

  3. Hugging Face

    使用 Strands Agents、LeRobot 与 Hugging Face Storage Buckets 在一处完成录制、训练和部署

    Strands Robots 将 LeRobotDataset 的录制、同步、流式训练和实体机器人部署整合到同一 Robot() 工作流,并让新演示数据返回同一 Storage Bucket。Xet 按字节去重,仅上传变化的数据块;stream_dataset() 可从 Hub 逐帧读取远程 Parquet 与 MP4 分片,无需完整下载。Bucket 仅支持流式读取且不保留修订记录,真实部署还需隔离数据与凭据、限制智能体权限并只加载可信组织的模型。

    mulation path) Python 3.12+, on Linux or macOS (Apple Silicon supported for the MuJoCo backend). A Strands-compatible model provider for the agent's reasoning. Amazon Bedrock with AWS credentials, the Anthropic API , OpenAI, or Ollama running locally. Strands Robots with the dataset extras: uv pip install -U "strands-robots[sim-mujoco,lerobot] =0.5.1" . The lerobot extra pulls in LeRobot ( =0.6.1), datasets , av , and torchcodec , so recording and video decode both work without further setup. Re

    匹配字段正文

  4. Figma AI / Agent / Make

    Figma 上线技能创作功能,并分享 10 个社区精选技能

    Figma 现在支持用户直接在 Figma 中为设计智能体编写技能(skills),即纯文本 markdown 文件,可让智能体按指定步骤或设计稿参考执行任务,并支持预览、迭代和手动编辑 markdown。部分技能也可用于第三方智能体中的 Figma MCP server,相关技能已发布在 GitHub 仓库。Figma 还从社区精选了 10 个技能,包括 /xray-scan、/find-animation-opportunities、/camera-pan-and-zoom、/better-interface、/component-handoff 和 /create-color 等。

    Abstract geometric artwork on a dark red background featuring bold orange, green, black, blue, yellow, and magenta shapes, including a curved ribbon, flower-like form, stacked circles, and a striped ring.

    over grids, color theory, image performance, and motion design. He treats screen recording videos like a fine art and documents design tips on social media. In his free time, he draws on coffee cups with Copic markers at cafes, and collects a wide range of synthesizers and drum machines. Related articles Got skills? Make the Figma agent a better collaboratorJuly 1, 2026By Sachi Shah, Sean LeeFrom sharing your best thinking to learning from your team’s go-to prompts, here’s what skills unlock in

    匹配字段正文

星期四

  1. Ben's Bites

    Grok Bot 上手体验:比 Codex 和 Claude 更简单的个人智能体

    Ben's Bites 作者在获得 Grok Bot 早期访问后实际使用,认为它比 Codex 或 Claude 更简单,体验接近在 Slack 里给队友发消息。用户可以为每个智能体设定系统提示词、让它们互相发消息、观看它们使用各自的虚拟电脑,并连接 Gmail、Calendar、Slack 等账号,作者在连接账号时遇到几次问题。目前只有 Cursor 或 Grok 的 200 美元套餐用户可以使用,Grok Bot 由新发布的 Grok 4.6 驱动,作者称 Grok 4.7 将在几周后推出。

    ps://substackcdn.com/image/fetch/$s !v7TF!,w 1028,c limit,f auto,q auto:best,fl progressive:steep/l play button usfui2,w 88,e colorize:0/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F ss-rehost tw-video-preview-13 2087040760793407488.jpg","link url":"https://t.co/vmeyicBwTk"}],"quoted tweet":{},"reply count":15,"retweet count":23,"like count":333,"impression count":104095,"expanded url":null,"video url":"https://video.twimg.com/amplify video/2087040760793407488/vid/avc1/

    匹配字段正文

  2. 張小珺 Xiaojùn

    NVIDIA 研究副总裁 Ming‑Yu Liu 谈 Cosmos 3、世界模型与物理世界 AI

    NVIDIA 研究副总裁兼 Cosmos 世界模型项目负责人 Ming‑Yu Liu 深入讨论了近期发布的 Cosmos 3、世界模型及物理世界 AI。访谈还涉及 NVIDIA 对竞争与组织理念的看法,以及 Jensen Huang,并讲述了 Liu 坚持武术训练的个人经历。

    th to help his body recover. With twenty years in the United States and ten years at NVIDIA, Liu still articulates technical ideas through simple Eastern wisdom. For example, he draws analogies between large models and traditional Chinese medicine. He also often quotes The Art of War — “To prepare everywhere is to be weak everywhere” — when contemplating the technical trends of world models. In this episode, we dive deep into the Cosmos series — especially the recently released Cosmos 3 — world‑

    匹配字段正文

  3. Hugging Face精选

    复现 ICML 2,200 篇论文后,我们学到了什么

    Hugging Face 与 alphaXiv 组织社区借助编码智能体复现了 2,226 篇 ICML 2026 论文,并公开实验日志、代码和产物。受检论文中,51% 至少有一项主张获独立验证,23% 至少有一项被证伪或存在争议;结果也显示,智能体会误判尺度效应或单位问题,可靠审查仍需人类引导。

    eflating perplexity roughly threefold. The abstract's "3.1% quality cost for 50% cache reduction" becomes roughly 9.4% once corrected. 🚨 False falsifications. Sometimes we found flaws in an attempted reproduction. One logboook claimed dramatically, "the paper's method is 2x slower than the baseline"; this turned out to be an arithmetic bug in the reproduction : per-trajectory time compared against per-batch-of-50 time. Correctly normalized, the participant's own data confirms the paper's claimed

    匹配字段正文

  4. The Register

    OpenWALDO 启动开源 AI 训练数据集项目,目前含 1673 亿 token

    CentOS 和 Rocky Linux 创始人 Gregory Kurtzer 发起 OpenWALDO(Open Weights, Artifacts, Licenses, Data, Origins)项目,由 CIQ 资助,目标是建立可公开贡献、可审计的开源 AI 训练数据集。目前该数据集包含 167.3 亿参考 token,来源包括政府记录、开源学术论文、邮件列表和公共领域文献,但相比前沿模型使用的数十万亿 token 仍微不足道。项目方称企业可将该语料及其物料清单作为经验证的基线,再叠加自有专有数据训练模型。

    tonking Stingray' beta swims out: Deeper Rust 'oxidization' plus LLM speech-to-text GNOME 51 is here, and Ubuntu interim releases are getting interesting again; Revive Raskin's 3 laws of software with humane work that considers a dev's duty of care All this, and angry NeoVim users; Firefox 157: Could bold new look be just what Mozilla needs? Yet another new version of Firefox, with a significant reskin; KDE turns 30 and someone's brought an AI-native desktop proposal Akademy talk imagines Plasma

    匹配字段正文