Microsoft Scout is a new AI personal assistant built on OpenClaw
The Verge AI53 min agoNew ToolMicrosoft is launching Scout, an always-on AI personal assistant built on OpenClaw. It integrates with Microsoft 365 apps including Outlook, OneDrive, and Microsoft Teams, enabling businesses to assign virtual assistants to employees. Mentioned tasks include calendar organization, expense reporting, and drafting emails, while the supplied excerpt does not fully explain how Scout differs from Copilot.
OpenAI launches new Codex tools for white-collar work
TechCrunch AI2h agoNew ToolOpenAI released new Codex capabilities intended to broaden the agentic tool's workplace uses and strengthen its appeal to enterprise customers. The company also published an internal report about how Codex is used for knowledge work. The provided excerpt does not specify the individual features or the report's detailed findings.
Anthropic scales Claude Mythos to critical infrastructure in 15+ countries★ 72
TechCrunch AI4h agoBusinessAnthropic is expanding its Project Glasswing security vulnerability program and access to Mythos. The rollout covers 150 organizations across 15 countries, focusing on power, water, healthcare, and communications infrastructure. The company is targeting sectors where a cyberattack could affect as many as 100 million people, although implementation details and participating organizations were not disclosed in the provided text.
Does your CEO have AI psychosis? Aaron Levie thinks most of them do.
TechCrunch AI4d agoCommentaryBox founder Aaron Levie calls some executive thinking around AI replacement “AI psychosis.” He argues that the people deciding AI can replace workers are often the least likely to understand what those jobs truly involve. The article frames this against ClickUp cutting 22% of staff for AI agents and 2026 tech layoffs nearly matching all of 2025.
Dcard Launches EntryDesk and VibeHost for Agent-Native Enterprise AI Workflows
INSIDE 硬塞 AI5d agoBusinessDcard introduced EntryDesk and VibeHost, products aimed at helping companies move toward Agent-Native operations. The first wave supports both cloud and on-premises deployment, with integration into internal enterprise systems. The article says Dcard’s method shortened process time by over 80%, but the provided text does not include detailed case data, pricing, or technical architecture.
Vertu wants CEOs to run companies from an AI foldable starting at $6,880
TechCrunch AI5d agoHardwareVertu has introduced a luxury AI foldable phone starting at $6,880, aimed at executives and CEOs. Built on the open-source Hermes project, it combines AI-agent workflows, enterprise integrations, and ultra-premium finishes. The available summary positions it as a high-end mobile business control hub, but does not specify supported enterprise platforms, model providers, hardware specs, or concrete agent capabilities.
Business Owners, Do You Really Know How to Use AI?
INSIDE 硬塞 AI6d agoOpinionThe article argues that many companies use AI mainly to improve efficiency, without creating meaningful revenue or strategic advantage. It proposes distributed AI, placing intelligence closer to where data is generated to reduce latency and support faster decisions. The key message is that firms should balance centralized and distributed architectures to strengthen competitiveness while preserving greater control over data and digital sovereignty.
Granite 4.1 LLMs: How They’re Built★ 75
Hugging Face Blog34d agoRelease本文深入探討 IBM 最新開源的 Granite 4.1 大語言模型家族。詳細介紹了其從數據清洗、模型架構設計（如優化的 Transformer 結構）到指令微調與安全對齊的完整構建流程。Granite 4.1 延續了 IBM 對於企業級安全與 Apache 2.0 開源協議的承諾，並在代碼生成、工具調用及多語言推理上展現出優異性能。
Granite 4.0 3B Vision: Compact Multimodal Intelligence for Enterprise Documents★ 75
Hugging Face Blog63d agoReleaseIBM 於 Hugging Face 發布全新 Granite 4.0 3B Vision 模型。這款僅有 30 億參數的輕量級多模態模型，專為企業級文件理解、圖表分析與 OCR 數據提取而設計。其小巧的體積不僅大幅降低了部署門檻與運算成本，更在處理複雜商業報表與 PDF 文件時展現出極高的實用性，是企業本地化部署的理想選擇。
IBM and UC Berkeley Diagnose Why Enterprise Agents Fail Using IT-Bench and MAST★ 80
Hugging Face Blog104d agoReleaseIBM 研究中心與柏克萊加州大學（UC Berkeley）合作發表了 IT-Bench 基準測試與 MAST 診斷框架。IT-Bench 模擬了真實的企業 IT 運維環境，而 MAST 則專門用來剖析 AI Agent 在執行多步驟任務時失敗的深層原因。研究指出，企業級 Agent 的失敗往往源於工具調用錯誤、狀態追蹤失效及錯誤累積，而非單純的 LLM 能力不足，這為未來 AIOps 的優化提供了明確方向。
Dell Enterprise Hub is all you need to build AI on premises★ 75
Hugging Face Blog375d agoReleaseHugging Face 與 Dell 合作推出 Dell Enterprise Hub，旨在簡化企業在本地端（on-premises）部署 AI 模型的流程。該平台整合了 Dell 的硬體優勢與 Hugging Face 的豐富模型庫，提供經優化的容器與自動化工作流。這讓企業能在確保數據安全與合規的前提下，輕鬆在自家伺服器上運行 Llama、Mistral 等主流開源模型。
Making AI Work: Leadership, Lab, and Crowd★ 80
One Useful Thing (Mollick)376d agoOpinion知名 AI 學者 Ethan Mollick 指出，企業要成功導入 AI 必須克服組織慣性。他提出「領導層、實驗室與群眾」三維度框架：領導層負責消除員工恐懼並制定安全規範；實驗室（Lab）專注於前沿技術研發與客製化工具；群眾（Crowd）則由基層員工發起，透過日常實踐找出真正有用的 AI 應用場景。唯有三者協同，才能打破 AI 導入的瓶頸。
Microsoft and Hugging Face expand collaboration★ 75
Hugging Face Blog379d agoBusinessMicrosoft 與 Hugging Face 宣布深化合作，將 Hugging Face 的開源模型庫與 Azure AI Foundry（前身為 Azure AI Studio）進行深度整合。開發者現在可以直接在 Azure 安全且具備合規性的企業級環境中，輕鬆搜尋、部署與微調數十萬個 Hugging Face 的開源模型，大幅簡化企業導入開源 AI 的工作流程。
Banque des Territoires (CDC Group) x Polyconseil x Hugging Face: Enhancing a Major French Environmental Program with a Sovereign Data Solution
Hugging Face Blog693d agoBusiness法國 Banque des Territoires（CDC 集團旗下）與技術夥伴 Polyconseil 及 Hugging Face 合作，針對其重大環境與生態轉型計畫推出「主權數據解決方案」。該方案旨在確保敏感的國土與環境數據在處理時符合歐洲嚴格的隱私法規（如 GDPR）。透過 Hugging Face 的開源模型生態系統與本地化部署，該計畫成功在不依賴非歐盟雲端服務的前提下，利用先進 AI 進行大規模文件分析與決策輔助，為公部門的 AI 主權化樹立了典範。
Leveraging Hugging Face for complex generative AI use cases
Hugging Face Blog1067d agoBusiness本案例研究探討了企業級生成式 AI 平台 Writer 如何與 Hugging Face 深度合作。Writer 透過 Hugging Face 的 Transformers、Accelerate 以及 Text Generation Inference (TGI) 等技術，成功開發並部署了專為企業設計的 Palmyra 系列模型。這不僅大幅降低了模型訓練與推論的成本，更在確保數據隱私與合規性的前提下，為企業客戶提供了高度客製化的 AI 解決方案。
Snorkel AI x Hugging Face: unlock foundation models for enterprises
Hugging Face Blog1153d agoBusinessSnorkel AI 與 Hugging Face 宣布合作，旨在解決企業應用基礎模型（Foundation Models）時面臨的領域數據不足挑戰。透過將 Snorkel Flow 的程式化標註技術與 Hugging Face 的開源模型庫結合，企業能快速標註私有數據並進行模型微調。此方案不僅能加速企業級 AI 應用的開發，還能透過模型蒸餾技術降低推理成本，實現高效且安全的落地部署。
Director of Machine Learning Insights
Hugging Face Blog1497d agoOpinion本報告源自 Hugging Face 針對多位企業機器學習主管（Director of ML）的調查與訪談。內容指出，將模型從實驗室原型轉化為穩定生產服務（Deployment Gap）仍是企業最大痛點。主管們強調了 MLOps 工具鏈整合、開源模型在企業級應用的崛起，以及跨團隊（數據科學與軟體工程）協作與人才招募的持續挑戰。