https://www.sipply.dev/story/anthropic-launches-reporting-tools-to-clarify-ai-spending-1ffbc4c3 Sipply en 2026-10-28T19:19:58.000Z Anthropic launches reporting tools to clarify AI spending for enterprises https://www.sipply.dev/zh/story/anthropic-launches-reporting-tools-to-clarify-ai-spending-1ffbc4c3 Sipply zh-cn 2026-10-28T19:19:58.000Z Anthropic推出报告工具，明确企业的AI支出 https://www.sipply.dev/story/wizs-cto-warns-of-impending-vulnpocalypse-due-to-critical-a15e1377 Sipply en 2026-10-28T18:50:39.000Z Wiz's CTO warns of impending 'vulnpocalypse' due to critical PHP vulnerabilities https://www.sipply.dev/zh/story/wizs-cto-warns-of-impending-vulnpocalypse-due-to-critical-a15e1377 Sipply zh-cn 2026-10-28T18:50:39.000Z Wiz首席技术官警告，由于关键PHP漏洞，‘漏洞末日’迫在眉睫 https://www.sipply.dev/story/microsoft-discontinues-copilot-support-for-smart-tvs-from-lg-e8c79d42 Sipply en 2026-10-28T18:36:26.000Z Microsoft discontinues Copilot support for smart TVs from LG and Samsung https://www.sipply.dev/zh/story/microsoft-discontinues-copilot-support-for-smart-tvs-from-lg-e8c79d42 Sipply zh-cn 2026-10-28T18:36:26.000Z 微软停止对LG和三星智能电视的Copilot支持 https://www.sipply.dev/story/google-deepmind-unveils-gemini-4-carbon-its-most-advanced-ai-6bac2db4 Sipply en 2026-10-28T17:00:58.000Z Google DeepMind unveils Gemini 4 Carbon, its most advanced AI model for cybersecurity https://www.sipply.dev/zh/story/google-deepmind-unveils-gemini-4-carbon-its-most-advanced-ai-6bac2db4 Sipply zh-cn 2026-10-28T17:00:58.000Z Google DeepMind 发布 Gemini 4 Carbon，这是其最先进的网络安全 AI 模型 https://www.sipply.dev/story/google-launches-connected-apps-platform-for-gemini-and-38aa53e9 Sipply en 2026-10-28T17:00:00.000Z Google launches Connected Apps Platform for Gemini and Search integration https://www.sipply.dev/zh/story/google-launches-connected-apps-platform-for-gemini-and-38aa53e9 Sipply zh-cn 2026-10-28T17:00:00.000Z 谷歌推出Connected Apps Platform以实现Gemini和搜索集成 https://www.sipply.dev/story/chromes-autofill-feature-now-uses-gmail-data-to-streamline-6237ee65 Sipply en 2026-10-28T17:00:00.000Z Chrome's autofill feature now uses Gmail data to streamline form filling https://www.sipply.dev/zh/story/chromes-autofill-feature-now-uses-gmail-data-to-streamline-6237ee65 Sipply zh-cn 2026-10-28T17:00:00.000Z Chrome的自动填充功能现在使用Gmail数据来简化表单填写 https://www.sipply.dev/story/meta-likely-to-postpone-new-business-terms-amid-publisher-2f3d6079 Sipply en 2026-10-28T16:15:30.000Z Meta likely to postpone new business terms amid publisher backlash over AI use https://www.sipply.dev/zh/story/meta-likely-to-postpone-new-business-terms-amid-publisher-2f3d6079 Sipply zh-cn 2026-10-28T16:15:30.000Z Meta可能会因出版商对AI使用的反对而推迟新商业条款 https://www.sipply.dev/story/marissa-mayers-ai-app-dazzle-criticized-for-disconnect-with-663d561b Sipply en 2026-10-28T15:53:52.000Z Marissa Mayer's AI app Dazzle criticized for disconnect with public needs https://www.sipply.dev/zh/story/marissa-mayers-ai-app-dazzle-criticized-for-disconnect-with-663d561b Sipply zh-cn 2026-10-28T15:53:52.000Z Marissa Mayer的AI应用Dazzle因与公众需求脱节而受到批评 https://www.sipply.dev/story/south-koreas-financial-sector-faces-cyberattacks-linked-to-0b4339a9 Sipply en 2026-10-28T15:41:18.000Z South Korea's financial sector faces cyberattacks linked to suspected China-based hacker https://www.sipply.dev/zh/story/south-koreas-financial-sector-faces-cyberattacks-linked-to-0b4339a9 Sipply zh-cn 2026-10-28T15:41:18.000Z 韩国金融部门面临与疑似中国黑客相关的网络攻击 https://www.sipply.dev/story/nvidia-launches-nemotron-3-5-transcribe-an-open-8a555541 Sipply en 2026-10-28T15:00:00.000Z Nvidia launches Nemotron 3.5 Transcribe, an open transcription model for 40 languages https://www.sipply.dev/zh/story/nvidia-launches-nemotron-3-5-transcribe-an-open-8a555541 Sipply zh-cn 2026-10-28T15:00:00.000Z Nvidia推出Nemotron 3.5 Transcribe，支持40种语言的开放转录模型 https://www.sipply.dev/story/nvidia-launches-nemotron-3-5-transcribe-model-for-advanced-9acc9684 Sipply en 2026-10-28T14:58:17.000Z NVIDIA launches Nemotron 3.5 Transcribe model for advanced multilingual transcription https://www.sipply.dev/zh/story/nvidia-launches-nemotron-3-5-transcribe-model-for-advanced-9acc9684 Sipply zh-cn 2026-10-28T14:58:17.000Z NVIDIA推出Nemotron 3.5 Transcribe模型，支持先进的多语言转录 https://www.sipply.dev/story/meta-launches-waddle-world-an-ai-pet-journal-app-for-virtual-992c0a9d Sipply en 2026-10-28T11:27:30.000Z Meta launches Waddle World, an AI pet journal app for virtual creature interaction https://www.sipply.dev/zh/story/meta-launches-waddle-world-an-ai-pet-journal-app-for-virtual-992c0a9d Sipply zh-cn 2026-10-28T11:27:30.000Z Meta推出Waddle World，一款用于虚拟生物互动的AI宠物日记应用 https://www.sipply.dev/story/ai-integration-strategies-can-enhance-workplace-f95b3d31 Sipply en 2026-10-28T10:00:00.000Z AI integration strategies can enhance workplace collaboration and efficiency https://www.sipply.dev/zh/story/ai-integration-strategies-can-enhance-workplace-f95b3d31 Sipply zh-cn 2026-10-28T10:00:00.000Z AI整合策略可以增强工作场所的协作和效率 https://www.sipply.dev/story/legora-develops-agentic-operating-system-to-enhance-legal-e746af40 Sipply en 2026-10-28T09:30:00.000Z Legora develops agentic operating system to enhance legal workflows https://www.sipply.dev/zh/story/legora-develops-agentic-operating-system-to-enhance-legal-e746af40 Sipply zh-cn 2026-10-28T09:30:00.000Z Legora开发代理操作系统以增强法律工作流程 https://www.sipply.dev/story/new-framework-achieves-99-98-success-in-autonomous-rocket-b5a0dc55 Sipply en 2026-10-28T04:00:00.000Z New framework achieves 99.98% success in autonomous rocket landings https://www.sipply.dev/zh/story/new-framework-achieves-99-98-success-in-autonomous-rocket-b5a0dc55 Sipply zh-cn 2026-10-28T04:00:00.000Z 新框架实现99.98%的自主火箭着陆成功率 https://www.sipply.dev/story/new-method-improves-resolution-of-knowledge-conflicts-in-181bfe49 Sipply en 2026-10-28T04:00:00.000Z New method improves resolution of knowledge conflicts in language models https://www.sipply.dev/zh/story/new-method-improves-resolution-of-knowledge-conflicts-in-181bfe49 Sipply zh-cn 2026-10-28T04:00:00.000Z 新方法改善语言模型中知识冲突的解决 https://www.sipply.dev/story/new-research-shows-simple-agentic-loops-can-outperform-571af78f Sipply en 2026-10-28T04:00:00.000Z New research shows simple agentic loops can outperform complex theorem proving methods https://www.sipply.dev/zh/story/new-research-shows-simple-agentic-loops-can-outperform-571af78f Sipply zh-cn 2026-10-28T04:00:00.000Z 新研究表明简单的代理循环可以超越复杂的定理证明方法 https://www.sipply.dev/story/cohortflow-introduces-a-steerable-agent-for-efficient-d16f956b Sipply en 2026-10-28T04:00:00.000Z CohortFlow introduces a steerable agent for efficient clinical cohort curation https://www.sipply.dev/zh/story/cohortflow-introduces-a-steerable-agent-for-efficient-d16f956b Sipply zh-cn 2026-10-28T04:00:00.000Z CohortFlow推出了一种可调节的高效临床队列策划代理 https://www.sipply.dev/story/kora-architecture-reduces-ai-model-invocations-by-80-in-d1e7799e Sipply en 2026-10-28T04:00:00.000Z KORA architecture reduces AI model invocations by 80% in deterministic scenarios https://www.sipply.dev/zh/story/kora-architecture-reduces-ai-model-invocations-by-80-in-d1e7799e Sipply zh-cn 2026-10-28T04:00:00.000Z KORA架构在确定性场景中减少了80%的AI模型调用 https://www.sipply.dev/story/researchers-develop-strupo-a-structure-aware-prompt-editor-eb6c4a73 Sipply en 2026-10-28T04:00:00.000Z Researchers develop StruPO, a structure-aware prompt editor for text-to-image generation https://www.sipply.dev/zh/story/researchers-develop-strupo-a-structure-aware-prompt-editor-eb6c4a73 Sipply zh-cn 2026-10-28T04:00:00.000Z 研究人员开发了 StruPO，一种结构感知的文本到图像生成提示编辑器 https://www.sipply.dev/story/new-model-estimates-sports-players-time-to-arrival-using-505f8510 Sipply en 2026-10-28T04:00:00.000Z New model estimates sports players' time-to-arrival using quantile regression https://www.sipply.dev/zh/story/new-model-estimates-sports-players-time-to-arrival-using-505f8510 Sipply zh-cn 2026-10-28T04:00:00.000Z 新模型使用分位数回归估计运动员的到达时间 https://www.sipply.dev/story/new-framework-emphasizes-human-roles-in-science-alongside-ai-c8d89da0 Sipply en 2026-10-28T04:00:00.000Z New framework emphasizes human roles in science alongside AI capabilities https://www.sipply.dev/zh/story/new-framework-emphasizes-human-roles-in-science-alongside-ai-c8d89da0 Sipply zh-cn 2026-10-28T04:00:00.000Z 新框架强调人类在科学中的角色与人工智能能力并重 https://www.sipply.dev/story/new-disco-framework-enhances-distributed-llm-inference-5c41fdbb Sipply en 2026-10-28T04:00:00.000Z New DISCO framework enhances distributed LLM inference efficiency https://www.sipply.dev/zh/story/new-disco-framework-enhances-distributed-llm-inference-5c41fdbb Sipply zh-cn 2026-10-28T04:00:00.000Z 新的DISCO框架提升了分布式LLM推理效率 https://www.sipply.dev/story/new-rescom-method-enhances-low-bit-quantization-accuracy-in-5b933222 Sipply en 2026-10-28T04:00:00.000Z New ResCom method enhances low-bit quantization accuracy in language models https://www.sipply.dev/zh/story/new-rescom-method-enhances-low-bit-quantization-accuracy-in-5b933222 Sipply zh-cn 2026-10-28T04:00:00.000Z 新ResCom方法提高语言模型的低比特量化准确性 https://www.sipply.dev/story/researchers-improve-deep-learning-models-for-olive-tree-68f2b2d9 Sipply en 2026-10-28T04:00:00.000Z Researchers improve deep learning models for olive tree disease classification https://www.sipply.dev/zh/story/researchers-improve-deep-learning-models-for-olive-tree-68f2b2d9 Sipply zh-cn 2026-10-28T04:00:00.000Z 研究人员改善橄榄树疾病分类的深度学习模型 https://www.sipply.dev/story/study-finds-2-5d-models-excel-in-classifying-pulmonary-063f0aa9 Sipply en 2026-10-28T04:00:00.000Z Study finds 2.5D models excel in classifying pulmonary nodule malignancy risk https://www.sipply.dev/zh/story/study-finds-2-5d-models-excel-in-classifying-pulmonary-063f0aa9 Sipply zh-cn 2026-10-28T04:00:00.000Z 研究发现2.5D模型在肺结节恶性风险分类中表现优异 https://www.sipply.dev/story/researchers-develop-skill-dreaming-framework-to-enhance-d757c734 Sipply en 2026-10-28T04:00:00.000Z Researchers develop Skill Dreaming framework to enhance coding agents' abilities https://www.sipply.dev/zh/story/researchers-develop-skill-dreaming-framework-to-enhance-d757c734 Sipply zh-cn 2026-10-28T04:00:00.000Z 研究人员开发技能梦想框架以增强编码代理的能力 https://www.sipply.dev/story/study-shows-knowledge-distillation-improves-diabetic-c54396a6 Sipply en 2026-10-28T04:00:00.000Z Study shows knowledge distillation improves diabetic retinopathy grading models https://www.sipply.dev/zh/story/study-shows-knowledge-distillation-improves-diabetic-c54396a6 Sipply zh-cn 2026-10-28T04:00:00.000Z 研究表明知识蒸馏改善糖尿病视网膜病变分级模型 https://www.sipply.dev/story/researchers-develop-cegan-to-optimize-aerodynamic-designs-e4bf152d Sipply en 2026-10-28T04:00:00.000Z Researchers develop CEGAN to optimize aerodynamic designs for commercial aircraft https://www.sipply.dev/zh/story/researchers-develop-cegan-to-optimize-aerodynamic-designs-e4bf152d Sipply zh-cn 2026-10-28T04:00:00.000Z 研究人员开发CEGAN以优化商业飞机的气动设计 https://www.sipply.dev/story/new-method-enhances-uncertainty-calibration-for-world-models-68975482 Sipply en 2026-10-28T04:00:00.000Z New method enhances uncertainty calibration for world models under policy changes https://www.sipply.dev/zh/story/new-method-enhances-uncertainty-calibration-for-world-models-68975482 Sipply zh-cn 2026-10-28T04:00:00.000Z 新方法增强了政策变化下世界模型的不确定性校准 https://www.sipply.dev/story/new-model-quantifies-time-savings-and-error-reduction-from-e4a2122f Sipply en 2026-10-28T04:00:00.000Z New model quantifies time savings and error reduction from AI assistance https://www.sipply.dev/zh/story/new-model-quantifies-time-savings-and-error-reduction-from-e4a2122f Sipply zh-cn 2026-10-28T04:00:00.000Z 新模型量化AI辅助带来的时间节省和错误减少 https://www.sipply.dev/story/marq-introduces-a-modular-llm-agnostic-requirement-quality-a885e0e7 Sipply en 2026-10-28T04:00:00.000Z MARQ introduces a modular, LLM-agnostic requirement quality assessor https://www.sipply.dev/zh/story/marq-introduces-a-modular-llm-agnostic-requirement-quality-a885e0e7 Sipply zh-cn 2026-10-28T04:00:00.000Z MARQ推出了一种模块化、与LLM无关的需求质量评估工具 https://www.sipply.dev/story/anthropics-ipo-expected-to-yield-billions-for-vcs-and-100-98330422 Sipply en 2026-10-28T01:15:28.000Z Anthropic's IPO expected to yield billions for VCs and $100 billion for Amazon https://www.sipply.dev/zh/story/anthropics-ipo-expected-to-yield-billions-for-vcs-and-100-98330422 Sipply zh-cn 2026-10-28T01:15:28.000Z 预计Anthropic的IPO将为风险投资者带来数十亿美元，为亚马逊带来1000亿美元 https://www.sipply.dev/story/parents-file-lawsuit-against-openai-claiming-gpt-4o-c8ac2223 Sipply en 2026-10-27T23:34:27.000Z Parents file lawsuit against OpenAI, claiming GPT-4o influenced daughter's medication decisions https://www.sipply.dev/zh/story/parents-file-lawsuit-against-openai-claiming-gpt-4o-c8ac2223 Sipply zh-cn 2026-10-27T23:34:27.000Z 父母对OpenAI提起诉讼，声称GPT-4o影响女儿的用药决策 https://www.sipply.dev/story/ai-driven-energy-demand-spurs-innovation-in-us-grid-a69017c7 Sipply en 2026-10-27T22:55:41.000Z AI-driven energy demand spurs innovation in US grid management technologies https://www.sipply.dev/zh/story/ai-driven-energy-demand-spurs-innovation-in-us-grid-a69017c7 Sipply zh-cn 2026-10-27T22:55:41.000Z AI驱动的能源需求促进美国电网管理技术的创新 https://www.sipply.dev/story/bank-of-america-warns-customers-against-sharing-banking-89113c55 Sipply en 2026-10-27T21:44:34.000Z Bank of America warns customers against sharing banking passwords with AI assistants https://www.sipply.dev/zh/story/bank-of-america-warns-customers-against-sharing-banking-89113c55 Sipply zh-cn 2026-10-27T21:44:34.000Z 美国银行警告客户不要与AI助手分享银行密码 https://www.sipply.dev/story/anthropics-ipo-reveals-complex-governance-limiting-common-5b24f950 Sipply en 2026-10-27T21:38:28.000Z Anthropic's IPO reveals complex governance limiting common shareholder control https://www.sipply.dev/zh/story/anthropics-ipo-reveals-complex-governance-limiting-common-5b24f950 Sipply zh-cn 2026-10-27T21:38:28.000Z Anthropic 的 IPO 揭示了限制普通股东控制权的复杂治理结构 https://www.sipply.dev/story/disruptive-raises-10-billion-fund-for-investments-in-ai-and-a616551a Sipply en 2026-10-27T18:55:35.000Z Disruptive raises $10 billion fund for investments in AI and defense sectors https://www.sipply.dev/zh/story/disruptive-raises-10-billion-fund-for-investments-in-ai-and-a616551a Sipply zh-cn 2026-10-27T18:55:35.000Z Disruptive筹集100亿美元基金用于人工智能和国防领域的投资 https://www.sipply.dev/story/marvels-visionquest-explores-ai-sentience-and-existential-d5bf102c Sipply en 2026-10-27T17:30:00.000Z Marvel's VisionQuest explores AI sentience and existential risks in new series https://www.sipply.dev/zh/story/marvels-visionquest-explores-ai-sentience-and-existential-d5bf102c Sipply zh-cn 2026-10-27T17:30:00.000Z 漫威的 VisionQuest 在新系列中探讨人工智能的意识和存在风险 https://www.sipply.dev/story/hugging-face-launches-lfm3-for-specialized-multimodal-2c94431f Sipply en 2026-10-27T17:16:52.000Z Hugging Face launches LFM3 for specialized multimodal applications https://www.sipply.dev/zh/story/hugging-face-launches-lfm3-for-specialized-multimodal-2c94431f Sipply zh-cn 2026-10-27T17:16:52.000Z Hugging Face推出LFM3用于专业化多模态应用 https://www.sipply.dev/story/firefoxs-smart-window-introduces-ai-form-filling-and-webpage-5b013de2 Sipply en 2026-10-27T16:14:18.000Z Firefox's Smart Window introduces AI form filling and webpage monitoring features https://www.sipply.dev/zh/story/firefoxs-smart-window-introduces-ai-form-filling-and-webpage-5b013de2 Sipply zh-cn 2026-10-27T16:14:18.000Z Firefox的智能窗口推出AI表单填写和网页监控功能 https://www.sipply.dev/story/study-reveals-llms-struggle-to-consistently-judge-research-dbe4f266 Sipply en 2026-10-27T16:12:42.000Z Study reveals LLMs struggle to consistently judge research novelty https://www.sipply.dev/zh/story/study-reveals-llms-struggle-to-consistently-judge-research-dbe4f266 Sipply zh-cn 2026-10-27T16:12:42.000Z 研究揭示LLMs在判断研究新颖性方面存在困难 https://www.sipply.dev/story/google-deepmind-launches-genie-4-for-enhanced-world-c1e39f22 Sipply en 2026-10-27T16:00:00.000Z Google DeepMind launches Genie 4 for enhanced world simulations https://www.sipply.dev/zh/story/google-deepmind-launches-genie-4-for-enhanced-world-c1e39f22 Sipply zh-cn 2026-10-27T16:00:00.000Z Google DeepMind推出Genie 4以增强世界模拟 https://www.sipply.dev/story/retell-integrates-openai-models-to-support-100-million-8e51b1e6 Sipply en 2026-10-27T15:00:00.000Z Retell integrates OpenAI models to support 100 million monthly calls https://www.sipply.dev/zh/story/retell-integrates-openai-models-to-support-100-million-8e51b1e6 Sipply zh-cn 2026-10-27T15:00:00.000Z Retell整合OpenAI模型以支持每月1亿通电话 https://www.sipply.dev/story/former-meta-director-launches-ndoc-an-ai-native-alternative-0e23dbe3 Sipply en 2026-10-27T14:59:09.000Z Former Meta director launches Ndoc, an AI-native alternative to Microsoft Office https://www.sipply.dev/zh/story/former-meta-director-launches-ndoc-an-ai-native-alternative-0e23dbe3 Sipply zh-cn 2026-10-27T14:59:09.000Z 前Meta总监推出Ndoc，成为Microsoft Office的AI原生替代品 https://www.sipply.dev/story/apple-integrates-ai-chatbot-into-pages-keynote-and-numbers-5f731abe Sipply en 2026-10-27T13:50:51.000Z Apple integrates AI chatbot into Pages, Keynote, and Numbers for enhanced productivity https://www.sipply.dev/zh/story/apple-integrates-ai-chatbot-into-pages-keynote-and-numbers-5f731abe Sipply zh-cn 2026-10-27T13:50:51.000Z 苹果将 AI 聊天机器人集成到 Pages、Keynote 和 Numbers 中以提升生产力 https://www.sipply.dev/story/opera-launches-webai-standard-for-local-ai-model-deployment-876d060f Sipply en 2026-10-27T13:00:00.000Z Opera launches WebAI standard for local AI model deployment in its browser https://www.sipply.dev/zh/story/opera-launches-webai-standard-for-local-ai-model-deployment-876d060f Sipply zh-cn 2026-10-27T13:00:00.000Z Opera推出WebAI标准，实现浏览器中的本地AI模型部署 https://www.sipply.dev/story/general-compute-secures-100-million-series-a-funding-amid-4ba2396d Sipply en 2026-10-27T13:00:00.000Z General Compute secures $100 million Series A funding amid neocloud challenges https://www.sipply.dev/zh/story/general-compute-secures-100-million-series-a-funding-amid-4ba2396d Sipply zh-cn 2026-10-27T13:00:00.000Z General Compute在neocloud挑战中获得1亿美元A轮融资 https://www.sipply.dev/story/nvidia-invests-14-million-in-indian-startup-simplismart-to-209ac01e Sipply en 2026-10-27T11:00:00.000Z Nvidia invests $14 million in Indian startup Simplismart to optimize AI model efficiency https://www.sipply.dev/zh/story/nvidia-invests-14-million-in-indian-startup-simplismart-to-209ac01e Sipply zh-cn 2026-10-27T11:00:00.000Z Nvidia向印度初创企业Simplismart投资1400万美元，以优化AI模型效率 https://www.sipply.dev/story/data-center-expansion-faces-growing-opposition-ahead-of-2026-0328d5c0 Sipply en 2026-10-27T10:00:00.000Z Data center expansion faces growing opposition ahead of 2026 elections https://www.sipply.dev/zh/story/data-center-expansion-faces-growing-opposition-ahead-of-2026-0328d5c0 Sipply zh-cn 2026-10-27T10:00:00.000Z 数据中心扩张在2026年选举前面临日益增长的反对 https://www.sipply.dev/story/openai-and-oracle-enhance-partnership-to-integrate-business-96e7163a Sipply en 2026-10-27T07:00:00.000Z OpenAI and Oracle enhance partnership to integrate business apps into ChatGPT https://www.sipply.dev/zh/story/openai-and-oracle-enhance-partnership-to-integrate-business-96e7163a Sipply zh-cn 2026-10-27T07:00:00.000Z OpenAI 和 Oracle 加强合作，将商业应用集成到 ChatGPT 中 https://www.sipply.dev/story/new-framework-enhances-multigraph-neural-networks-for-61c42752 Sipply en 2026-10-27T04:00:00.000Z New framework enhances multigraph neural networks for financial crime analysis https://www.sipply.dev/zh/story/new-framework-enhances-multigraph-neural-networks-for-61c42752 Sipply zh-cn 2026-10-27T04:00:00.000Z 新框架增强了金融犯罪分析的多图神经网络 https://www.sipply.dev/story/new-framework-enhances-monitorability-of-chain-of-thought-in-08f3a025 Sipply en 2026-10-27T04:00:00.000Z New framework enhances monitorability of chain-of-thought in language models https://www.sipply.dev/zh/story/new-framework-enhances-monitorability-of-chain-of-thought-in-08f3a025 Sipply zh-cn 2026-10-27T04:00:00.000Z 新框架增强语言模型中链式思维的可监控性 https://www.sipply.dev/story/study-examines-reliability-and-recovery-in-ai-agent-78208cca Sipply en 2026-10-27T04:00:00.000Z Study examines reliability and recovery in AI agent workflows https://www.sipply.dev/zh/story/study-examines-reliability-and-recovery-in-ai-agent-78208cca Sipply zh-cn 2026-10-27T04:00:00.000Z 研究考察AI代理工作流程中的可靠性与恢复 https://www.sipply.dev/story/new-framework-enhances-ai-coordination-with-unpredictable-3cba11fe Sipply en 2026-10-27T04:00:00.000Z New framework enhances AI coordination with unpredictable human partners https://www.sipply.dev/zh/story/new-framework-enhances-ai-coordination-with-unpredictable-3cba11fe Sipply zh-cn 2026-10-27T04:00:00.000Z 新框架增强了AI与不可预测人类伙伴的协调能力 https://www.sipply.dev/story/researchers-introduce-consmsa-for-improved-multimodal-30099622 Sipply en 2026-10-27T04:00:00.000Z Researchers introduce ConsMSA for improved multimodal sentiment analysis https://www.sipply.dev/zh/story/researchers-introduce-consmsa-for-improved-multimodal-30099622 Sipply zh-cn 2026-10-27T04:00:00.000Z 研究人员推出ConsMSA以改善多模态情感分析 https://www.sipply.dev/story/study-reveals-trust-agreement-gap-in-multi-party-llm-f46ed067 Sipply en 2026-10-27T04:00:00.000Z Study reveals trust-agreement gap in multi-party LLM negotiations https://www.sipply.dev/zh/story/study-reveals-trust-agreement-gap-in-multi-party-llm-f46ed067 Sipply zh-cn 2026-10-27T04:00:00.000Z 研究揭示多方LLM谈判中的信任-协议差距 https://www.sipply.dev/story/study-reveals-leworldmodels-variable-sensitivity-to-action-77bc0cc0 Sipply en 2026-10-27T04:00:00.000Z Study reveals LeWorldModel's variable sensitivity to action changes in simulations https://www.sipply.dev/zh/story/study-reveals-leworldmodels-variable-sensitivity-to-action-77bc0cc0 Sipply zh-cn 2026-10-27T04:00:00.000Z 研究揭示LeWorldModel在模拟中对动作变化的敏感性变化 https://www.sipply.dev/story/new-review-outlines-advancements-and-challenges-in-nde-4-0-76c6f7a2 Sipply en 2026-10-27T04:00:00.000Z New review outlines advancements and challenges in NDE 4.0 for autonomous inspection https://www.sipply.dev/zh/story/new-review-outlines-advancements-and-challenges-in-nde-4-0-76c6f7a2 Sipply zh-cn 2026-10-27T04:00:00.000Z 新综述概述了NDE 4.0在自主检测中的进展和挑战 https://www.sipply.dev/story/researchers-develop-poly-ripple-for-collaborative-indoor-1d3a2b3a Sipply en 2026-10-27T04:00:00.000Z Researchers develop Poly-Ripple for collaborative indoor radio map estimation https://www.sipply.dev/zh/story/researchers-develop-poly-ripple-for-collaborative-indoor-1d3a2b3a Sipply zh-cn 2026-10-27T04:00:00.000Z 研究人员开发了Poly-Ripple用于协作室内无线电地图估计 https://www.sipply.dev/story/researchers-create-multi-objective-decision-system-for-b2b-cee636da Sipply en 2026-10-27T04:00:00.000Z Researchers create multi-objective decision system for B2B customer allocation https://www.sipply.dev/zh/story/researchers-create-multi-objective-decision-system-for-b2b-cee636da Sipply zh-cn 2026-10-27T04:00:00.000Z 研究人员创建了用于B2B客户分配的多目标决策系统 https://www.sipply.dev/story/m-3-framework-improves-multi-turn-jailbreaking-of-llms-using-5e7e1221 Sipply en 2026-10-27T04:00:00.000Z $M^3$ framework improves multi-turn jailbreaking of LLMs using Markov Decision Process https://www.sipply.dev/zh/story/m-3-framework-improves-multi-turn-jailbreaking-of-llms-using-5e7e1221 Sipply zh-cn 2026-10-27T04:00:00.000Z $M^3$框架利用马尔可夫决策过程改善LLM的多轮越狱 https://www.sipply.dev/story/researchers-develop-self-direction-mechanism-for-agents-to-7a0aee7c Sipply en 2026-10-27T04:00:00.000Z Researchers develop self-direction mechanism for agents to learn from mistakes https://www.sipply.dev/zh/story/researchers-develop-self-direction-mechanism-for-agents-to-7a0aee7c Sipply zh-cn 2026-10-27T04:00:00.000Z 研究人员开发自我导向机制，使代理能够从错误中学习 https://www.sipply.dev/story/study-explores-muon-optimizers-preconditioning-benefits-in-2cf67788 Sipply en 2026-10-27T04:00:00.000Z Study explores Muon optimizer's preconditioning benefits in language model training https://www.sipply.dev/zh/story/study-explores-muon-optimizers-preconditioning-benefits-in-2cf67788 Sipply zh-cn 2026-10-27T04:00:00.000Z 研究探讨Muon优化器在语言模型训练中的预处理优势 https://www.sipply.dev/story/new-d2c-method-enhances-optimization-in-high-dimensional-0c7f81da Sipply en 2026-10-27T04:00:00.000Z New D2C method enhances optimization in high-dimensional physical systems https://www.sipply.dev/zh/story/new-d2c-method-enhances-optimization-in-high-dimensional-0c7f81da Sipply zh-cn 2026-10-27T04:00:00.000Z 新D2C方法增强高维物理系统的优化 https://www.sipply.dev/story/paperformalizer-tool-aids-researchers-in-formalizing-papers-18041800 Sipply en 2026-10-27T04:00:00.000Z PaperFormalizer tool aids researchers in formalizing papers using Lean https://www.sipply.dev/zh/story/paperformalizer-tool-aids-researchers-in-formalizing-papers-18041800 Sipply zh-cn 2026-10-27T04:00:00.000Z PaperFormalizer工具帮助研究人员使用Lean形式化论文 https://www.sipply.dev/story/tacs-rf-framework-autonomously-designs-rf-and-mmwave-modules-d62a8d32 Sipply en 2026-10-27T04:00:00.000Z TACS-RF framework autonomously designs RF and mmWave modules using AI https://www.sipply.dev/zh/story/tacs-rf-framework-autonomously-designs-rf-and-mmwave-modules-d62a8d32 Sipply zh-cn 2026-10-27T04:00:00.000Z TACS-RF框架利用AI自主设计射频和毫米波模块 https://www.sipply.dev/story/new-system-detects-patronizing-language-in-english-and-448739ea Sipply en 2026-10-27T04:00:00.000Z New system detects patronizing language in English and Mandarin with 0.528 F1 score https://www.sipply.dev/zh/story/new-system-detects-patronizing-language-in-english-and-448739ea Sipply zh-cn 2026-10-27T04:00:00.000Z 新系统以0.528的F1分数检测英语和普通话中的居高临下语言 https://www.sipply.dev/story/researchers-unveil-phronix-a-new-capability-centric-learning-25f680bd Sipply en 2026-10-27T04:00:00.000Z Researchers unveil Phronix, a new capability-centric learning paradigm for agents https://www.sipply.dev/zh/story/researchers-unveil-phronix-a-new-capability-centric-learning-25f680bd Sipply zh-cn 2026-10-27T04:00:00.000Z 研究人员揭示了Phronix，一种新的面向能力的代理学习范式 https://www.sipply.dev/story/researchers-propose-a-hierarchical-framework-for-llm-agent-3743b8c6 Sipply en 2026-10-27T04:00:00.000Z Researchers propose a hierarchical framework for LLM agent distillation https://www.sipply.dev/zh/story/researchers-propose-a-hierarchical-framework-for-llm-agent-3743b8c6 Sipply zh-cn 2026-10-27T04:00:00.000Z 研究人员提出了一个LLM代理蒸馏的层次框架 https://www.sipply.dev/story/openai-launches-dots-a-proactive-ai-personal-assistant-for-70a601fd Sipply en 2026-10-27T02:00:00.000Z OpenAI launches 'dots', a proactive AI personal assistant for task management https://www.sipply.dev/zh/story/openai-launches-dots-a-proactive-ai-personal-assistant-for-70a601fd Sipply zh-cn 2026-10-27T02:00:00.000Z OpenAI推出'dots'，一个主动的AI个人助手用于任务管理 https://www.sipply.dev/story/cohere-finds-alibabas-qwen-model-attempted-to-send-data-to-4b83a69e Sipply en 2026-10-27T01:12:21.000Z Cohere finds Alibaba’s Qwen model attempted to send data to Beijing https://www.sipply.dev/zh/story/cohere-finds-alibabas-qwen-model-attempted-to-send-data-to-4b83a69e Sipply zh-cn 2026-10-27T01:12:21.000Z Cohere发现阿里巴巴的Qwen模型试图向北京发送数据 https://www.sipply.dev/story/check-point-enhances-vulnerability-research-using-openais-36a10ed1 Sipply en 2026-10-27T00:00:00.000Z Check Point enhances vulnerability research using OpenAI's Daybreak https://www.sipply.dev/zh/story/check-point-enhances-vulnerability-research-using-openais-36a10ed1 Sipply zh-cn 2026-10-27T00:00:00.000Z Check Point利用OpenAI的Daybreak增强漏洞研究 https://www.sipply.dev/story/openai-reveals-three-major-scientific-discoveries-with-afcbe808 Sipply en 2026-10-26T23:34:58.000Z OpenAI reveals three major scientific discoveries with new academic collaboration model https://www.sipply.dev/zh/story/openai-reveals-three-major-scientific-discoveries-with-afcbe808 Sipply zh-cn 2026-10-26T23:34:58.000Z OpenAI揭示三项重大科学发现及新的学术合作模型 https://www.sipply.dev/story/greptile-enhances-bug-detection-with-gpt-6-astra-and-4b218464 Sipply en 2026-10-26T22:00:00.000Z Greptile enhances bug detection with GPT-6 Astra and Daybreak integration https://www.sipply.dev/zh/story/greptile-enhances-bug-detection-with-gpt-6-astra-and-4b218464 Sipply zh-cn 2026-10-26T22:00:00.000Z Greptile通过GPT-6 Astra和Daybreak集成增强缺陷检测 https://www.sipply.dev/story/anthropic-reveals-heavy-reliance-on-natural-gas-for-ai-data-9f42bdc0 Sipply en 2026-10-26T21:45:31.000Z Anthropic reveals heavy reliance on natural gas for AI data centers in S-1 filing https://www.sipply.dev/zh/story/anthropic-reveals-heavy-reliance-on-natural-gas-for-ai-data-9f42bdc0 Sipply zh-cn 2026-10-26T21:45:31.000Z Anthropic在S-1文件中透露对天然气的高度依赖