OrangeBot.AI Digest — 2025-11-28

🎯 告别信息过载，AI 助你看懂新闻资讯热点，简单的舆情监控分析 - 多平台热点聚合+基于 MCP 的AI分析工具。监控35个平台（抖音、知乎、B站、华尔街见闻、财联社等），智能筛选+自动推送+AI对话分析（用自然语言深度挖掘新闻：趋势追踪、情感分析、相似检索等13种工具）。支持企业微信/个人微信/飞书/钉钉/Telegram/邮件/ntfy/bark/slack 推送，30秒网页部署，1分钟手机通知，无需编程。支持Docker部署⭐ 让算法为你服务，用AI理解热点

An open-source, code-first Go toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control.

所有小初高、大学PDF教材。

[Support 0.49.x]（Reset Cursor AI MachineID & Bypass Higher Token Limit） Cursor Ai ，自动重置机器ID ， 免费升级使用Pro功能: You've reached your trial request limit. / Too many free trial accounts used on this machine. Please upgrade to pro. We have this limit in place to prevent abuse. Please let us know if you believe this is a mistake.

Node Version Manager - POSIX-compliant bash script to manage multiple active node.js versions

The Cloud Native Application Proxy

[EMNLP2025] "LightRAG: Simple and Fast Retrieval-Augmented Generation"

A list of open source games.

verl: Volcano Engine Reinforcement Learning for LLMs

Open-Source Memory Engine for LLMs, AI Agents & Multi-Agent Systems

Curated coding interview preparation materials for busy software engineers

Send a phone call from AI agent, in an API call. Or, directly call the bot from the configured phone number!

Run Windows Subsystem For Android on your Windows 10 and Windows 11 PC using prebuilt binaries with Google Play Store (MindTheGapps) and/or Magisk or KernelSU (root solutions) built in.

Powerful web graphics runtime built on WebGL, WebGPU, WebXR and glTF

Collection of publicly available IPTV channels from all over the world

Reward feedback learning (ReFL) has proven effective for aligning image generation with human preferences. However, its extension to video generation faces significant challenges. Existing video reward models rely on vision-language models designed for pixel-space inputs, confining ReFL optimization to near-complete denoising steps after computationally expensive VAE decoding. This pixel-space approach incurs substantial memory overhead and increased training time, and its late-stage optimization lacks early-stage supervision, refining only visual quality rather than fundamental motion dynamics and structural coherence. In this work, we show that pre-trained video generation models are naturally suited for reward modeling in the noisy latent space, as they are explicitly designed to process noisy latent representations at arbitrary timesteps and inherently preserve temporal information through their sequential modeling capabilities. Accordingly, we propose Process Reward Feedback Learning~(PRFL), a framework that conducts preference optimization entirely in latent space, enabling efficient gradient backpropagation throughout the full denoising chain without VAE decoding. Extensive experiments demonstrate that PRFL significantly improves alignment with human preferences, while achieving substantial reductions in memory consumption and training time compared to RGB ReFL.

While modern diffusion models excel at generating high-quality and diverse images, they still struggle with high-fidelity compositional and multimodal control, particularly when users simultaneously specify text prompts, subject references, spatial arrangements, pose constraints, and layout annotations. We introduce Canvas-to-Image, a unified framework that consolidates these heterogeneous controls into a single canvas interface, enabling users to generate images that faithfully reflect their intent. Our key idea is to encode diverse control signals into a single composite canvas image that the model can directly interpret for integrated visual-spatial reasoning. We further curate a suite of multi-task datasets and propose a Multi-Task Canvas Training strategy that optimizes the diffusion model to jointly understand and integrate heterogeneous controls into text-to-image generation within a unified learning paradigm. This joint training enables Canvas-to-Image to reason across multiple control modalities rather than relying on task-specific heuristics, and it generalizes well to multi-control scenarios during inference. Extensive experiments show that Canvas-to-Image significantly outperforms state-of-the-art methods in identity preservation and control adherence across challenging benchmarks, including multi-person composition, pose-controlled composition, layout-constrained generation, and multi-control generation.

Instruction-guided image editing offers an intuitive way for users to edit images with natural language. However, diffusion-based editing models often struggle to accurately interpret complex user instructions, especially those involving compositional relationships, contextual cues, or referring expressions, leading to edits that drift semantically or fail to reflect the intended changes. We tackle this problem by proposing MIRA (Multimodal Iterative Reasoning Agent), a lightweight, plug-and-play multimodal reasoning agent that performs editing through an iterative perception-reasoning-action loop, effectively simulating multi-turn human-model interaction processes. Instead of issuing a single prompt or static plan, MIRA predicts atomic edit instructions step by step, using visual feedback to make its decisions. Our 150K multimodal tool-use dataset, MIRA-Editing, combined with a two-stage SFT + GRPO training pipeline, enables MIRA to perform reasoning and editing over complex editing instructions. When paired with open-source image editing models such as Flux.1-Kontext, Step1X-Edit, and Qwen-Image-Edit, MIRA significantly improves both semantic consistency and perceptual quality, achieving performance comparable to or exceeding proprietary systems such as GPT-Image and Nano-Banana.

Large multimodal models (LMMs) are increasingly adopted as judges in multimodal evaluation systems due to their strong instruction following and consistency with human preferences. However, their ability to follow diverse, fine-grained evaluation criteria remains underexplored. We develop Multi-Crit, a benchmark for evaluating multimodal judges on their capacity to follow pluralistic criteria and produce reliable criterion-level judgments. Covering both open-ended generation and verifiable reasoning tasks, Multi-Crit is built through a rigorous data curation pipeline that gathers challenging response pairs with multi-criterion human annotations. It further introduces three novel metrics for systematically assessing pluralistic adherence, criterion-switching flexibility, and the ability to recognize criterion-level preference conflicts. Comprehensive analysis of 25 LMMs reveals that 1) proprietary models still struggle to maintain consistent adherence to pluralistic criteria--especially in open-ended evaluation; 2) open-source models lag further behind in flexibly following diverse criteria; and 3) critic fine-tuning with holistic judgment signals enhances visual grounding but fails to generalize to pluralistic criterion-level judgment. Additional analyses on reasoning fine-tuning, test-time scaling, and boundary consistency between open-source and proprietary models further probe the limits of current multimodal judges. As a pioneering study, Multi-Crit lays the foundation for building reliable and steerable multimodal AI evaluation.

Embodied cognition argues that intelligence arises from sensorimotor interaction rather than passive observation. It raises an intriguing question: do modern vision-language models (VLMs), trained largely in a disembodied manner, exhibit signs of embodied cognition? We introduce ENACT, a benchmark that casts evaluation of embodied cognition as world modeling from egocentric interaction in a visual question answering (VQA) format. Framed as a partially observable Markov decision process (POMDP) whose actions are scene graph changes, ENACT comprises two complementary sequence reordering tasks: forward world modeling (reorder shuffled observations given actions) and inverse world modeling (reorder shuffled actions given observations). While conceptually simple, solving these tasks implicitly demands capabilities central to embodied cognition-affordance recognition, action-effect reasoning, embodied awareness, and interactive, long-horizon memory from partially observable egocentric input, while avoiding low-level image synthesis that could confound the evaluation. We provide a scalable pipeline that synthesizes QA pairs from robotics simulation (BEHAVIOR) and evaluates models on 8,972 QA pairs spanning long-horizon home-scale activities. Experiments reveal a performance gap between frontier VLMs and humans that widens with interaction horizon. Models consistently perform better on the inverse task than the forward one and exhibit anthropocentric biases, including a preference for right-handed actions and degradation when camera intrinsics or viewpoints deviate from human vision. Website at https://enact-embodied-cognition.github.io/.

Language understanding entails not just extracting the surface-level meaning of the linguistic input, but constructing rich mental models of the situation it describes. Here we propose that because processing within the brain's core language system is fundamentally limited, deeply understanding language requires exporting information from the language system to other brain regions that compute perceptual and motor representations, construct mental models, and store our world knowledge and autobiographical memories. We review the existing evidence for this hypothesis, and argue that recent progress in cognitive neuroscience provides both the conceptual foundation and the methods to directly test it, thus opening up a new strategy to reveal what it means, cognitively and neurally, to understand language.

MLLMs exhibit strong reasoning on isolated queries, yet they operate de novo -- solving each problem independently and often repeating the same mistakes. Existing memory-augmented agents mainly store past trajectories for reuse. However, trajectory-based memory suffers from brevity bias, gradually losing essential domain knowledge. More critically, even in truly multimodal problem-solving settings, it records only a single-modality trace of past behavior, failing to preserve how visual attention and logical reasoning jointly contributed to the solution. This is fundamentally misaligned with human cognition: semantic memory is both multimodal and integrated, preserving visual and abstract knowledge through coordinated but distinct representational streams. We thus introduce ViLoMem, a dual-stream memory framework that constructs compact, schema-based memory. It separately encodes visual distraction patterns and logical reasoning errors, enabling MLLMs to learn from their successful and failed experiences. Following a grow-and-refine principle, the system incrementally accumulates and updates multimodal semantic knowledge -- preserving stable, generalizable strategies while avoiding catastrophic forgetting. Across six multimodal benchmarks, ViLoMem consistently improves pass@1 accuracy and substantially reduces repeated visual and logical errors. Ablations confirm the necessity of dual-stream memory with explicit distraction--hallucination separation, demonstrating the value of error-aware multimodal memory for lifelong and cross-domain agentic learning. Our project page will be available at https://weihao-bo.github.io/ViLoMeo-page.

发表在《科学》期刊上的一项新实验，采用了独立于 X 平台算法的由 AI 驱动的浏览器扩展程序来重新排序 X/Twitter 上的信息流，结果表明即使所接触的仅是敌对政治内容的微小变化，也能在数天内显著影响用户对反对党的观感。这些发现为算法所控帖子排名对用户社交媒体信息流会有影响提供了直接的因果证据。社交媒体已成为全球许多人获取政治信息的重要来源。然而平台算法对我们在浏览过程中所接触的内容施加了强大的影响力，因为这些算法能以种种难以理解的方式悄然左右人们的思想、情绪和行为。尽管人们就这些排名算法影响我们的方式提出了诸多解释，但要验证这些理论却异常困难。这是因为平台运营方独自掌控着其专有算法的运行方式，而且只有他们才能尝试不同信息流设计并评估其因果效应。为规避这些难题，研究人员创建了浏览器扩展，在人们浏览社交媒体信息流时对信息流实时进行重新排序，这一过程无需获得平台本身的许可。研究表明算法介导的与政治敌意内容的接触既能塑造情感极化，也能实时调控用户在使用平台过程中的瞬时情绪反应。

KDE Plasma 团队宣布即将发布的 v6.8 将只支持 Wayland，停止支持 X11。KDE Plasma 6.7 系列的最后一个版本预计会在 2027 年初发布，对 X11 会话的支持将会持续到 2027 年初。如果用户想要继续使用 X11，可以选择支持 Plasma X11 的长期支持发行版如 AlmaLinux 9，它会一直支持到 2032 年。而 X11 应用仍然可以通过 Xwayland 兼容层运行。

美国五角大楼建议将阿里巴巴、百度和比亚迪纳入协助中国军方的1260H名单。1260H名单虽无直接法律效力，但对美国投资者具有重要警示作用。美国国防部副部长 Stephen Feinberg 在 10 月 7 日的信函中表示，这三家企业连同另外五家，包括成都新易盛通信技术、华虹半导体、RoboSense速腾聚创、药明康德以及中际旭创，都应被列入1260H名单。阿里巴巴在一份声明中说，将该公司列入名单“毫无依据”，“阿里巴巴既非中国军事企业，也未参与任何军民融合战略” 。阿里巴巴进一步指出，由于公司不从事与美国军事采购相关的业务，被列入1260H清单不影响它在美国或全球任何地区正常开展业务。

沙盒游戏《Garry's Mod》和生存游戏《Rust》的开发商 Facepunch Studios 宣布在 MIT 许可证下开源其正在开发的新一代沙盒引擎 S&box，游戏代码托管在 GitHub 上。开发者表示此举受到了开源游戏引擎 Godot 成功的启发。S&box 是基于 Valve 的 Source 2 引擎，底层的 Source 2 引擎并没有开源（需要由 Valve 决定是否开源），s&box 开源的是底层之上的编辑器、网络、场景系统、用户界面等，这些部分都是用 C# 开发的。

在印度首都德里一间手术室内，医生们准备切除患者的胆囊。患者全身麻醉，但却戴着播放音乐的耳机。虽然麻醉期间患者的大脑大部分区域不活跃，但听觉通路仍然维持部分活跃。根据印度科学家在《Music and Medicine》期刊上发表的一项研究，全身麻醉期间播放音乐能显著减少用药量和改善术后恢复。研究主要针对通过腹腔镜技术进行的胆囊切除手术。这是一种标准的微创切除手术，持续时间短，通常不到一小时，因此需要患者迅速且“头脑清醒”的恢复。在试验中，佩戴耳机播放音乐的患者需要的药物剂量更低，术后恢复更平稳，压力激素水平更低，且在手术过程中血压得到更好的控制。

根据发表在《自然》期刊上的一项研究，毅力号火星漫游车在两个火星年的观测中记录到了 55 次闪电现象。闪电是由于大气湍流使粒子相互碰撞摩擦产生电荷。电荷积累到一定程度后以放电形式释放。闪电在地球上无处不在。科学家认为火星也存在放电现象，虽然火星大气层主要由二氧化碳组成，比地球更稀薄干燥。毅力号漫游车配备了能探测闪电迹象的仪器。法国图卢兹大学行星科学家 Baptiste Chide 领导的团队分析了漫游车 SuperCam 收集的数据，发现了 55 个事件，其中 7 个事件完整捕捉到了放电特征。大多数放电事件都非常微弱，能量仅为 0.1-150 纳焦耳。第七次事件最大，能量达到 40 毫焦耳。相比下，地球的一次云地闪电释放的能量约为 10 亿焦耳。

戴尔 COO Jeffrey Clarke 称 Windows 11 PC 的更新换代速度比 Windows 10 慢。如果以上一代操作系统停止支持这一时间点进行比较，Windows 11 普及率比 Windows 10 落后 10 到 12 个百分点。由于微软提高了硬件需求，现有的 Windows 10 PC 很多无法升级到 Windows 11。Clarke 表示有 5 亿台 PC 无法运行 Windows 11，还有同样数量的 PC 能升级到 Windows 11。戴尔三季度营收为 270 亿美元，同比增长 11%，预计四季度营收将达到 315 亿美元，2026 财年营收将达到 1117 亿美元，分别同比增长 32% 和 17%。

苹果有望时隔十多年再次成为全球最大的智能手机制造商。根据研究机构 Counterpoint Research 的预测，苹果今年推出的 iPhone 17 在美国本土以及中国都取得了成功，推动 iPhone 年销量超过三星手机。Counterpoint Research 预测苹果 iPhone 在 2025 年出货量增长 10%，而三星手机增速预计为 4.6%。苹果 iPhone 年出货量预计达到 2.43 亿部，略高于三星手机的 2.35 亿部。苹果的市场占有率达到 19.4%，三星约为 18.7%。

SEO 公司 Graphite 上个月发表了一份报告，称互联网上逾半数内容是 AI 生成的。Graphite 分析了 2020 年 1 月至 2025 年 5 月间发表的 65,000 篇英文文章的随机样本，使用了 AI 检测工具 Surfer 进行评估，如果一篇文章的内容有五成或更多部分被认为是大模型撰写的，那么这篇文章就被视为是 AI 生成。对 AI 撰写文章的分析显示大部分属于大众兴趣类文章：新闻更新、指南、生活方式、评论和产品介绍。此类文章主要是说服读者或为读者提供信息，不涉及表达原创。也就是说 AI 擅长处理低级别的模式化的写作，如周末游清单、求职信和商业文案等。此类工作过去由自由职业者操手，现在大模型的普及使得这些工作急剧减少。由人类创作的原创类文章可能比以前更富有价值。即使互联网上的大部分内容不再由人类撰写，作家、记者和知识分子的工作并不会变得多余。

中国是美国之后最大的新药研发国，中国药企去年开展的临床试验占到了全球的三分之一，而十年前这一比例仅为 5%。中国生物科技公司在癌症等领域正迅速崛起，相关领域企业的股价今年以来飙升了 110%，是美国同行涨幅的三倍多。2025 年上半年大型药企签署的全球授权协议中，近三分之一涉及到中国企业，是 2021 年的四倍。辉瑞公司于 5 月同意向三生制药（3SBio）支付 12.5 亿美元收购一种实验性抗癌药物；葛兰素史克于 6 月与恒瑞医药达成了潜在总金额 120 亿美元的组合授权协议。

德国研究人员回顾性对比了 648 名歌手的死亡风险，其中一半是明星，另一半则未成名。研究人员将 324 名明星与名气较小的同行在年龄、性别、国籍、种族、音乐流派以及是否为乐队独唱/主唱等方面进行了匹配。数据分析显示，著名歌手的平均寿命为 75 岁，而名气较小的歌手平均寿命为 79 岁。尽管与独唱艺人相比，乐队成员的死亡风险低 26%，但并未影响名气的整体效应——著名歌手的早逝风险仍比名气较小的同行高 33%。研究人员提出，一个可能解释是“名气带来的独特社会心理压力，例如强烈的公众审视、表演压力和隐私丧失”。“这些压力源可能引发心理困扰和有害的应对行为，使名气成为一种慢性负担，加剧了已有的职业风险。”

科学家可能首次观察到了“看不见”的暗物质。日本东京大学研究团队利用 NASA 费米伽马射线望远镜的最新观测数据探测到了与假想的“弱相互作用大质量粒子（WIMP）”碰撞湮灭时所预测的高能光子。科学家怀疑暗物质由 WIMP 构成。这些粒子被认为比质子重，与普通物质极少相互作用。理论预测，当两个 WIMP 粒子碰撞时，它们会相互湮灭并释放出包括伽马射线光子在内的高能粒子。东京大学的 Tomonori Totani 教授报告探测到了与理论预测结果相符的 20 GeV 伽马射线光晕。研究报告本周发表在《Journal of Cosmology and Astroparticle Physic》期刊上，预印本 7 月发表在 arXiv 上。

欧洲议会周三呼吁欧盟设定儿童使用社交媒体的最低年龄限制，以应对青少年因过度接触社交媒体而导致的心理健康问题日益增多的现状。此前澳大利亚通过了全球首个针对 16 岁以下儿童的社交媒体禁令，丹麦和马来西亚也计划效仿。欧洲议会以 483 票赞成、92 票反对、86 票弃权通过决议呼吁欧盟范围内禁止 16 岁以下儿童在未经家长同意的情况下访问在线平台、视频分享网站和人工智能助手，并彻底禁止 13 岁以下儿童使用。决议还呼吁禁止“战利品箱”以及针对未成年人的基于用户参与度的推荐算法，并要求制定相关法律，规定内容设计必须符合儿童的年龄特点。

马来西亚柔佛州停止批准一级和二级数据中心，理由这些数据中心用水量太高。柔佛是东南亚数据中心枢纽之一。截至 2025 年 11 月它批准了 51 个数据中心项目，其中 17 个已投入运营，11 个正在建设中，23 个是今年新获批准的。一级和二级数据中心每天用水量在 4000-5000 万升之间。相比下三级和四级数据中心每天用水量约为 20 万升，与普通工业负荷相当。柔佛希望所有数据中心能达到更高、更可持续、更节能的标准，与国际标准接轨。马来西亚官员称，美国佐治亚州的一家数据中心投入运营后，居民区频繁遭遇供水中断；乌拉圭民众抗议数据中心可能会影响农田供水。

根据发表在《Nature Communications》期刊上的一项研究，对猕猴前额叶皮层神经元活动的分析揭示了大脑如何调配有限的工作记忆资源。分析发现，当记忆负荷增加时，大脑并不是不断招募新的神经元，而是反复“再利用”已有的神经元。这些神经元既能稳定保留早先的信息，又能灵活调整去编码新的信息，并尽量减少不同记忆项目之间的干扰。随着需要记住的项目增多，这些神经元在保留旧信息、编码新信息以及减少信息间干扰之间进行权衡，体现了前额叶皮层对有限资源的主动调配。