OrangeBot.AI Digest — 2026-11-04

Check your exposure to critical CVEs with one agent prompt

Drop files easily across your local network

Fund 45+ top AI models using your idle CPU/GPU compute

Hidden AI buttons on your MacBook’s trackpad

Real or not? Check files before your AI trusts them.

Small, fast models for reranking, embeddings and decisions

Plan, build and review software with your coding agents

GNOME 51: Wayland fixes, oo7 keys, and modern toolchains.

A little bird that whispers what to say in your calls

Know when to trust the weather forecast

A hand-gesture instrument and MIDI controller for Mac

Take payments in apps you build in ChatGPT and Claude

Auto pause video when you look away

Make and publish 3D games with the first AI game engine

Personalized sleep and development app for parents

Artificial general intelligence (AGI) requires a method that can understand, generate, reason, and act across diverse domains and modalities. Rather than relying on a single model to master every capability, a promising path toward AGI now is to orchestrate a heterogeneous ecosystem of specialized models, tools, and policies (candidates) into a unified AI system. At the heart of this vision lies a fundamental question: how can an AI system dynamically select and coordinate the most capable candidate for each task while balancing performance and computational cost? In this work, we present Route Anything, a unified framework for studying model and policy routing across general-purpose AI systems. We develop a unified theoretical foundation that characterizes optimal routing under quality--cost trade-offs and establishes how the information available to a router determines its attainable performance. To systematically investigate routing across diverse capabilities, we first construct an omni-modal trajectory matrix spanning 25 benchmarks across language understanding, image, video, and speech generation, coding and workplace agents, and robotic manipulation, comprising 374,401 scored runs from 452 candidate configurations. Building on this framework, we train and test different routers to investigate the potential of omni-modal model routing, the capabilities of learned routers, and the conditions under which routing improves efficiency. Our results reveal substantial opportunities for capability orchestration: per-input oracle routing outperforms the best fixed candidate by 5.4 points, and learned routers outperform it by 0.3 points across 16 benchmarks, by 2.1 points on video generation, and by 1.1 points on robot manipulation while using 24% less policy time. More importantly, routing raises efficiency by 8.82x on average compared with the reference model.

Psychotherapy clients may experience distress before they can coherently describe it, and their expression changes with emotion, perceived safety, and trust. Large language model-based client simulators often rely on clinically organized case descriptions, which can encourage responses explaining the client's difficulties more clearly than their current understanding warrants. We propose Pandora, a psychotherapy client simulation framework that models how psychological experience becomes situated, individual expression. Pandora updates psychological and relational states in response to therapist interventions, constructs subjective appraisals and attributions, and plans response content and strategy. A separate expression planner combines these decisions with language patterns distilled from the client's utterances to guide linguistic realization. We evaluate Pandora against five baselines on 115 cases from four counseling and motivational interviewing datasets. It achieves the highest expression similarity and overall disclosure-alignment score among the compared systems. In a four-evaluator study of 45 matched disclosure items, Pandora receives AI-like judgments in 7.2% of ratings, compared with 43.3--67.2% for the baselines and 35.6% without expression planning. The ablation also reduces expression similarity. These results support explicit expression planning in client simulation to improve alignment with individual language patterns and reduce perceived AI-like expression in the evaluated setting.

Large reasoning models achieve strong performance on challenging tasks, but their long chain-of-thought often incurs substantial inference costs by continuing to deliberate after sufficient information has been obtained. Existing reasoning compression methods primarily operate on reasoning trajectories, paying less attention to the local decision of whether further reasoning is necessary given the current state. We propose LoCo, a local preference learning method for reasoning compression. LoCo detects candidate overthinking states from the model's own traces, freezes the corresponding reasoning prefixes, regenerates shorter verified continuations from the same states, and trains on the resulting state-conditioned preference pairs with DPO. This formulation teaches the model when to continue reasoning and when to conclude, rather than simply encouraging shorter trajectories overall. We construct training data entirely from self-generated mathematical traces and evaluate LoCo on five mathematical and three code generation benchmarks. LoCo reduces completion length by 41.7\%--45.4\% on mathematics and 23.4\%--36.3\% on code, while maintaining or improving accuracy and achieving the best AUC across all code benchmarks. LoCo consistently improves the accuracy-efficiency trade-off on mathematics and transfers effectively to code generation despite math-only training, demonstrating the cross-domain generality of local preference learning for reasoning compression. The code and model will be publicly released at https://github.com/MuyaoYuan/LoCo.

Long-form multi-shot video editing requires an intended edit to remain coherent across abrupt changes in scene, viewpoint, and subject appearance, while processing substantially longer visual context than conventional single-shot editors. We introduce MoviEdit, an end-to-end framework for scalable long-form multi-shot video editing that jointly addresses the challenges of data, modeling, and evaluation. To provide cross-shot supervision data at scale, we develop automatic data-generation pipelines spanning synchronized multi-view footage, real-world multi-shot videos, and generated multi-shot content, yielding 87K paired editing videos without manual annotation. Architecturally, MoviEdit employs a cascaded bidirectional-autoregressive design: a lightweight bidirectional stage first establishes a globally coherent editing draft over a spatiotemporally compressed representation, which is then autoregressively propagated and refined over the full video. We further introduce MoviEdit-Bench, a dedicated benchmark for systematic evaluation of long-form multi-shot video editing. Extensive experiments show that MoviEdit substantially outperforms existing open-source models and performs on par with industry frontier models across text-guided and image-guided editing. Visit our project page for complete results and to access the benchmark: https://multi-shot-edit.github.io/

Video generation has progressed from isolated clips to coherent short-form audiovisual sequences, and the bottleneck is shifting from rendering to semantic control. This report summarizes the observations and insights on multimodal understanding gained from successive generations of Kling. We argue that generation quality is the projection of multimodal understanding through the entire data--conditioning--alignment--evaluation chain, and we examine this chain module by module. Kling-MLLM, a shared multimodal foundation model that unifies image, video, audio, synchronized audio--video, and interleaved inputs, is specialized for each downstream module. Because a DiT is trained on far fewer training samples than a language model, we treat captioning as a semantic contract that maximizes description density, and extend it to reference-grounded captioning that makes reference bindings explicit. Prompt enhancement translates user intent into this caption distribution: basic PE is derived from captions of compelling videos and refined with text-level and generator-in-the-loop reinforcement learning, while agentic PE refines the story and retrieves compelling examples. We show that a reliable text encoder comes from perceptual training rather than from stronger language ability, and that candidates can be screened from text alone before end-to-end training. Reward modeling pairs a general reward model for explainable, localized critiques with a decision reward model for fast multidimensional scoring, treating perceptual bandwidth as a first-class design parameter. Finally, we review task-specific automatic evaluation, focusing on protocols that expose scoring shortcuts and help trace failures to their upstream source. We hope these lessons offer practical guidance for building the understanding stack behind controllable video generation.

Learning a universal action representation from video requires modeling dynamics that are consistent across different executions of the same action yet invariant to changes in embodiment and appearance. Existing latent action models infer actions from visual transitions, but their representations often entangle the underlying action with irrelevant visual variations, limiting transfer across embodiments. We introduce DreamX, which repurposes a pretrained video diffusion model as a universal latent action learner. Our approach extracts compact latent actions from the video model's rich spatiotemporal features, compressing each latent video frame by 300times (1{,200 rightarrow 4} tokens). In addition, by leveraging the video model's generative prior, these latent actions capture action-relevant dynamics disentangled from visual appearance. The resulting latent space serves as a shared action interface across embodiments (human and robot), environments, and action parameterizations. On LARY-Bench, DreamX achieves state-of-the-art performance among latent action models. Moreover, actions encoded in one setting remain effective in another: the latent space supports any-to-any video transfer and zero-shot action transfer across diverse embodiments and environments. Furthermore, this shared interface also enables policy learning in a compact latent action space, reducing the computational cost of video-diffusion-based policies.

Composed Image Retrieval (CIR) aims to retrieve a target image given a reference image and a modification text. Although recent approaches leveraging Multimodal Large Language Models (MLLMs) have shown significant progress, we argue that the MLLM's compositional understanding remains underutilized. This stems from directly mapping composed queries into a retrieval embedding via contrastive supervision alone, without explicitly resolving the intended visual modifications. In this work, we introduce Factorized MLLM Embeddings (FactEmbed), which decouples compositional query comprehension from building discriminative embeddings. Rather than directly projecting the composed query into a single embedding without internal reasoning, FactEmbed employs a set of factorized latents that simulate the intended visual modifications via an intermediate target imagination objective. This process encourages the latents to ground into target semantics before forming discriminative representations for retrieval. Subsequently, to propagate these factorized representations throughout the model for retrieval, we optimize a factor-level contrastive loss while distilling the decomposed representations into a deeper attention layer of the MLLM. Finally, for efficient inference, we retain the [EOS] token to establish a global-level embedding on top of these factorized latents, while we exploit the mean-pooled factorized latents to rerank a short candidate list at negligible cost, as both are produced by the same forward pass. Extensive evaluations demonstrate that FactEmbed outperforms prior works across CIR benchmarks.

We introduce Axion 1.0, a fully open agent foundation model that combines strong and broad agentic capabilities with a remarkably simple training recipe. Axion 1.0 is trained in only three stages: two lightweight supervised fine-tuning stages followed by reinforcement learning, without dedicated mid-training, direct preference optimization (DPO), or on-policy distillation (OPD). The first SFT stage develops strong self-contained reasoning through model-aware data construction, where we synthesize challenging problems and adversarially strengthen them based on the base model's own successful solution strategies. The second stage extends these capabilities to tool-dependent reasoning through a scalable pipeline for constructing diverse and automatically verifiable tasks grounded in real-world data sources and deterministic computation. Finally, reinforcement learning further strengthens long-horizon reasoning and interaction with external tools and environments. Despite this compact training recipe, Axion 1.0 achieves strong performance across a broad suite of agentic benchmarks and is competitive with leading commercial models on several of them. Equally importantly, Axion 1.0 is open beyond model weights: we release the complete recipe, including data construction, generation, filtering, verification, prompts, and training code, enabling the entire development process to be reproduced and extended. Axion 1.0 demonstrates that strong agent foundation models can be both open and simple, offering a practical and reproducible path toward advancing open agentic intelligence.

Large language models are deployed as long-horizon agents that compose diverse tools across MCP services into executable workflows. Existing MCP environment-scaling methods are data-centric: they first synthesize a database and scenario, then derive tools and tasks from them. As a result, tool interfaces may not match real ones, tool compositions are not grounded in execution, and tasks rarely span more than one environment. We propose ToolSpace, a tool-centric framework that reverses this order. ToolSpace mines data, state, and retry dependencies from adjacent calls in execution traces of real MCP services and open agent corpora, using deterministic rules rather than LLM judgment, and clusters the resulting weighted tool-composition graph into environments. Databases are synthesized only for state-mutating environments, while query and stateless tools replay recorded returns. ToolSpace then performs state-aware random walks with rollback and decompiles each chain into a natural-language task with assertion-based checks, keeping only samples that replay successfully from the initial state. Applied to AutomationBench and tau3-Bench Banking, the pipeline produces harder Pro versions on which 7 frontier models drop by up to 38.2 points. Across seven long-horizon tool-use benchmarks, ToolSpace raises Qwen3.6-35B-A3B from 23.0 to 41.0 average score (43.2 with RL) and GLM-5.2 from 44.6 to 56.6, on par with Claude Opus 5. Ablations show that execution-witnessed edges clearly outperform LLM-judged ones, and performance keeps improving with data and environment scale.

Unified Multimodal Models (UMMs) integrate visual understanding, generation, and editing within a single unified framework. However, existing unified multimodal datasets and benchmarks remain dominated by single-turn tasks. Even in recent multi-turn datasets, many turns can be solved independently without tracking information across turns. Consequently, they provide limited supervision and evaluation for multimodal stateful interaction, which requires models to understand, maintain, update, and faithfully realize evolving dialogue states across turns. To address this gap, we introduce YANAMI, the first large-scale multimodal dialogue dataset and benchmark for stateful interaction, comprising YANAMI-400K and YANAMI-Bench. YANAMI-400K contains 400K multi-turn multimodal dialogues, with more than 2M images and 2.6M dialogue turns. The dialogues span four representative scenarios: Multi-Image Fusion, Image Editing, World Modeling, and Visual Reasoning, and require models to continuously track, preserve, and revise evolving multimodal dialogue states across turns. YANAMI-Bench comprises 400 multi-turn dialogues for systematically evaluating multimodal stateful interaction. It jointly evaluates Dialogue State Understanding and Visual Realization, characterizing the gap between models' ability to understand evolving dialogue states and their ability to faithfully realize them in visual outputs. Experiments show that fine-tuning BAGEL on YANAMI-400K improves image editing, generation, and multimodal stateful interaction by 10.3%, 2.0%, and 13.23%, respectively, while preserving its general visual understanding capabilities. Evaluations on YANAMI-Bench further reveal that existing UMMs still struggle to maintain, update, and visually realize evolving multimodal dialogue states.

The advancement of multimodal learning is heavily constrained by the scarcity and high annotation cost of high-quality instruction data. While raw videos offer an abundant and dynamic source of visual knowledge, transforming them into precise instruction data remains highly challenging. To bridge this gap, we present VidUEU-Agent, an automated pipeline that synthesizes instruction data from raw videos for multimodal understanding, editing, and unified tasks. VidUEU-Agent follows a four-stage pipeline: it performs global perception to filter low-quality shots, mines keyframes and multimodal signals, constructs metadata and routes the metadata to synthesize instruction samples. More importantly, our agent framework can autonomously iterate the data distribution based on downstream training results, thereby optimizing data quality, balance, and training effectiveness. Scaling this pipeline, we construct VidUEU-100K, a large-scale dataset of 100K training samples spanning multiple domains and tasks, providing a robust foundation for developing comprehensive multimodal capabilities. Extensive experiments demonstrate that fine-tuning baseline models on VidUEU-100K yields significant and consistent performance gains across diverse representative benchmarks, validating the effectiveness of our proposed agent framework and dataset.

LLM-based multi-agent systems (MAS) rely on collaboration among specialized agents to solve complex tasks. Existing methods improve MAS by refining the prompts, skills, and execution strategies of individual agents based on task feedback. However, improving individual agents does not necessarily lead to better team collaboration. Recent studies have therefore extended evolution to the collaboration level, using collective execution experience to adjust team organization, and inter-agent interactions. Nevertheless, evolving collaboration across heterogeneous tasks remains challenging, as existing approaches often constrain team reorganization within a shared structure and reuse interaction rules across incompatible team contexts. In this paper, we propose EvoTeam, a collaboration-centric evolution framework that jointly evolves team organization, interaction rules, and individual agent skills through execution experience. It maintains a library of reusable teams, selecting or creating agent teams and guiding collaboration through team-specific interaction rules. Collaboration experience accumulates within compatible team contexts, while individual skills remain reusable across teams. Experiments across four benchmarks demonstrate its effectiveness. Our code is available at https://github.com/DEEP-JLU/EvoTeam.

GUI web agents require realistic and interactive environments that remain stable during both evaluation and training. Live website evaluation suffers from changing content, access restrictions, and limited coverage of authenticated workflows, while existing synthetic environments for training often lack functional depth and visual fidelity. We propose WebHarbor, a suite of 153 executable and dockerized website environments that faithfully reconstruct the visual interfaces, content, and functional workflows of real-world web services under deterministic control. We build WebHarbor through task-driven human-agent collaboration, where environments evolve as new tasks expand their functional scope, coding agents implement the required changes, and human contributors audit visual fidelity, functional correctness, and task validity. The resulting suite comprises 3,518 tasks spanning multi-step and post-authentication workflows, where existing works often fail to cover. Beyond reliable evaluation, WebHarbor provides high-fidelity synthetic web environments for cost-efficient reinforcement learning (RL). We demonstrate that a Qwen 4B-level GUI agent trained with RL in WebHarbor generalizes to unseen tasks and performs competitively with leading open-source GUI web agents across four benchmarks. The code and data are publicly available at https://github.com/aiming-lab/WebHarbor.

Speculative decoding accelerates memory-bound LLM decoding with a lightweight drafter, and parallel drafters extend this speedup by drafting a whole block in one forward pass. However, it remains unclear whether their released systems (stock) attain the full speedup that parallel drafting permits. We identify three challenges. First, with one request, each step accepts few tokens but reads every target weight. Second, with batching, per-request verification work roughly halves the speedup over autoregressive (AR) decoding. Third, speculative decoding is deployment-specific, and no universal optimization fits every serving regime, drafter, target model and GPU. To address these challenges, we present SpecPilot, an autonomous agentic loop that discovers algorithms and systems optimizations for each deployment. In each round, a cost model locates the largest gaps between the current system and its ceiling. Parallel agents explore directions and integrate the successful ones. Token forcing checks every output token against stock to detect reward hacking. A structured memory carries validated mechanisms to new rounds and deployments. SpecPilot discovered a general lossless 12-bit weight format, tree verification for hybrid recurrent targets, two-level verification and three batching mechanisms. We evaluate it across four target models, three drafters, two GPU platforms and five benchmarks. Without training or quality loss, it raises the speedup over AR from stock's 4.46times to 10.87times with one request, and from 1.94times to 3.93times with batching.

Harness optimization can substantially improve LLM agents without modifying model weights, but existing methods primarily optimize response quality. Extending them to joint quality--cost optimization introduces a challenge: applying the deployment preference during intermediate updates can prematurely reject candidates that are valuable for later improvements. We introduce ParetoSaddler, which addresses this mismatch by separating intermediate candidate promotion from final deployment selection. At each iteration, an optimizer updates a parent harness to produce a candidate, and then compares the two over repeated executions and uses Bayesian bootstrap Pareto support to promote candidates showing credible progress in either quality or cost. Promoted candidates receive broader development evaluation and remain available as parents for later optimization steps. We evaluate ParetoSaddler on ClawMark and GAIA2 using domain-disjoint splits. On ClawMark, ParetoSaddler improves quality from 70.86\% to 81.35\% while reducing cost from \1.17 to 1.01 per task; on GAIA2, it improves quality from 56.22\% to 62.89\% while reducing cost from \0.78 to 0.45 per task. The Pareto-selected harnesses also achieve higher quality at lower cost than quality-only harness optimization. Analysis shows that Pareto-aware optimization retains intermediate candidates that would be discarded by deployment preference but can support later improvements, while Bayesian bootstrap screening filters out candidates whose gains are weakly supported across runs.

Deborah Mary Sophia / Reuters : Figma reports Q3 revenue up 43% YoY to $391.4M, above est., gross margin of 80%, below 81% est., and lifts its 2026 revenue outlook; FIG drops 6%+ after hours —  Figma (FIG.N) is leaning on cashing in on its AI features through credits and using its first-party model to power tasks …

Spencer Soper / Bloomberg : Snap reports Q3 revenue up 14% YoY to $1.72B, in line with est., and DAUs up 4% to 495M, vs. 499M est.; SNAP drops 9%+ after hours —  Snap Inc. fell after the company reported third-quarter ad revenue and user growth that missed estimates, with Chief Executive Officer Evan Spiegel acknowledging that …

Reuters : Arm reports Q2 revenue up 27% YoY to $1.44B, vs. $1.38B est., royalty revenue up 8% to $670M, vs. $720.8M est., and licensing revenue up 50% to $773M —  Arm Holdings (O9Ty.F), beat Wall Street estimates for second-quarter revenue on Wednesday, driven by higher royalty payments and licensing fees …

Ian King / Bloomberg : Qualcomm reports Q4 revenue down 8% YoY to $10.3B, vs. $10.1B est., and handsets revenue down 26% to $5.14B, vs. $5.25B est.; QCOM drops 5%+ after hours —  Qualcomm Inc., the largest maker of smartphone processors, gave a lackluster forecast for profit in the current quarter …

Bloomberg : Meta raised €9B from its debut euro bond sale that drew a record €53B of orders, paying ~1.4x the risk premium Alphabet offered for similar bonds this week —  Meta Platforms Inc. raised €9 billion from a debut euro debt bond sale that drew a record €53 billion of orders …

Mark Gurman / Bloomberg : Sources: Apple cut some engineers' daily AI spending cap to $200 from $300 and plans cuts to as little as $100, as cost cutting stokes staff fears of layoffs —  Apple Inc. is reducing employee budgets for artificial intelligence usage as part of a cost-cutting effort, even as the company seeks …

Murad Hemmadi / Globe and Mail : Filing: AMD will pay ~$2.9B in cash and stock for AI chip startup Taalas, with $1.4B dependent on unspecified conditions; the deal was announced in August —  Advanced Micro Devices Inc. AMD-Q said Wednesday that it will pay about US$2.9-billion for Toronto-based semiconductor startup Taalas Inc. …

Ryan Christoffel / 9to5Mac : Apple says it will check iCloud items shared to “Anyone with the link” for “harmful content” in 27.3 OS versions, and may delete the content or disable accounts —  iOS 27.3 adds lots of new iPhone features, and it also makes a change to how public iCloud links work …

Madison Mills / Axios : Meta acquires Conductor, which lets developers run multiple AI coding agents in parallel; CEO Charlie Holtz says the team will join Meta Superintelligence Labs —  Meta has acquired Conductor, an AI coding startup that lets software engineers manage multiple coding agents simultaneously …

Bloomberg : Shares of US tech giants jump after Democrats' midterm sweep as investors bet that a Democrat-controlled Congress will slow AI capex that's draining cash flow —  The shares of US technology giants led the market higher on Wednesday as investors bet that a Democrat-controlled Congress will slow …

Connie Loizos / TechCrunch : Sequoia raised ~$10B in new funds, and announces it is shutting down its Arc accelerator and Scouts program, merging its early-stage and growth teams, and more —  Sequoia Capital posted a memo to X on Wednesday morning titled “The Sequoia of Tomorrow” that's full of surprises.

Isaac Kasahara / Pangram : Pangram releases Pangram Image 1.0, which it says is state-of-the-art in AI image and AI video detection, and the smaller Image Flash 1.0 for AI video detection —  State-of-the-art AI image and video detection with our latest Pangram Image models. … A few months ago, we released Pangram Image as a research preview.

Ivan Mehta / TechCrunch : Titan OS, which monetizes smart TVs via ads and FAST deals, raised $46M in equity and debt, taking its total funding to $110M, and says it now has 30M users —  Barcelona-based smart TV operating system maker Titan OS has raised $46 million in new funding, consisting of both equity and debt, the company announced on Wednesday.

Shakeel Hashim / Transformer : What the blue tsunami means for AI policy —  America is waking up to an astonishing Democratic sweep.  As I write, Dems may end up with 56 Senate seats and as many as 261 House seats — more than almost anyone predicted.  States long considered unwinnable have flipped …

Katherine Schwartz / Bloomberg : CoreWeave's stock and unsecured bonds fall after sweeping Democratic midterm victories, as Texas Governor-elect Gina Hinojosa calls for a data center moratorium —  CoreWeave Inc.'s bonds tumbled on Wednesday, in some cases to new lows, after Democratic candidates in favor of slowing …

AMD 周二公布了 2026 财年第三季度财报，营收同比增长 45% 达到 134 亿美元。其中数据中心业务贡献了 85 亿美元营收同比增长 97%。客户端与游戏业务部门的营收下滑了 8% 降至 37 亿美元。作为 AMD 在 x86 市场的主要竞争对手，英特尔下一季度营收预计为 150-160 亿美元，其增长几乎停滞。相比下 AMD 预计下一季度营收将达到约 170 亿美元。如果达到这一目标，这将标志着 AMD 营收首次超过英特尔。

韩国游戏《胜利女神：新的希望（NIKKE）》的 PC 启动器被植入恶意代码。恶意代码是通过纂改 Level Infinite CDN 分发的四个官方 JavaScript 文件实现的，篡改发生于 2026 年 9 月 27 日前后，攻击则始于 2026 年 10 月 23 日到 11 月 3 日，在此期间打开 PC 启动器的用户可能会被植入木马。恶意程序会窃取浏览器保存的密码、Windows 凭据管理器信息、SSH 密钥以及云服务或 AI 服务的访问令牌，还能窃取 28 种基于浏览器的加密货币钱包数据、执行任意命令、读取文件以及屏幕截图。攻击期间打开过启动器的玩家可以打开游戏文件夹，搜寻 debug.log，打开该文件后寻找 k2f_chan，如果找到且显示 registered，那么最好立即卸载游戏并用杀毒软件进行扫描。

法国国民互助信贷银行（Crédit Mutuel Arkéa）被黑客通过欺诈性交易盗走 1.18 亿欧元。被盗走的钱随后被转移到多个加密货币账号。该银行证实遭到网络攻击，但尚未证实 1.18 亿欧元的数字，只是表示攻击已得到控制，客户不会遭受任何经济损失。国民互助信贷银行暂停了其应用的部分转账功能。

Mozilla 正在新版 Firefox Nightly 中测试迷你窗口（Mini Window）功能。该功能类似视频画中画功能，将一个标签页作为独立窗口一直显示在当前标签页之上，这个窗口隐去了地址栏、各类控件、滚动条等常见标签页元素，但如果用户将鼠标悬停在迷你窗口内，各种标签页元素将会显示。迷你窗口方便用户在当前标签页上工作的同时参考迷你窗口的内容，免于在两个标签页之间反复切换。用户可以在专注于当前标签页的同时查看迷你窗口。

富士胶片表示它将大幅增加磁带产能，原因是磁带备份成为 AI 热的一个意外受益者。普通消费者可能早在几十年前就已不再使用磁带，但业界却从未将其抛弃。云服务巨头依靠磁带对海量数据进行备份。富士胶片称，大型云服务商和数据中心运营商正在扩大对磁带的使用，因为磁带能以相对较低的成本存储海量数据。磁带备份可以轻松与系统断开连接，存放在抽屉、保险库或独立设施内。富士胶片强调，磁带介质能实现物理隔离（air-gap），为用户数据提供最后一道防线。

最大 PC 游戏 Mod 分享托管平台 Nexus Mods 公布了 AI 生成内容政策。Nexus Mods 没有禁止 AI 生成 Mod，但对此类 Mod 的分享设定了更严格的规定。新上传 Mod 必须披露是否使用 AI，AI 辅助 Mod 必须说明如何使用 AI 以及 AI 的使用范围。现有 Mod 需在 2027 年 1 月 2 日前符合新要求。如果 Mod 作者无视平台方提出的要求，可能会面临权限受限或被下架的后果。它计划针对 AI 使用、AI 模型训练、分支和衍生 Mod 引入新的授权选项，默认选项为“需征求作者同意（Ask me）”。Nexus Mods 承诺绝不会利用用户上传的 Mod 训练 AI，也不会将其出售或授权给任何 AI 公司。Nexus Mods 严禁利用生成式 AI 生成虚假的游戏截图。

苹果从 11 月 4 日起开始在中国推出 AI 功能 Apple Intelligence，中国用户可通过最新的 iOS 开发者 Beta 版进行测试，AI 功能将于 12 月随 iOS 27.3 的发布正式推出。中国版本的 Apple Intelligence 是苹果和阿里巴巴合作推出的，基于阿里巴巴的千问模型。首批 AI 功能包括实时翻译、写作工具、视觉智能、照片 app 的消除工具、智绘表情、照片 app 中的自然语言搜索、在照片 app 中创建回忆影片、识别优先邮件、邮件摘要、智能回复、通知摘要、首要通知、减少干扰、备忘录 app 中的音频转录摘要、通话转录文本的摘要、输⼊⽂字与 Siri 对话、Siri 能接入千问（需登录千问）、创作等。

根据发表在《自然》上的一项研究，出生前后大脑会清理突变神经元。对 2300 多个神经细胞进行的分析显示，在从人类胎儿大脑中采样的神经元中，超过 1/4 出现了染色体片段增加或缺失的情况。相比之下，出生后婴儿大脑中的神经元中，只有约 6% 存在此类异常。研究结果揭示了一种新的机制，该机制不仅调控着一个人的神经元数量，还决定了哪些胎儿神经元能够在大脑中存活下来。针对小鼠的研究表明，在早期发育过程中，哺乳动物会通过激活一个导致细胞死亡的分子程序，清除大量脑细胞。分析结果显示，在胎儿发育过程中，存在数量惊人的、具有广泛拷贝数变异的细胞。然而出生后，这些异常神经元的比例下降至原来的 1/4，表明它们在发育过程中已被淘汰。大脑是如何分辨哪些神经元正常、哪些不正常呢？一个可能的线索是那些拷贝数变异水平较高的脑细胞，其与其他神经元建立连接（即突触）的基因也往往表现异常。

德国开源软件公司 SUSE 现有员工 1900 人，它计划裁掉十分之一的员工。CEO Dirk-Peter van Leeuwen 在电子邮件里表示，自动化和 AI 降低了某些岗位的需求。他表示做出这一决定是非常痛苦的，但也是必要的。Van Leeuwen 称，在被瑞典私募股权投资公司殷拓集团收购后，公司在仔细审视了其未来战略、所需能力及团队架构后做出了裁员决定,“显而易见自动化和 AI 正从根本上改变软件公司的运营方式。AI 技术正呈指数级发展，迫使我们彻底重新构想工作模式。无论是所在行业，还是在更广泛的经济领域，我们都已经看到了这一趋势。这意味着 SUSE 目前所依托的组织架构和成本基础，已无法支撑未来的成功发展。”受影响的员工将在 11 月 4 日前收到通知，员工代表将于次日选出，一对一沟通会议则将于下周初启动。

在 2018 年宣布《上古卷轴VI》八年之后，Bethesda 宣布了游戏的副标题《上古卷轴VI：哨兵》。知名游戏记者 Jason Schreier 称，该游戏至少还要两年时间才会面世。Bethesda 是在庆祝工作室诞生四十周年的视频中公布这一消息的。《上古卷轴VI》是与《星空》一同宣布的，2023 年 9 月《星空》发售后游戏才进入正式开发阶段，预计至少到 2028 年 Bethesda 才会公布更多游戏相关信息。

币安官方博客披露了黑客如何窃取价值 63 亿美元加密货币的初步调查报告。报告称，攻击始于今年六月： 2026 年 6 月——供应链攻击。钱包开发运维团队的工程师 A 为安装一款热门 AI 编程工具，通过搜索引擎查找并下载了一个遭篡改的 Homebrew 安装程序。我们的安全团队主动检测到了这一异常，更换了该设备，同时轮换了所有相关凭证。尽管如此，工程师 A 的岗位信息，以及其钱包基础设施团队中的相关同事信息，仍已暴露给攻击者。 9月——社会工程攻击。软件工程师 B 通过一个正规的求职平台，收到了一名冒充项目总监人员的联系。工程师B在该平台上的个人资料中提及了自己在币安的工作经历。对方以项目总监的身份安排了第一轮面试，由一名人力资源代表参与；随后又安排了第二轮面试，面试者自称是该虚构公司的 CEO。 在第一轮面试中，冒充人力资源代表的人员与工程师B进行了实时视频面试。为了建立信任，面试者还向工程师B分享了自己的社交媒体账号，其平台帐号有个人生活照片和视频。这种方式比以往的社会工程攻击手段更复杂，也更具欺骗性。 第二轮面试通过一款热门社交聊天应用进行。面试接近尾声时，工程师B被要求将一份含代码仓库的“作业”下载到工作电脑中。该代码仓库包含恶意代码，在安装过程中便会执行，进而触发了一系列未经授权的操作。 该恶意软件采用了复杂的代码混淆和规避检测技术，包括用恶意版本替换某个正规开发工具组件，以执行命令并与攻击者控制的外部服务器建立通信。当时，终端安全防护工具未能检测到这一零日恶意软件。攻击者还采取了隐藏恶意行为、清除入侵痕迹等措施，进一步增加了检测和后续调查的难度。 10 月 20 日——利用内部 AI 工具作为攻击跳板。攻击者利用遭入侵的工程师 B 的电脑，通过内部 AI 工具控制了钱包开发运维工程师 A，以及另一名钱包软件工程师 C 的设备。攻击者很快发现，工程师 A 的电脑上存有一个具有容器部署权限的令牌。 10 月 30 日——渗透钱包签名操作。10 月 30 日 01:15 至 03:56（东八区时间）期间，攻击者登录了软件工程师 C 的电脑，使用从工程师 A 处窃取的凭证部署了一个容器，并执行恶意脚本，进而发起此次攻击。

《纽约时报》、《芝加哥论坛报》和《奥兰多哨兵报》等美国出版商起诉 Brave Software，指控其未经许可抓取网站内容，并帮助用户绕过付费墙。Brave Software 被控将其抓取内容的索引数据库 Brave Index 提供给 Anthropic、Perplexity、Mistral、xAI 和 Cohere 等 AI 公司，通过 Paywall blocker 功能帮助用户绕过数百家新闻媒体的付费墙。《纽约时报》发言人 Graham James 在一份声明中表示：“我们支持合乎道德且负责任的使用与开发 AI，但 Brave 公然蔑视现行版权法，对《纽约时报》多次要求其停止侵权行为的诉求置之不理。我们提起此项诉讼，旨在追究 Brave 的责任——它窃取了我们的内容，不仅用于支撑其自身产品，还将其提供给部分全球最知名、资源最雄厚的生成式 AI 公司。”

Google 宣布美国 K-12 学校学生使用的软件将默认关闭所有 AI 功能，除非学校管理人员自己选择启用。此举旨在回应教育工作者和家长的诉求，他们希望在孩子与 AI 互动上有更多的掌控权。今天 AI 功能已成为 Google 商用搜索引擎的标配，Google 负责教育的副总裁 Chris Phillips 表示，搜索与 AI 已深度融合，如果学校禁用了 AI，那么学生将无法使用搜索。美国 K-12 学校广泛使用 Google 的软件，包括 Gmail、Docs 以及作业平台 Classroom。Google 的 Chromebook 是美国学校最广泛使用的笔记本电脑。

Fedora Linux 项目释出了 v45。Fedora Linux 项目同时释出了 Fedora Workstation、Fedora KDE Plasma Desktop、Fedora Server、Fedora Cloud、Fedora CoreOS 和 Fedora IoT，以及 Atomic Desktops 等众多衍生版本。Fedora Linux 45 的主要变化包括：支持通过 Anaconda 或 Kickstart 在 Stratis 存储上进行安装，移除 NetworkManager 不再维护的 VPN 插件，GNOME 和 KDE ​​Plasma 桌面环境的最新版本，完善 Wayland 支持，改进分数缩放，gcc 16.2、binutils 2.47、glibc 2.44、gdb 17.2、Fontconfig 2.18、LLVM 23 等开发工具的新版本。更多变化可浏览 ChangeSet。

血细胞并不容易形成化石。大多数血细胞在死亡后不久便会分解，这使得找到令人信服的、非常古老的血液痕迹非常困难。然而澳大利亚科廷大学的研究团队在 PNAS 上发表论文称，凭借多条证据，他们在巴西一只 1.15 亿年前的恐龙遗骸中发现了最古老的红细胞化石。这项研究的对象是巴西恐龙“阿拉塔龙”的骨骼化石。研究人员估计，这只恐龙体长约 3 米、重约 34 公斤，死亡后遗体沉入水底，微生物生长并将其掩埋，从而保存了下来。在显微镜下，研究人员在这只恐龙的骨骼中发现化石细胞的迹象：这些结构呈圆形，中心凹陷，每个细胞平均长 23 微米、宽 20 微米，嵌在曾经让血管穿过骨骼的通道中。下一步则是寻找化学线索，以证明这些结构确实含有血细胞，而不仅仅是形似血细胞的矿物质。在他们进行的近 30 项试验中，有一项是将骨样本研磨成粉后用激光照射，从样本散射回来的光显示出血红素的存在。血红素是一种环状化合物，其中的铁元素有助于红细胞运输氧气。研究人员还发现了大量降解形式的血红素。