Weekly Digest — 2025-W38

Implement a ChatGPT-like LLM in PyTorch from scratch, step by step

Python tool for converting files and office documents to Markdown.

PowerShell for every system!

FULL Augment Code, Claude Code, Cluely, CodeBuddy, Cursor, Devin AI, Junie, Kiro, Leap.new, Lovable, Manus Agent Tools, NotionAI, Orchids.app, Perplexity, Poke, Qoder, Replit, Same.dev, Trae, Traycer AI, VSCode Agent, Warp.dev, Windsurf, Xcode, Z.ai Code, dia & v0. (And other Open Sourced) System Prompts, Internal Tools & AI Models

An AI Hedge Fund Team

G-code generator for 3D printers (Bambu, Prusa, Voron, VzBot, RatRig, Creality, etc.)

Run LLMs with MLX

基于大模型和 RAG 的智能问数系统。Text-to-SQL Generation via LLMs using RAG.

DeepResearchAgent is a hierarchical multi-agent system designed not only for deep research tasks but also for general-purpose task solving. The framework leverages a top-level planning agent to coordinate multiple specialized lower-level agents, enabling automated task decomposition and efficient execution across diverse and complex domains.

A cryptocurrency trading API with more than 100 exchanges in JavaScript / TypeScript / Python / C# / PHP / Go

Engagement and motivation are crucial for second-language acquisition, yet maintaining learner interest in educational conversations remains a challenge. While prior research has explored what makes educational texts interesting, still little is known about the linguistic features that drive engagement in conversations. To address this gap, we introduce IntrEx, the first large dataset annotated for interestingness and expected interestingness in teacher-student interactions. Built upon the Teacher-Student Chatroom Corpus (TSCC), IntrEx extends prior work by incorporating sequence-level annotations, allowing for the study of engagement beyond isolated turns to capture how interest evolves over extended dialogues. We employ a rigorous annotation process with over 100 second-language learners, using a comparison-based rating approach inspired by reinforcement learning from human feedback (RLHF) to improve agreement. We investigate whether large language models (LLMs) can predict human interestingness judgments. We find that LLMs (7B/8B parameters) fine-tuned on interestingness ratings outperform larger proprietary models like GPT-4o, demonstrating the potential for specialised datasets to model engagement in educational settings. Finally, we analyze how linguistic and cognitive factors, such as concreteness, comprehensibility (readability), and uptake, influence engagement in educational dialogues.

Generating 3D shapes at part level is pivotal for downstream applications such as mesh retopology, UV mapping, and 3D printing. However, existing part-based generation methods often lack sufficient controllability and suffer from poor semantically meaningful decomposition. To this end, we introduce X-Part, a controllable generative model designed to decompose a holistic 3D object into semantically meaningful and structurally coherent parts with high geometric fidelity. X-Part exploits the bounding box as prompts for the part generation and injects point-wise semantic features for meaningful decomposition. Furthermore, we design an editable pipeline for interactive part generation. Extensive experimental results show that X-Part achieves state-of-the-art performance in part-level shape generation. This work establishes a new paradigm for creating production-ready, editable, and structurally sound 3D assets. Codes will be released for public research.

Does continued scaling of large language models (LLMs) yield diminishing returns? Real-world value often stems from the length of task an agent can complete. We start this work by observing the simple but counterintuitive fact that marginal gains in single-step accuracy can compound into exponential improvements in the length of a task a model can successfully complete. Then, we argue that failures of LLMs when simple tasks are made longer arise from mistakes in execution, rather than an inability to reason. We propose isolating execution capability, by explicitly providing the knowledge and plan needed to solve a long-horizon task. We find that larger models can correctly execute significantly more turns even when small models have 100\% single-turn accuracy. We observe that the per-step accuracy of models degrades as the number of steps increases. This is not just due to long-context limitations -- curiously, we observe a self-conditioning effect -- models become more likely to make mistakes when the context contains their errors from prior turns. Self-conditioning does not reduce by just scaling the model size. In contrast, recent thinking models do not self-condition, and can also execute much longer tasks in a single turn. We conclude by benchmarking frontier thinking models on the length of task they can execute in a single turn. Overall, by focusing on the ability to execute, we hope to reconcile debates on how LLMs can solve complex reasoning problems yet fail at simple tasks when made longer, and highlight the massive benefits of scaling model size and sequential test-time compute for long-horizon tasks.

Arbitrary resolution image generation provides a consistent visual experience across devices, having extensive applications for producers and consumers. Current diffusion models increase computational demand quadratically with resolution, causing 4K image generation delays over 100 seconds. To solve this, we explore the second generation upon the latent diffusion models, where the fixed latent generated by diffusion models is regarded as the content representation and we propose to decode arbitrary resolution images with a compact generated latent using a one-step generator. Thus, we present the InfGen, replacing the VAE decoder with the new generator, for generating images at any resolution from a fixed-size latent without retraining the diffusion models, which simplifies the process, reducing computational complexity and can be applied to any model using the same latent space. Experiments show InfGen is capable of improving many models into the arbitrary high-resolution era while cutting 4K image generation time to under 10 seconds.

The Retrieval-Augmented Generation (RAG) approach enhances question-answering systems and dialogue generation tasks by integrating information retrieval (IR) technologies with large language models (LLMs). This strategy, which retrieves information from external knowledge bases to bolster the response capabilities of generative models, has achieved certain successes. However, current RAG methods still face numerous challenges when dealing with multi-hop queries. For instance, some approaches overly rely on iterative retrieval, wasting too many retrieval steps on compound queries. Additionally, using the original complex query for retrieval may fail to capture content relevant to specific sub-queries, resulting in noisy retrieved content. If the noise is not managed, it can lead to the problem of noise accumulation. To address these issues, we introduce HANRAG, a novel heuristic-based framework designed to efficiently tackle problems of varying complexity. Driven by a powerful revelator, HANRAG routes queries, decomposes them into sub-queries, and filters noise from retrieved documents. This enhances the system's adaptability and noise resistance, making it highly capable of handling diverse queries. We compare the proposed framework against other leading industry methods across various benchmarks. The results demonstrate that our framework obtains superior performance in both single-hop and multi-hop question-answering tasks.

Spoken language models (SLMs) have emerged as a unified paradigm for speech understanding and generation, enabling natural human machine interaction. However, while most progress has focused on semantic accuracy and instruction following, the ability of SLMs to adapt their speaking style based on spoken instructions has received limited attention. We introduce Voice Style Adaptation (VSA), a new task that examines whether SLMs can modify their speaking style, such as timbre, prosody, or persona following natural language spoken commands. To study this task, we present VStyle, a bilingual (Chinese & English) benchmark covering four categories of speech generation: acoustic attributes, natural language instruction, role play, and implicit empathy. We also introduce the Large Audio Language Model as a Judge (LALM as a Judge) framework, which progressively evaluates outputs along textual faithfulness, style adherence, and naturalness, ensuring reproducible and objective assessment. Experiments on commercial systems and open source SLMs demonstrate that current models face clear limitations in controllable style adaptation, highlighting both the novelty and challenge of this task. By releasing VStyle and its evaluation toolkit, we aim to provide the community with a foundation for advancing human centered spoken interaction. The dataset and code are publicly available at https://junzhan2000.github.io/VStyle.github.io/{project's homepage}.

The field of 4D world modeling - aiming to jointly capture spatial geometry and temporal dynamics - has witnessed remarkable progress in recent years, driven by advances in large-scale generative models and multimodal learning. However, the development of truly general 4D world models remains fundamentally constrained by the availability of high-quality data. Existing datasets and benchmarks often lack the dynamic complexity, multi-domain diversity, and spatial-temporal annotations required to support key tasks such as 4D geometric reconstruction, future prediction, and camera-control video generation. To address this gap, we introduce OmniWorld, a large-scale, multi-domain, multi-modal dataset specifically designed for 4D world modeling. OmniWorld consists of a newly collected OmniWorld-Game dataset and several curated public datasets spanning diverse domains. Compared with existing synthetic datasets, OmniWorld-Game provides richer modality coverage, larger scale, and more realistic dynamic interactions. Based on this dataset, we establish a challenging benchmark that exposes the limitations of current state-of-the-art (SOTA) approaches in modeling complex 4D environments. Moreover, fine-tuning existing SOTA methods on OmniWorld leads to significant performance gains across 4D reconstruction and video generation tasks, strongly validating OmniWorld as a powerful resource for training and evaluation. We envision OmniWorld as a catalyst for accelerating the development of general-purpose 4D world models, ultimately advancing machines' holistic understanding of the physical world.

Graphical User Interface (GUI) agents have demonstrated remarkable progress in automating complex user interface interactions through reinforcement learning. However, current approaches face a fundamental dilemma: offline RL enables stable training on pre-collected trajectories, but struggles with multi-step task execution for lack of trajectory-level reward signals; online RL captures these signals through environment interaction, but suffers from sparse rewards and prohibitive deployment costs. To address it, we present Semi-online Reinforcement Learning, a novel paradigm that simulates online RL on offline trajectories. During each rollout process, we preserve the original model output within the multi-turn dialogue, where a Patch Module adaptively recovers the divergence between rollout and expert trajectories. To capture long-term training signals, Semi-online RL introduces discounted future returns into the reward computation and optimizes the policy with weighted step-level and episode-level advantages. We further introduce Semi-Online Performance (SOP), a metric that aligns better with true online performance, serving as a practical and effective proxy for real-world evaluation. Experiments show that ours Semi-online RL achieves SOTA performance among 7B models across four dynamic benchmarks, with significant gains over the base model (e.g., +12.0% on AndroidWorld, +23.8% on AITW), demonstrating significant progress in bridging the gap between offline training efficiency and online multi-turn reasoning. The code is available at https://github.com/X-PLUG/MobileAgent/tree/main/UI-S1.

The advancement of Embodied AI heavily relies on large-scale, simulatable 3D scene datasets characterized by scene diversity and realistic layouts. However, existing datasets typically suffer from limitations in data scale or diversity, sanitized layouts lacking small items, and severe object collisions. To address these shortcomings, we introduce InternScenes, a novel large-scale simulatable indoor scene dataset comprising approximately 40,000 diverse scenes by integrating three disparate scene sources, real-world scans, procedurally generated scenes, and designer-created scenes, including 1.96M 3D objects and covering 15 common scene types and 288 object classes. We particularly preserve massive small items in the scenes, resulting in realistic and complex layouts with an average of 41.5 objects per region. Our comprehensive data processing pipeline ensures simulatability by creating real-to-sim replicas for real-world scans, enhances interactivity by incorporating interactive objects into these scenes, and resolves object collisions by physical simulations. We demonstrate the value of InternScenes with two benchmark applications: scene layout generation and point-goal navigation. Both show the new challenges posed by the complex and realistic layouts. More importantly, InternScenes paves the way for scaling up the model training for both tasks, making the generation and navigation in such complex scenes possible. We commit to open-sourcing the data, models, and benchmarks to benefit the whole community.

Supervised Fine-Tuning (SFT) is essential for training large language models (LLMs), significantly enhancing critical capabilities such as instruction following and in-context learning. Nevertheless, creating suitable training datasets tailored for specific domains remains challenging due to unique domain constraints and data scarcity. In this paper, we propose SearchInstruct, an innovative method explicitly designed to construct high quality instruction datasets for SFT. Our approach begins with a limited set of domain specific, human generated questions, which are systematically expanded using a large language model. Subsequently, domain relevant resources are dynamically retrieved to generate accurate and contextually appropriate answers for each augmented question. Experimental evaluation demonstrates that SearchInstruct enhances both the diversity and quality of SFT datasets, leading to measurable improvements in LLM performance within specialized domains. Additionally, we show that beyond dataset generation, the proposed method can also effectively facilitate tasks such as model editing, enabling efficient updates to existing models. To facilitate reproducibility and community adoption, we provide full implementation details, the complete set of generated instruction response pairs, and the source code in a publicly accessible Git repository: [https://github.com/mostafaamiri/SearchInstruct](https://github.com/mostafaamiri/SearchInstruct)

The reliance on implicit point matching via attention has become a core bottleneck in drag-based editing, resulting in a fundamental compromise on weakened inversion strength and costly test-time optimization (TTO). This compromise severely limits the generative capabilities of diffusion models, suppressing high-fidelity inpainting and text-guided creation. In this paper, we introduce LazyDrag, the first drag-based image editing method for Multi-Modal Diffusion Transformers, which directly eliminates the reliance on implicit point matching. In concrete terms, our method generates an explicit correspondence map from user drag inputs as a reliable reference to boost the attention control. This reliable reference opens the potential for a stable full-strength inversion process, which is the first in the drag-based editing task. It obviates the necessity for TTO and unlocks the generative capability of models. Therefore, LazyDrag naturally unifies precise geometric control with text guidance, enabling complex edits that were previously out of reach: opening the mouth of a dog and inpainting its interior, generating new objects like a ``tennis ball'', or for ambiguous drags, making context-aware changes like moving a hand into a pocket. Additionally, LazyDrag supports multi-round workflows with simultaneous move and scale operations. Evaluated on the DragBench, our method outperforms baselines in drag accuracy and perceptual quality, as validated by VIEScore and human evaluation. LazyDrag not only establishes new state-of-the-art performance, but also paves a new way to editing paradigms.

Among generative models, diffusion models are uniquely intriguing due to the existence of a closed-form optimal minimizer of their training objective, often referred to as the optimal denoiser. However, diffusion using this optimal denoiser merely reproduces images in the training set and hence fails to capture the behavior of deep diffusion models. Recent work has attempted to characterize this gap between the optimal denoiser and deep diffusion models, proposing analytical, training-free models that can generate images that resemble those generated by a trained UNet. The best-performing method hypothesizes that shift equivariance and locality inductive biases of convolutional neural networks are the cause of the performance gap, hence incorporating these assumptions into its analytical model. In this work, we present evidence that the locality in deep diffusion models emerges as a statistical property of the image dataset, not due to the inductive bias of convolutional neural networks. Specifically, we demonstrate that an optimal parametric linear denoiser exhibits similar locality properties to the deep neural denoisers. We further show, both theoretically and experimentally, that this locality arises directly from the pixel correlations present in natural image datasets. Finally, we use these insights to craft an analytical denoiser that better matches scores predicted by a deep diffusion model than the prior expert-crafted alternative.

IDC 的数据显示，在 2025 年 1～6 月扫地机器人全球市场份额居前 5 的企业中，中国企业占 4 家：北京石头世纪科技、科沃斯、追觅科技市场份额均超过 10%，小米 7.4%。2010 年代曾占据三成以上份额的 iRobot 仅占 5.8% 跌至第 5 位。韩国的两大巨头三星电子和 LG 电子也都有扫地机器人产品，但相比中国产品零售价在 965~1448 元，韩国产的扫地机器人普遍在 4826~9653 元之间，且性能无法让消费者认为值得付出更高的代价。由于无法在价格上与中国产品竞争，韩国厂商通过强调安全性、高档感和易用性，力求在价格以外实现差异化，这种战略与日本厂商也有相似之处。

伊斯坦布尔妇产科医生 Furkan Kayabasoglu 过去经常会为同一个家庭多次接生，如今绝大部分家庭只生一胎。去年土耳其的总和生育率低至 1.48，已经远低于 2.1 的人口更替率。而联合国人口司此前预测土耳其要到 2100 年总和生育率才会到达如此低的水平。土耳其并非例外。世界各地，从发展中国家到中等收入国家到发达国家，生育率的下降幅度都远超预期。哥伦比亚首都波哥大的总和生育率只有 0.91，甚至低于日本东京。印度的总和生育率已低于人口更替率，中国的人口已经开始萎缩。墨西哥的总和生育率为1.6，与美国相当。2024 年法国出生人口数低于 1806 年，而当时人口不到今天的一半。意大利出生人数创下了 1861 年统一以来的最低水平。非洲出生率仍然远高于全球平均水平，但人口下降速度也远超预期。这一切意味着世界人口可能比专家预测的更早达到峰值，且峰值水平要低得多。世界人口可能到 2050 年代就会停止增长，而不是原来预测的 2084 年，且总数不会超过 90 亿。生育率下降的可能因素包括了女性受教育程度提高、城市化、避孕普及、养育成本急剧上升以及社会观念的转变等。

日本总务省周日公布了人口估算数据，65 岁以上老年人为 3619 万人，占总人口的比例为 29.4%，创下新高，该比例也是人口 4000 万以上国家中最高。老年人就业人数为 930万 ，连续 21 年增加，也创下新高，除了有更多老年人健康良好外，少子化导致人手短缺也是原因之一。人口估算显示，65 岁以上男性为 1568 万人，女性 2051 万人，总数较上年减少 5 万人。这是有可比数据的 1950 年以来继 2023 年后第二次减少。主要原因是新达到 65 岁的人数较少。国立社会保障和人口问题研究所估算，由于第二次婴儿潮（1971-1974 年）出生的一代逐渐进入老年，估计 2040 年老年人口将达到 3928 万人，占总人口的 34.8%。

基因编辑的猪和绵羊等动物正逐渐在农业领域获得认可。这些技术可提升动物的性状表现，为人类提供更安全、优质的肉类产品。但经 CRISPR 技术改造的马，却被马球比赛拒之门外。专家强调，必须严格追踪并确保基因编辑动物的安全性，审慎推进相关应用。CRISPR 基因编辑技术能够精准切割基因组特定位置，改变基因表达，从而赋予生物新的性状。Kheiron 公司以阿根廷一匹冠军马为原型，利用克隆技术培育出五匹遗传背景完全一致的克隆马。在此基础上，研究人员进一步应用 CRISPR 技术，靶向抑制了肌生成抑制素基因的表达。该基因天然存在于动物体内，作用是限制肌肉过度发育。通过精准下调其活性，团队增加了马匹体内负责爆发性运动的肌纤维数量，从而将它们培育成更出色的“短跑健将”。阿根廷马球协会明确禁止基因编辑马参赛。协会主席表示，这项技术“会剥夺育种的魅力与魔法”。

日本百岁老人数量接近 10 万，创历史新高，其中最年长老人年龄 114 岁。根据日本厚生劳动省上周五公布的数据，截至 9 月，日本百岁老人数量 99,763 人，连续 55 年创新高。其中女性 87,784 名占 88%，男性 11,979 名。预期寿命提高主要归因于心脏病和常见癌症如乳腺癌和前列腺癌死亡人数减少。由于日本饮食中红肉摄入量较少、鱼类和蔬菜摄入量较高，肥胖率较低，而肥胖是导致上述两类疾病的主要因素。女性的肥胖率尤其低，这或许可以解释为什么日本女性预期寿命远高于男性。此外日本人在晚年仍然保持活跃，比美国和欧洲老年人更多地步行和使用公共交通工具。

新闻评级公司 NewsGuard 调查了 10 款领先的生成式 AI 工具，分析了它们在回复中生成虚假新闻信息的比例。结果显示，2025 年 8 月，10 款 AI 工具在新闻主题上重复虚假信息的比例超过三分之一（35%），高于 2024 年 8 月的 18%。AI 公司并未兑现让 AI 更安全更可靠的承诺。生成虚假信息比例翻一倍的一大原因是今天的 AI 工具都支持联网查询，不再拒绝回答提问，它们不回复比例从 2024 年 8 月的 31% 下降到 2025 年 8 月的 0%，结果就是更多虚假信息。NewsGuard 认为攻击者正利用 AI 这一特点用各种方法洗白虚假信息，让 AI 模型无法区分内容农场和可信新闻渠道。

根据发表在《自然》期刊上的一项研究，一种名为伊比利亚收获蚁（Messor ibericus）的蚁后可以产下两个不同物种的蚂蚁。蚁群里的个体通常分为三类，包括蚁后，雄蚁以及工蚁。研究发现，伊比利亚收获蚁的工蚁体内带有另一种“工匠收获蚁”（Messor structor）的基因，但这并非杂交的结果，而是蚁后从工匠收获蚁雄蚁的精子直接克隆出来的“复制品”。这一“异种生殖”现象不仅刷新了人类对蚂蚁生殖机制的认知，还首次提供了雌性主动“繁殖”另一物种的直接证据。研究人员推断，这个演化故事的起点是一种被称为“精子寄生”（sperm parasitism）的现象。在数百万年前，伊比利亚收获蚁的蚁后因某种未知原因失去了生产本物种工蚁的能力。为了维持蚁群的运作，它们不得不“借用”生活在附近的工匠收获蚁雄蚁的精子，与自己的卵结合，生产出兼具两个物种基因的混血工蚁。然而这种生殖策略严重依赖于建造收获蚁种群的地理邻近性，对于蚁后来说，寻找异种雄蚁交配是一件既耗时又不稳定的“麻烦事”。为了摆脱这种束缚，伊比利亚收获蚁演化出了一种更为高效且惊人的策略——“性驯化”（sexual domestication）。它们不再需要在野外寻找工匠收获蚁雄蚁，而是直接利用储存在自己体内的异种精子，通过一种特殊的克隆过程来生产它们。

由 Amazon、Cisco, Google、Intel、Microsoft、Mozilla 和 Netflix 等联合创办的开放媒体联盟 AOMedia 宣布将于年底发布 AV1 的后继者 AV2 编解码器。AOMedia 声称，AV2 是开放视频编码的一次世代飞跃，旨在满足全球日益增长的流媒体需求，压缩性能显著优于 AV1。AV2 增强了对 AR/VR 应用的支持，支持多节目分屏播放，改进屏幕内容处理，能在更宽的视觉质量范围内运行。

开源游戏引擎 Godot 释出了 v4.5 版本。主要新特性包括：模板缓存（stencil buffer），内置屏幕阅读器支持以改进可访问性，着色器烘焙器（shader baker）提供更好的着色器编译处理加速启动，改进物理功能等等。Linux 版 Godot 4.5 支持原生 Wayland 子窗口，基于 WebAssembly 的 Web 版本支持 SIMD 显著提升了性能。

微软宣布从 10 月份起，在欧盟经济区（EEA）外的 Microsoft 365 应用将强制安装 Copilot Chat。Word、 Excel、PowerPoint、Outlook 和 OneNote 都将更新包含 Copilot Chat 侧边栏。用户利用 Copilot Chat 可以起草文档、分析电子表格和制作幻灯片。该功能可以免费使用，Copilot 的付费用户则可以访问更高级的功能如对工作数据进行推理、支持上传文件和生成图像，以及使用最新模型如 GPT-5。如果企业不想要该功能，IT 管理员可以在 Apps Admin Center 中修改设置退出 Copilot Chat，方法是 Customization > Device Configuration > Modern App Settings，选择 Microsoft 365 Copilot app，移除 Enable 的勾选框。

过去十年 Google 每个月都会发布 Android Security Bulletin，详细介绍了当月释出的 Android 安全更新所修复的漏洞，给出漏洞的危险程度。但 2025 年 7 月它打破了这一惯例，首次没有列出任何漏洞。而 2025 年 9 月的 Security Bulletin 披露了多达 119 个安全漏洞。Google 已经将过去十年采用的 Android 每月安全更新重组为“基于风险的更新系统”，区分高优先级补丁和常规修复。每个月的 Security Bulletin 将只列出正被活跃利用或是已知漏洞利用链中的漏洞，大部分补丁将累积到季度发布。此举将大幅减少 OEM 厂商月度更新的工作量。Google 也不再发布月度安全更新源代码，将自定义 ROM 开发限制在季度周期内。

美国财长 Scott Bessent 宣布中美两国就 TikTok 美国业务达成初步框架协议，两国元首将在周五敲定细节。TikTok 母公司字节跳动必须在 9 月 17 日期限之前，向非中国买家出售或剥离其美国业务，否则 TikTok 面临在美下架。特朗普此前已三次延长期限。中国国务院副总理何立峰表示，中方维护自身正当权益的决心坚定不移，将坚决维护国家利益和海外中资企业的合法权益。对于TikTok问题，中方将依法依规开展技术出口审批。同时，中国政府充分尊重本国海外企业意愿，支持企业在符合市场原则基础上，与合作方开展平等商业谈判。