Weekly Digest — 2026-W21

AI computer screen and voice control with custom automation

Build AR/VR Apps in React Native + ship directly to devices

Your Chief Agent Operator for multi-agent work

Scrape emails from socials and maps by location

Your Mac wallpaper that listens to your music & weather

Sandboxes for your AI agents

Cursor’s most powerful model yet

A multi-agent world model you can play

Auto-organize files on your Mac

Noise-cancelling for your  screen

Chat-based calorie tracker that actually knows desi food

The modern way to teach what you build

Multimodal Large Language Models (MLLMs) have significantly advanced document understanding, yet current Doc-VQA evaluations score only the final answer and leave the supporting evidence unchecked. This answer-only approach masks a critical failure mode: a model can land on the correct answer while grounding it in the wrong passage -- a critical risk in high-stakes domains like law, finance, and medicine, where every conclusion must be traceable to a specific source region. To address this, we introduce CiteVQA, a benchmark that requires models to return element-level bounding-box citations alongside each answer, evaluating both jointly. CiteVQA comprises 1,897 questions across 711 PDFs spanning seven domains and two languages, averaging 40.6 pages per document. To ensure fidelity and scalability, the ground-truth citations are generated by an automated pipeline-which identifies crucial evidence via masking ablation-and are subsequently validated through expert review. At the core of our evaluation is Strict Attributed Accuracy (SAA), which credits a prediction only when the answer and the cited region are both correct. Auditing 20 MLLMs reveals a pervasive Attribution Hallucination: models frequently produce the right answer while citing the wrong region. The strongest system (Gemini-3.1-Pro-Preview) achieves an SAA of only 76.0, and the strongest open-source MLLM reaches just 22.5. Ultimately, towards trustworthy document intelligence, CiteVQA exposes a reliability gap that answer-only evaluations overlook, providing the instrumentation needed to close it. Our repository is available at https://github.com/opendatalab/CiteVQA.

Vision-language-action models have advanced rapidly, but robot trajectories alone provide limited coverage for learning broad physical understanding. PhysBrain 1.0 studies a complementary route: converting large-scale human egocentric video into structured physical commonsense supervision before robot adaptation. Our data engine extracts scene elements, spatial dynamics, action execution, and depth-aware relations, then turns them into question-answer supervision for training PhysBrain VLMs. The resulting physical priors are further transferred to VLA policies through a capability-preserving and language-sensitive adaptation design. Across multimodal QA benchmarks and embodied control benchmarks, including ERQA, PhysBench, SimplerEnv-WidowX, LIBERO, and RoboCasa, PhysBrain 1.0 achieves SOTA results and shows especially strong out-of-domain performance on SimplerEnv. These results suggest that scaling physical commonsense from human interaction video can provide an effective bridge from multimodal understanding to robot action.

Reusable skills have become a core substrate for improving agent capabilities, yet most existing skill packages encode reusable behavior primarily as textual prompts, executable code, or learned routines. For visual agents, however, procedural knowledge is inherently multimodal: reuse depends not only on what operation to perform, but also on recognizing the relevant state, interpreting visual evidence of progress or failure, and deciding what to do next. We formalize this requirement as multimodal procedural knowledge and address three practical challenges: (I) what a multimodal skill package should contain; (II) where such packages can be derived from public interaction experience; and (III) how agents can consult multimodal evidence at inference time without excessive image context or over-anchoring to reference screenshots. We introduce MMSkills, a framework for representing, generating, and using reusable multimodal procedures for runtime visual decision making. Each MMSkill is a compact, state-conditioned package that couples a textual procedure with runtime state cards and multi-view keyframes. To construct these packages, we develop an agentic trajectory-to-skill Generator that transforms public non-evaluation trajectories into reusable multimodal skills through workflow grouping, procedure induction, visual grounding, and meta-skill-guided auditing. To use them, we introduce a branch-loaded multimodal skill agent: selected state cards and keyframes are inspected in a temporary branch, aligned with the live environment, and distilled into structured guidance for the main agent. Experiments across GUI and game-based visual-agent benchmarks show that MMSkills consistently improve both frontier and smaller multimodal agents, suggesting that external multimodal procedural knowledge complements model-internal priors.

Human-centric video customization, particularly at the garment level, has shown significant commercial value. However, existing approaches cannot support low-latency and interactive garment control, which is crucial for applications such as e-commerce and content creation. This paper studies how to achieve interactive multi-garment video customization while preserving motion coherence using only single-garment video data. We present FashionChameleon, a real-time and interactive framework for human-garment customization in autoregressive video generation, where users can interactively switch garment during generation. FashionChameleon consists of three key techniques: (i) Instead of training on multi-garment video data, we train a Teacher Model with In-Context Learning on a single reference-garment pair. By retaining the image-to-video training paradigm while enforcing a mismatch between the reference and garment image, the model is encouraged to implicitly preserve coherence during single-garment switching. (ii) To achieve consistency and efficiency during generation, we introduce Streaming Distillation with In-Context Learning, which fine-tunes the model with in-context teacher forcing and improves extrapolation consistency via gradient-reweighted distribution matching distillation. (iii) To extend the model for interactive multi-garment video customization, we propose Training-Free KV Cache Rescheduling, which includes garment KV refresh, historical KV withdraw, and reference KV disentangle to achieve garment switching while preserving motion coherence. Our FashionChameleon uniquely supports interactive customization and consistent long-video extrapolation, while achieving real-time generation at 23.8 FPS on a single GPU, 30-180times faster than existing baselines.

On-policy distillation (OPD) has emerged as an efficient post-training paradigm for large language models. However, existing studies largely attribute this advantage to denser and more stable supervision, while the parameter-level mechanisms underlying OPD's efficiency remain poorly understood. In this work, we argue that OPD's efficiency stems from a form of ``foresight'': it establishes a stable update trajectory toward the final model early in training. This foresight manifests in two aspects. First, at the Module-Allocation Level, OPD identifies regions with low marginal utility and concentrates updates on modules that are more critical to reasoning. Second, at the Update-Direction Level, OPD exhibits stronger low-rank concentration, with its dominant subspaces aligning closely with the final update subspace early in training. Building on these findings, we propose EffOPD, a plug-and-play acceleration method that speeds up OPD by adaptively selecting an extrapolation step size and moving along the current update direction. EffOPD requires no additional trainable modules or complex hyperparameter tuning, and achieves an average training acceleration of 3times while maintaining comparable final performance. Overall, our findings provide a parameter-dynamics perspective for understanding the efficiency of OPD and offer practical insights for designing more efficient post-training methods for large language models.

Achieving human-level manipulation requires dexterous robotic hands capable of complex object interactions. Advancing such capabilities further demands standardized benchmarks for systematic evaluation. However, existing dexterous benchmarks lack tasks that reflect the unique manipulation capabilities of dexterous hands over parallel grippers, as well as comprehensive evaluation pipelines. In this paper, we present DexJoCo, a benchmark and toolkit for task-oriented dexterous manipulation, comprising 11 functionally grounded tasks that evaluate tool-use, bimanual coordination, long-horizon execution, and reasoning. We develop a low-cost data collection system and collect 1.1K trajectories across these tasks, with support for domain randomization to assess robustness. We benchmark modern models under diverse settings, including visual and dynamics randomization, multi-task training, and action-head adaptation. Through extensive empirical analysis, we identify several important insights and common limitations of current policies in dexterous manipulation, highlighting key challenges for future research in dexterous hand robot learning. Project page available at: https://dexjoco.github.io

Recent large language models (LLMs) have demonstrated strong capabilities in understanding and generating code, from competitive programming to repository-level software engineering. In emerging agentic systems, code is no longer only a target output. It increasingly serves as an operational substrate for agent reasoning, acting, environment modeling, and execution-based verification. We frame this shift through the lens of agent harnesses and introduce code as agent harness: a unified view that centers code as the basis for agent infrastructure. To systematically study this perspective, we organize the survey around three connected layers. First, we study the harness interface, where code connects agents to reasoning, action, and environment modeling. Second, we examine harness mechanisms: planning, memory, and tool use for long-horizon execution, together with feedback-driven control and optimization that make harness reliable and adaptive. Third, we discuss scaling the harness from single-agent systems to multi-agent settings, where shared code artifacts support multi-agent coordination, review, and verification. Across these layers, we summarize representative methods and practical applications of code as agent harness, spanning coding assistants, GUI/OS automation, embodied agents, scientific discovery, personalization and recommendation, DevOps, and enterprise workflows. We further outline open challenges for harness engineering, including evaluation beyond final task success, verification under incomplete feedback, regression-free harness improvement, consistent shared state across multiple agents, human oversight for safety-critical actions, and extensions to multimodal environments. By centering code as the harness of agentic AI, this survey provides a unified roadmap toward executable, verifiable, and stateful AI agent systems.

Long-horizon LLM agents leave traces that could become reusable experience, but raw trajectories are noisy and hard to govern. We treat Agent Skills as an experience schema that couples executable scripts, with non-executable guidance on procedures. Yet open skill ecosystems contain redundant, uneven, environment-sensitive artifacts, and indiscriminate updates can pollute future context. We present SkillsVote, a lifecycle-governance framework for Agent Skills from collection and recommendation to evolution. SkillsVote profiles a million-scale open-source corpus for environment requirements, quality, and verifiability, then synthesizes tasks for verifiable skills. Before execution, SkillsVote performs agentic library search over structured skill library to expose instructional skill context. After execution, it decomposes trajectories into skill-linked subtasks, attributes outcomes to skill use, agent exploration, environment, and result signals, and admits only successful reusable discoveries to evidence-gated updates. In our evaluation, offline evolution improves GPT-5.2 on Terminal-Bench 2.0 by up to 7.9 pp, while online evolution improves SWE-Bench Pro by up to 2.6 pp. Overall, governed external skill libraries can improve frozen agents without model updates when systems control exposure, credit, and preservation.

We present LongLive-2.0, an NVFP4-based parallel infrastructure throughout the full training and inference workflow of long video generation, addressing speed and memory bottlenecks. For training, we introduce sequence-parallel autoregressive (AR) training, instantiated as Balanced SP, which co-designs the efficient teacher-forcing layout with SP execution by pairing clean-history and noisy-target temporal chunks on each rank, enabling a natural teacher-forcing mask with SP-aware chunked VAE encoding. Combined with NVFP4 precision, it reduces GPU memory cost and accelerates GEMM computation during training, the proportion of which increases as video length grows. Moreover, we show that a high-quality infrastructure and dataset enable a remarkably clean training pipeline. Unlike existing Self-Forcing series methods that rely on ODE initialization and subsequent distribution matching distillation (DMD), LongLive-2.0 directly tunes a diffusion model into a long, multi-shot, interactive auto-regressive (AR) diffusion model. It can be further converted to real-time generation (4 to 2 denoising steps) with standalone LoRA weights. For inference on Blackwell GPUs, we enable W4A4 NVFP4 inference, quantize KV cache into NVFP4 for memory savings, and boost end-to-end throughput with asynchronous streaming VAE decoding. On non-Blackwell GPU architectures, we deploy SP inference to match the speed on Blackwell GPUs, while the quantized KV cache can lower inter-GPU communication of SP. Experiments show up to 2.15x speedup in training, and 1.84x in inference. LongLive-2.0-5B achieves 45.7 FPS inference while attaining strong performance on benchmarks. To our knowledge, LongLive-2.0 is the first NVFP4 training and inference system for long video generation.

We present Lance, a lightweight native unified model supporting multimodal understanding, generation, and editing for both images and videos. Rather than relying on model capacity scaling or text-image-dominant designs, Lance explores a practical paradigm for unified multimodal modeling via collaborative multi-task training. It is grounded in two core principles: unified context modeling and decoupled capability pathways. Specifically, Lance is trained from scratch and employs a dual-stream mixture-of-experts architecture on shared interleaved multimodal sequences, enabling joint context learning while decoupling the pathways for understanding and generation. We further introduce modality-aware rotary positional encoding to mitigate interference among heterogeneous visual tokens and boost cross-task alignment. During training, Lance adopts a staged multi-task training paradigm with capability-oriented objectives and adaptive data scheduling to strengthen both semantic comprehension and visual generation performance. Experimental results demonstrate that Lance substantially outperforms existing open-source unified models in image and video generation, while retaining strong multimodal understanding capabilities. The homepage is available at https://lance-project.github.io.

AI-assisted research is crossing a threshold: fully automated systems can now generate research papers for as little as $15, while long-horizon agents can execute experiments, draft manuscripts, and simulate critique with minimal human input. Yet this productivity frontier exposes a deeper integrity problem: under scientific pressure, even frontier LLMs still fabricate results, miss hidden errors, and fail to judge novelty reliably. Studying developments through April 2026, we present an end-to-end analysis of AI across the complete research lifecycle, organized into four epistemological phases: Creation (idea generation, literature review, coding & experiments, tables & figures), Writing (paper writing), Validation (peer review, rebuttal & revision), and Dissemination (posters, slides, videos, social media, project pages, and interactive agents). We identify a sharp, stage-dependent boundary between reliable assistance and unreliable autonomy: AI excels at structured, retrieval-grounded, and tool-mediated tasks, but remains fragile for genuinely novel ideas, research-level experiments, and scientific judgment. Generated ideas often degrade after implementation, research code lags far behind pattern-matching benchmarks, and end-to-end autonomous systems have not yet consistently reached major-venue acceptance standards. We further show that greater automation can obscure rather than eliminate failure modes, making human-governed collaboration the most credible deployment paradigm. Finally, we provide a structured taxonomy, benchmark suite, and tool inventory, cross-stage design principles, and a practitioner-oriented playbook, with resources maintained at our project page.

End-to-end automation of realistic healthcare operations stresses three capabilities underrepresented in current benchmarks: policy density, decisions must be grounded in a large library of medical, insurance, and operational rules; Multi-role composition: a single task requires the agent to play multiple roles with handoffs; and multilateral interaction: intermediate workflow steps are multi-turn dialogs, such as peer-to-peer review and patient outreach. We introduce χ-Bench, a benchmark of long-horizon healthcare workflows across three domains: provider prior authorization, payer utilization management, and care management. Each task hands the agent a clinical case in a high-fidelity simulator of 20 healthcare apps exposed via 87 MCP tools, which it must drive to a terminal status through tool calls and writing the role's artifacts, guided by a 1,290+ document managed-care operations handbook skill. Across 30 agent harness/models configurations, the best agent resolves only 28.0% of tasks, no agent clears 20% on strict pass^3, and executing all tasks in a single session slumps the performance to 3.8%. These results raise the hypothesis that similar gaps are likely to surface in other policy-dense, role-composed, irreversible enterprise domains.

Eli Tan / New York Times : Internal memo: Meta is reassigning 7,000 workers to four new units focused on building AI tools, two days before it is set to lay off 10% of its workforce —  The company announced the changes two days before it plans to lay off 10 percent of its work force, or about 8,000 employees.

Bloomberg : Sources: AI chip designer Tenstorrent has drawn takeover interest from Intel and Qualcomm; Tenstorrent could be valued at more than $5B in a potential deal —  Artificial intelligence chip startup Tenstorrent Inc. is drawing early takeover interest from prospective buyers at a moment …

Amrith Ramkumar / Wall Street Journal : Anthropic last week began letting Mythos users share cybersecurity threats with others who may face similar vulnerabilities, modifying its previous stance —  How to restrict access while still allowing users to share threat information is a major challenge facing AI companies

Elon Musk / @elonmusk : Musk v. Altman: Elon Musk says the judge and jury “never actually ruled on the merits of the case, just on a calendar technicality” and he will file an appeal —  Regarding the OpenAI case, the judge & jury never actually ruled on the merits of the case, just on a calendar technicality. There is no question to anyone following the case in detail that Altman & Brockman did in fact enrich themselves by stealing a charity. The only question

David Morris / Bloomberg : Akamai is seeking to raise $2.6B in a convertible bond offering, and plans to use $350M of the offering to buy back its common stock from buyers of the bonds —  Akamai Technologies Inc. is seeking to raise $2.6 billion in a convertible bond offering to fund spending on cloud computing infrastructure.

Grant Bourzikas / Cloudflare : Cloudflare tests Mythos against 50+ repositories, highlights its ability to chain bugs into a single exploit, and details a vulnerability discovery harness —  For the last few months, we've been testing a range of security-focused LLMs on our own infrastructure.

OpenAI : OpenAI introduces Guaranteed Capacity, a new offering that lets customers guarantee access to OpenAI's compute through one- to three-year commitments —  Guarantee long-term access to OpenAI compute for the products, agents, and customer workflows that matter most.  —  Plan for capacity Contact sales

New York Times : Meta begins laying off 8,000 employees, or 10% of staff, in a push to become an AI-first company; another 7,000 workers will be reassigned to AI initiatives —  Meta told employees last month that it would carry out mass layoffs on May 20, as the Silicon Valley giant tries to transform into an A.I.-first company.

Bobby Allyn / NPR : Minnesota Gov. Tim Walz has signed the nation's first law banning prediction market sites from operating in the state; the CFTC has sued Minnesota in response —  Minnesota Gov. Tim Walz has signed the nation's first law banning prediction market sites from operating in the state …

Bloomberg : Sources: Google DeepMind has reached a ~$100M deal to hire 20+ researchers from Contextual AI, including CEO Douwe Kiela, and license its technology —  Google DeepMind has reached a deal to hire more than 20 researchers from artificial intelligence startup Contextual AI and license its technology …

Mat Smith / Engadget : Google unveils Pics, an AI image editor in Workspace that lets users edit specific elements and modify text, rolling out this summer to AI Pro and Ultra users —  It's not Photoshop, but it could be better than what's in Google Photos.  —  Alongside an array of updates across its Workspace apps …

Meir Orbach / CTech : Ocean, which uses AI agents to detect email attacks, raised a $20M Series A led by Lightspeed, following an $8M seed in 2024 —  The cybersecurity startup says its platform replaces legacy phishing defenses with AI agents that analyze intent, not just patterns.

根据发表在《Cell》期刊上的一项研究，研究人员通过分析欧洲、东亚和南亚的 322 名健康人去构建迄今最详尽的遗传祖先和环境如何塑造人类生物学特征的图谱。通过招募居住在不同大洲、具有相同遗传背景的人群，科学家得以以前所未有的清晰度，将 DNA 的影响与环境的影响区分开来。研究人员发现，无论搬到哪里，种族背景会对免疫系统、新陈代谢和肠道菌群产生深远影响。南亚人表现出更高的病原体暴露水平。欧洲人的肠道微生物多样性更丰富，且与心脏病风险相关的化合物含量更高。跨州迁移会改变主要的代谢途径，改变肠道微生物的平衡。研究的一大发现是你生活的地点与你衰老的速度相关。居住在亚洲外的东亚人比东亚人生物年龄更大。欧洲人则相反，居住在欧洲外的欧洲人生物年龄更小。

伊朗军方发言人 Ebrahim Zolfaghari 在 X 上宣布对通过霍尔木兹海峡的海底光缆收费。暂时不清楚伊朗只是发出一种威胁，还是会将威胁付诸实施。伊朗的计划将要求 Google、微软、Meta和亚马逊等公司遵守其法律，同时海底光缆公司被要求支付通行许可费，而维修和维护权则完全授予伊朗公司。海底光缆传输着欧洲、亚洲和波斯湾之间的网络和金融流量，破坏光缆将会引发数字灾难，威胁到从银行系统、军事通信和 AI 云基础设施到远程办公、在线游戏和流媒体服务。

安全研究员本月初披露，Edge 内置的密码管理器会在浏览器启动时候解密所有密码然后加载到内存里。他联络了微软，结果收到的回应是“源于设计（by design）”，认为不是安全隐患。微软当时强调这是应用的预期功能。然而仅仅过了几天，微软就改变了主意，宣布未来版本的 Edge 不会再在启动时加载密码。相关补丁已经释出到 Edge Canary 版本，将包含在 Edge build 148 或更新版本中。

独立沙盒游戏《Terraria》的开发商宣布游戏上市 15 年共售出了七千万份拷贝，其中 PC 平台销量最高 3960 万份，主机 1070 万份，移动 1970 万份，Mod 工具 tModLoader 下载量 1230 万次。过去一年《Terraria》PC 版本日均玩家 46.1 万最高 140 万，PC 玩家平均游戏时长 101 小时 18 分钟。开发商表示会继续更新《Terraria》。在史上最畅销的游戏中，《Terraria》排在第 7 位。销量最高的是《我的世界（Minecraft）》（如果不考虑《俄罗斯方块》），售出超过 3.5 亿份拷贝，其次是《Grand Theft Auto V》的 2.25 亿份，《Wii Sports》的 8290 万份、《Red Dead Redemption 2》的 8200 万份，《马里奥赛车8》的 7954 万份， 《绝地求生》的 7500 万份等。

三星电子工会已经宣布将于 21 日启动为期 18 天的总罢工，双方就绩效奖金的上限存在分歧。目前韩国政府对此事表达了高度关注，总理金民锡周日表示若罢工对国民经济造成巨大损失，政府将为保护国民经济而采取包括行使紧急调整权在内的所有可行手段。周一三星电子劳资双方展开了最新一轮谈判。韩国法院同一天就三星电子资方针对工会提出的禁止其进行违法集体斗争行为的申请作出裁定，支持资方大部分诉求，要求工会即便罢工也不得耽误生产。该决定或给劳资谈判以及工会的罢工计划带来不小影响。专家估计每罢工一天造成的损失最高达到 20 亿美元，18 天总罢工将接近 170 亿美元。而 JPMorgan 估计损失最高将达到 280 亿美元。

NASA 在 48 年前先后发射了两艘旅行者号探测器，当年曾为旅行者号写代码的工程师如今早已白发苍苍，甚至已经去世。旅行者号机载计算机运行的是汇编语言，是专为通用电气开发的处理器编写的。探测器上有三个计算机系统：计算机指令子系统（CCS）、姿态调节控制子系统（ACS）以及飞行数据子系统（FDS）。其底层飞行工作依赖于专门的汇编语言，地面系统和早期任务工具使用了 Fortran 语言。探测器上的计算机内存非常小，总容量仅为 64-70 KB。几十年来，地面控制团队成员不断减少，也逐渐老去。更糟的是很多原始文档遗失或支离破碎。项目文件大多是纸质的，每次项目搬迁办公室，会有更多的文件丢失。NASA JPL Interplanetary Network Directorate 项目主任 Suzy Dodd 在 2024 年称建造探测器的人都已经不在人世。Larry Zottarelli 是最后一位仍在工作的原始团队工程师，他于 2016 年 80 岁时退休。目前旅行者号维护团队大多数人都年过八旬，团队还依赖于一份退休工程师名单，以便在紧急情况下呼叫。该名单每年都在缩小。

上月底，PostgreSQL 备份恢复项目 pgBackRest 的维护者 David Steele 宣布项目存档停止维护。pgBackRest 被广泛视为是 PostgreSQL 生态系统最流行的运维工具之一。Steele 解释说，过去 13 年 pgBackRest 是他倾注热情的项目，幸运的是大部分时间里他都有企业资助，他的长期赞助商是 Crunchy Data 公司，但这家公司被 Snowflake 收购了，而新东家无意资助他继续从事相关工作，因此他过去几个月一直在寻找继续这项工作的职位但没有成功，获得的赞助也远远未能达到维持项目运营所需的金额，因此只能宣布停止维护。在这一声明公布数周之后，他更新了消息，宣布将继续开发 pgBackRes：因为一个赞助商联盟同意为项目持续提供资金，给予了 pgBackRes 开发所需的长期稳定性，他对此表示了感谢。

负责索尼 PS 工作室业务的高管 Hermen Hulst 周一证实了此前的流言：取消将 PS 独占单人游戏移植到 PC 的计划。索尼过去几年将此前的独占 PS 单人游戏如 God of War 系列、Spider-Man 系列、Ghost of Tsushima、The Last of Us 系列和 Horizon Zero Dawn 系列移植到了 PC 平台，但最近一段时间移植频率下降，引发了索尼改变移植战略的流言。Hermen Hulst 周一在员工大会上宣布了公司的战略调整计划。索尼据称是担心稀释 PlayStation 品牌影响力。此举意味着索尼最近推出的单人游戏 Ghost of Yotei 和 Saros 将会无缘登陆 PC。索尼的战略调整针对的是第一方工作室的单人游戏，多人游戏以及第三方工作室的单人游戏仍然会登陆 PC。

人类中的大多数是右撇子，左撇子占约十分之一。为什么会出现这一倾向？研究人员分析了 41 种灵长类动物，共计 2025 只猴子与猿类的数据，逐一分析了工具使用、食性、栖息环境、体型、社会结构、脑容量、行动方式等各类影响因素。人类的用手倾向与其他灵长类动物存在明显差异。当研究人员将两个关键特征纳入模型中，情况就发生了变化。这两个特征分别是大脑大小及臂长与腿长的比例，这一比例常作为衡量两足行走能力的指标。纳入上述因素后，人类不再被视为特殊的进化产物。研究结果表明，直立行走与脑容量增大的共同作用，或是人类形成强烈右手使用偏好的核心原因。研究人员认为，惯用右手的进化分为两个阶段。首先，直立行走使双手从运动中解放出来，偏爱更专业和不对称的手部使用；其次，随着人类大脑变得更大且更为复杂，对右手的偏好变得愈发强烈且更为普遍。

Mozilla 释出了 Firefox 151。主要新特性包括：更新内置 VPN 支持，改进隐私浏览，Firefox PDF 查看器支持直接合并多个 PDF 文件，Linux 和 macOS 本地配置文件备份支持跨平台恢复，文档画中画 API——提供了比目前的视频画中画 API 更多功能体验，等等。JPEG-XL 原生图像解密器推迟到了下个版本。

根据发表在《国家科学评论》期刊上的一项研究，中科院研究团队汇总 588 个湖泊的高精度实测水下地形和水深数据。研究发现，我国湖泊水深受地形地貌影响，西部高海拔内流湖盆区受构造断陷和冰川侵蚀影响，形成了深水湖泊，而东部平原因长期泥沙淤积，形成浅碟形湖泊。全国湖泊总蓄水量约 1081-1285 立方公里，其中淡水约 335 立方公里，咸水约 839 立方公里。约 65% 的湖泊淡水储存于青藏高原等西部内流湖盆区少数几个深水开放型湖泊。学界对我国淡水湖的关注多聚焦于东部平原区及云贵高原，但本研究发现，青藏高原不仅拥有塔若错、玛旁雍错、吴如错等超大型深水淡水湖，其湖区天然湖泊的淡水总储量超过东部平原湖区：青藏高原湖泊区人均储量约为 20680 立方米，而东部平原湖泊区人均储量仅为 65 立方米，两者相差近 330 倍。

Kubernetes 联合创始人、微软副总裁 Brendan Burns 在北美开源峰会上突然宣布了一个通用 Linux 发行版。微软以前发布过 Linux 应用，针对边缘计算设备的 Azure Sphere，Linux 容器软件平台 CBL-Marnier——后更名为 Azure Linux，但此前从未发布过通用发行版。微软 Azure 开源团队首席项目经理 Lachlan Everson 表示，通过 Azure Linux 4.0，微软正致力于将 Azure Linux 转变成一个功能完整的通用云发行版。Azure Linux 4.0 是基于 Fedora Linux 发行版，已发布在 GitHub 上，使用 Fedora 的 RPM 包管理系统，深度整合 Azure 云平台。开发者可以通过 WSL 在 Windows 11 上运行 Azure Linux 4.0，但没有 GUI。微软承诺为 Azure Linux 每月释出补丁，如果出现重要漏洞，微软也承诺及时释出补丁。