OrangeBot.AI Digest — 2026-07-21

Instant, hardware Isolated Sandboxes for AI agents

The video AI agent that understands your whole library

The AI agent that handles checkout.

The easiest way to build AI agents for all your tasks

Clone any website into clean code. Free & open source

Turn any webpage into an action manifest for AI agents

Control work with your voice. The Siri for work.

Comment on your AI agent's code & collaborate on changes

Launch a third-party blog to win AEO with just two clicks

A terminal, a real browser, and Claude Code under one roof

Simultaneous desktop and mobile streaming, the easy way.

Private Voice Ring for dictation across iOS & Mac

Track your music repertoire and practice

Your Mac tidies itself: screenshots, installers, Downloads

Find, research, and reach the right people

Video multimodal large language models (MLLMs) can describe what happens in a video, but rarely identify when the supporting evidence occurs. We study generalist video temporal grounding, in which one model predicts a variable-cardinality set of evidence intervals across video lengths, domains, query forms, and viewpoints. Existing training strategies are misaligned with this set-valued task: long-video labels often rely on brittle one-pass annotation, while reinforcement-learning rewards either fail to distinguish non-overlapping predictions or require fragile segment matching. TimeLens2 treats temporal evidence as an interval set throughout supervision and optimization. TimeLens2-93K constructs reliable multi-span supervision through caption-derived proposals, independent localization, cross-agent consensus, semantic verification, and boundary refinement. Our temporal Wasserstein reward computes exact one-dimensional \(W_1\) between uniform distributions over merged interval supports, providing dense, matching-free feedback under unequal cardinalities and equivalent fragmentation; temporal IoU complements it with precise-overlap feedback. Across seven benchmarks, TimeLens2-2B outperforms all size-matched baselines on every benchmark, while the 4B and 8B variants achieve state-of-the-art performance, surpassing open-source models with up to 397B parameters. The 2B, 4B, and 8B variants improve over their Qwen3-VL backbones by 14.2, 13.0, and 18.1 mIoU points, respectively.

This paper introduces EvolvingWorld, a framework and benchmark for character and world co-evolution in interactive literary worlds. Existing systems either treat interactive literary simulation as static persona imitation or isolated scene generation, failing to capture how characters and worlds evolve together over time. To address this, EvolvingWorld models literary simulation as a long-horizon process where characters interact, scenes progress, and character and world states are persistently updated. Unlike prior systems relying on fixed schemas, EvolvingWorld adopts an open-schema framework to support simulation across diverse literary worlds. The framework consists of two coupled modules: a Character Agent for multi-character role-play and persistent profile evolution, and an LLM-based World Model for global and location/entity-level state maintenance and scene progression. Based on this architecture, we formulate 7 trainable tasks for scene initialization, interaction generation, and state update. We construct a dataset from 57 books, producing 138,596 supervised training samples and 222 snapshots for testing. Furthermore, we introduce a trajectory-level LLM-as-Judge evaluation protocol spanning 10 dimensions and 20 metrics. Experiments show that EvolvingWorld can improve long-horizon simulation by effectively maintaining persistent, coherent character and world development.

Training tool-use agents to improve from their own experience remains challenging, as supervised fine-tuning relies on fixed teacher-distilled trajectories, while sparse-reward reinforcement learning provides weak supervision for long-horizon interactions. We present DeepSearch-Evolve, a self-distillation framework for web agents built on DeepSearch-World, a deterministic and verifiable environment with reproducible search and page-reading tools. DeepSearch-World contains 420K multi-hop QA tasks constructed from entity-level random walks and supports key agentic cognitive behaviors useful for self-evolving, including progress verification, grounded reflection, and failure recovery. DeepSearch-Evolve iteratively performs trajectory generation, filtering, data mixing, and fine-tuning to train stronger agents. Without distillation from more capable models, DeepSearch-World-9B achieves competitive performance compared with open-source agents, reaching 31.2% on BrowseComp, 61.5% on GAIA, and 93.4% on HotpotQA, showing that verifiable environments enable scalable self-evolution for long-horizon web agents. We will release the environment, 420K training pool, validation set, model, and code to facilitate future research on self-improving deep search agents.

Pruning long context for coding agents has been a vital technology for efficient context management. While existing context pruning methods such as SWE-Pruner realize this by attaching a separate code classifier, we find the agent itself encodes internal representations indicating the relevance of code context when reading tool output. Based on this finding, we propose SWE-Pruner Pro, which prunes tool outputs directly inside the agent. Concretely, a small head turns the agent's own internal representations into a keep-or-prune label for each line, with a length-aware embedding keyed to each tool output's line count. Across two open-weight backbones and four multi-turn benchmarks, SWE-Pruner Pro saves up to 39% of prompt and completion tokens while preserving task quality, with bounded inference overhead. Notably, on MiMo-V2-Flash SWE-Pruner Pro additionally raises the SWE-Bench Verified resolve rate by +3.8% and the long-context Oolong accuracy by +2.2 points.

Human-object centric video personalization (HOCVP) is a core task within subject-driven video generation. However, existing methods suffer from two key limitations. First, most approaches focusing on inter-subject personalization still struggle to strike a balance between high subject fidelity and accurate interaction patterns between humans and diverse objects, especially when objects represent abstract concepts such as logos. Second, while intra-subject references (e.g., OCR maps, multi-view inputs) are expected to enhance subject fidelity, most existing works lack mechanisms to understand such latent correspondence. To address both challenges, we propose HOMIE, an HOCVP framework that tackles both inter- and intra-subject input settings in a unified manner. Compared to previous approaches, HOMIE proposes a better MLLM integration strategy to extract knowledge of reference-level relationships without compromising the controllability of text encoders or incurring costly re-alignment. Specifically, we introduce global multimodal guidance within self-attention to better align MLLM-derived semantic features with VAE tokens. Furthermore, we propose modality-reference embedding to differentiate tokens from MLLM features and VAE tokens and associate intra-subject reference image tokens. Extensive experiments validate that our method achieves state-of-the-art performance across various HOCVP tasks. Project Page: https://yiyangcai.github.io/homie-page.github.io/

Modern video generation models are increasingly hailed as emerging world models with an internalized grasp of physical law. Yet existing benchmarks largely evaluate physical plausibility only at the output level, without verifying whether the model arrives there through a faithful, law-grounded reasoning process. We introduce Apple-PI, the first benchmark that anchors video-model evaluation explicitly in physical laws. Apple-PI comprises three components. 1) Orchard: a dataset of 400 videos covering ten canonical tasks in classical mechanics. It separates single-law tasks for confounder-free diagnosis from multi-law tasks for probing generalization. 2) Benchmark Protocol: a three-stage protocol based on scientific reasoning, including Perception, Formulation, and Deduction. It uses chain-of-frames prompting on infographic-annotated first frames, treating the generated video as the model's visible reasoning trace. 3) Evaluation Suite: a hybrid evaluation suite that combines MLLM-based subjective scoring with physics-law-grounded objective measures. This enables stage-resolved diagnosis of not only whether a model fails, but where it fails. Benchmarking 11 models shows that current video models remain far from reliable law-grounded world simulators, with the best video model scoring only 0.473. Our stage-, pillar-, and source-resolved analyses further expose a Perception-to-Formulation-to-Deduction bottleneck, weak multi-law state transfer, and a persistent Sim-to-Real gap. These findings position Apple-PI as a diagnostic foundation for guiding future video models toward world models with law-grounded physical intelligence.

We present RynnBrain 1.1, a family of embodied foundation models spanning 2B, 9B, and 122B-A10B scales. Trained with a unified spatio-temporal and physically grounded framework, RynnBrain 1.1 supports embodied perception, spatial reasoning, localization, and planning. Compared with RynnBrain 1.0, it further introduces contact-point prediction across the model family and native 3D grounding for the 2B and 9B models, yielding representations and outputs that are more directly aligned with robot manipulation. We also develop RynnBrain-VLA with a unified cross-embodiment action space and embodiment-specific masking, and deploy it on Unitree G1, Astribot-S1, and Tianji-Wuji. RynnBrain 1.1 achieves strong results on embodied cognition, localization, and 3D grounding, with the 122B-A10B model outperforming all evaluated proprietary and open-source models on VSI-Bench, MMSI, and RefSpatial-Bench. Real-robot experiments show that RynnBrain-initialized policies outperform Qwen-based and representative generalist VLAs, while joint multi-task and multi-embodiment training improves process scores and success rates over per-task training.

Temporal grounding in long recordings remains challenging for audio-conditioned LLMs. We present a time-aware audio LLM that answers questions with explicit timestamps over up to 120 minutes of input. Our approach interleaves periodic time markers with continuous audio tokens using large-scale synthetic supervision from a cascaded pipeline. Our model achieves strong temporal-grounding accuracy on short and long benchmarks and supports time-anchored fragment descriptions and summaries. Extensive ablations examine how time representation, marker frequency, tokenization, and duration-mixture design affect accuracy and computational cost. We release model weights and datasets to support further research on time-aware audio understanding, available at https://huggingface.co/ai-sage/GigaChat3.1-Audio-10B-A1.8B.

Despite recent scaling successes, multilingual ASR performance remains highly uneven, with long-tail languages suffering from severe data scarcity. This work addresses the challenge of building robust foundation models for underrepresented Central Asian languages (Kazakh, Kyrgyz, Uzbek). We present GigaAM Multilingual, a Conformer encoder pre-trained on 2M hours of audio using a HuBERT-style objective. Crucially, we introduce a cluster-level data balancing strategy during pre-training and a domain-aware sampling method during fine-tuning to mitigate head-language dominance. In controlled comparisons, our approach outperforms strong open pretrained encoders (Whisper Large v3, Omnilingual-1B) on target languages, achieving significant gains on spontaneous speech while maintaining efficiency. We release the foundation encoder and ASR model, offering a proven recipe for effective multilingual adaptation under realistic data imbalance.

Egocentric videos of human manipulation provide scalable supervision for embodied intelligence, yet existing resources rarely combine low-cost continuous capture, manipulation-level structured annotations, and reusable tools for robot learning. We present Open-AoE, an open, community-oriented egocentric manipulation dataset and toolchain spanning the full pipeline from smartphone capture to model training. Its first release contains approximately 2,000 hours of manipulation video collected in natural environments by 500+ contributors using 400+ smartphones. The dataset provides text annotations, MANO-based hand poses, camera trajectories, and temporally localized atomic actions. Open-AoE further includes a data processing pipeline that transforms raw recordings into structured samples through temporal action segmentation, semantic annotation, hand reconstruction, and camera trajectory reconstruction. Meanwhile, we provide a separate downstream toolchain supports visualization, cross-embodiment retargeting, model-specific data conversion, and training recipes for VLA policies, WAMs, and World Models. By integrating scalable capture, structured processing, and downstream adaptation, Open-AoE reduces the barriers to both data contribution and reuse, providing practical open infrastructure for embodied model training, human-to-robot transfer, and world modeling.

Building assistants that can continually watch the world, remember what they see, and reason over their accumulated experience is a long-standing goal, and recently multimodal agents equipped with long-term memory over video streams have attracted increasing interest. Unfortunately, existing systems either keep their memory inside the model context or in a flat feature store, and organize it around frames rather than around the persistent entities a stream is really about, which confines them to bounded videos and weakens their ability to track who and what reappears over time. In this paper, we propose ReflectWorld-MM, an entity-oriented multimodal memory system for open-ended video streams. It consists of three parts. The first is a perception front-end that turns an audiovisual stream into entity-resolved observations under a bounded short-term memory. The second is a hierarchical long-term memory, grounded in human memory theory, that couples a multi-scale episodic memory, an evolving entity-centric semantic memory, and a procedural memory. The third is a complete realization, built for real-world operation, that ingests arbitrary streams and plugs into off-the-shelf assistants. Across six long-video and lifelong-memory benchmarks, ReflectWorld-MM achieves the best accuracy on all six, outperforming strong memory agents and a frontier model.

In line with the prevailing direction of vision research, we explore the integration of both generation and editing capabilities for video and image modalities within a single model. Current approaches to collecting video editing data typically depend on labour-intensive, time-consuming curated procedures--involving object mask annotation, the use of error-introducing pair synthesis via I2V model and ControlNet-like guidance, and VLM-based quality filtering or refinement--and demonstrate limited task scalability. As a result, the diversity of editing tasks remains substantially narrower than that available for image editing models. We develop a pixel-pair temporal warped flow field that can directly generate corresponding video editing samples in real time from image editing samples, and we demonstrate across multiple levels of video editing tasks that a model can learn video editing using only such data. We regard the image modality as a particular form of the video modality. Accordingly, we design a modality mimic generation loss and a modality mimic editing loss to relatively align the capabilities--and thereby the output distributions--of the two modalities through mutual imitation. Moreover, language-based visual editing entails the comprehension of the editing instruction and the reference visual content, the localization of the region corresponding to that instruction within the reference visual contents, and the modification of that region alone. Existing approaches predominantly rely on external aids, such as fine-tuning an additional MLLM or explicitly supplying a mask sequence as auxiliary input during inference. In contrast, we aspire for the model to internalize this capability. To that end, we introduce sense-related tasks--for instance, referring expression segmentation--along with corresponding editing-region-aware latent-level loss and attention-level loss.

Entropy control has become an effective tool in reinforcement learning (RL) of large language models (LLMs), helping balance exploration-exploitation trade-off during alignment process. Such RL paradigm is often conducted on mixtures of heterogeneous tasks, which induce distinct entropy regimes under the same policy, making global or token-level entropy regulation insufficient to corresponding heterogeneous needs of exploration. This heterogeneity further makes GRPO-style normalized advantages induce an entropy-dependent bias, making advantage signals across prompt groups statistically non-comparable. To address this issue, we propose Group Entropy-Controlled Policy Optimization (GEPO), a lightweight extension to GRPO that uses group entropy, estimated from existing grouped samples to perform entropy-conditioned asymmetric advantage shaping. GEPO attenuates positive advantages in low-entropy groups to reduce over-exploitation, and negative advantages in high-entropy groups to preserve exploration, with adaptive thresholds derived from historical entropy statistics. Extensive experiments on two base models across thirteen benchmarks spanning mathematics, physics, science, code generation, and instruction following show that GEPO consistently outperforms GRPO and recent entropy-controlled methods, delivering balanced cross-task improvements while preserving task-specific exploration levels throughout training.

Training API-calling large language model (LLM) agents demands massive amounts of high-quality trajectories. However, collecting such data at scale typically requires fully implemented environments with executable APIs and realistic, pre-populated backend databases, creating a major bottleneck for scalability. To overcome this, we propose an environment-free synthetic data generation approach that leverages LLMs as on-the-fly digital world models. Given only API specifications, our method generates trajectories mimicking interactions between an agent and a stateful environment. Specifically, an LLM first generates diverse tasks solvable with the provided APIs. A teacher agent then iteratively solves each task while an LLM simulator generates coherent synthetic API responses conditioned on the task context and simulation history. Finally, an LLM judge filters the trajectories to ensure the quality of the resulting dataset. We evaluate our approach on the challenging AppWorld and OfficeBench benchmarks, which include both information-retrieval and state-changing tasks. Fine-tuning models on our synthetic data yields significant performance gains, demonstrating that effective supervision for API-calling agents can be generated without any executable environment. Our results establish LLM-based API simulation as a practical, scalable solution for training agents across diverse API ecosystems.

Structure-based drug design (SBDD) leverages the 3D structure of protein targets, often complemented by other spatial constraints, to generate candidate binding molecules. While diffusion models have dominated as a leading paradigm for high-quality 3D molecule generation, LLM-based methods are rapidly emerging in molecular design and have shown competitive performance in pocket-conditioned molecular generation. However, their ability to reason about physics and 3D spatial environments is largely underexplored. In this work, we systematically analyze whether current general-purpose LLMs are capable of navigating complex 3D constraints compared to established baselines such as specialized diffusion models. We consider 3D ligand generation conditioned on protein pockets together with ligand- and interaction-derived spatial constraints, including anchor fragments, pharmacophore points, and mandatory pocket-ligand interactions. To enable this evaluation, we introduce 3D-Fit - a token-efficient benchmarking strategy for assessing LLM performance on multi-conditioned spatial molecule generation. Our findings reveal a clear pattern in LLM spatial capabilities: while they still lag behind state-of-the-art approaches, they are promising and can handle multiple spatial constraints simultaneously, enabling scaling to heterogeneous setups.

Sara Fischer / Axios : OpenAI renews its investment in the American Journalism Project by committing an additional $5M in funding and $3M in tech credits over the next two years —  OpenAI is renewing its investment in the American Journalism Project by committing an additional $5 million in funding and $3 million …

Wall Street Journal : OpenAI adds Nubank founder and CEO David Vélez and BNY CEO Robin Vince to its board of directors, as the company moves toward a public listing —  AI company names David Vélez of Nubank and Robin Vince of BNY as directors  —  OpenAI is adding David Vélez …

OpenAI : OpenAI says its models chained vulnerabilities across its research environment and Hugging Face's infrastructure to find solutions for the ExploitGym benchmark —  Last week, Hugging Face disclosed a new kind of security incident(opens in a new window) after they detected and contained …

Sam Sabin / Axios : Cisco releases Antares-350M and Antares-1B, two open-weight AI models to find known vulnerabilities within a codebase, and plans to release Antares-3B soon —  - Cisco says the models performed similarly to much larger models, including OpenAI's GPT-5.5 and Z.ai's GLM-5.2 …

Juby Babu / Reuters : Super Micro says Q4 orders have topped $60B and expects gross margins of 15% to 17%, up from its prior forecast of 8.2% to 8.4%; SMCI jumps 13%+ after hours —  Super Micro Computer (SMCI.O) said on Tuesday it has received orders in excess of $60 billion in the fourth quarter …

Mark Landler / New York Times : France becomes the first European country to pass a law banning social media access for users under 15, potentially going into effect as soon as September 1 —  Its new law, the first of its kind enacted by a European country, will ban access to some accounts for anyone under 15.

Ina Fried / Axios : OpenAI says its models, including GPT-5.6 Sol and “an even more capable pre-release model”, breached Hugging Face while OpenAI tested their cyber capabilities —  OpenAI said Tuesday that models it was testing escaped their sandbox and compromised parts of AI platform Hugging Face's production infrastructure last week.

Iain Martin / Forbes : Poolside launches Laguna S 2.1, an 118B open-weight model built for agentic coding and long-horizon work, that it says competes with larger open models —  Chinese open-source models like Kimi, Deepseek and Qwen are now dominant AI tools.  American upstarts like Mira Murati's Thinking Machines …

New York Times : Some Instagram and Facebook users say Meta's AI moderation deleted their accounts; Meta says AI makes 13% fewer errors and finds 10% more violations than humans —  Users said the technology mistakenly deleted their accounts.  To resolve the issue, they still had to rely on A.I.

Chris Best / The Substack Post : Substack partners with AI-detection tool Pangram, allowing users to scan text longer than 100 words for an estimate of how much was written with AI assistance —  Building trust in the AI age  —  It's getting harder to tell what's real on the internet.  —  This is less true on Substack, and we aim to keep it that way.

Block : Block launches Buzz, an open source collaboration workspace on Nostr protocol for humans and AI agents to share messages, code repositories, workflows, and more —  Open Source, Decentralized Technology, Developers, AI, Technology  —  TODAY WE'RE RELEASING Buzz, A FREE …

Tracey Tully / New York Times : NJ Governor says a software error at DMV led to the improper voter registration of ~6,600 noncitizens between June 2023 and June 2024; 400 of those cast ballots —  A 2023 software glitch at the state's Motor Vehicle Commission led to roughly 6,600 noncitizens being registered to vote during …

Luke Fountain / CNBC : Federal lobbying disclosures: Anthropic spent $1.97M in Q2, up 26% QoQ, and OpenAI spent $1.2M, up 18%; Meta remained the largest spender at $5.99M, down 15% —  OpenAI and Anthropic increased their federal lobbying spending to record levels in the second quarter of 2026 …

Scott Stein / CNET : Meta adds the Xbox Game Pass starter edition to its Horizon Plus plan, which costs $60/year or $8/month; the pass only allows 10 hours of Xbox gaming per month —  Meta's throwing Xbox cloud gaming Into Its Horizon Plus Quest subscription, and converting its controllers to work with it.

AI Security Institute : Analysis: every frontier AI model tested in cybersecurity evaluations attempted to “cheat”, led by GPT-5.4 at 14.1% of tasks; Mythos cheated the least, at 7.8% —  Can you trust an AI model to do what you intended?  This is a central question both for those deploying AI systems …

去年美国在全球大规模征收关税，任天堂据此提高了 Switch 2 等相关设备和配件的售价。今年 2 月美国最高法院裁决征收关税违法，任天堂随后起诉美国政府要求退还关税。美国玩家则跟着提起了一项拟议中的集体诉讼，认为任天堂应该将美国关税退款退给消费者。任天堂周一请求法庭驳回该诉讼，称它无法律义务将美国关税退款退给消费者。任天堂的观点十分简单，任天堂或其零售合作伙伴设定了最终价格，而买家是自愿支付的。

FBI 逮捕了一名 21 岁的佛罗里达居民 Zyaire Dontaevious Zamarion Wilkins，他涉嫌与同伙通过在 Steam 游戏中植入恶意程序去窃取玩家的加密货币。该团伙在近两年时间内通过在至少八款 Steam 游戏中植入恶意程序，感染了约 8000 台电脑，在 2024 年 5 月至 2026 年 2 月间从约 80 个加密货币钱包中窃取了价值至少 22 万美元的加密货币。被植入恶意程序的 Steam 游戏包括 BlockBlasters、Dashverse、Lunara 和 PirateFi。其中仅 BlockBlasters 一款游戏就从 261 至 478 名受害者手中窃取了价值约 15 万美元的加密货币，受害者包括了 Twitch 主播 RastalandTV，他被盗走了 3.2 万美元，他正接受癌症治疗，这笔钱是观看者捐赠的。

在生成式 AI 时代，流行公民科学数据库如 iNaturalist 和 Macaulay Library 也免不了被 AI slop 入侵。研究人员在《自然》上报告，他们在记录野生动物物种的流行公民数据库内发现了数百张 AI 生成的虚假图像。问题的真实规模尚不清楚，因为有许多假图可能未被发现。文章作者 Alexander Lees 博士称他在 Facebook 上看到的野生动物照片基本上都是 AI 生成的。他指出公民科学数据库里彻头彻尾的假图还比较罕见，问题主要是照片上传者会用 AI 美化下图像，结果生成式 AI 给图像添加了不存在的内容。iNaturalist 上逾 6.1 亿张图像只有 1400 张被标记为可能使用 AI。iNaturalist 社区支持总监 Tony Iwane 认为大多数假图并非是恶意的，他同时呼吁用户保持警惕，因为信息的准确性至关重要。

WordPress 上周释出补丁修复了两个高危漏洞，它督促运行 WordPress 软件的网站立即更新，漏洞是如此严重以至于 WordPress 在条件允许的情况下启用了强制更新。安全公司随后警告，黑客正利用漏洞入侵未更新的 WordPress 网站。WordPress 是最流行的博客软件和内容管理平台，存在漏洞的 WordPress 网站数量可能多达数千万。存在漏洞的 WordPress 版本为 v6.9.0-6.9.4，以及 v7.0.0-7.0.1。根据 WordPress 的官方统计，逾 4 亿个网站运行存在漏洞的 WordPress 版本，但其中一部分网站可能已经打上了补丁。组合利用两个漏洞，黑客能远程控制存在漏洞的网站。

欧盟委员会对阿里巴巴旗下面向消费者的电商平台 AliExpress 处以 5.5 亿欧元的罚款，原因是它在打击非法商品销售上做得不够——这是根据 Digital Services Act（DSA）开出的最高罚单。布鲁塞尔去年曾警告 AliExpress，它在阻止平台上的卖家兜售非法商品以及确保其推荐引擎不推广这些非法商品上做得不够。欧盟委员会根据 DSA 要求其整改，并详细列出了需要采取的措施。一年后，欧盟委员会认定 AliExpress 的措施不足，宣布处以巨额罚款。委员会认为，AliExpress 未能妥善评估其是否有足够的人员审查潜在的非法产品，也未能充分评估其推荐和广告系统如何加剧非法产品的扩散。DSA 允许对商家处以全球营业额最高 6% 的罚款，阿里巴巴集团截至 3 月 31 日的财年收入为 1480 亿美元。布鲁塞尔原本可要求近 90 亿美元的罚款。

加州大学圣地亚哥的科学家警告地球海洋和淡水系统中的水体溶解氧迅速减少，可能会给地球带来短期内无法逆转的变化。研究人员认为水体溶解氧也应纳入地球限度理论框架，该框架突出了人类行为对地球系统的冲击的限度，若超过该限度，自然环境将丧失正常的自我调节能力，现代人类所处的相对稳定的全新世将会结束。现有的地球限度理论包含了 9 个限度：气候变化、海洋酸化、大气平流层臭氧消耗、生物地质化学循环中氮循环和磷循环、全球淡水使用、土地系统变化、生物圈完整性的侵蚀、化学污染和空气污染。水体溶解氧迅速减少的原因包括人为全球暖化、富营养污染、深层水体的运动和通气过程，等等。

鸡蛋的价格过去几个月快速上涨，以至于被称为“火箭蛋”。根据一项追踪主要产蛋省份的指数，鸡蛋价格比一年前上涨了 40% 以上。另一项追踪一处批发市场价格的基准指数显示，蛋价最近创下了 10 年来的新高，这让业内人士大感意外。研究鸡蛋市场的分析师表示，两年前，在连续几年的丰收之后，养殖户开始扩大规模，2024 年至 2025 年间新增的蛋鸡数量创下历史新高。突然之间，鸡蛋供过于求，价格暴跌。随后，养殖户们发现自己没有足够的资金来喂养所有的鸡。于是他们开始大量淘汰蛋鸡，专家认为这样的淘汰速度过快了。最终，蛋鸡数量不足以满足需求，导致价格上涨。与此同时，市场需求却格外强劲。

日经的调查显示，因 AI 投资激增美国五大科技巨头 Google/Alphabet、微软、亚马逊、Meta 和甲骨文的隐性债务高达 1.65 万亿美元，超过了其实际债务。其中 Meta 的表外债务达到了 4200 亿美元，是其账面债务的几乎三倍。甲骨文隐性债务约 2733 亿美元。为支持 AI 研发这些科技巨头在迅速扩充其数据中心等 AI 计算资源，签订了 GPU 和服务器的长期采购协议。

开源媒体服务器 Jellyfin 项目联合创始人 Joshua Boniface 以及 Anthony Lavado 宣布辞去项目领导人以及核心团队成员职务，退出项目，将该项目交给其他团队成员。Joshua 称自己面临严重的职业倦怠，无法胜任工作，是时候离开了，他创建 Jellyfin 项目时以为它最多会有几百或几千用户，但七年半后它成为了排名第一的开源媒体服务器，服务于数以百万计的用户，是 Plex 的可行替代。Anthony 称自己主要负责后端管理工作，自己的生活发生了变化，有其他优先事项需要处理。

医学界完成了一项前沿手术—— 一名男性通过移植自己童年取出并冷冻保存16年的睾丸组织，重新恢复了精子生成能力。这批睾丸组织样本是他在接受化疗前采集冷冻的。化疗会严重损伤生殖功能，导致不育。这一里程碑式成果有望开启全新的人类不育治疗浪潮。2008 年比利时布鲁塞尔自由大学医院的医生为一名患病的 10 岁男孩摘除部分睾丸组织并冷冻保存。他当时罹患镰状细胞贫血，即将接受化疗，为后续造血干细胞移植做准备。10 余年后，这名男孩成年，希望拥有自己的后代。医护人员对其进行了两年监测，确认他已完全无法正常生成精子。随后，团队将冷冻保存 16 年的11个睾丸组织片段，分别移植至其剩余睾丸内部及阴囊皮下。移植组织在成年男性的激素环境与生理条件下培育生长一年。一年后，研究团队取出移植组织并开展检测分析。检测结果显示，多个移植组织中均发现生精干细胞，且存在活跃的精子生成迹象。在其中一个移植样本中，研究人员成功检出一枚成熟精子。

在月之暗面宣布首个 3 万亿参数级别开放权重模型 Kimi K3 之后，前特朗普 AI 高级顾问、现 OpenAI 战略主管 Dean Ball 在 X 平台上表示，开放权重模型的战略将导致“AI 共产主义”，可能会“减缓 AI 的发展”，因为会“阻碍 AI 领域的资本支出”。他认为政府可以通过 FUD 策略（即 fear, uncertainty, and doubt）阻止美国公司使用开放权重模型。他后来澄清说这只是预测不是建议，声称他支持 AI 开源/开放权重。美国两大前沿 AI 公司 OpenAI 和 Anthropic 仍然有领先优势，但优势在缩小，它们采用的都是闭源模式。

Mozilla 释出了 Firefox 153.0。它是 ESR 版本，将提供至少 15 个月的支持。Firefox 153.0 主要新特性包括：Vulkan 视频解码支持，Linux 版支持 Video Acceleration API (VA-API) ，增加对 JPEG-XL 图像格式的支持，Windows 版支持 HDR 视频播放，PDF 功能增强，等等。

黑客入侵了罗马尼亚的土地管理机构 Agenția Națională de Cadastru și Publicitate Imobiliară（ANCPI)，在勒索失败之后删除了整个土地登记数据库以及备份，导致全国房地产市场瘫痪，官方应用和网站已离线一周。公证员无法记录新的交易，公民也无法获取所有权证明或详细的土地记录。黑客是在 7 月 14 日删库的，一天后 ANCPI 部分被盗数据在黑客论坛上出售，泄露的数据包括员工凭证、内部文件以及该机构的 IT 网络详细信息。ANCPI 表示正在重建系统。虽然黑客声称删除了备份，但 ANCPI 表示它还有备份副本。

体细胞突变随年龄增长而积累，能导致细胞死亡。根据发表在《npj Aging》期刊上的一项研究，科学家开发出一种方法量化体细胞突变对人类寿命的影响。研究只考虑体细胞突变，没有考虑其它任何衰老特征。研究发现，肝脏等增殖性组织通过细胞更新能在数千年内维持功能，而神经元和心肌细胞等有丝分裂后细胞则是长寿的重要瓶颈。综合多个器官预测的人类寿命中位数是 146-194 岁，约为目前人类寿命的两倍。这意味着体细胞突变是衰老的重要驱动因素，但无法单独解释目前的人类寿命，驱动衰老的还有其它因素。

法国和意大利研究人员发现，有了 AI 之后人们愈来愈不愿说不知道了，AI 抑制了人们的批判性思维，使得人们更容易重复 AI 生成的虚假信息。研究人员将参与者分成两组，一组必须在没有 AI 的帮助下回答问题，另一组可以向 AI 寻求建议。结果发现没有 AI 帮助时 44% 的人会回答说不知道答案，而有了 AI 的帮助后只有 3% 的人这样做。研究人员让参与者使用了一个更可能出错的模型 Step 3.5 Flash。结果显示，不使用 AI 正确率为 27%，使用 AI 后降至了 9%，原本会给出正确答案的人在听取了 AI 建议后给出了错误答案。尽管正确率下降了，但参与者的自信心却增加了。