OrangeBot.AI Digest — 2026-08-13

Stops AI agents before they break something

Ask any question, get a video back instantly

The only mechanic is waiting. The only currency is time.

A 360 camera, gimbal, and 3D scanner in one

Browser for your agent. Runs local in a window you can see

Publish interactive docs that answer and report back

AI code review that runs your code

Audit AI search structure. Dry run fixes. Receipts.

Create AR experiences with AI, no code required!

The AI that refuses to let you drop the ball

Test firmware on a virtual chip with no hardware needed

Where humans and agents come to work.

Natural language browser & mobile app tests from terminal

The gate between your AI agents and the real world

Move AI chats between ChatGPT, Claude & Gemini in one click

AI agents operate in persistent environments where early state changes can influence decisions far into the future. Unlike conventional language-model interactions, agent behavior is mediated through a shared state that is repeatedly modified and reused across long-horizon workflows. Current safety benchmarks often fail to capture these cumulative risks because they focus on short, static tasks. To address these limitations, we introduce OpenART, an open-ended arena for scalable agent red teaming through environment evolution. OpenART provides over 10,000 validated stateful scenarios across 50 domains, drawing from a pool of more than 500,000 tools and skills. These tasks require a median of 97 tool calls and enable unified evaluation across 75 different agent-model configurations. To systematically explore these evolving attack surfaces, we propose the Evolutionary Markov Hypergraph Attack (EMHA). EMHA is a black-box policy that performs feedback-driven environment evolution by coordinating authorized state transitions without requiring parameter updates. Throughout the evaluation, task objectives remain fixed while only the environment state changes. Across all configurations, EMHA achieves a pooled Attack Success Rate (ASR) of 85.0%. Its advantage over instruction-only evolution increases from approximately 2% on simple environments to over 17% on the most complex ones, demonstrating that environment evolution increasingly exposes safety failures as task complexity grows. Furthermore, our analysis shows that the specific runtime implementation of an agent explains a significant portion of safety variation beyond the underlying model's capabilities. These results establish OpenART as a scalable foundation for studying agent safety in complex, evolving environments.

Turning a research idea into a complete paper requires more than text generation: the system must retrieve literature, design and execute experiments, revise claims according to evidence, produce publication-ready figures, and maintain consistency across a long generation process. We present Spark-to-Paper, an end-to-end research paper generation system implemented as thirteen composable skills inside an existing coding assistant, without requiring a separate agent platform or orchestration service. Spark-to-Paper separates model-based judgment from deterministic operations that can be directly executed and checked. It further separates experiment planning from reporting, so that required evidence is specified before results are observed and manuscript claims are revised according to measured outcomes. To improve reliability over long research trajectories, the system combines deterministic integrity checks with self-critique and bounds a failure mode we call the Self-Refutation Loop, in which repeated experiments continue to reject the original research objective. Spark-to-Paper also produces editable vector figures through programmatic plotting for experimental results and code-based reconstruction for generated method diagrams. Across eight controlled research topics, Spark-to-Paper achieves 99.5% citation validity and 96.4% figure editability. A controlled ablation increases fabrication detection from 14% for a single-pass draft to 92% with the full integrity and review stack, while adversarial review achieves 74% precision. The full system uses 11.9M tokens, costs $8.1 per manuscript, and requires 3.2 hours on average. These results show that end-to-end research paper generation can be implemented as a lightweight, composable workflow inside existing coding assistants while keeping experimental evidence central to how claims are accepted, revised, or abandoned.

Recent work on distillation transfers the capabilities of large models to smaller ones often by updating the latter's parameters, through teacher forcing, on-policy distillation, and related training-time methods. In this paper, we ask whether such transfer can instead occur at test time. We study strong-to-weak scaffolding: whether a stronger builder model can construct inference-time harnesses that help a weaker target model solve tasks more reliably without any parameter updates. Using four representative Theory-of-Mind benchmarks, each builder model uses 5% of the data as a validation set to iteratively refine its harness over multiple rounds, after which the finalized harness is evaluated on the full test set. Empirically, this form of test-time capability transfer is highly effective, nearly doubling average target-model performance from 0.49 to 0.91. Our analysis shows that the gains come primarily from offloading unstable model reasoning into deterministic code, benchmark-specific routing, and strict answer-format enforcement, rather than from encouraging the target model to reason more extensively or sample more broadly. We further find that builder-model reasoning effort improves harness quality monotonically, platform effects are modest relative to the builder model's own capability, and weaker target models receive the largest gains. These results suggest that inference-time harness design is an important complement to conventional training-time distillation, enabling strong models to transfer cognitive structure to weaker models without retraining.

AI models have achieved remarkable success across diverse domains, yet the mechanisms underlying their capabilities and the risks they may pose remain poorly understood. As AI development becomes faster and increasingly automated, mechanistic exploration remains largely manual, widening the gap between what models can do and our ability to understand and control them. To bridge this gap, we introduce Mechanist, an agentic system that uses AI as a scientific instrument for the autonomous discovery of mechanisms underlying AI intelligence. To support autonomous mechanistic discovery, we construct an interpretability-focused knowledge graph of approximately 13,000 papers and integrate it with a multidisciplinary database of 43 million papers spanning 26 fields. We further curate a library of 32 foundational methods for mechanism analysis, causal intervention, and validation. Compared with Claude Code and existing AI-scientist systems, Mechanist generates more valuable mechanism hypotheses and executes experiments more reliably. Mechanist also demonstrates a progression from discovering model behaviors to explaining and controlling AI models. Specifically, Mechanist first uncovers a counterintuitive safety risk in scientific laboratories, showing that unsafe traits can transfer across modalities through apparently safe training data. Mechanist then develops a mechanism theory of belief, revealing how models represent world knowledge, form beliefs, infer the beliefs of others, and how these mechanisms emerge during pretraining. Finally, Mechanist translates these mechanistic insights into practical interventions that improve model performance across diverse scenarios and steer scientific foundation models toward generating DNA sequences with specified properties.

Large Language Models (LLMs) increasingly act as agents whose procedural knowledge is stored in reusable skill packages and loaded at inference time. As skill libraries grow, a central challenge is to expose the smallest sufficient executable context under a limited context budget. Existing systems struggle to reuse routines below the whole-skill level, preserve procedural contracts during compression, keep compressed routines executable and expandable, and update the compressed library as skills evolve. These challenges reveal a unit mismatch: skills are retrieved as packages, compressed as text, and converted into execution graphs only after retrieval, whereas reliable reuse requires a contract-bearing procedural unit. We propose SkillZip, an execution-aware procedural abstraction framework that performs contract-preserving compression over section-level graphs. SkillZip rewrites recurring contract-valid motifs into reversible ported macros while preserving boundary signatures, dependency closure, verifier reachability, and source-level expansion. At inference time, it hydrates a compact, dependency-closed context and expands macros only when required. ReZip further integrates new skills and revises risky macros using execution evidence. Comprehensive experiments1 on technical and embodied agent benchmarks show SkillZip consistently outperforms the strongest baseline by up to 12.2 points, while achieving a 3.46x compression ratio with 99.2% dependency preservation and 98.7% verifier reachability. Scaling analyses further confirm robust retrieval across skill libraries ranging from 200 to 100K skills.

The rapid advancement of Large Language Models (LLMs) is revolutionizing AI for Games by enabling open-ended and fluid interactive storytelling. However, existing research has largely overlooked the critical challenge of maintaining long-horizon logical consistency and narrative integrity against unconstrained user interventions. To address this, we formulate this challenge as Narrative Commitment Preservation (NCP), and take interactive narrative as our testbed. We introduce NCP-Bench, a benchmark of 100 narrative environments derived from movie synopses. Each environment includes a structured narrative specification (trajectory, commitments, and initial facts) that we can automatically check throughout the interaction between the player agent and the narrator agent. Experiments across state-of-the-art LLMs reveal a substantial long-horizon consistency gap: high linguistic quality does not guarantee commitment preservation; even strong models frequently generate logically conflicting content under adversarial interventions, with the best-performing model (GPT-5.2) achieving only 42% survival rate after 20 turns and fact conflict rates ranging from 40% to 68% across models, and only isolated runs satisfying all achievement commitments within the 100-turn limit.

Previsualization is an intermediate layer between ideas and production in film, games, architecture, and urban design. It lets creators iteratively refine scenes, actions, cameras, and spatial-temporal dynamics. Yet existing generative methods rely on simple prompts to jointly control all of these factors through one-shot image or video synthesis, offering weak controllability and limited support for iterative editing. Fundamentally, a world comprises multiple elements with geometry, appearance, and other attributes, together with cameras. Different frames are produced through local modifications or recombinations of this shared state, which is otherwise largely reused. Therefore, we argue that the missing component is an explicit and persistent working state. To address this, we present StateFlow, a state-centric framework for generative previsualization. Rather than generating videos in one shot, StateFlow uses an editable 3D world to organize scene structure, evolution, and cameras, while off-the-shelf video models enhance visual quality when higher fidelity is desired. This world is maintained as a persistent structured 3D state of scene elements and camera configurations, serving as the core working representation for previsualization. Built on this insight, StateFlow has three stages to construct, evolve, and access the world state. State construction lifts generated 2D content into a coherent 3D world through prior-guided, conflict-aware dual-view initialization, while State evolution translates user intent into structured state transitions while preserving world memory, avoiding full-scene regeneration for each edit. State access uses render-feedback reflection to refine camera plans into visually feasible trajectories, avoiding reliance on VLM semantics alone. Experiments show that StateFlow produces high-quality 3D worlds for video creation and game-like prototyping.

Recent Vision Foundation Models (VFMs) predict depth, camera pose, and pointmap in a single forward pass without per-scene optimization, achieving strong generalization. However, enforcing explicit multi-view geometric consistency, e.g., through bundle adjustment, is computationally costly and is thus not imposed during VFM pretraining, so such inconsistency can arise. To address this, implicit self-consistency derived from model outputs (e.g., pointmaps, features), though enforced at test-time in prior work, delivers inherently limited performance gain, especially on scenes where the pretrained VFM is highly inaccurate. In contrast to this implicit signal, we propose Self-Geometry, a plug-and-play test-time adaptation pipeline that directly imposes explicit multi-view geometric constraints using 2D pixel correspondences as pseudo ground-truth. Our proposed Self-Geometry consists of Geometric Disentanglement Optimization, which combines Multi-View Consistency and Epipolar Consistency losses with Gradient Disentanglement to prevent gradient conflict; Frame Angular-Neighbor, a view sampler based on SO(3) geodesic distances for lightly imposing these constraints; and Lightweight TTA, which adapts VFMs via LoRA. Our method achieves consistent improvements in both pose and geometry estimation across six VFMs (VGGT, π^3, DA3-Giant/Large/Base/Small) and four benchmarks (7Scenes, ETH3D, ScanNet++, HiRoom).

Embodied agents are increasingly built as systems around foundation models, where performance depends not only on model weights but also on the skills, context, action interfaces, and execution harness surrounding the model. While supervised fine-tuning and reinforcement learning can adapt agents to new environments, they require additional data, rewards, and training runs; meanwhile, many train-free code-centric approaches rely on programmable robot APIs that may be unavailable in fixed-interface settings. We propose SHAPER, a self-evolving framework for train-free embodied adaptation that keeps model parameters frozen and improves the non-parametric agent system by evolving reusable skills and a context-code harness through target-environment rollouts. In SHAPER, the same frozen model can serve as both planner and optimizer, refining its external skills and context-code harness without parameter updates. We evaluate SHAPER on VLABench and ESI-Bench, covering embodied agents with different low-level action interfaces, and compare against pure execution, supervised fine-tuning, and test-time-scaling baselines such as verifier-free selection and voting. Our results suggest that skill-and-harness optimization is a practical route to self-evolving embodied agents when model training is expensive, unavailable, or undesirable.

The "thinking-with-images" paradigm equips multimodal LLMs with active visual operations such as crop-and-zoom. However, models using these operations often achieve only marginal or negative gains over direct inference at substantially higher token cost. They may also repeatedly crop irrelevant regions and fail on questions that direct inference answers correctly. We ask whether the returned visual evidence causally affects the answer. To answer this question, we formulate visual tool-use as a causal graph that separates observation-mediated paths from action-induced shortcuts. We then audit it through interventions at the three levels: policy (comparing tool-use with direct inference), trajectory (corrupting all observations during rollout), and step (counterfactually replacing one individual observation under a fixed prefix). Our step-level estimand, Visual Evidence Gain, isolates the contribution of each returned observation. Across six representative models and five fine-grained perception benchmarks, we uncover policy miscalibration with two failure modes. In Calling Without Looking, returned observations have no causal effect on the answer. In Looking Without Planning, observations are informative but the call schedule is incoherent. A trajectory-level diagnostic decomposes the policy-level accuracy gain and shows that the gain is concentrated in a Calibrated minority. We term this discrepancy the illusion of visual tool-use: despite aggregate accuracy gains, visual tool-use is not causally effective across a broad range of rollouts. The code is available at https://github.com/OpenCausaLab/CauAudit.

Videos captured through glass often contain reflections that degrade visual quality and interfere with downstream vision tasks. Although single-image reflection removal has been extensively studied, video reflection removal remains largely underexplored due to the lack of paired video data, temporally coherent removal models, and dedicated evaluation benchmarks. We present a closed-loop framework that unifies physics-grounded reflection simulation, diffusion-based video dereflection, and benchmark evaluation. Our S2R-Synthesis pipeline generates paired reflected and reflection-free videos by performing physics-grounded augmentation in the structure space and rendering realistic reflected videos with a trained video diffusion renderer; the augmentation models key glass-related effects including roughness-induced blur, thickness-induced ghosting, and reflectance variation. Based on the synthesized data, we introduce S2R-Removal, the first diffusion-based video reflection removal model, which adapts a pretrained video diffusion prior through reflection-aware latent adaptation and one-step pixel-geometric refinement, recovering the clean transmission in a single denoising step. We further build S2R-Bench, the first benchmark for video reflection removal, supporting both full-reference evaluation and real-world human perceptual assessment. Experiments on S2R-Bench and multiple public image benchmarks demonstrate state-of-the-art performance and faster inference than even non-diffusion baselines, and validate the effectiveness of S2R-Synthesis. Project page: https://codingwzp.github.io/VideoDereflection_S2R.

Large language model (LLM) agents integrated with external tools are vulnerable to indirect prompt injections embedded in environmental states. However, existing studies largely rely on manually implemented or reused environments, stochastic LLM-based tool simulation, and predefined injection locations, limiting scalable security research across broader domains. To bridge this gap, we propose **ToolHazard**, a scalable adversarial environment synthesis framework that reduces human engineering and supports expansion with additional seed domains and compute. Through an Environment Simulator, an Attacker Agent, and a User Simulator, ToolHazard synthesizes executable stateful environments, discovers viable injection points and generates environment-specific payloads, and constructs state-grounded long-horizon tasks. Based on ToolHazard, we build **ToolHazard-Bench** for stress-testing agents under complex workflows and diverse environmental attacks. Experiments reveal substantial agent vulnerabilities and show that injection timing and placement affect attack effectiveness. Moreover, ToolHazard-generated alignment data improves security on both ToolHazard-Bench and AgentDojo while preserving benign task utility.

Discrete diffusion models for categorical generation are defined by a corruption kernel, which determines the intermediate state space and the associated reverse prediction problem. We study uniform discrete diffusion and ask whether its training objective and reverse transitions can be enriched without changing the underlying categorical corruption process. We introduce Simplax, an exact Dirichlet--categorical augmentation that couples each corrupted categorical state with an auxiliary simplex-valued variable while preserving the original uniform diffusion process as its categorical marginal. This augmentation yields a tractable Rao--Blackwellized reverse-bridge objective and a corresponding stochastic reverse sampler, while retaining the corrupted categorical state as the denoiser input. Empirically, Simplax improves the generative perplexity--entropy tradeoff on unconditional OpenWebText generation. On Sudoku, a model trained exclusively on 30-clue puzzles achieves the highest accuracy among the compared methods across all evaluated clue densities, including the minimum uniquely solvable 17-clue regime, and also achieves the highest validity in unconditional generation.

The dominant paradigm treats AI safety as a property to be instilled during model training via RLHF, DPO, or Constitutional AI. We argue this is structurally insufficient for autonomous agents that execute code, mutate files, send messages, and modify databases. Agent safety should be a runtime contract enforced by the harness, and the contract has two complementary faces. The preventive face blocks dangerous actions before they happen via sandboxes, permission gates, output filters, and trajectory monitors. The evidential face requires verifiable proof that good actions actually happened, gating task submission on hard evidence such as test runs, log captures, file diffs, and citation grounding. We ground the position in four lines of public evidence, with row-level protocols and data released in the supplementary JSON files: a survey of 52 documented AI-agent and LLM safety incidents, a false-completion audit with 31 non-contested core cases plus one disputed illustrative case, a trajectory-schema audit of 12 public agent systems and harnesses, and a title-level audit of all 28,560 papers accepted at NeurIPS, ICML, and ICLR 2023-2025 showing a pooled 8-12x imbalance between training-time and deployment-time publication. Two prior communities that needed to enforce safety, computer security and the experimental sciences, converged on runtime contracts with both preventive and evidential elements; agentic AI is now under the same pressure. We formalize an Agent Trajectory Schema and Evidence Chain, state a compositional gating proposition based on standard monitor composition, and outline a research agenda. The right unit of safety in agentic AI is the trajectory-with-checkable-evidence, not the model.

Complex software systems develop over timescales that exceed the lifespan of any individual coding agent. Most agentic software systems preserve continuity through persistent sessions, memories, managers or shared context. We introduce EvoX Genesis (hereafter, Genesis), which instead makes the software project persistent while allowing local agents to remain finite-lived. Genesis represents software as a persistent recursive world: each local world is situated by an accepted version and a repository path, finite-lived agents propose local changes, recursive delegation moves work across paths, and only accepted consequences advance the persistent version history. We evaluate this organization across formation, continuation and redevelopment. Starting from a repository with no compiler implementation, Genesis used DeepSeek V4 Flash to build a Rust-based C compiler with about 250k tracked lines; the run lasted over 120 hours, archived over 1,000 agent episodes and incurred only US$44 in model-token charges. The compiler passed the complete c-testsuite and most LLVM and Csmith tests. In a separate compiler world generated with GLM 5.2, development continued after repeated agent replacement while retaining full test performance. Genesis also reimplemented 13 MESA modules with over 100k Fortran lines as a Rust workspace with nearly 90k Rust lines; across six numerical workloads, it achieved median speedups of 1.55--6.87x. These results show that long-horizon software development can be organized around a persistent project rather than a persistent agent.

Jonathan Vanian / CNBC : Reddit will join the S&P 500 on August 18, becoming only the second pure-play social media company in the index after Meta; RDDT jumps 10%+ after hours —  Reddit shares rose 11% in extended trading on Thursday on the social media company's inclusion in the S&P 500.

Frederic Lardinois / The New Stack : OpenAI launches Computer History, an opt-in feature that turns recent computer activity on macOS into memories and a timeline that ChatGPT and Codex can use —  OpenAI's opt-in Computer History captures a local timeline of Mac activity so ChatGPT Work can recover context across selected apps and sites.

Stephen Totilo / Game File : Netflix is closing game studios Night School Studio, which it acquired in 2021, and Moonloot, and is also cutting jobs from its internal games team —  Oxenfree Night School and Moonloot closing, as streaming giant continues to shrink internal game studio operation

Rebecca Bellan / TechCrunch : Anthropic details multiagent experiments showing Claude agents can wage a “turf war” over incompatible goals, fail to coordinate, collude on prices, and more —  What happens when you pit AI agents against each other?  According to Anthropic's testing, things get messy fast.

Milana Vinn / Reuters : Sources: Silver Lake is in talks to acquire HR and financial management software maker Workday, which has a market value of ~$43B; WDAY jumps 18%+ —  Private equity firm Silver Lake is in talks to acquire Workday (WDAY.O), which has a market value of about $43 billion …

Krisztian Sandor / CoinDesk : Tether says KPMG completed Tether's first full audit and found its 2025 financial statements fairly presented its financial position in all material respects —  Tether, the company behind the world's largest stablecoin USDT USDT  —  , said Thursday that it had completed a full financial audit …

Reuters : Sources: Vantage Data Centers, a major hyperscale developer, is exploring options including a sale or an IPO at a valuation of about $100B as soon as next year —  Vantage Data Centers, a major hyperscale developer and operator, is exploring options including an IPO or sale as soon as next year …

Chris Welch / Bloomberg : US-based robot vacuum maker Matic rolls out a firmware update that lets users control its $1,245 vacuum with voice commands in more than 70 languages —  The device, from Silicon Valley brand Matic, can be controlled fully by voice instructions, helping it stand apart from the competition.

Zac Hall / 9to5Mac : OpenAI previews Ultrafast, an API tier powered by Cerebras that runs GPT-5.6 Sol up to 14× faster and generates up to 750 output tokens per second —  OpenAI is previewing a new way to run its most capable GPT-5.6 model at dramatically higher speeds.  The company says its new Ultrafast …

Tulsee Doshi / Google : Google unveils Gemini 3.7 Flash, its “most intelligent workhorse model” for coding and agents, pricing it at $0.75/1M input and $3.75/1M output tokens at launch —  Our most intelligent workhorse model yet for coding and agents.  —  Senior Director, Product Management, on behalf of the Gemini team

Sarah Perez / TechCrunch : X makes the source code for its For You timeline available on GitHub and adds tools to let users see if X's ranking systems “shadowbanned” them —  X is significantly expanding its open-source codebase, which includes the app's “For You” algorithm and its core ranking engine …

Ara Kharazian / Ramp : Ramp's July AI index: Anthropic's market share hit 43.5%, widening its lead over OpenAI; Fable 5 is only 6% of tokens businesses bought, likely due to high cost —  Dear Colleagues: Today's letter includes my monthly update of Ramp AI Index, our flagship research using spend data from Ramp to track how American businesses are using AI.

Shirin Ghaffary / Bloomberg : OpenAI hires Dali Rajic, COO of Alphabet's Wiz, as chief revenue officer, replacing ex-Slack CEO Denise Dresser, who was hired in December 2025 and will leave —  OpenAI is naming its second new chief revenue officer in less than a year, a sign of efforts to bolster sales growth ahead of a highly anticipated Wall Street debut.

Bloomberg : HPE secures court approval for its settlement with the US DOJ over its takeover of networking company Juniper, defeating a challenge by Democratic state AGs —  Hewlett Packard Enterprise Co. secured court approval for its settlement with the US Justice Department over its takeover …

Wall Street Journal : Tech hardware companies report the highest US CBP tariff refunds at $2.5B across six companies, including $2.2B in refunds to Apple alone; Amazon received $640M —  Apple, Nike and FedEx are just a few companies recovering big sums relatively quickly, and some are passing at least a share on to customers

NASA 工程师找到一个方法，可以释放旅行者 2 号探测器的额外电力，帮助这台服役将近 50 年的航天设备延长科学探测工作的时间。代号为“大爆炸”的调整工程由 NASA 喷气推进实验室完成。团队同步关闭部分耗电组件，替换为功耗更低的备选方案，同时保证旅行者2号维持足够温度，保障设备正常运转。旅行者2号及其姊妹探测器旅行者1号，依靠放射性同位素热电发生器供电。该装置利用钚元素衰变释放的热量产生电能。随着钚燃料不断消耗，两台探测器每年可用电力都会减少约 4 瓦。每台探测器原本搭载 10 个仪器。自2024 年起，由于电力持续下降，团队被迫在两台探测器上各自关闭两个科学仪器。其余仪器此前就已经关停，它们仅用于任务初期的行星飞越探测阶段。如果没有这次“大爆炸”工程的节电改造，旅行者系列探测器项目团队原本需要在 2026 年底前再关闭旅行者2号上的一个仪器。此次改造释放出来的电力，预计可以让探测器剩余3个仪器至少再多工作一年。NASA计划对距离地球更远的旅行者1号实施同样的节电改造工程。

消费级神经技术正从实验室迅速走向工作场所和家庭，包括苹果、Meta 和 Snap 在内的公司开发能监测或解读大脑活动的产品。研究人员警告，神经数据可能成为下一个隐私争夺焦点。部分神经学家开始呼吁“神经权利”以保护“精神隐私”、“身份”和个人“自主权”。过去二十年研究人员利用 fMRI 技术构建日益详细的哺乳动物脑皮层结构图谱和相关连接清单。得益于机器学习 AI 的巨大进步，fMRI 扫描现在可用来识别从抑郁情绪到嫉妒和幸灾乐祸等情绪。

根据发表在《American Journal of Cardiovascular Drugs》期刊上的一项研究，爬楼梯能显著降低死于心血管疾病的风险。研究期间，相比与不常爬楼梯的人，常爬楼梯的人死于心脏相关疾病的风险降低 39%，全因死亡率降低 24%。常爬楼梯的人患心脏病、中风和心力衰竭等严重心血管疾病的风险也较低。研究人员称，心血管疾病是全球首要死因，1990-2019 年间相关病例数几乎翻了一番。研究人员分析了来自 9 项高质量研究的逾 48 万名参与者的数据，中位随访时间为 14 年。研究人群包括健康参与者和有心脏病史的参与者。年龄范围为 35-84 岁，其中 53% 为女性。“与去健身房或进行锻炼不同，爬楼梯很容易融入日常生活，无论是在家、在工作场所还​​是外出。对于没有太多时间或不方便锻炼的人而言，这是一个不错的选择。”“所以，如果可以选择走楼梯或乘电梯，那就选择楼梯，因为它对你的心脏有益。”

2025 年 NASA 毅力号火星车利用激光对 Jezero 陨石坑周围浅色岩石进行分析时，意外发现了此前在火星上从未发现过的物质：刚玉。部分刚玉颗粒还含有铬——在地球上掺有铬的刚玉被称为红宝石。火星岩石中发现刚玉令人惊讶，因为火星条件很难解释它们是如何形成的。毅力号是在三块富含斜长石的岩石块中发现刚玉颗粒，这三块岩石分别被称为 Hampden River、Coffee Cove 和 Smiths Harbour，它们都是浮岩——即不再附着在形成它们的基岩上的岩石，它们可能来自其他地方。对于刚玉如何形成，研究人员猜测可能是几十亿年前火星表面遭到撞击后产生的，地球以及月球的撞击蚀变岩石中都发现了刚玉。

Django Python Web 框架项目宣布采用每年发布一个版本的发布模式，放弃以前使用的短期支持、长期支持版本交叉发布方案。Django 不再有长期支持版本（LTS），每个功能版本都有相同的支持承诺。每个功能版本都有三年支持时间，第一年为主流支持，之后两年为安全更新和数据丢失 bug 修正。Django 项目将在任何时候支持三个版本，用户可以按照自身需求在支持期间升级。从 2028 年 1 月发布的 Django 2028 起开始采用这一发布模式。

扫描且不破坏珍本图书是一件费时费劲的工作。根据互联网档案馆 2021 年发表的一篇文章，该机构尝试过自动化图书扫描，但对于易碎珍本图书而言不可行，干净干燥的人手仍然是最佳的翻页方式。自 2010 年起就在互联网档案馆工作的图书扫描员 Eliza Zhang 表示这项工作需要高度集中注意力。她当时已扫描了逾 300 万页、14000 张折页和 18000 件物品。AI 公司如果要扫描图书显然不太可能使用这种方式。最近发表的多篇报道引发了对 AI 公司收购然后销毁珍本图书的担忧。《每日电讯报》称硅谷销毁了数百万册珍稀图书并“撕碎了原件”，404 Media 报道称名为 ISBNdb 的图书数据库公司宣传其可以帮助 AI 公司大量获取图书。2025 年的图书侵权诉讼披露 Anthropic 内部启动了扫描数百万册图书的 Project Panama 项目。但没有证据表明 Anthropic 曾销毁过珍本图书，该公司也公开声明否认。包括马斯克旗下 xAI 在内的 AI 公司也公开表示，他们不会为了训练 AI 而销毁珍本图书。与此同时，珍本图书零售商则在标记可疑订单，因为他们知道，大多数真心想购买珍本图书的人通常只会购买单本，而不是一次性订购大量不同种类的图书。爱尔兰书店 Kennys 报告了一笔令人匪夷所思的订单，该订单订购了 5000 本冷门图书。买家甚至没有试图还价。

Google 本月初重组了其 AI 部门 DeepMind 的领导层，知情人士表示此举旨在赶上竞争对手。Google 的 Gemini 模型去年 11 月一度短暂超越竞争对手，但 Anthropic 和 OpenAI 的新模型使其再次陷入追赶状态。Google 推迟了 Gemin 新模型的发布，原因是内部测试显示它在编程等领域仍然落后于竞争对手。Google 联合创始人 Sergey Brin 在今年 4 月的一次员工大会上督促加快 AI 研发。通过削弱 DeepMind 的自主权，Google 正集中更多精力在 Gemini 模型的开发上。

位于银河系中心的超大质量黑洞 Sgr A* 周围，存在极端强烈的辐射与复杂的星际环境，长久以来被认为不利于尘埃与分子长时间存在。最新研究利用韦伯太空望远镜（JWST）的中红外线仪器（MIRI），观测距离 Sgr A* 投影约 0.17 秒差距的恒星 IRS 3。研究团队在 2025 年的观测中取得 4.9 至 27.9 微米的中红外线光谱，首次获得 IRS 3 完整且经前景消光修正的连续光谱，借此分析其周围尘埃包层的化学组成与物理环境。更令人瞩目的是，研究团队首次在 IRS 3 的包层中确认明确的水吸收特征，模型显示水可能位于距离恒星约 900 天文单位的内侧包层。这项发现显示，即使在 Sgr A* 周围由强烈辐射主导的极端环境中，基本分子仍可能存活。超大质量黑洞附近并非完全无法形成或保存尘埃与分子。

自 2026 年 7 月下旬起，约有 75,000 名移民从摩洛哥越过边境，进入西班牙位于北非的自治市休达。休达政府为此宣布进入“绝对的人道与社会紧急状态”，请求西班牙中央政府宣布国家紧急状态，派军队支援。这一切是如何发生的？在数万移民涌入休达两天前，当地的一家报纸《El Faro de Ceuta》在 TikTok 和 Instagram 上发布了一则视频，展示了两名身穿潜水衣的年轻女性在城市中行走，头发仍然湿漉漉的。虽然未明说，但意思非常明显：两名女性是刚从摩洛哥游泳穿越边境过来的。报纸宣称：“休达再也承受不了了。”在边境的另一侧，这则视频引发了截然不同的反响。被裁剪后用阿拉伯语转发的视频看起来像是一封邀请函。视频中的女性面带微笑，比着和平手势和大拇指，最重要的一点是，她们在自由行走。这段视频似乎验证了社媒上流传数周的传言——西班牙最高法院最近的一项裁决意味着非法抵达的移民可以留在该国。正如 Facebook 上一个账号错误声称：“休达正变成一扇敞开的大门”，但实际上移民在司法审查后仍可能被驱逐。此事成为了社媒直接导致现实悲剧的最典型案例之一。

安全研究员 Cory Solovewicz 分别在 2020 年和 2024 年购买了两个域名 noreply.us 和 noreply.net，他原本计划用于过滤该域名地址收到的邮件，结果发现有很多企业的邮件系统也会向该域名发送邮件，而且数量非常庞大。他无意中打造了一个蜜罐。他在本月举行的 Defcon 安全大会公布了结果：noreply.net 自 2024 年 12 月以来收到了 401,796 封邮件，平均每天 699.99 封；noreply.us 数量没有这么多，自 2020 年以来发送了 37255 封邮件。他指出，邮件是企业邮件系统自动发送的，并非人工撰写，这些都是内部系统配置错误导致的。另一名安全研究员 Mike Sheward 在购买了 deleteduser.com 域名之后也有类似的发现，企业没有真的彻底删除用户账号，而只是改了电邮地址。Sheward 透露，一家 AI 公司向该域名发送了大量邮件，该公司利用目标识别技术检测中东工业场所中可能违反安全规程的工人，他从这家公司收到了数千张 CCTV 监控录像截图。

Google 宣布了 Pixel 11 系列智能手机，一大变化是所有型号的起售价上涨 100 美元。现已开放预购， 8 月 20 日发售。Pixel 11 系列沿用了 Pixel 9 的设计风格，主要变化包括：通知 LED 的回归，但仅限于 Pro 型号，基础款没有；屏幕亮度提升，基础款仍然是 3000 尼特，但 Pro 型号提高到 3600 尼特；电池容量略有下降，但 Google 声称由于新 SoC Tensor G6 效能提升了 20%，电池续航时间与上一代差不多；Tensor G6 细节未公开，泄露信息称它是一款 7 核 Arm 芯片，比 Tensor G5 少一个 CPU 核心；摄像头像素与上一代一致，但变焦倍数提升至 30 倍，Pro 版提升至 120 倍；其它是 AI 功能的增强，等等。Pixel 11 起售价 899 美元，Pixel 11 Pro Fold 最高售价 1899 美元。

安徽滁州一名 67 岁农民疑因过度相信 AI 提供的建议，误用除草剂导致大批芝麻苗死亡，150 亩农田一夜之间几乎全毁，损失惨重。吴姓农民约 1 年前开始接触 AI 软件，从农时安排、施肥到农药使用，几乎大小事都会询问 AI。起初他对相关建议半信半疑，但长期使用后逐渐建立信任，甚至将 AI 提供的方案直接付诸实行。吴大伯 7 月 10 日向 AI 询问芝麻如何除草、除虫，AI 随即生成一套“百亩芝麻飞防除草＋除虫全套方案”，其中推荐使用“高效氟吡甲禾灵”及“氟磺胺草醚”除草，另搭配“噻虫嗪”及“甲维盐”等药剂。吴大伯未进一步向农技人员确认，便依照方案喷药。没想到隔天田里便出现严重异常，芝麻苗与杂草大面积枯死。事后他再次询问 AI，AI 才指出配方中的“氟磺胺草醚”可能是造成芝麻死亡的主要原因。当地农技人员表示，氟磺胺草醚主要用于大豆田防除阔叶杂草，并不适用于芝麻田；即使部分其他作物可使用，也必须采取定向喷洒，不能全田施用，否则可能造成药害。吴大伯认为，AI 提供方案时未明确提醒风险，若当时多一句警告，他可能不会照做。不过，AI 对话页面上方其实标示“AI 生成可能有误，注意核实”的提醒，只是他此前从未留意。涉事 AI 软件客服表示，系统没有独立知识库，回答主要是根据网络公开信息整合生成，目前将进一步核查信息来源。至于农户损失，客服表示已完成登记并会向内部反映。

城市由于黑色沥青和混凝土吸收太阳辐射而变得比周边地区更热。这种被称为“城市热岛效应”的现象，是研究城市升温的基础。气象和卫星数据普遍显示，过去几十年来这一效应已不断加强。一项新研究发现，城市未来很快将不再比周围地区升温得更快。到 2050 年，尽管城市平均气温仍高于周边乡村，“但城市的农村周边地区升温速度更快”，除寒冷地区外，城市热岛效应正在减弱。此前的研究表明，随着气温升高导致乡村地区更加干燥，蒸发降温空气的能力下降，因此乡村的升温速度将加快，从而缩小与城市的温差。这一结果并不意味着气候变化对城市的升温影响会消失，只是说明城市与乡村之间的温差可能减小。

移民涌入会导致犯罪率上升吗？加州尔湾的研究人员分析了 2010-2018 年间美国 11500 个社区 4600 万居民的犯罪趋势，结果显示犯罪率与无证移民无关联。与很多人的说法相反，非法移民比例上升的社区，财产犯罪率下降幅度更大，暴力犯罪率没有显著变化。研究发现，非法移民人口不断增长的社区抢劫案略微上升，但不是移民抢劫本地居民，而是这些移民成为抢劫对象。研究人员指出，非法移民更容易成为抢劫的目标，是因为他们更多地参与现金交易。很多非法移民无法获得银行账户和正规金融服务，他们更有可能携带现金，这使他们成为抢劫的目标。对被驱逐出境的恐惧也使他们不愿向警方报案，从而更容易受到伤害。移民被称为“行走的 ATM 机器”。

美国国家海洋和大气管理局（NOAA）称，7 月是美国有记录以来最热的月份。美国本土 48 州 7 月的平均气温为 24.94 摄氏度，比 20 世纪平均气温高出 0.6 摄氏度，略高于 1936 年 7 月气温。2021 年和 1936 年的夏季是美国历史上最热的夏季。NOAA 表示，此次破纪录高温的最大因素是夜间最低气温更高。夜间气温升高被广泛认为是人为气候变化的典型迹象。NOAA 的计算基于 24小 时平均气温，而不仅仅是最高气温或最低气温。