OrangeBot.AI Digest — 2026-07-11

Your org changes. Access keeps up.

A distraction-free breathing app for labor preparation

AI video & audio effects SDK for real-time apps

Drop your folder in browser & deploy instantly on Cloudflare

Partner for your most ambitious work

A startup tycoon game

SoundPipe is a mixing board for your Mac

The terminal and Mac menu bar companion for Kickbacks.ai

Your AI video editor in ChatGPT, desktop, and web

A new standard for intelligence and efficiency

Open-source workspace for AI agents and workflows

Dictate, rewrite, translate, and an agent in a single device

An AI coworker that lives in your Slack to get work done

Beautiful Mac battery alerts, health insights & charge limit

AI digital business card that remembers everyone you meet

We introduce Vidu S1, a real-time interactive video generation model supporting voice control of digital characters. Users can control video generation content at any moment through voice instructions. Vidu S1 supports infinite-length real-time video generation without blurring, drift, or visual distortion. Built with TurboDiffusion and TurboServe, Vidu S1 outputs 540p real-time videos at up to 42 FPS on regular consumer GPUs. Users can upload custom images of real people, anime, and pets, and choose different voice tones for personalized experiences. Experiments show that Vidu S1 achieves the best performance across all test metrics while fully meeting real-time inference requirements. A playable online demo is available at https://vidu.com/vidu-stream.

The inherent complexity of video understanding makes it difficult to determine whether Video-LLM benchmark performance stems from visual perception, linguistic reasoning, or knowledge priors. While many benchmarks have emerged to assess high-level reasoning, shared criteria for evaluating video understanding remain largely overlooked. Instead of introducing yet another benchmark, we take a step back to re-examine the criteria for evaluating video understanding. In this work, we introduce Video-Oasis, a sustainable diagnostic suite for systematically auditing existing video understanding benchmarks. This audit reveals that 55\% of existing benchmark samples are solvable without visual input or temporal context. After filtering these shortcuts, the remaining video-native challenges expose a substantial capability gap: state-of-the-art models perform only marginally above random guessing. Building on these findings, we use the distilled challenges as a testbed to investigate which algorithmic design choices contribute to robust video understanding. We hope our work provides a practical foundation for constructing rigorous video benchmarks and evaluating future Video-LLMs. Code is available at https://github.com/sejong-rcv/Video-Oasis.

Zero-Shot Compositional Action Recognition (ZS-CAR) requires recognizing novel verb-object combinations composed of previously observed primitives. In this work, we tackle a key failure mode: models predict verbs via object-driven shortcuts (i.e., relying on the labeled object class) rather than temporal evidence. We argue that sparse compositional supervision and verb-object learning asymmetry can promote object-driven shortcut learning. Our analysis with proposed diagnostic metrics shows that existing methods overfit to training co-occurrence patterns and underuse temporal verb cues, resulting in weak generalization to unseen compositions. To address object-driven shortcuts, we propose Robust COmpositional REpresentations (RCORE) with two components. Co-occurrence Prior Regularization (CPR) adds explicit supervision for unseen compositions and regularizes the model against frequent co-occurrence priors by treating them as hard negatives. Temporal Order Regularization for Composition (TORC) enforces temporal-order sensitivity to learn temporally grounded verb representations. Across Sth-com and EK100-com, RCORE reduces shortcut diagnostics and consequently improves compositional generalization.

Scientific ideas rarely start from a blank page. They inherit mechanisms, repair known limitations, and recombine pieces of earlier work, much like biological genomes. Current benchmarks still say little about whether AI systems can follow this inheritance structure. We present IdeaGene-Bench (IG-Bench), a benchmark for scientific lineage reasoning and lineage-grounded idea generation. IG-Bench is organized around the IdeaGene framework: each paper or proposal is represented as a set of minimal, typed, evidence-grounded Idea Genome objects, and a GenomeDiff aligns these objects to record inheritance, mutation, loss, external import, and novel insertion under six operational evolutionary dynamics. The benchmark contains 1,961 golden lineage traces, 1,085 curated Idea Genome objects, and 920 pairwise GenomeDiff records across 10 scientific domains. It supports two evaluations. IG-Exam (42 task types, 1,029 instances) tests closed-form lineage reasoning across Idea Genome abstraction, inheritance tracing, evolutionary reasoning, and lineage verification. IG-Arena evaluates generation with a lineage-conditioned Population-Evolution Score(PES), asking whether a proposal can be inserted as a coherent descendant of a given lineage population: it should inherit the right Idea Genome objects, vary meaningfully from nearby work, and offer selection value for future research. Experiments on 14 LLM-based scientists expose a compositional bottleneck. The strongest system reaches only 27.3% exact accuracy on lineage reasoning, and structured lineage context reshuffles system rankings rather than helping every participant uniformly.

The rapid development of large language models and multimodal large language models has accelerated the emergence of proactive agents capable of operating everyday tools and assisting users in real-world environments. However, existing benchmarks struggle to evaluate such agents effectively, as they often rely on sandboxed environments and single-turn evaluation paradigms. Moreover, their scenario-based task taxonomies mix multiple model capabilities within the same task category, making it difficult to identify the root causes of agent failures. To address these limitations, we introduce UniClawBench, the first capability-driven benchmark designed to evaluate proactive agents in dynamic, real-world settings. UniClawBench is built around five foundational model capabilities: Skill Usage, Exploration, Long-Context Reasoning, Multimodal Understanding, and Cross-Platform Coordination. Based on these capabilities, we design 400 bilingual real-world tasks. Unlike previous benchmarks that rely on static, pre-recorded answers, our benchmark evaluates agents in live Docker containers using fine-grained, step-by-step completion checkpoints. Furthermore, we design a closed-loop evaluation strategy comprising an executor agent, a hidden supervisor agent, and a user agent to simulate realistic multi-turn human feedback without leaking grading criteria. To disentangle base model capabilities from framework-level design choices, we evaluate state-of-the-art models under multiple agent frameworks. Through comprehensive comparisons across both models and frameworks, we show how base model capabilities and agent framework designs jointly shape performance in real-world environments. To facilitate future research, we make our benchmark and code publicly available at https://github.com/HKU-MMLab/UniClawBench.

Recovering high-quality video from sparse event streams is a challenging task. Regression methods often blur textures, while existing generative models struggle with long-term stability. We propose LongE2V, a novel approach that leverages pre-trained video diffusion priors to jointly handle event-based video reconstruction, prediction, and frame interpolation. By fine-tuning a foundational video model, our approach achieves high data efficiency and superior perceptual quality. We introduce Autoregressive Unrolling and Adaptive Context Switching to mitigate temporal drift in extremely long sequences. We also propose Reencoding Alignment with Cross Residual Correction to ensure precise bidirectional consistency during frame interpolation. Furthermore, Event Voxel Density Augmentation ensures robustness across varying sensor resolutions. Extensive experiments on real-world benchmarks demonstrate that LongE2V outperforms state-of-the-art methods across all three tasks, exhibiting exceptional temporal coherence and zero-shot generalization. Project page: https://cdfan0627.github.io/LongE2V-page/

Current computational approaches for drug design typically focus on generating molecules conditioned on specific targets or general molecular properties, often neglecting the influence of disease context on target behavior and therapeutic outcomes. To address this gap, we introduce DrugGen-2, a novel generative model that designs small molecules conditioned on both disease ontology and target protein sequences. DrugGen-2 was developed by fine-tuning a pre-trained GPT-2 model on a curated dataset of approved drugs linked to their diseases and targets, using a two-step strategy of supervised fine-tuning followed by reinforcement learning via group relative policy optimization (GRPO). This process was guided by reward functions optimizing for chemical validity, novelty, diversity, and high predicted binding affinity. When evaluated on five protein targets relevant to diabetic nephropathy, DrugGen-2 significantly outperformed baseline models (DrugGPT and DrugGen). It demonstrated a superior capacity to generate unique molecules, exhibited greater structural similarity to approved drugs, and achieved improved predicted binding affinities across all targets. Molecular docking analyses further supported these findings, identifying candidate ligands with strong binding potential, including compounds with predicted affinities (-9.917, -9.485, and -9.367) exceeding those of reference drugs such as enalapril for angiotensin-converting enzyme (-8.283). By integrating disease-specific context into molecular generation, DrugGen-2 advances AI-assisted drug discovery, offering a powerful tool for de novo design and drug repurposing that accounts for the complex interplay between diseases and molecular targets.

In this work, we present Canvas360, a two-stage framework for in-context panoramic generation that combines geometry-aware pretraining with downstream task-specific fine-tuning. To address the lack of large-scale, high-quality training data tailored to in-context panoramic tasks, we propose Canvas360Dataset, a collection of 1M high-quality paired panoramic samples for style transfer, inpainting, outpainting, and editing, enabling effective supervision across diverse in-context generation scenarios. On the modeling side, Canvas360 enhances text-to-panorama generation through parallel depth generation, velocity circular padding, and similarity loss regularization, enabling the model to learn geometry-aware representations, capture object distortion details, and improve geometric consistency and global coherence. Furthermore, empowered by strong panoramic priors, Canvas360 enables a unified in-context panoramic generation framework that supports diverse downstream tasks via token-level concatenation, surpassing prior methods in both task coverage and modeling flexibility. Extensive experiments show that Canvas360 improves panoramic image fidelity, achieving particularly strong performance on the panorama-specific FAED metric and competitive or leading results across the reported quantitative evaluations. More information can be found on our project page: https://zry000.github.io/Canvas360/

Modern LLMs are increasingly deployed in long-context applications such as retrieval-augmented generation, repository-level coding, and agentic workflows whose accumulated reasoning and tool traces routinely push the input an order of magnitude past the pretraining window, making zero-shot context extension the dominant deployment path for open-weight checkpoints. Most existing zero-shot methods fix a single rescaling factor up front, so an aggressive factor sacrifices short-context fidelity while a conservative one breaks down at long contexts. We propose Jet-Long, a tuning-free zero-shot method that pairs a local RoPE-faithful window with a long-range window whose rescaling factor adapts dynamically to the current sequence length, recovering the base model exactly at short inputs while extrapolating cleanly at long ones. An inclusion-exclusion attention merge and an on-the-fly RoPE correction rotation make the bifocal construction essentially free at inference; fused into a single CuTe kernel, long-context prefill reaches up to 1.39times FA2 throughput on H100 (approaching the Hopper-only FA4), and single-batch generation incurs le 4% overhead at every length. On Qwen3-1.7B/4B/8B up to 128K context, Jet-Long leads RULER by +4.79/+2.18/+2.03~pp over the strongest baseline at 1.7B/4B/8B, achieves the best overall accuracy on HELMET-RAG (a benchmark identified by HELMET as the most efficient predictor of downstream long-context performance) and attains the lowest PG-19 perplexity. Jet-Long also generalizes to hybrid attention architectures such as Jet-Nemotron for further long-context improvement without retraining, and remains hyperparameter-resilient for ease of deployment.

The growing demand for image-to-video creation on mobile devices has increasingly focused on cinematic motion effects like bullet time, dolly zoom, slow motion, etc. While Diffusion Transformers (DiTs) exhibit strong performance in video generation, their large parameter sizes and multi-step iterative denoising processes lead to substantial computational overhead, making efficient generation on mobile devices challenging. We propose CineMobile to bridge the gap. In particular, CineMobile adopts a three-fold optimization strategy: (1) leveraging a distillation-guided pruning approach to derive a compact yet efficient model that retains the essential video generation capabilities required for cinematic effects; (2) optimizing the compressed model into a 4-step generator via a combination of diffusion distillation and reinforcement learning; (3) employing a hybrid post-training quantization strategy to compress the model footprint to under 1 GB. Experimental results show that compared to the teacher model with the Wan 2.1 architecture, CineMobile achieves a 40x speedup in generation while maintaining comparable visual quality. Specifically, CineMobile generates 49-frame 480p videos with a per-step denoising latency of 0.6s on an NVIDIA H200 GPU and 20s on the MediaTek Dimensity 8400 Ultimate 5G platform, with a peak memory usage of 1.8 GB, demonstrating its practical applicability for mobile-based image-to-video creation.

Reasoning has become a core capability for large models, especially when reliable decisions require understanding logical consequences. Recent video generation models offer a reasoning path distinct from previous Chain-of-Thought (CoT): reasoning can unfold through temporally connected frames, known as Chain-of-Frame (CoF) reasoning. However, existing video generators are primarily trained on general video corpora, still lacking diverse supervision and dedicated designs for CoF reasoning. To address this gap, we introduce OpenCoF, a framework comprising the OpenCoF-17K dataset, a reasoning video dataset spanning 11 task families, and Wan-CoF, a fine-tuned video model for studying whether diverse temporal supervision improves CoF behavior. Across four video reasoning benchmarks, Wan-CoF achieves considerable gains over the Wan2.2-I2V-A14B baseline. Building on this, we empirically explore more advanced designs for CoF capabilities, i.e., equipping the model with visual and textual reasoning tokens. This mechanism respectively captures low-level visual cues and high-level semantic priors for spatial and temporal reasoning. Through performance comparisons and attention analysis, we examine how these tokens contribute across model depth, denoising steps, space, and time. Our results suggest that stronger video reasoning requires both broad temporal supervision and explicit mechanisms for organizing intermediate reasoning state. We open-source the dataset, model, and code to facilitate future research on reasoning-oriented video generation.

Self-attention lets each token retrieve information from the full context, but its quadratic cost in sequence length limits training and inference at long context. This paper presents a comparative study of softmax attention and four recent recurrent linear-attention architectures: DeltaNet, Gated DeltaNet, Kimi Delta Attention, and Gated DeltaNet-2. We express these mechanisms in a common recurrent-memory notation, making explicit how they differ in expressivity, memory decay, erase and write control, training throughput, and implementation complexity. Our experiments center on 350M-parameter models trained for 15B tokens, and include optimizer and learning-rate comparisons, hybrid-versus-pure stack comparisons, sequence-length runtime measurements, larger DeltaNet runs at 1.3B and 3B parameters, and a small set of downstream evaluations. The reported speed results measure training throughput and iteration time; we do not provide an empirical inference-speed benchmark. Within the reported 350M-parameter, 15B-token sweep, Kimi Delta Attention with Muon reaches the lowest final validation loss, a pure Gated DeltaNet stack trained with AdamW has the highest normalized training throughput, hybrid stacks generally improve loss at a throughput cost, and Muon consistently lowers final validation loss relative to AdamW in the matched architecture settings we evaluate. We introduce and evaluate lightweight cross-layer routing mechanisms for DeltaNet-style memories. The most natural DeltaNet-inspired formulation, forwarding a lower layer's delta-rule write error into the next layer's value target, does not improve over matched baselines. Routing into the aligned hidden stream and forwarding the write value instead yields a modest improvement in the matched runs we report: Cross-Layer Value Routing (CLVR) lowers final validation loss for both DeltaNet and Gated DeltaNet.

In long-horizon tasks, decision-relevant state is often scattered across an expanding trajectory, while the action agent must surface it and act. As trajectories grow, task requirements, environment facts, prior attempts, diagnoses, and open subgoals can be buried in the context window or pushed beyond it, failing to influence decisions when needed. We call this failure mode "behavioral state decay". We study memory as an active intervention mechanism rather than passive retrieval. A separate memory agent runs alongside an unmodified action agent, updating a structured memory bank from the recent trajectory and deciding whether to inject a memory-grounded reminder or remain silent. The module is plug-and-play with frontier action agents and existing agent harnesses. Across Terminal-Bench 2.0 and τ^2-Bench, it improves pass@1 for both weaker and stronger action agents, with gains of +8.3 pp on Terminal-Bench and +6.8 pp on τ^2-Bench. Ablations show that selective intervention outperforms passive bank exposure, always-on injection, advisor-only guidance, and general retrieval. As an early step toward open-weight memory policies, we train Qwen3.5-27B on SETA using SFT and GRPO, improving validation reward and achieving partial transfer to Terminal-Bench.

Reinforcement learning (RL) has become the standard paradigm for enhancing the complex reasoning capabilities of large language models (LLMs). To achieve sample efficiency, modern RL frameworks rely on importance sampling (IS). However, these algorithms suffer from an exploration-stability dilemma. Pure IS often leads to catastrophic training instability, while standard clipping mechanisms used to mitigate this instability strictly constrain the policy update budget. By formalizing the concept of Probability Capacity (Cap), we reveal that conservative clipping structurally stifles exploration by prematurely truncating the update budget for correct but low-confidence reasoning paths. To break free from these constraints, we propose Unbounded Positive Asymmetric Optimization (UP), a universal and plug-and-play objective. UP theoretically restructures the optimization process by anchoring the policy to its current state via the stop-gradient operator. This asymmetric design unleashes unclipped, stable gradients for positive advantages to maximize exploration, while maintaining standard clipping safeguards for negative advantages to prevent training instability. Furthermore, our formulation readily extends across different optimization granularities, including token-level (GRPO, DAPO) and sequence-level (GSPO) frameworks. Extensive experiments demonstrate that UP enhances exploration capacity and achieves superior reasoning accuracy across diverse RL algorithms (DAPO, GSPO, and GRPO), model architectures (Dense, MoE, and vision-language), and training modalities (language and multimodal), validating UP as a truly universal plug-and-play enhancement for RL-based training.

Generating realistic 3D human motions in real-time within interactive applications is key for animation, simulation, and humanoid robotics. While recent offline motion generation approaches offer precise control via text and kinematic constraints, they lack the inference speed required for interactive settings. Conversely, existing online methods enable real-time synthesis but often sacrifice controllability or struggle with complex text semantics and long-horizon goals due to limited context windows. In this work, we introduce ARDY, a streaming generation framework that bridges this gap by enabling high-fidelity motion generation controllable via online text prompts and flexible kinematic constraints. ARDY employs a hybrid representation that combines explicit root features with a latent body embedding, balancing precise trajectory control with efficient generative learning. We propose a two-stage autoregressive transformer denoiser that features variable history context and supports conditioning on flexible, long-horizon kinematic constraints. By training on a large-scale motion capture dataset and being directly conditioned on text labels and kinematic constraints sampled from ground truth poses, ARDY natively learns controllable generation that supports online prompting and flexible long-horizon goals. Extensive evaluations on the HumanML3D benchmark and the large-scale, high-fidelity Bones Rigplay dataset demonstrate ARDY's high motion quality and constraint adherence, validating the efficacy of our key architectural decisions. Finally, we demonstrate the method's practical versatility through an interactive demo featuring dynamic text control, diverse keyframe pose constraints, path following, and interactive locomotion control via mouse and keyboard. Supplementary video results, code, and model releases can be found at https://research.nvidia.com/labs/sil/projects/ardy/.

Kaitlyn Tiffany / The Atlantic : Psychologist Peter Gray argues that school stress, not smartphone use, is the main driver of the teen mental health crisis, challenging Jonathan Haidt's thesis —  An evolutionary psychologist is challenging the popular understanding of kids and technology.

M.G. Siegler / Spyglass : Apple's lawsuit could sidetrack OpenAI's hardware aspirations for years, or possibly forever, as the startup gets into yet another controversy and messy divorce —  If their lawsuit is successful, that ChatGPT device may be no more...  We all know the old saying “don't poke the bear.”

Mark Gurman / Bloomberg : Apple's OpenAI lawsuit follows months of simmering tensions and highlights OpenAI's hardware chief Tang Tan's strained relationship with former boss John Ternus —  When iPhone engineer Chang Liu quit for a job at OpenAI's nascent hardware division, Apple Inc. says he left with more than just years of experience.

Ernesto Van der Sar / TorrentFreak : Filing: Google urged the European Commission not to target DNS resolvers, VPNs, or IPs to fight piracy, calling the measures ineffective and easily circumvented —  Google has told the European Commission that pirate site blocking causes “significant harm” and should not target DNS resolvers, VPNs or shared IPs.

Guillermo Gallacher / Indeed Hiring Lab : US software development job postings on Indeed have grown by ~15% since the launch of Claude Code in February 2025, while overall job postings fell by 7% —  Agentic AI may be flipping the relationship between AI exposure and job posting growth.  —  Key points:

New York Times : How members of the extremist group Boko Haram are using AI chatbots to design explosives, fix or upgrade weapons, and brainstorm attack ideas —  A.I. chatbots are not just a propaganda tool for violent extremists but are aiding in bomb construction and attack planning, new research finds.

Bloomberg : SK Hynix's historic US stock market listing is a bet that the AI boom is breaking the memory chip industry's decades-long boom-and-bust cycle —  South Korean memory chipmaker SK Hynix Inc. just pulled off the largest public listing by a foreign company in US market history.

Max Spero / Pangram Labs : An analysis of 1M+ social media posts from April 24 to June 30: ~25% of longform posts with 250+ words were fully AI-generated; on LinkedIn, the figure was 41% —  A first look at analytics from our Pangram Chrome extension  —  Two months ago, we launched our Chrome extension to help combat the rising slop problem on social media.

Robbie Whelan / Wall Street Journal : Sources detail the Trump admin's heavy-handed intervention to aid Intel, including pushing it to expand local capacity and pressuring Apple to use Intel's fabs —  The chip maker's business is improving, with government twisting the arms of potential customers and partners including Apple and Nvidia

Bloomberg : Sources: activist investor Elliott has built a large stake in car insurance software maker CCC, which is exploring a potential sale and has a ~$3.5B market cap —  Elliott Investment Management has built a large stake in car-insurance software provider CCC Intelligent Solutions Holdings Inc. …

Maxwell Zeff / Wired : OpenAI's head of safety, Johannes Heidecke, is leaving as OpenAI integrates its research and safety teams; Mia Glaese will become VP of research and safety —  Johannes Heidecke's departure comes as OpenAI tries to further integrate its research and safety teams.

Thinking Machines Lab : Thinking Machines says its mission is to build AI that people and organizations can shape and make their own, and that “extends human will and judgment” —  The mission of Thinking Machines is to build AI that extends human will and judgment.  —  Artificial intelligence …

Reuters : SK Hynix CEO Kwak Noh-Jung says the memory industry is heading for its worst-ever supply shortage in 2027 and demand will outstrip supply beyond 2030 —  SK Hynix (000660.KS) Chief Executive Kwak Noh-jung said the global memory industry is heading for its worst-ever supply shortage in 2027 …

Corbin Bolies / Variety : Meta says it will discontinue a feature that allowed users to generate images in Meta AI using public Instagram accounts, following days of criticism —  Meta said it will discontinue an AI feature that allowed users to generate images using public Instagram accounts following days of criticism …

Marcus Mendes / 9to5Mac : In response to Apple's trade secret theft lawsuit, OpenAI says “we have no interest in other companies' trade secrets” —  OpenAI has issued a formal statement in response to Apple's lawsuit accusing the company of trade secret theft.  Read it below.  —  OpenAI denies Apple's allegations

在去年 12 月学校发生枪击案之后，布朗大学经济学教授 Roberto Serrano 首次让学生在家中完成期中考试。期中考试成绩显示大部分学生获得了满分或接近满分，他怀疑学生大规模使用 AI 作弊。因此决定将期末考试改为线下进行。他没有立即宣布期中考试无效，而是决定先看看期末考试的得分。如果期末考试得分的分布与期中考试的分布基本相同，那么他会将期中考试得分计入成绩。结果是 18 名学生退课，9 名学生没有参加期末考试。3 名学生得了零分，期末考试的平均分只有 48.6%——这是他所教班级迄今的历史最低分。此前期末考试的平均分从未低于 65%。只有少数学生的期末考试成绩与期中考试的成绩相近。

苹果起诉 OpenAI，指控该公司窃取其商业机密。这起诉讼的两位核心人物是 OpenAI 首席硬件官 Tang Tan 以及前苹果工程师 Chang Liu。苹果指控称，Tang Tan 会指导跳槽到 OpenAI 的苹果前员工如何规避苹果针对离职员工的安全流程。而 Chang Liu 被控秘密访问并下载了数十份苹果的机密硬件相关文件，其中包括大量关于未发布产品的详细信息、工程演示文稿、技术规格和私有项目数据。Liu 被控没有归还苹果配发的笔记本电脑，他还访问和使用了前同事的笔记本电脑，利用认证漏洞访问了苹果的共享网络文件夹。在发现漏洞之后，Liu 在前同事的笔记本电脑上留言说“LOL”。

日本宇航研究开发机构(JAXA) 7 月 11 日宣布，成功测试了一枚可回收火箭 RV-X。本次测试总共持续了约 40 秒，火箭上升至约 11 米的高度，期间完成了升空、悬停、水平移动、直立着陆共 4 个动作。目前旨在大幅降低航天任务成本的可回收火箭是许多国家的研发重点。RV-X 计划将取代不可回收的 H-3 火箭。下一步 JAXA 还将测试 RV-X 的升空一百米回收测试。法国、德国也参与到了日本的可回收火箭合作研发项目中。作为日本的战略竞争对手，中国在可回收火箭研发方面的进展要快得多。

根据一项新研究，具有特定设计元素的人造环境可能会给大脑带来过度负担，导致视觉不适和压力。视觉不适是指人在看到某些图像或环境时所体验到的不适感，可能表现为眼睛疲劳、偏头痛、阅读困难，或者在他人毫无问题的情况下感到不堪重负。条纹图案、凌乱的内饰、高对比度颜色、闪烁的灯光，甚至是超市中密集的货架，都可能导致视觉不适，这有助于解释为何某些空间会让人感到不舒服。现代人造环境与视觉系统在演化过程中高效处理的自然场景存在显著差异。研究还发现，现代环境对敏感人群影响更大，对感官输入更敏感的人群（如偏头痛、自闭症、注意力缺陷多动障碍、阅读障碍或癫痫患者）可能受到的影响更强烈。

新冠疫情加速了被称为安静辞职(quiet quitting)的现象，年轻一代的上班族将工作热情不高的态度视为某种形式的“辞职”，他们还想继续领工资，但仅完成最低工作要求，把精力放在工作之外的事情上。韩国嘉泉大学的研究人员调查了权威型领导如何推动中国中小企业员工的安静辞职现象（即躺平）。他们收集了 363 名中国中小企业员工数据。结果显示权威型领导通过增加工作倦怠间接导致躺平，而非自愿出勤通过放大倦怠加速躺平。研究有助于更深入地理解权威型领导带来的有害后果，阐明躺平在中国文化背景下的出现和演变。

根据发表在《自然》期刊上的一项研究，人形机器人成功完成切除活体动物胆囊的手术。人形机器人并不具有自主能力，它们并不会取代医生，而是由外科医生远超操作。远程操控的人形机器人为活猪完成两例微创胆囊切除手术。如果该方法被证实适用于临床，外科医生就可以利用人形机器人，在资源有限、无法安装专业且昂贵手术机器人的小型医院和诊所远程开展机器人辅助手术。人形机器人相比专业手术机器人更便宜，占用的空间更小，而且易于部署，能部署到从偏远地区到战场甚至太空。研究人员利用宇树科技的 G1 人形机器人。最便宜的 G1 基础型号起售价 13,500 美元，配置灵活机械手以及运费之后费用会超过 6.7 万美元。相比下 Intuitive Surgical 的达芬奇手术机器人费用在 50 万到数百万美元之间。宇树科技机器人的缺陷是需要频繁重校准，耗时更长。

长征十号乙运载火箭于 7 月 10 日 12 时 15 分在海南商业航天发射场发射升空，火箭一二级分离约 6 分钟后，一子级垂直返回，在海上回收平台通过网系捕获方式成功回收。此次回收的一子级预计将在今年年底前完成复用飞行。长征十号乙运载火箭由中国航天科技集团一院抓总研制，是 5 米直径两级串联构型的大型液体运载火箭。火箭芯一级沿用长征十号甲运载火箭一子级状态，采用液氧煤油推进剂，芯二级采用液氧甲烷推进剂；全箭起飞推力约 890 吨，起飞重量约 760 吨；首飞箭全箭长度约 63 米，重复使用状态下近地轨道运载能力 16 吨。该火箭可满足低轨卫星互联网星座部署、大型商业卫星发射等各类任务需求，复用状态下可大幅降低发射成本。

OpenRouter 的数据显示：美国企业每周调用中国 AI 的占比按代表数据处理量的“词元”统计，自 2 月起突破30%，峰值达 46%。相比下 2025 年上半年仅 4% 左右。背景是美国本土 AI 使用成本升高。美国OpenAI 及美国 Anthropic 等高端模型的性能不断提高，可自动处理长时间、高复杂度的任务。但随着企业在内部业务、面向客户服务中广泛使用 AI，词元的消耗量增加，使用成本增加。OpenRouter 平台拥有 800 万用户，以工程师群体为主，每输出 100 万词元（约对应 70 万英文词汇）的收费标准方面，Anthropic 的“Claude Opus 4.7”收费 25 美元，而最热门的 DeepSeek V4 Flash 仅收费 0.18 美元，成本不足前者的 1%。

AI 检测平台 Pangram 的研究显示，LinkedIn 和 X 等平台上四分之一的长文完全是 AI 撰写的。Pangram 对长文定义是包含至少 250 个字符，对 LinkedIn、Medium、Substack、X 和 Reddit 等平台帖子的分析显示，长文受 AI slop 影响最大，这些平台四分之一长文完全是 AI 生成，这里的“完全”并不包含用 AI 润色文字。研究显示，LinkedIn 的长文 AI 生成比例最高，达到了 41%，该平台包含 50-250 字的帖子 AI 生成比例也高达 30%。LinkedIn 上 55.2% 的长文是人类撰写的，4.3% 是在 AI 帮助下撰写的。X 上四分之一的推文完全由 AI 撰写，23.2% 的推文是在 AI 辅助下完成的，52.7% 的推文则是由人类撰写。Medium 上约三分之一的文章是 AI 撰写或 AI 辅助撰写，Substack 上有 21.9% 的文章是 AI 撰写或 AI 辅助撰写。Reddit 上 11.6% 的帖子是 AI 撰写或 AI 辅助撰写，98.1% 的评论是人类撰写的。

Google 表示世界杯期间其搜索量创下历史纪录，每秒查询量的峰值是在阿根廷与埃及比赛中间射入制胜球之后。 这一里程碑式的成就正值 Google 试图证明其传统搜索引擎在 AI 聊天机器人日益普及的时代仍能保持其重要性之际。Google 仍然占据着九成的搜索市场份额，其股价过去一年翻了一番多，第一季度营收增速是自 2022 年以来最快的。Google 表示赛后搜索量最高的查询是“阿根廷 vs 埃及（argentina vs egypt）”。在全球范围内，用户还搜索了“阿根廷 vs 哥伦比亚”和“梅西在世界杯上进了多少球”，以及“比赛中一名球员撞击其他球员叫什么”和“这是梅西的最后一届世界杯吗”。

美国国会议员正在调查美国公司使用中国 AI 模型。议员们担心审查、安全风险，以及美国 AI 公司的模型是否过于昂贵或限制过多。Cursor 和 Airbnb 等公司是调查重点。众议院国土安全委员会和众议院中国问题特别委员会致函 Cursor 和 Airbnb 询问他们使用中国 AI 模型的风险。美国部分政府部门已禁止使用 DeepSeek 等中国模型，但美国公司并未被禁止使用中国 AI 模型。有很多美国公司一直用中国 AI 模型降低使用成本，其中就包括了即将被 SpaceX 公司以 600 亿美元收购的 Cursor 公司，该公司的 Composer 2 模型是基于北京月之暗面（Moonshot AI）的 Kimi 模型。美国国会还在调查美国公司的开源 AI 模型战略，确保美国公司不需要在昂贵或受到限制的本国模型与廉价且功能强大的中国 AI 模型之间做选择。

根据发表在《Frontiers in Psychology》期刊上的一项研究，父母对屏幕和智能手机的上瘾可能会对孩子的发育和心理造成长期的负面影响。研究显示，对设备管理不当的看护者可能会加剧“不安全依恋”，使得人际关系变得更加焦虑和回避。这项研究基于美国 600 名 12-17 岁的未成年人的调查，儿童表示他们感到被盯着屏幕的父母边缘化或忽视。研究人员表示，缺乏安全依恋的孩子可能会缺乏自信或表现出较低的自我意识；在人际关系和亲密关系方面表现出困难；并不愿意承担取得成功所必需的风险。

拟议中的集体诉讼披露了一起使用 xAI 的 Grok 生成数千张 CSAM（child sex abuse materials）图像的案件。 一名女孩的继父利用一张继女在 11 岁时拍摄的照片使用 Grok 生成了 7000 张 CSAM 图像。诉讼称 Grok 在整个过程中没有标记任何有害行为。Grok 的儿童安全系统只在这名男子输入“gang rape”后才介入，向 NCMEC 发送了举报，NCMEC 随后向执法部门报告了这起案件。法律规定 CSAM 被标记后需要共享 IP 等用户信息。但 xAI 被指控拒绝合作。最终警方通过搜查令扣押了继父的设备随后将其逮捕。继父还被指控网上出售这些 CSAM 材料，换取其他儿童性犯罪者的 CSAM 材料。该男子获准保释两天后便开枪自杀，而他的继女则饱受焦虑和抑郁的折磨，诉讼称她的生活在一夜之间被摧毁了，生活变成了一场噩梦。

广西南宁横州市爆发了严重洪灾，网上视频显示，救援人员借助无人机，转移被困群众。视频中的无人机是深圳大疆的农业无人机 FlyCart 100，其载重量为 85 公斤，能悬挂起正常体重的民众。这其实并非第一次用无人机救灾民。去年中国就有农业无人机从洪水中救出被困男子的案例。无人机已经在深圳等城市广泛用于外卖和快递。去年 3 月，中国民用航空局批准亿航智能和合肥合翼航空两家公司开展商用载人无人机服务。

根据发表在《Morbidity and Mortality Weekly Report》期刊上的一项研究，美国部分地区居民多达三成携带导致红肉过敏的抗体。这一比例远超此前的预计，意味着红肉过敏的美国人比以前认为的多得多。该过敏反应被称为α-半乳糖综合征（alpha-gal syndrome），其特点是发作缓慢，通常在餐后 2-6 小时之间出现，使得很难将过敏反应与食物联系起来。症状包括荨麻疹、恶心、呕吐、腹部绞痛、腹泻，或严重过敏反应如呼吸困难、喉咙紧缩、舌头或嘴唇肿胀、头晕、脉搏微弱和血压下降等迹象。研究人员分析了美国 10 个州 3000 份献血样本，结果显示田纳西州、阿肯色州等五个州的抗体比例最高达到 31.2%，平均 24%。