OrangeBot.AI Digest — 2026-07-07

The AI agent that works your Salesforce Pipeline

Screen recordings that edit themselves

AI agents collect peer reviews to generate proof of work

Your AI inbox that writes like you + replies on autopilot

The product that makes AI recommend your brand

Gather people instantly by AI txt, video, or ringing phones

The trust layer for online shopping

Your AI coworker, in Slack. Just tag @O.

AI moderated interviews that read how people feel

Social media scraper API for every major platform

Smarter way to manage work and data.

1.6T MoE trained entirely on AI ASICs

AI notes for in-person meetings

Real-time structured search trusted by agents and developers

Private AI autocomplete for every app on your Mac

Optimizer selection for large-scale model training has become a system-level design decision constrained jointly by compute, memory, tuning budget, and task diversity, yet the landscape of over one hundred methods remains fragmented. We therefore present OmniOpt, a unified survey and benchmark cookbook of optimizers for the research community. OmniOpt rests on four coupled components. First, we treat every optimizer update as a structured transformation through a five-stage meta-pipeline, and show that most methods engage only one or two of these stages. Second, we use norm-constrained linear minimization oracles (LMOs) to unify different optimizers. Third, these two views ground a dual-dimension taxonomy, one dimension assigning each method to a mechanism family and the other recording the measurable training objectives it aims to improve. Fourth, and at the core of this paper, we instantiate the full taxonomy in a unified cross-domain benchmark spanning representative optimizers, model scales, and training regimes from language model pretraining to image classification, systematically analyzing each method family across multiple effect objectives and laying out their trade-offs. OmniOpt thus supplies the research community with an operational coordinate system for selecting optimizers under explicit mechanism and objective assumptions, and charts a direction for the future development of the optimizer community.

Recent advances in multimodal foundation models and agent systems have driven GUI agents from single-platform task execution toward cross-platform interaction. However, building multi-platform GUI agents remains challenging. On one hand, high-quality and executable cross-platform interaction trajectories are still scarce, and existing data often suffer from limited platform coverage. On the other hand, different platforms exhibit distinct interaction conventions, making joint or continual training prone to behavioral pattern mixing, platform-specific capability degradation, and catastrophic forgetting. To address these challenges, we construct Uni-GUI, a high-quality cross-platform GUI interaction dataset, and propose UI-MOPD, the first method that incorporates multi-teacher on-policy distillation into continual learning for GUI agents. UI-MOPD dynamically selects a platform-specific teacher according to the current environment and transfers platform-specific behavioral priors to a shared policy through platform-conditioned distillation, enabling adaptation to new platforms while preserving capabilities on existing ones. Experiments on OSWorld and MobileWorld show that UI-MOPD achieves task success rates of 38.2% and 12.0%, respectively, demonstrating its effectiveness in balancing cross-platform capability retention and new-platform adaptation. Project page: https://elispectre.github.io/UI-MOPD/.

Research dissemination, turning a paper into a poster, a talk video, and a blog post, is still a manual last mile. Prior automation treats each artifact in isolation that each re-extract the paper from scratch, usually ship one-way renders the author cannot reopen in PowerPoint or Word, and gates quality on soft VLM-preference scores that plateau while load-bearing sections still read as empty. We argue this last mile is best built as a composition of skills: thin agent-readable contracts that share one upstream extractor and wrap deterministic primitives in a measured-fill loop whose exits are hard pass/fail render gates. We instantiate this as ResearchStudio-Reel, five Claude Code and Codex skills organized into one shared extractor (Paper2Assets), three editable generators (Paper2Poster, Paper2Video, Paper2Blog), and one interactive convergence layer (Paper2Reel). Paper2Assets extracts each paper once into a shared bundle that can be reused by every downstream skill; The three generators produce a print-ready poster, a synchronized talk video, and a bilingual blog that stay factually consistent and round-trip through PowerPoint or Word; Paper2Reel then binds all three into a self-contained HTML viewer whose section-level clicks jump the video, slides, captions, and blog to matching content. On the Paper2Poster benchmark, our posters lead every aesthetic and information sub-criterion against both prior automated systems and single-shot frontier LLMs, surpassing the authors' own on aesthetics under two held-out VLM judges and winning overall on 84% to 93% of papers; capability audits further show that, by uniquely pairing narration-aligned on-slide highlights with a bilingual blog gated by layout-aware DOCX repair, ResearchStudio-Reel is the only pipeline to ship all three editable artifacts. Project is available at https://aka.ms/ResearchStudio

3D reconstruction and generation are commonly tackled by separate paradigms: pixel-based regression for reconstruction, and latent diffusion for generation. Recent works attempt to unify them in latent space, but with notable drawbacks: the diffusion objective is defined on latent features rather than the underlying 3D representation, and both branches suffer from information loss introduced by latent encoding, while requiring a pretrained Variational Autoencoder (VAE) or Representation Autoencoder (RAE). In this paper, we reformulate these two tasks under a unified pixel-space diffusion paradigm and introduce PixWorld, a single model that jointly addresses 3D reconstruction and generation. By supervising diffusion directly on rendered images, PixWorld removes the above limitations and aligns optimization with 3D scene fidelity. Beyond photometric and perceptual supervision that operates at the 2D image level and lacks 3D geometric awareness, we further introduce a geometry perception loss that aligns rendered views with their ground truth in the geometry-aware feature space of a pretrained 3D foundation model, providing 3D structural supervision. PixWorld consistently outperforms prior latent-space generation methods and matches state-of-the-art reconstruction methods, demonstrating the superiority of a unified pixel-space approach.

Large language models have made research ideation increasingly accessible, yet effective idea development requires more than generating candidate directions. Researchers must ground a problem in current literature, identify meaningful bottlenecks, differentiate from existing solutions, and evaluate risks before committing to implementation. We present ResearchStudio-Idea as a reusable skill suite for this first mile of research ideation. The suite includes Paper-Search, a standalone multi-source literature search skill; Scoop-Check, a standalone prior-art collision checker for novelty claims; and IdeaSpark, the end-to-end skill that composes evidence grounding, pattern-guided generation, collision retrieval, audit, and idea-card rendering into one workflow. IdeaSpark is constructed from a corpus of 1,947 machine learning conference papers collected from ICLR, ICML, and NeurIPS between 2021 and 2025, including Oral papers, a separately tracked high-citation subset, and rejected submissions. Analysis of these outcomes reveals 31 recurring ideation sub-patterns, consolidated into 15 reusable ideation patterns. Each pattern is operationalized as a structured card containing research contexts, bottleneck types, differentiation strategies, supporting precedents, and common failure modes. Given a research problem and an evidence bundle, IdeaSpark evaluates evidence readiness, reconstructs the surrounding research context, identifies unresolved bottlenecks, selects relevant patterns, instantiates one candidate direction, retrieves potentially conflicting prior work, and performs outcome-informed auditing. This workflow transforms reusable ideation patterns into traceable research proposals. Blind automated-judge evaluations show that IdeaSpark consistently produces stronger research proposals than no-skill and generic-skill baselines while maintaining competitive novelty.

Concept erasure aims to remove a target concept from a representation while preserving the other information encoded in it. This is difficult because representations encode many concepts that are often correlated with the erasure target, so removing the target risks damaging them. We propose the Manifold Constraint Hypothesis (MCH): if natural representations concentrate on a structured, lower-dimensional manifold, then interventions should be constrained to that manifold and better preserve other information encoded in the representation during interventions. We instantiate MCH in a new concept erasure method: MANifold aware Concept Erasure (MANCE). MANCE performs iterative updates to the representations using signals from a classifier that predicts a target concept. We estimate the manifold using representations obtained from natural inputs, and then we project the concept removal update to the estimated manifold. We perform extensive evaluation on 119 settings spanning text and vision, including 13 language models, three NLP concepts, and 40 CelebA-CLIP attributes. Employing MANCE on top of previous methods shows consistent improved leakage results. We also introduce MANCE+ and MANCE++, which prepend a closed-form erasure algorithm before employing MANCE, achieving better leakage--surgicality tradeoffs relative to matched full-space updates. MANCE++, our best method, achieves state-of-the-art results on nonlinear concept erasure. These results support MCH in the erasure setting: interventions should be constrained to the natural representation manifold.

Evaluating embodied robot foundation models remains a critical bottleneck; unlike large language models efficiently assessed via digital benchmarks, robotic policies require slow, costly real-world rollouts limited by hardware and human supervision, which has driven interest in world models as surrogate policy evaluators, yet the key properties that make a world model reliable for policy assessment remain poorly understood. This work presents a systematic study of world models for robotic policy evaluation and introduces WMBench, a benchmark constructed from real-robot teleoperation data and matched policy rollouts covering diverse manipulation tasks to enable controlled comparisons across model families, action encodings, rollout horizons, and evaluation metrics. Using WMBench, we analyze 7 video world models, 4 action representation schemes, and over 324,000 simulated policy rollouts paired with real robot executions, further enriching our analysis with large-scale community submissions from the CVPR 2026 GigaBrain Challenge, curated synthetic trajectories, and a training videos spanning more than 12,000 hours. Our experiments deliver three core insights: evaluator quality is dominated by long-horizon, action-faithful rollout consistency rather than short-term visual realism; pretraining gains stem not only from data scale but from balancing general world knowledge with robot-specific controllability; and architectural choices including action encoding, memory design, and evaluator-focused post-training strongly determine alignment with real-world robot behavior. Drawing on these results, we derive a practical design roadmap and realize it in GigaWorld-1, a world model specially optimized for policy evaluation, and we fully release our code, models, datasets, and toolkits to advance scalable evaluation research for embodied foundation models.

Dense spatial perception is essential for physical intelligence, where visual systems are expected to recover structured, metric, and actionable representations from pixel observations. Modern visual foundation models tend to prioritize semantic invariance, often at the expense of detailed spatial understanding. In this work, we study vision pretraining through a boundary-centric lens, motivated by the premise that boundaries and shape discontinuities offer essential cues for perceiving geometric properties. Concretely, we propose masked boundary modeling, a self-supervised paradigm that dynamically learns sub-pixel boundary representations and subsequently leverages the discovered boundary-bearing tokens as masked targets to facilitate dense visual token learning. By scaling this framework, we develop LingBot-Vision and demonstrate its efficacy across a diverse set of downstream vision tasks with DINOv3 as a strong baseline. Remarkably, LingBot-Vision drives the progression from LingBot-Depth 1.0 to LingBot-Depth 2.0 for depth completion, and thereby yields enhanced depth estimation, a key pillar for embodied artificial intelligence. Our findings reveal that boundary modeling goes beyond simple line segments and instead serves as a scalable pretraining principle for learning spatially structured visual representations.

We present Wan-Streamer v0.2, a latency-preserving upgrade of the native-streaming, end-to-end audio-visual interaction model. v0.2 keeps the v0.1 modeling formulation, but raises the interactive output stream from 192x336 to 640x368 while preserving approximately 200 ms model-side signal-to-signal latency at 25 FPS. The higher-resolution stream supports scene-grounded mid-shot agents whose posture, gaze, hands, nearby objects, and local scene layout remain legible during real-time conversation. To support the larger visual stream without adding user-visible delay, v0.2 keeps the thinker as a single-GPU low-latency path for streaming perception, the short language/state Transformer pass that builds the generation cache, and final decoding. The performer becomes a multi-GPU Ulysses-style context-parallel group for the expensive next-unit latent generation. Each performer rank writes incoming K/V into a pre-sharded local cache. The long high-resolution latent video sequence is split across ranks for denoising and gathered through Ulysses communication, while the much shorter audio latent sequence is generated without sequence sharding. In this split, the thinker's language/state computation reaches the performer only as K/V conditioning, so no separate language sequence has to be communicated inside the performer group. This concentrates additional hardware on visual generation while preserving the compact thinker-performer boundary, keeping total remote interaction latency at approximately 550 ms when a 350 ms bidirectional network budget is included.

We present EVA-Client, an open-source framework for deployment, data collection, and evaluation of trained manipulation policies on real robots. Sitting between a policy server and the physical hardware, EVA-Client unifies the real-robot stages of the policy iteration loop within a single codebase. It makes three contributions. First, a component-decoupled architecture in which robot backends, inference strategies, and transport middlewares form an orthogonal grid: adding a robot or a strategy touches only its own layer. Second, inspectable execution through Debug, Collect, and Eval workflows, with modes ranging from open-loop simulation to continuous real-time control. Third, every evaluation run doubles as a data collection, recording full rollouts in training-ready format alongside exhaustive logs and a side-by-side comparison viewer, so each evaluation feeds the next round of training rather than ending as an unrecorded impression. EVA-Client further consolidates major real-time inference strategies, synchronous and asynchronous execution, ACT-style temporal ensembling, Real-Time Chunking, and a naive-async ablation baseline, behind a single configuration surface.

Unified models for robot manipulation aim to equip one policy with both the semantic priors of pretrained VLMs and the physical dynamics learned through future prediction. In practice, existing designs tend to erode the semantics of the pretrained backbone, suffer interference among heterogeneous objectives, and learn future prediction from scratch in pixel space, leaving the dynamics priors of pretrained video generators unexploited. We present InternVLA-A1.5, which builds the policy on a native VLM backbone that keeps training on VQA and subtask prediction, and attaches a lightweight unified expert for continuous action generation. Future prediction is recast as a latent-querying problem, where a small set of learnable foresight tokens condenses the task-relevant future into a compact latent code under the supervision of a frozen pretrained video generation model, so the policy inherits world-model dynamics priors without ever learning pixel-level generation. The video branch is discarded at inference, keeping real-time control. Pretrained on 1.2M robot episodes and 3M multimodal samples, InternVLA-A1.5 achieves the best overall results on all six simulation benchmarks. In the real world, the preserved semantics deliver the strongest compositional generalization on held-out instruction bindings, and the two designs together sustain long-horizon execution.

Key-value (KV) cache growth is a major bottleneck in autoregressive decoding, as memory and bandwidth scale linearly with context length. Existing KV eviction methods often rely on static heuristics or proxy scores, which poorly track future token utility and cause brittle eviction as relevance shifts. To address this, we introduce KVpop, which learns a fixed-budget KV eviction policy by directly supervising the keep-or-drop decision. The scorer is trained against a novel future-attention target, computed efficiently without materializing dense attention maps. We further introduce a delayed memory-based scorer that, uniquely among learned eviction methods, defers scoring for a fixed number of steps to exploit near-future context. On AIME and HMMT mathematical reasoning, KVpop retains 98% of full-attention performance on Qwen3-4B at 75% KV cache compression and 97% at 88% compression, consistently outperforming established eviction baselines. Qwen3-8B shows even stronger results, reaching near-full teacher performance. These results show that supervising eviction with future-attention signals cuts memory costs while maintaining quality.

Scientific literature search often requires more than retrieving papers from a single query: users' intents are underspecified, preference-dependent, and evolve through interaction. Existing search agents typically rely on fixed pipelines or implicit language-only reasoning, making their search strategies difficult to control, inspect, and refine. We introduce PaperPilot, a multi-turn literature search agent that frames scientific search as workflow induction. Given an anchor paper and a user query, PaperPilot constructs an executable DAG of paper-search operators, including keyword search, citation expansion, filtering, scoring, reranking, and evidence extraction. User feedback is then used to refine both the query and the workflow itself. We train PaperPilot with supervised workflow imitation and preference optimization over controlled workflow corruptions. Experiments show that PaperPilot-9B improves over the base Qwen3.5-9B toolset agent under multi-turn interaction, increasing Hit@5 from 58.0 to 77.0, MRR from 47.5 to 59.4, and nDCG@10 from 26.8 to 32.5, while reducing workflow execution errors from 9.5% to 0%. These results show that explicit, editable search workflows provide an effective and controllable interface for aligning literature search agents with complex scientific intent.

Multi-vector vision-language retrieval preserves fine-grained visual evidence through maximum-similarity late interaction, but dense image-side tokens make storage and scoring expensive. Existing token compression methods reduce this cost, yet they can remove or collapse object- and region-level evidence that future query tokens may need to select. We propose SaMer, an object-aware token merging framework that compresses image-side post-projector tokens into K representative centroids while preserving the original late-interaction interface. SaMer uses object annotations only during training as a merge prior to discourage cross-instance mixing, requires no ground-truth bounding boxes or detectors at inference time, and adapts only the shared projection layer with frozen vision and language backbones. With K=64, SaMer removes more than 93% of image-side tokens and reduces ColPali storage by 16.09times, while improving R@1 on Flickr30K and MSCOCO. These gains arise because object-aware merging preserves query-selectable object evidence that pruning or feature-only pooling can remove or collapse. SaMer also outperforms compression baselines and shows stronger phrase-level grounding, suggesting that efficient multi-vector retrieval depends not only on reducing token count, but on preserving the evidence future query tokens need to select.

Diffusion large language models (dLLMs) generate text by iteratively denoising a masked sequence, offering a parallel alternative to autoregressive models, but eliciting strong reasoning through post-training remains difficult: supervised fine-tuning is off-policy and suffers from exposure bias, while reinforcement learning gives only sparse, sequence-level rewards and is hard to apply without tractable sequence likelihoods. On-policy self-distillation (OPSD) offers a promising alternative, using one model as both student and teacher to provide dense, token-level, on-policy supervision, but its effectiveness hinges on giving the teacher privileged information (PI) - typically an instance-specific ground-truth reference unavailable at inference - so the student ends up distilling a weak PI-free consensus policy that yields little improvement on dLLM reasoning. We introduce dOPSD, which instead derives the teacher's privilege directly from the student's own denoising trajectory, evaluating masked positions using later, more-decoded steps of that same trajectory rather than an external label, so the teacher's advantage emerges from the model's own decoding process; on Dream and LLaDA, dOPSD improves both in-domain math reasoning and out-of-domain code generation, outperforming supervised and on-policy baselines.

New York Times : Anthropic plans to lease a 16-story building in Lower Manhattan and double its NYC workforce to 1,000 people this year, as AI companies expand in the city —  Anthropic, the artificial intelligence company, plans to lease a 16-story building in Hudson Square and to double its work force in New York City to 1,000 people this year.

Maxwell Zeff / Wired : OpenAI Chief Futurist Joshua Achiam is leaving the company later this month after nearly nine years —  Joshua Achiam spent nearly nine years at OpenAI researching AI safety and made a memorable appearance in the Musk v. Altman trial.  —  OpenAI's chief futurist Joshua Achiam notified colleagues …

Grace Kay / The Information : Memo: SpaceXAI and Cursor plan to launch their first jointly developed AI model as soon as Wednesday, after pushing back the launch to improve model efficiency —  SpaceXAI and Cursor plan to launch their first jointly developed AI model as soon as Wednesday, according to a memo sent to staff.

Meta : Meta launches Muse Image in Meta AI, Instagram, and WhatsApp, and previews Muse Video, the first media generation models from its Superintelligence Labs —  We're excited to launch Muse Image and preview Muse Video, the first media generation models developed by Meta Superintelligence Labs.

Ian Allison / CoinDesk : Source: Kraken is pursuing a full banking license in Europe, with a focus on Lithuania as the jurisdiction to secure it —  Kraken, the cryptocurrency exchange planning to go public in the U.S., is pursuing a full banking license in Europe, with a focus on Lithuania as the jurisdiction to secure it …

Zac Hall / 9to5Mac : Anthropic extends Claude Fable 5 access to all paid plans through July 12; access to the model was set to shift to token-based usage on July 7 —  Anthropic announced today that it is expanding Claude Cowork capabilities from the desktop app to the web and mobile.

Meta Newsroom : Introducing Muse Image: Image Generation Built for Your World —  Today, Meta is rolling out Muse Image, the company's first image generation model from Meta Superintelligence Labs, now available in Meta AI.  Muse Spark made Meta AI a smarter assistant; now, Muse Image acts as the creative partner …

Abner Li / 9to5Google : Google announces a Pixel event on August 12 in NYC, where it is expected to launch Pixel 11 series of phones and Pixel Watch 5 —  Made by Google 2026 is officially taking place on Wednesday, August 12.  This is a week earlier compared to last year, with Google once again opting for an earlier Pixel hardware unveil.

Brody Ford / Bloomberg : Sources: Microsoft, looking to reduce AI costs, is starting to replace models from OpenAI and Anthropic with its MAI models in products like Excel and Outlook —  Microsoft Corp., looking to reduce AI costs, is starting to replace OpenAI and Anthropic with its own models in software products like Excel and Outlook.

William Earl / Variety : Netflix signs deals with Penske Media, Condé Nast, Hearst, People, and other publishers to carry a range of programming from 2 to 20+ minutes, starting August 3 —  Netflix is further expanding the mix of content available on its platform through licensing partnerships with a range of top publishers …

David Gewirtz / ZDNET : Anthropic expands Claude Cowork to web and mobile in beta for Max plan subscribers, and says 90%+ of Cowork usage is unrelated to software development —  ZDNET's key takeaways  — Claude Cowork is becoming a cloud-based AI helper.  — Desktop still offers the fullest Cowork experience.

Mike Wheatley / SiliconANGLE : Tangos, which uses AI to conduct financial crime investigations at scale, raised a $20M seed led by Red Dot —  Tangos AI Inc., a startup that uses artificial intelligence to conduct financial crime investigations at unprecedented scale, said today it has closed on $20 million in seed funding.

Lily Mae Lazarus / Fortune : Super.com, which offers a savings app for lower-income consumers, raised a $65M Series D led by TPG at a $1.2B valuation and says it is profitable —  In 2022, Hussein Fazal flew 200 of his employees to Las Vegas to cash a $200 check at a payday loan shop.  He then asked them to buy a week …

Mike Lennon / SecurityWeek : Keyfactor, which manages billions of machine identities for more than 2,500 client organizations, raised $1B+ led by Summit Partners —  The investment will accelerate Keyfactor's machine identity, PKI, and cryptographic security platform as enterprises prepare for AI-driven and post-quantum threats.

Jay Peters / The Verge : Google launches “platform properties” in the Google Search Console, letting creators and website owners see which search terms lead to their social platforms —  A new feature will let creators and publishers see what search terms take people to their social platforms.

学习多种语言与大脑衰老延缓相关。一项研究发现，双语者的大脑比单语者年轻约六岁，掌握四种语言的人的大脑则年轻多达十三岁。随着年龄的增长，大脑的连接性会下降，记忆力和思维速度会减慢。此前的研究发现，掌握多种语言的欧洲国家居民衰老更慢。最新研究则测量了多语言能力对个体大脑的影响。研究对象是西班牙巴斯克（Basque）地区的居民，当地居民会说西班牙语、巴斯克语、法语和/或英语。研究人员用脑磁图（magnetoencephalography）测量了 728 名不同年龄和语言能力水平的人的大脑活动。结果显示，掌握了多种语言的人其大脑年龄比实际年龄更年轻。研究结果表明，学习第二、第三甚至第四门语言有助于大脑保持年轻状态，而且越早开始学习效果越好。

LineageOS 项目谈论了即将生效的 Google Android 开发者认证计划对该项目意味着什么。Android 开发者认证计划意味着所有 Android 安装路径都将通过 Google 控制的基础设施，授予 Google 一个“终止开关”可封禁全球任何 Android 应用或开发者。开发者认证对 LineageOS 项目没有直接影响，但会影响到运行任何官方 ROM 的用户。因为验证开发者的基础设施是 Google Mobile Services（GMS）的一部分，而 LineageOS 默认并不提供也无计划提供 GMS。除非 LineageOS 用户选择安装包含开发者验证基础架构的 GApps 包，否则该功能不会造成任何影响。如果 Google 将该功能集成到 Play Services 中，那么 LineageOS 项目将会全局禁用该功能。

根据分析机构 Omdia 的数据，2026 年第一季度 400美元以下智能手机的物料清单中，内存成本几乎占到了六成，而且此后情况并未好转。市场观察机构 TrendForce 上月预测，2026 年 DRAM 价格还将上涨 50% 以上，这使得廉价手机制造商不可避免将组件成本上涨的压力转嫁给消费者。为了抵消不断上涨的内存成本，制造商尝试转向更便宜的显示面板、传感器或射频模块，但低端手机本就建立在极其紧凑的成本结构上，几乎没有进一步压缩的空间。这与入门级 PC 的情况类似。Omdia 预计 2026 年 400 美元以下智能手机的出货量将同比下降 22%。不过 Omdia 认为，虽然 2026 年全球智能手机市场整体将下滑 12%，但 400 美元以上的中高端市场将保持韧性，出货量有望增长 5.7%。智能手机制造商正将生产重心转向中高端机型。部分中国制造商正在某些升级至新型 LTPO 技术的机型中重新采用 LTPS 显示面板，将 LTPO 保留给高端机型。这可以为每台设备节省 3-5 美元的成本。 其它措施还包括减少摄像头数量、使用更小的图像传感器，或改用上一代 SoC，这些措施可将成本降低约 30% 至40%。

研究人员发现，连续六周每晚少睡 80 分钟的人体重平均增加一磅，且更长时间久坐。研究团队招募了 95 名睡眠时间为 7-8 小时的成年人。参与者被要求在一个为期六周阶段内将正常就寝时间推迟 90 分钟，在另一个为期六周阶段内保持正常睡眠时间。研究人员使用腕式监测器测量了每个阶段的睡眠和活动水平，体重、腰围、身体成分和几种已知能增加或抑制食欲的激素的空腹水平变化。结果显示仅仅六周体重就出现了增长。此外，缺觉阶段的受试者久坐时间平均每天增加 17 分钟，男性和绝经后女性的久坐时间平均每天增加近 30 分钟。

空气污染可能通过影响儿童控制冲动能力而导致其肥胖。研究发现，在出生第一年接触较高浓度 PM2.5 的婴儿在儿童后期更容易出现冲动控制障碍。PM2.5 颗粒物常见人为来源包括交通和化石燃料燃烧，它与一系列健康问题相关，如痴呆症和中风。此前的研究表明 PM2.5 会扰乱新陈代谢并与体重增加有关。研究人员分析了墨西哥城的 434 名儿童的数据，他们多数出生于 2007-2008 年间。研究人员模拟了孕期和儿童出生后第一年的环境 PM2.5 水平。之后研究人员对这些儿童的冲动性和肥胖进行了评估。PM2.5 暴露水平最高的儿童组表现出较高的冲动性，反映出其抑制控制能力存在显著缺陷。

在智能手机和社媒时代，人们将越来越多的时间花在屏幕上。调查显示，美国人每天的平均社交时间从 20 年前的 45 分钟减少 10 分钟至 35 分钟。这一趋势在所有年龄段都存在，其中 15-24 岁人群下降幅度最大，每天社交时间从约 1 小时降至 35 分钟。与此同时，青少年平均每天花费 4.8 小时在 TikTok、Instagram 和 Snapchat 等社媒应用上。聚会场所的减少也是造成这一趋势的原因之一。过去十年美国各类休闲场所——从图书馆到咖啡馆到博物馆——都出现了大范围关闭，甚至教堂也存在类似现象。

微软裁员 4800 人，其中游戏业务 Xbox 裁员 1600 人，多个游戏工作室或者出售或者独立。微软执行副总裁兼首席人力资源官 Amy Coleman 强调：被裁掉的职位不会被 AI 取代，但 AI 确实在改变工作的方式。部分日常工作可以自动化了，意味着人人都需要不断学习，不断提升技能，并不断调整自己。Xbox CEO Asha Sharma 称这是 Xbox 历史上最重大的重组。Xbox 将缩小业务范围，放弃回报率低的业务，专注于回报率高的业务如《我的世界》的 Mojang 以及《糖果粉碎传奇》的手游开发商 King。作为重组的一部分，Double Fine 和 Compulsion 两大工作室将独立。Ninja Theory 和 Undead Labs 工作室出售，Arkane 工作室则在评估中可能独立可能出售，Obsidian 工作室裁掉四分之一员工，id Software 也大规模裁员。

美国最高法院允许德州要求移动应用验证用户年龄。德州的新法律要求应用商店在未成年人下载应用前验证用户年龄并获得家长同意。代表 Google 和苹果的行业组织 Computer & Communications Industry Association 认为该法律覆盖范围太广，会限制青少年访问各种数字内容，会对数字言论保护产生深远影响。在该法律生效前几天，联邦地区法院去年底阻止了该法律的实施。保守派的第五巡回上诉法院在今年六月搁置了该裁决，允许德州继续推行这一法律。美国最高法院维持第五巡回上诉法院的裁决，它并没有做出决定，但允许德州在诉讼期间继续执行该法律。

美国大学协会 55 所研究型大学的数据显示，2026 年秋季招收的博士生人数比 2025 年减少了 15%。博士生减少引发了对美国科研能力下降的担忧。造成这一现象的一大原因是特朗普政府混乱且难以预测的联邦拨款环境。美国大学协会包含了 69 所最知名的研究型大学，全美半数的研究型博士学位由其这些大学授予。该协会的高级副总裁 Toby Smith 表示，由于支持学生的能力下降，美国正面临失去整整一代新人才的风险。博士生人数下降的原因包括国立卫生研究院（NIH）和国家科学基金会（NSF）等重要联邦机构在减少科研经费，而最富有的大学还面临对其捐赠基金征收的新联邦税。

中国国家航天局周一宣布，天问二号探测器历经约 400 天、行程约 10 亿千米飞行后于近日与小行星 2016HO3 成功交会，到达距离小行星 20 千米处，开始科学探测。在抵近小行星过程中，探测器获得小行星影像数据。同时任务团队利用探测器抵近过程中获得的光学导航数据，改进了小行星星历，将之前仅依靠地基观测所确定的小行星位置误差，由上百千米减小到千米量级。天问二号探测器于 2025 年 5 月 29 日在西昌卫星发射中心成功发射。2026 年 6 月 6 日，探测器首次捕获到小行星；6 月 7 日，在距离小行星 3 万千米处实施捕获控制，实现与小行星共面飞行；6 月 19 日，到达距离小行星 2000 千米处。后续探测器将逐步开展更精细科学探测，获取小行星形貌、物质成分、内部结构等信息，为做好采样准备提供支撑。

被社媒用户指控是 AI 创作的短篇小说《The Serpent in the Grove》赢得英联邦短篇小说奖。这篇小说由 Jamir Nazir 创作，在五月份赢得了地区奖项，随后立即被 X 等平台的社媒用户指控是 AI 创作。原因是它包含了典型的 AI 写作模式，包括“不是 x 而是 y”结构以及三列表结构。在引发争议之后，Commonwealth Foundation 检查了所有地区奖得主的草稿、含时间戳的文件和笔记等内容。该基金会之后宣称他们没有发现作品是 AI 生成的证据。 Nazir 称其作品深受 VS Naipaul 和 Derek Walcott 等作家作品的影响。他说这篇小说写了六七稿，他解释说其手机屏幕一次只能显示三四行文字，所以他会写下一行前对前面的文字细致修改，最终使故事显得“高度润色”。

亚马逊众包平台 Mechanical Turk 发表公告，宣布将于 7 月 30 日起停止接受新用户。亚马逊 AWS 称，该决定是在“慎重考虑”后做出的，“现有用户可以继续正常使用该服务。AWS 将继续投资改进 Mechanical Turk 的安全性和可用性，但我们不打算推出新功能。”Mechanical Turk 暂时不会关闭，但其未来岌岌可危。亚马逊从 2018 年起将 Mechanical Turk 变成训练神经网络的标注数据服务。但讽刺的是 2023 年的研究发现，该平台 33% 到 46% 的众包工作者使用大模型去完成任务，引发了对标注数据的可靠性以及是否真的需要人类参与的质疑。社交平台的用户称，由于大量的机器人和欺骗行为，研究人员已经放弃了该平台，它的关闭可能只是时间问题。

有没有类似经历：走进一间房间，却忘记来干什么？你可能是来拿钥匙，在来的路上还记住钥匙，但在进门的刹那钥匙从你的记忆里消失了。这种现象有时被称为门口效应（Doorway Effect），通常发生在走进一间陌生的房间时。它与工作记忆密切相关。当我们执行当前任务（如记住去拿钥匙）时，信息会被存储在工作记忆中。门口效应表明，当信息从工作记忆中移除时，它似乎会立即从意识中消失。这也表明工作记忆中的信息很容易被遗忘。工作记忆内部包含众多不同子系统，分别负责特定的任务，包括视觉和空间推理（如解魔方）以及存储信息块（如电话号码）。甚至还有一个“中央执行系统”。中央执行系统就像一个冷酷的老板，负责将任务分配给工作记忆内的不同系统，确保一切井然有序。工作记忆的一个独特之处是它的容量非常有限。尽管它拥有丰富的可用信息，但工作记忆在任何时刻只能存储极少量的信息。它很难记住非常复杂的物体。工作记忆的有限容量能解释为什么门口效应会发生：当新信息进入时旧信息必须被清除。进门的动作会触发信息消除。从演化论角度看，遗忘旧信息有助于适应新环境。这种遗忘也表明工作记忆与意识密切相关。一种观点认为，意识的产生源于工作记忆和注意力的协同作用。

C# 和 TypeScript 等语言的创始人 Anders Hejlsberg 在 2025 年 3 月宣布他的团队正在开发用 Go 语言编写的 TypeScript 编译器。现在首个基于 Go 语言的 TypeScript 7.0 发布了 RC 版本。正式版本预计会在下个月发布。微软表示，得益于原生代码速度和共享内存并行处理能力，TypeScript 7.0 的速度通常比 TypeScript 6.0 快约 10 倍。不同于 TypeScript 6.0，TypeScript 7.0 的许多步骤（如解析、类型检查和发射）都可以并行执行。其中一些步骤（如解析和发射）在不同文件之间可以基本独立完成。并行处理在较大的代码库中能自动实现良好扩展，且开销相对较小。但微软也表示 TypeScript 构建中并非每个步骤都能轻松实现并行处理。

美国司法部、FBI 与芬兰国家调查局合作逮捕了 19 岁的网络勒索组织 Scattered Spider 成员 Peter Stokes，该组织至今已勒索逾 1 亿美元赎金。Stokes 已引渡到美国并于 6 月 30 日出庭，目前处于羁押中。微软在识别其身份过程中起了很大的作用。微软向 FBI 提供了 GDID（Global Device Identifier 或全球设备标识符），GDID 是分配给每个 Windows 的唯一标识符，用于跟踪特定设备的遥测数据。法庭文件显示，Stokes 使用的是 Windows 系统，他的网络活动、游戏历史记录、IP 地址、工具使用情况（包括 Ngrok）、Azure 状态等信息都被记录下来并带有时间戳，这些信息由微软提供给了调查人员。他拥有美国和爱沙尼亚双重国籍，其身份早在 2024 年就被确定，但由于他当时未成年且居住在爱沙尼亚和阿联酋，因此他一直处于被监视状态，直到最近试图从芬兰赫尔辛基搭乘航班飞往日本。