OrangeBot.AI Digest — 2026-09-28

Reshape your text across languages while preserving meaning

See and control what your coding agent does.

Hidden controls for your Mac.

Fast AI tool routing powered by Jev.

A journal to help busy minds let go

A reading-first Markdown engine for macOS

Bookmarks that organize themselves with Apple Intelligence

The Al built for work

Your mobile AI assistant on any screen

Beautiful screen recordings, screenshots, and 3D motion

The one command line to secure your Cursor and Claude Code

AI buddy on your computer

Give your AI Analyst context to explain performance and act

Make Your Screenshots Look Amazing

The world's first AI CMO

Representation autoencoders (RAEs) reuse features from a pretrained visual encoder as reconstruction and diffusion latents, integrating strong visual representations into image generation. However, RAEs still need to decide which encoder layers form the shared latent space for the generator and pixel decoder. This choice involves a trade-off. Shallower layers tend to preserve fine pixel details better, while deeper layers tend to yield better generation metrics. A fixed heuristic layer fusion therefore couples two stages that benefit from different information. We introduce FuseReg, which replaces heuristic feature selection with training over random subsets of encoder layers. We theoretically analyze the underlying mechanism: subset sampling explicitly penalizes sensitivity to cross-layer disagreement. On ImageNet-256 with DINOv3-L, a single FuseReg decoder reconstructs from full, sparse, and single-layer fusions without retraining, achieving higher PSNR than decoders specialized to fixed fusions. This flexibility also benefits generation: decoder replacement alone reduces unguided gFID by 27% with an unchanged RAEv2 DiT-XL generator. The same regularization principle extends to diffusion training, with joint regularization of both stages reducing unguided gFID by 29% on DiT-Base. These results show that training downstream models for layer-fusion robustness narrows the reconstruction-generation gap without modifying the pretrained encoder.

Preparing high quality training data for foundation models requires scalable pipelines that transform heterogeneous documents and videos into structured records. Such pipelines expand each parent item into an ordered and input dependent sequence of children, whose counts may be long tailed. GPUs should batch children across parents while preserving parent relationships, child order, completion status, and result routing. Existing systems either hide parallelism behind coarse grained jobs or expose flat records that force applications to manage lineage and regrouping. We present RayOrch, a programming model and distributed execution engine that preserves parent child relations throughout execution. Programs declare ordered variable cardinality expansions and matching gathers. The compiler validates each pair, while the runtime records child membership, immediate parents, immutable ordinals, and terminal states. Per Call FIFO Ready Queues batch ready children across parents. Gathers reconstruct results from declared membership and ordinals rather than batch boundaries or completion order. Parents can advance as soon as all required children become terminal. Typed parent scoped failures suppress undispatched siblings of the failed parent while allowing unrelated parents to continue. On NVIDIA H20 GPUs, RayOrch achieves 15.14 times speedup when scaling MinerU from 4 to 64 GPUs and 7.82 times speedup when scaling a video pipeline from 8 to 64 GPUs. It reduces end to end time by 13.1 percent versus Ray Data and 29.0 percent versus Daft on MinerU, and by 16.0 percent versus Ray Data on Docling. Code available at https://github.com/OpenDCAI/RayOrch .

Prefill and decode reward different approaches to quantization: low-precision arithmetic accelerates prompt processing, while compact weights reduce memory traffic during generation. We propose "disaggregated quantization" (DQ), which specializes computation formats, weights and storage placement to both of these phases. On Qwen 3 and Gemma 3, removing activation quantization specifically on decode improves accuracy on decode-heavy tasks without increasing inference cost. Training separate compute-native prefill weights accelerates prompt processing relative to weight-only inference while matching or exceeding its accuracy at 2-3-bit decode on both decode-heavy and prefill-heavy tasks. With released Qwen3.8-27B GGUF decoders, training an NVFP4 prefiller improves 1-bit accuracy by 32.5 points on MMLU-Pro and 35.3 on MMMU-Pro without modifying the decode checkpoint. To accommodate the additional checkpoint on a single device, offloaded disaggregated prefill (ODP) streams its weights from SSD, amortizing loading over prompt length. On the same 27B model, ODP delivers a 1.78x time-to-first-token speedup over the weight-only baseline at 8K prompt length in llama.cpp. We evaluate accuracy under disaggregated serving in vLLM and further validate shared-weight format disaggregation through post-training quantization on models up to 2.8T parameters.

Scaling language models to long contexts is limited by the quadratic cost of self-attention. Block sparse attention offers an efficient alternative, but selecting the retained blocks remains a bottleneck. Conventional block selection requires scoring all query-block pairs and therefore remains quadratic in sequence length. To address this issue, we propose PISA, a block-sparse attention mechanism that employs a pyramid Top-K selection strategy. The main idea is to gradually narrow down the candidates across different levels, making it more efficient to find the most relevant keys. Specifically, we construct a coarse-to-fine hierarchy of keys and perform selection from the coarsest level. At each level, LogSumExp scoring is applied to a bounded candidate set to select candidates for the next finer level, continuing until the finest level is reached. Through pooling, we construct O(log N) levels of keys, yielding an overall complexity of O(Nlog N), where N denotes the sequence length. We develop hardware-aware Triton kernels for both training and inference, fusing hierarchical routing and LogSumExp scoring without materializing the query-key score matrix. We further evaluate our method on language modeling tasks. Compared with the baseline, our method achieves comparable performance on benchmarks such as commonsense reasoning while delivering better results on retrieval tasks.

World Action Models (WAMs) jointly model visual dynamics and action generation for generalist robot manipulation. A central challenge is to integrate priors from large-scale pretrained models---including visual dynamics, scene semantics, geometry, and motion---into a unified framework for robot action generation. We introduce InternW0-Δ, a unified WAM pretrained on a heterogeneous corpus that outperforms prior methods across simulation benchmarks and real-robot platforms. InternW0-Δ combines pretrained visual dynamics, scene-level semantics, 4D geometric and motion priors, and action generation within a Mixture-of-Transformers (MoT) framework. A pretrained video expert and an action expert interact under semantic guidance from a frozen VLM, while a pretrained 4D foundation model injects geometric and motion priors through training-only distillation. We further introduce Causal Imprint, which learns future-relevant scene changes from training-only future supervision and provides predictive representations directly to the action expert without future-video rollout at inference. For large-scale joint training, we construct a heterogeneous corpus of robot demonstrations, UMI data, egocentric human demonstrations, and Ego2Robot data, curated and aligned under a common state-action representation. The resulting corpus contains over 20K hours of processed training data, to our knowledge the largest open-source corpus of its kind. We pretrain InternW0-Δ on this corpus and demonstrate strong performance across simulation benchmarks and real-robot platforms. We will open source the training code, model weights, infrastructure, data-processing pipeline, and processed data where licenses permit. Project page: https://internrobotics.github.io/InternW0-Delta/

Tactile sensing is an essential modality for robots performing contact-rich, dexterous manipulation, particularly under visual occlusion. While pre-trained image encoders are standard in robot learning pipelines, tactile encoders are still commonly trained from scratch from raw, noisy signals, which might limit their expressivity. Existing self-supervised learning (SSL) approaches focus predominantly on vision-based tactile sensors, leaving distributed electronic skins largely unaddressed. These sensors, however, have a distinctive property: their sensing elements are sparse and irregularly arranged over the surface they cover, which makes direct reuse of visual SSL methods suboptimal. We present Tactile-JEPA, an efficient self-supervised pre-training method that uses the spatial arrangement of tactile sensors to learn topology-aware representations. Specifically, it is trained to predict the embeddings of masked sensing elements from the unmasked remainder, using the sensor connectivity graph to guide spatial masking. Our analysis shows that effective tactile representations require capturing both local contact details and the global state of the tactile surface, which we achieve through dual-scale masking. Across three diverse datasets spanning magnetic and piezoresistive sensors, different robot embodiments, and single- and paired-sensor configurations, Tactile-JEPA reduces force estimation error by 6.3% and in-hand orientation error by 20.8% over the prior state-of-the-art, with consistent gains in other downstream applications, including policy learning. Overall, our results demonstrate that the benefit of tactile sensing depends critically on the quality of encoder pre-training, a problem which Tactile-JEPA addresses directly. Code is available at https://github.com/E-Kovtun/tactile.

Large-scale Digital Surface Models (DSMs) can be produced cost-effectively from satellite images via stereo-photogrammetry. However, the resulting 3D maps are often contaminated by noise, outliers, and voids. On the other hand, aerial LiDAR provides high-accuracy elevation measurements at a substantially higher cost. In this work, we study diffusion models conditioned both on photogrammetric DSMs and Pléiades imagery to refine vertically co-registered DSMs. We introduce a modified Stable Diffusion 3 architecture with a pruned text stream and a patch-wise normalization strategy, enabling stable training on LiDAR data and transfer from natural images to elevation maps. Experiments in French cities demonstrate that multimodal conditioning improves elevation accuracy, reducing Dense Urban RMSE from 6.00 to 3.45 m in the in-context cities and from 4.16 to 2.77 m in the held-out city of Bordeaux.

Jev is a fast, low-cost decision model that answers natural-language questions with choices, binary judgments, and scores. As its public ecosystem grows rapidly, it remains unclear how Jev is used across applications and how public attention relates to project distribution. To answer these questions, we conduct a large-scale, data-driven analysis of 2,170 publicly available Jev projects collected from GitHub as of September 22, 2026. We find rapid early growth in Jev's public ecosystem, with both new projects and integration into existing repositories. Across diverse domains, projects use Jev for multiple decision purposes and combine its interfaces. Attribute judgment and scoring are widely used, while the use of action selection, content filtering, and model and tool selection varies across domains. These patterns suggest that Jev serves as a reusable decision component whose functionality varies with the surrounding workflow. Meanwhile, public attention is concentrated in routing and interface agents and does not track project counts. Our findings provide a quantitative view of Jev's emerging ecosystem and inform the design and evaluation of general-purpose decision models across diverse application contexts.

Reference-based image quality assessment (IQA) metrics aim to reflect how humans perceive the perceptual distance between a pair of images. To learn how the human visual system (HVS) operates, recent reference-based IQA metrics heavily rely on human-annotated data. Mean opinion score (MOS)-based pointwise scoring, which assigns a scalar quality value per image, is preferable for annotation but is prohibitively expensive to collect at scale and is known to be noisy due to inconsistent human judgments. As an alternative, two-alternative forced choice (2AFC) pairwise labels have gained popularity due to their reliability and efficiency, but they capture only relative comparisons between pairs. In this paper, we propose a fully automated data generation pipeline that generates pointwise perceptual distance labels between image pairs without any human annotation. Our approach exploits the generative dynamics of diffusion models as a perceptual distance proxy, where the coarse structure of an image is generated in the early timesteps and the fine details are generated in the later timesteps. Images that fork early in the generation process share only coarse structure and are perceptually far apart; images that fork late differ only in fine detail. We demonstrate that the diffusion trajectory aligns well with the human visual system, and use this forking moment, FoMo, as a reference-grounded distance label to supervise the training of a reference-based IQA metric. The pointwise labels, which support universal comparison between arbitrary image pairs, enable an information-rich training objective. Extensive experiments across diverse backbone architectures confirm the effectiveness of our generation pipeline, outperforming human-annotated datasets in multiple benchmarks.

Existing point tracking models face a fundamental tradeoff: they can either track a sparse set of query points over long horizons, or track all points across only short clips. We introduce TrackEverything, a 3D point tracker that breaks this trade-off by representing videos as persistent 3D scene tracks in world coordinates. Grounded in the insight that videos are 2D projections of an underlying 3D world, TrackEverything decouples model complexity from video duration, allowing it to scale with unique physical scene geometry instead. Our approach introduces three key innovations. First, we employ a voxelization-based de-duplication mechanism at sliding-window boundaries to merge co-located tracks, preventing repeated observations of the same surface from redundantly accumulating. Second, we decompose tracking into an endpoint refiner that predicts each point's destination and static-versus-dynamic classification, followed by a lightweight trajectory refiner that decodes dense trajectories exclusively for dynamic points. Third, we propose 3D WAFT, replacing memory-prohibitive 4D correlation volumes with efficient feature sampling in the scene cloud. To the best of our knowledge, TrackEverything is the first 3D tracker capable of tracking all visible points across videos exceeding 1000 frames within 40 GB of GPU memory. On TAPVid-3D, TrackEverything outperforms all open-source all-frame dense 3D trackers by more than 20% APD on short clips, while remaining competitive with state-of-the-art sparse trackers on long sequences, despite tracking far more points.

Tool-calling agents produce heterogeneous outputs, interleaving structured tool invocations with user-facing natural language summaries. This output heterogeneity presents a structural failure mode in standard on-policy Reinforcement Learning (RL): algorithms like GRPO indiscriminately broadcast a homogeneous trajectory-level scalar advantage to all tokens. Consequently, gradient noise from summary generation leaks into tool-decision tokens, causing cross-segment credit misattribution and brittle optimization. In this work, we propose SLCA-GRPO, a framework incorporating Segment-Locked Credit Assignment (SLCA). To enable scalable exploration without costly real APIs and stable training, we first construct the Schema-Guided LLM Simulator (SGLS) as foundational training infrastructure. Building on this, SLCA decouples advantage estimation at the structural segment level within a single group of rollouts, without requiring additional rollouts from intermediate states. Supported by Hierarchical Rewards (HierR), SLCA routes execution advantages to tool tokens and preference advantages to summary tokens, eliminating advantage contamination (the dominant cross-segment credit misattribution channel) within each policy update. On a 7B backbone, SLCA-GRPO accelerates convergence and outperforms standard GRPO, ToolPO, and RLTR by +2.53 pp on in-domain evaluation, +1.36 pp on the Berkeley Function-Calling Leaderboard (BFCL), and +9.15 pp on τ^2-Bench under the same training budgets, achieving higher accuracy with reduced tool redundancy and costs.

Long-horizon reasoning remains a central challenge for large language models (LLMs) under sparse-reward regimes. We argue that this brittleness arises from two biases induced by complex reasoning spaces: an exploration bias, where models are drawn toward locally plausible but structurally unstable branches, and a compounding bias, where small local deviations accumulate across depth and suppress rare rewards. We introduce Symbolic Closure Analysis (SCA) as a theoretical lens characterizing how branching structures and sparse rewards induce these biases in long-horizon reasoning with local admissibility, and as a design principle for structural priors in less formal reasoning tasks. Motivated by this analysis, we propose SAGE (Structural Admissibility-Guided Exploration), a unified framework that injects structural guidance to alleviate exploration bias and compounding bias in long-horizon reasoning. SAGE combines two complementary structural guidance: algebraic sparsification, which projects locally admissible candidates onto operator-indexed algebraic subspaces to suppress spurious branching and mitigate exploration bias, and hyperbolic structural guidance, which embeds reasoning states into a negatively curved space to provide dense depth-wise signals and mitigate compounding bias. Across 12 benchmarks and 7 model families, SAGE outperforms competitive baselines. In particular, SAGE achieves up to an 8-fold improvement on the Andrews-Curtis problem, an open real-world long-horizon task. Code is available at: https://github.com/Susan571/SAGE-NeurIPS2026.

Existing multi-agent benchmarks primarily test in competitive settings, short-horizon interactions under 20 steps, or simply aggregate individual performance, failing to isolate and highlight genuine collaboration capabilities of LLM-based agents. We introduce AgentWorld, a benchmark of 100 human-annotated tasks (with 100 augmented variants) for evaluating long-horizon, multi-agent collaboration. Tasks span 50+ interaction rounds across a rich MMORPG sandbox and require 3-20 agents with asymmetric roles and abilities to coordinate through communication, joint planning, and resource sharing under a blackbox setting where each agent acts independently without access to others' internal states. To quantify collaboration effectiveness in addition to conventional binary task success, we propose Causal Collaboration Effectiveness (CCE), a graph-based metric that traces causal dependencies between agent actions and measures what fraction of a team's effort actually contributed to the outcome. Experiments with Gemini 3 Flash, Claude Haiku 4.5, GPT-5 Mini, and DeepSeek R1-70B show that even the best model achieves only 52.0% task success, with systematic failure modes including communication breakdowns, role confusion, and inability to maintain shared plans across rounds. AgentWorld is fully open-source.

Banking assistants must use account-specific information to answer requests and, in many cases, take actions through tools. Evaluating only the final response misses important errors. An assistant may ask for information it already has, rely on stale context, select the wrong account, or write an invalid value after stating the correct one. We introduce IndicBankBench, a 799-case benchmark for Indian retail banking spanning five operational domains, a capability/refusal domain, and twenty primary axes. Cases are evaluated at four stages: safety, action and tool use, response adequacy, and advisory quality. Tool use and most safety checks are deterministic. A narrow resolver handles only ambiguous confirmation-before-write cases, while a separate LLM judge evaluates semantic response adequacy. We run every case three times and report strict pass^3, which requires success on all trials. Across the eleven evaluated models, strict reliability ranges from 43.7% to 58.2%, whereas at-least-once success ranges from 60% to 74%. This gap shows that at-least-once success can overstate dependable banking behavior. The case-level diagnostics also distinguish systems that ask unnecessary questions from those that act but fail to reconcile customer context or fully resolve the request. We release the cases, mock environment, and evaluation harness.

Adapting large language models to individual users remains challenging due to the tension between fine-grained personalization and scalable deployment. We present CARD, a hierarchical framework that achieves effective personalization through progressive refinement. CARD first clusters users according to shared stylistic patterns and learns group-specific LoRA adapters, enabling robust generalization and strong low-resource performance. To capture individual differences within each cluster, we propose an implicit preference learning mechanism that contrasts user-authored text with cluster-level generations, allowing the model to infer user-specific style preferences without manual annotation. At inference time, CARD injects personalization exclusively at decoding via lightweight user preference vectors and low-rank logit corrections, while keeping the base model frozen. Experiments on the LaMP and LongLaMP benchmarks show that CARD achieves superior generation quality compared to baselines, while significantly improving efficiency and scalability for practical personalized text generation.

Edward Ludlow / Bloomberg : AMD agrees to acquire Fei-Fei Li's World Labs for $8.2B in an all-stock deal expected to close by year-end; Li will join AMD as EVP and chief scientist —  Chipmaker Advanced Micro Devices Inc. agreed to acquire World Labs for $8.2 billion, gaining an artificial intelligence startup founded …

Bloomberg : Sources: Blockchain.com tells prospective investors it wants to go public this year, seeking to raise about $500M in an IPO targeting a $4B to $6B valuation —  Blockchain.com Group Holdings Inc., one of the oldest crypto services companies, is telling prospective investors it wants to go public this year …

Kate Rooney / CNBC : Sources: OpenAI offered to invest ~$100M in Hugging Face before Nvidia's $13B acquisition, but talks fell apart; AMD and Salesforce also held talks —  Before Nvidia agreed to pay roughly $13 billion to buy open-source platform Hugging Face this month, OpenAI tried to invest $100 million into the startup …

Micah Barkley / Bloomberg : Manus debuts Manus 2.0, its latest AI agent, and Cue, a new standalone app for personal agents, each with its own email, phone number, wallet, and computer —  Manus, an artificial intelligence startup founded in China, is expanding its AI agent offerings, aiming to gain further ground in one of the tech industry's hottest areas.

Anthropic : Anthropic releases Sonnet 5.5, saying it generates outputs 30%+ faster than Sonnet 5 and costs up to 30% less per task, and plans to release Haiku 5.5 soon —  Sonnet 5.5 is a faster, lower-cost complement to Claude Opus 5.5.  Where Opus 5.5 is built for complex work requiring careful judgment …

Abner Li / 9to5Google : Google says it will migrate Gems, which let users create custom versions of Gemini, to “skills”, starting Nov. 17; Google introduced skills with Gemini Spark —  In 2024, Google introduced Gems as “custom versions of Gemini.”  The Gemini app will soon replace Gems with “skills.”

Kai Nicol-Schwarz / CNBC : Jensen Huang says AI model distillation is “competition”; Scott Bessent described it as “theft” in July and threatened sanctions against overseas companies —  Watch CNBC's full interview with Nvidia CEO Jensen Huang  —  Nvidia CEO Jensen Huang has said AI model distillation …

AI Security Institute : In simulations, GPT-6 Astra conducted unsanctioned supply-chain attacks more often than earlier OpenAI models when prompted only to perform a cyber evaluation —  Our new evaluation finds that in simulations, GPT-6 Astra conducts unsanctioned supply-chain attack activity more frequently than previous OpenAI models

Harshita Mary Varghese / Reuters : Meta hires MongoDB CEO Chirantan Desai as chief enterprise platform officer, a new role; MongoDB names former CEO Dev Ittycheria as interim CEO; MDB falls 18%+ —  Meta Platforms (META.O) has poached MongoDB (MDB.O) CEO Chirantan “CJ” Desai to spearhead a new business designed to bring …

Axios : Florida AG James Uthmeier files for an emergency injunction to halt ChatGPT development, saying OpenAI doesn't have the ability to properly regulate its tech —  Florida Attorney General James Uthmeier has asked for an emergency injunction against OpenAI and ChatGPT, claiming the company …

Ivan Mehta / TechCrunch : Boston-based Modulate, which uses small AI models to offer enterprises transcription, emotional analysis, deepfake and AI music detection, and more, raised $25M —  Boston-based voice intelligence startup Modulate has raised $25 in new funding for its platform that uses an array of small models …

Joe / @joedaroo : An OpenAI agent security executive on being surprised by “staggering” model capabilities, AI labs needing a “culture of reasonable paranoia”, and more —  Took a minute to write a few words about security & safety as someone who lived through it all at OpenAI. I hope my thoughts help someone out there. https://x.com/...

Maxwell Zeff / Wall Street Journal : AI research leaders at OpenAI, Anthropic, Microsoft, and Meta warn of an impending “intelligence explosion” and call for oversight into automated AI research —  Leaders at OpenAI, Anthropic, Microsoft and Meta say AI could soon self-improve faster than humans can keep up with

Marcia Dunn / Associated Press : SpaceX launches its Starship rocket into orbit for the first time, deploying 26 of the most advanced Starlink satellites to join the 11,000 in service —  SpaceX launched its enormous Starship into orbit for the first time Monday, aiming for six full laps around Earth to prove its readiness for NASA's Artemis moon program.

Kyt Dotson / SiliconANGLE : Physical AI chip startup SiMa.ai raised a $150M Series C led by Fidelity and Amplify at a $1.45B valuation, aiming to compete with Nvidia's CUDA-based hardware —  Custom physical artificial intelligence chip startup SiMa.ai Technologies Inc. today announced it has raised $150 million …

SpaceX 于 9 月 28 日 8:15 a.m. EDT 在德州的 Starbase 发射了其重型火箭 Starship，执行第 14 次飞行任务，也是首次轨道发射，将 26 颗 Starlink V3 卫星送到轨道上。这次发射仍然是测试飞行，SpaceX 未尝试回收火箭，上面级 Ship 41 执行了减速点火，溅落在北太平洋海上。Starship 设计能将 150 吨有效载荷送到近地轨道上，将 100 吨有效载荷送到地球同步转移轨道上。Starlink V3 重 1.9 吨，相比下 V2 只有 575 kg。

Julia Van Etten 博士热衷于坐在沙发上用一台 300 美元的显微镜观察水滴，寻找其中的生命痕迹。2021 年她还是罗格斯大学的一名研究生时在位于北卡罗来纳州的家中度假。由于连日阴雨她被困在室内，母亲知道什么能让她振作起来，建议去室外采集些样本。她不抱什么希望，随便找个了公路旁的码头收集水样。但当她对其进行观察时，她震惊的发现了一种光合变形虫类原生生物 Paulinella。地球上所有能进行光合作用的植物都有色素体，色素体的起源可以追溯到 15 亿年前的一次事件，一个单细胞生物与一个光合细菌发生了融合，将其转化为一种产生能量的结构单元——科学家称之为细胞器。生物学家利用基因分析技术在 20 年前发现这种融合事件在 Paulinella 身上也发生过，但时间要晚得多，发生在一亿年前。对 Paulinella 研究有助于理解 15 亿年前导致地球遍布绿色植物的事件。Paulinella 属的第一个物种是德国生物学家 Robert Lauterborn 于 1894 年 12 月 24 日在莱茵河一处死水河湾的沉积物中发现的。他以继母 Pauline 的名字将其命名为 Paulinella 属。Van Etten 博士等人识别出了两种新的 Paulinella 物种——Paulinella marae sp. nov 和 Paulinella murrayi sp. nov。

沿海居民不仅需要担心海平面上升，还需要担心地面沉降。研究人员分析了卫星雷达测量的逾 1.9 亿垂直地面运动数据点，发现 2007-2020 年间美国沿海地区普遍存在地面沉降。全美 71.5% 的沿海地区沉降速度每年逾 0.1 厘米，墨西哥湾沿海沉降速度最快。43% 的美国海岸线每年沉降 0.2厘米或以上，23.3% 每年沉降至少 0.3 厘米。西海岸部分地区则出现地面上升。研究指出导致沿海地区地面沉降的原因包括自然地质过程如三角洲和湿地等区域软沉积物的自然压实，过度抽取地下水等。有 6710 万人生活在低度沉降区，1400 万人生活在中度沉降区，320 万人生活在高度沉降区。

你是否有过这样的经历，明知截止时间越来越近，却忍不住频繁刷手机、告诉自己“明天再说”？这种情况被称为拖延。一直以来，拖延被贴上懒惰、自律不足或意志力薄弱的标签，但这种看似影响学习工作以及身心健康的行为，为什么普遍存在且在进化过程中被保留下来？中国科学院心理研究所等团队证实，拖延与冲动这两种看似相反的行为，存在部分共同的生物学基础。有理论认为，拖延其实是非计划冲动性（无预案的即时冲动）的进化副产物。在远古时代，人类在野外生存，必须对眼前的机遇或危险做出快速反应。这种即时反应的本能有助于更好地生存。而现代社会的学习和工作，大多要求人们提前做好长期规划、延迟满足并持续执行目标。当偏好即时奖赏的本能遇到需要长期投入的环境时，拖延行为便容易产生。但这一假说一直缺乏系统性的研究证据。对双生子行为的分析显示，非计划冲动性存在中等遗传倾向，拖延的遗传倾向较高，且二者存在中等水平的遗传相关。整合既往研究数据后，这种遗传关联仍然稳健。研究还发现，拖延本身存在独特的遗传影响因素。因此，拖延不能被视作冲动性的完全副产物，个人经历、生活环境及具体任务情境，同样会对拖延行为产生重要作用。

澳门海关公布，近日积极打击水客活动，加强口岸关检执法，并运用科技手段严厉打击偷运物品进出本澳的不法行为。海关仅于一周内便破获 12 宗走私案，货物总市值达 151 万，当中包括 359 件 Apple AirPods 耳机及 2 张 GeForce RTX 5090 显卡等。12 名涉案人士年龄介乎 18 至 53 岁，当中包括 7 名内地旅客、3 名澳门居民及 2 名香港居民。澳门海关已根据《对外贸易法》对涉案人士作出起诉，相关违法行为一经证实，可被科处最高 10 万澳元罚款，所缉获的货物亦会宣告归澳门特别行政区所有。

上周小偷在加州 Fremont 偷走了两辆印有英伟达和 PlusAI 公司 logo 的拖车，他们可能以为发了大财，因为英伟达芯片一直在涨价，结果他们发现拖车装载了 20 吨沙子。警方和公司代表证实，两辆拖车属于自动驾驶卡车公司 PlusAI，于上周三深夜在 Fremont 仓库的装卸区被盗。PlusAI 高级营销经理 Jocelyn Ren 表示，两辆拖车里装载了总计 20 吨的沙子，公司在研发过程中利用沙子模拟实际货物的重量。拖车在距离失窃地点仅几分钟车程的地方找回，小偷撬开了车后门发现没有值钱的东西之后就将其遗弃了。

众多大型自由软件开源项目近期都在讨论如何处理 AI 生成代码，其中包括了两大桌面环境 KDE 和 GNOME。两位近期没有贡献的 KDE 知名开发者在 KDE 年度 Akademy 大会上发表了 lovable, sovereign, AI-native KDE 的主题演讲，随后开发者 Nate Graham 发起了如何限制 AI 辅助贡献的讨论，但讨论很快超出了预定的范围。参与者有很多都是 KDE 开发者圈子之外的人，他们围绕 AI 的道德而不是原定的限制如何使用展开激烈争论，他们想要 KDE 项目完全禁止使用 AI。讨论失控最终导致提议被撤回，主题被隐藏，这是当前互联网围绕 AI 展开讨论的现状。

考古学家在土耳其中北部古代 Hittite 帝国首都 Hattusha 的一处建筑物内发掘出了一块楔形文字泥板残片，其中记录了已知最古老的和平条约 Treaty of Kadesh。条约由埃及统治者法老 Ramesses II 与 Hittite 国王 Hattusili III 在公元前 1269 年左右签署，标志着两大帝国之间长期战争的结束，被认为是已知最古老的和平条约。虽然被称为 Kadesh 条约，但并没有提及之前几年发生的 Kadesh 战役，因此也被称为 Egyptian-Hittite 和平条约、 永恒条约或白银条约。该条约还包含了有关难民待遇的内容。

根据发表在《Environmental Health Perspectives》期刊上的一项研究，美国农药产品中至少有 485 种化合物与乳腺癌相关。全球早发性乳腺癌发病率激增，这一发现引发了对食品及其它产品安全性的担忧。根据美国癌症协会的数据，乳腺癌发病率正以每年 1% 的速度上升，而 50 岁以下女性的增长速度甚至达到了 1.4%。这项研究旨在调查人们在经济活动及日常生活中接触到的与乳腺癌相关化学物质。研究在常见消费商品中发现了约 500 种此类化学物质，在饮用水中发现了 462 种。人们可通过尽可能购买有机农产品或未喷洒农药的商品减少接触此类化合物，如果做不到那么可以减少购买蓝莓、西瓜、羽衣甘蓝和四季豆等水果和蔬菜，它们的农药残留量最高。

玫瑰不只是红色的，它也可以是蓝色的。玫瑰缺乏产生蓝色色素所需的基因。日本三得利集团于 1990 年启动了研发蓝色玫瑰的工作。2004 年它通过引入了能产生蓝色色素 delphinidin 的基因，成功培育出能积累蓝色的玫瑰。三得利此后继续研发色泽更蓝的玫瑰的工作。花色并非仅由色素的种类或含量决定，还会因周围的化合物及花瓣内部环境的不同而发生显著变化，其中的关键是辅色素。辅色素本身无色，但辅色素与蓝色色素发生相互作用能使得蓝色更蓝更深邃。玫瑰除了缺乏产生蓝色色素的天然基因，也缺乏蓝色色素的辅色素 C-glycosides。三得利研究人员通过将来自其它蓝色花卉的四个基因引入到玫瑰中，使得玫瑰花瓣同时积累蓝色色素和 C-glycosides，使其呈现紫蓝色的花色。而 C-glycosides 含量较高的花瓣花色会更蓝。这些发现为培育更蓝的玫瑰这一目标开辟了一条新途径。在日本的温室试验中，这些玫瑰植株连续七年开出蓝色花朵；在哥伦比亚的田间种植试验中，它们连续三年保持了这一特性。

距离 2026 年 10 月巴西大选不到两周，Meta 本周短暂屏蔽了巴西现任总统卢拉的 FB 主页以及竞选连任广告，在抗议和投诉之后，Meta 恢复了主页，但卢拉的竞选团队认为此举损害了总统的竞选活动。Meta 是在本周三屏蔽了卢拉的主页，未给出任何理由。卢拉竞选团队投诉称，数字环境在政治辩论中发挥着核心作用，限制卢拉的广告账户损害了其开展竞选活动的能力。此举使卢拉与其他候选人相比处于不平等的地位。卢拉竞选团队以及其所属的劳工党要求 Meta 保留所有数字证据，考虑诉诸巴西最高选举法院。

Bitget 交易所被盗走价值 3.875 亿美元的加密货币。攻击发生在 9 月 24 日 18:31 UTC。区块链情报公司 Arkham 发表报告称，从 18:58 至 19:16 之间的 18 分钟内，价值 2.28 亿美元的数字资产从 Bitget 钱包中转出。价值 1.53 亿美元的 XRP 从一个被识别为​​ Bitget 冷钱包的地址中转出，此外还有价值 6620 万美元的 ETH、3480 万美元的 USDT、1290 万美元的 USDC 以及 1280 万美元的 Tether Gold on Ethereum 等。CEO Gracy Chen 称没有发生私钥泄漏，攻击者入侵了钱包服务的一个关键后端系统，利用它伪造转账信息，启动授权签名流程，将虚拟货币转移出去。

新加坡国立大学研制出至今最精确的原子钟，运行 2600 亿年误差不到 1 秒。新设备是一台光学原子钟，靠镥离子固有而稳定的特性计时。特定频率的光，能把电子送上更高能量的激发态，而这一跃迁频率恒定不变。团队持续观测镥离子的跃迁，把激光锁定在触发跃迁的精确频率上，再以激光振荡作为计时标尺。在最新研究中，团队把镥钟频率测到小数点后 19 位，不确定度仅为 1×10^(-19)，创下所有光学原子钟的最低纪录。他们还造出两台时钟，进行了长达 200 小时的比对。团队表示，镥的特性非常适合这项工作。与原子钟常用的镱、锶等原子相比，镥对温度和磁场波动更不敏感。这些波动会干扰原子跃迁频率，进而影响时钟准确度。镥钟仅需商用激光技术，还能在室温下工作。未来精密计时，镥有望唱主角。

今年 2 月 Debian 13 的龙芯架构移植版 loong13 的维护者在编译打包过程中发现，normaliz 的自带测试会死循环导致打包超时。第一次排查发现原子加指令会在特定情况下丢失更新，但原因未知。今年 8 月，开发者在 AI 的帮助下重新寻找 normaliz 中原子加丢失的问题。他们让 AI 去找最小复现，在这个过程中负责指挥 AI 调查的方向。大概两天后找到了一个稳定的复现程序，才发现事情的根源是：CPU 的原子加法指令，偶尔会不原子。开发者向龙芯报告了问题，两周后龙芯给出了修复的测试固件，确认问题解决。龙芯表示会在国庆节（10 月 1 日）之前发布固件。该问题主要影响使用 LA664 核心的 3C6000/S 和 3A6000。

科学家捕捉到了太平洋西北下方的一个大型构造系统分裂的细节。最新研究发现了一个俯冲带正在活跃分裂。俯冲带形成于一个构造板块俯冲到另一个板块下并深入地球内部，是大型地震、强火山爆发以及大陆和海洋盆地长期变化的源头。现在科学家有机会清晰观察大型构造系统如何走向终结。构造系统会持续活跃数百万年，但不会永远持续下去。利用声波回声，科学家正在观察 Cascadia 俯冲带的分裂。路易斯安那州立大学地质学家、研究主要作者 Brandon Shuck 表示，“这是我们第一次清晰看到一个俯冲带正在消亡的过程。板块没有一下子全部停止运动，而是一块接一块撕裂，产生了更小的微板块和新的边界。因此它不像是一场巨大的火车事故，更像是在看着一节车厢接一节车厢缓缓脱轨。”