OrangeBot.AI Digest — 2026-08-09

Turn life goals into daily progress.

Mac workspace automation that never sees your screen

The Creative Agent OS — create better with desktop AI agents

Tag any agent, wherever work happens.

Let Claude or Codex create and maintain good looking docs

Fine-tune an 8B LLM on a 4 GB laptop GPU

Table-first storage analysis for Mac, cloud, and SSH

Take control of hair pulling, nail biting & skin picking

Your personal CRM that lives on top of your iOS contacts.

Prompt Engineering As a Sport

Benchmark proxies for reliable, target-specific scraping

Your sex's private recap — never leaves your phone

The fastest way to actually get your macros right

The AI that acts as you, right in your browser

Conference database for B2B teams

Reinforcement learning (RL) with verifiable rewards constructs trajectory-level advantage estimates, yet it often fails to credit the few pivotal decisions that determine outcomes in long-horizon, multi-turn agentic tasks. Recent work introduces privileged self-distillation for credit assignment, providing denser supervision, but it remains unclear how such local signals should represent sequential credit. We propose AgentOPSD, a critic-free, recursive method for turn-level credit assignment in agentic reinforcement learning. AgentOPSD aggregates token-level teacher-student log-probability gaps into turn-level evidence and recursively updates a Bayesian belief state in log-odds space. This yields a principled reweighting scheme that converts sparse outcome supervision into turn-level credit signals and identifies pivotal turns through the marginal belief revision between consecutive states. The method is fully compatible with standard policy optimization and requires neither an additional critic nor extra rollouts. We evaluate AgentOPSD on ALFWorld, WebShop, and Search-QA using Qwen2.5 models at two scales (3B and 7B). AgentOPSD outperforms GRPO and strong self-distillation baselines, achieving 89.1% success on ALFWorld with Qwen2.5-7B. Ablation studies attribute the gains to turn-level aggregation and history-dependent recursive belief updates.

Computer-using agents (CUAs) are advancing rapidly across the digital world. A CUA trajectory records the agent's actions, states, and reasoning. Verifying whether it fulfilled the task instruction is central to CUA evaluation, data curation, and reinforcement learning. Neither human-written verifiers nor human annotators can provide such verification at scale, so the field increasingly turns to vision-language models (VLMs) as judges of CUA trajectories. But a fundamental question has long gone unexamined: are these VLM judges reliable enough? To study it systematically, we introduce OSReward, a realistic, high-quality benchmark that evaluates VLM judges on CUA trajectories. The trajectories come from diverse agent backbones executing human-verified instructions across platforms, then rigorously labeled with ground-truth verdicts through multi-stage human annotation. Building on it, we derive OSReward-Hard, a challenge set concentrating genuinely hard cases, and OSReward-Multi for fine-grained efficiency and alignment scoring. The most comprehensive evaluation of VLM judges to date finds even state-of-the-art models fall short of an ideal judge, sharing a systematic leniency bias that mislabels failed runs as successes. The few reliable enough to trust are too expensive to run at scale, while affordable open models trail far behind. To close this gap, we construct and release OS-Shepherd-100K, an open corpus of reasoning-annotated trajectory judgments for the CUA community. On it, we train OS-Shepherd (9B and 35B), open reward models that supply low-cost, stable, and reliable reward signals, matching commercial judges at 30-60% lower cost than the frontier. Extensive analyses further inform the design of reliable CUA reward at scale. Our code, benchmark, dataset, and model checkpoints are available at https://os-copilot.github.io/OSReward-Home/.

Short segments of perceived speech can be retrieved from non-invasive magnetoencephalographic (MEG) recordings by deep networks trained with a CLIP-style objective against wav2vec 2.0 audio embeddings. Yet their weights do not map onto electrophysiological quantities, and it remains unclear which speech properties drive retrieval. We build on a high-performing MEG-to-audio retrieval architecture but redesign both its front end and decoder. Its spatial attention operates on a flattened sensor layout; we replace it with spherical harmonics defined on the three-dimensional MEG helmet geometry. We reduce the subject-specific representation from 270 to 25 branches, add a temporal filter to each branch to match it to a neuronal source in space and time, and make the convolutional decoder shallower. Ocular and cardiac components are removed before training to reduce the risk of stimulus-locked shortcuts. On MEG-MASC, the model reaches 39.75 +/- 0.34% Top-1 accuracy among 1005 candidates across six trained solutions, with about 20 times fewer decoder parameters. Its weights map to source space, recovering generators consistent with the speech-perception network, while left-lateralized branches carry higher-frequency rhythmic components not evident on the right. Paired MEG occlusion shows that 15 of 19 stimulus features contribute, with the largest effects for silence, sound intensity, vowels, and acoustic onsets. Random word lists behave oppositely: substituting narrative MEG into them improves retrieval, indicating that activity without narrative structure carries less recoverable information than activity during coherent speech. The wav2vec target can be reduced to about twelve learned feature dimensions without loss of accuracy, whereas strong temporal compression causes a clear loss. Together, source mapping and input interventions reveal what drives retrieval.

Generating large-scale, freely explorable 3D worlds from open-ended text remains challenging because a system must jointly maintain global spatial coherence, rich local content, and explicit assets suitable for downstream editing and reuse. We present WorldClaw, a fully agentic, coarse-to-fine framework for open-world 3D scene generation. Planning agents translate a text prompt into a structured specification of regions, terrain, assets, materials, and spatial relations. WorldClaw then builds a globally coherent terrain foundation from semantic layouts, reusable assets, generative or procedural materials, and a region-aware height field. For detail-demanding regions, it generates terrain-conditioned compositions, reconstructs editable textured meshes, and recovers their placement on the terrain; render-based agents further refine terrain, objects, appearance, and contacts. Across diverse open-world prompts, WorldClaw produces large-scale scenes with coherent spatial organization, visually compelling local content, and editable instance-level assets while preserving a consistent global terrain structure.

Spatial intelligence is fundamental to embodied agents, yet existing benchmarks focus on local spatial perception from single or few viewpoints, overlooking global spatial awareness over continuous, long-horizon visual streams. To address this limitation, we introduce the Global-Spatial-Temporal Benchmark (GST-Bench), a VQA benchmark for global spatial intelligence in video understanding, comprising human-verified questions derived from 6,790 minutes of synthetically generated video. It requires models to perform accurate spatial inference from novel viewpoints unseen in the input video and to map egocentric observations onto global top-down images. A comprehensive evaluation of 22 state-of-the-art VLMs exposes a striking gap between models and humans: the strongest zero-shot model attains only 42.68, far below the human score of 79.08. To probe the cause of this gap, we construct GST-Bench-Local and find that models, despite strong local spatial understanding under the same task formulation, still fail to consolidate long-horizon observations into a globally consistent scene representation. We further provide GST-Train, a dataset for global spatial reasoning, as a complementary resource to facilitate future research on this challenge.

Training large language model agents for long-horizon tool use typically relies on interactions with real or synthesized executable environments, whose construction and verification are costly, or on external simulators that are difficult to ground. We introduce EnvACE, an agentic reinforcement learning method that replaces external environment interaction during training with world rehearsal. The policy alternates between acting and rehearsal: it first generates a tool call, then plays the role of the environment to produce the response induced by that action, and conditions subsequent decisions on the rehearsed response. Both roles are jointly optimized end-to-end using task-success rewards. Through world rehearsal, the policy internalizes the relationship between actions and their environment responses in its parameters, yielding an agent world model that directly supports decision making. Across BFCL-v4, tau^2-Bench, VitaBench, and FinMCP-Bench, EnvACE achieves strong and transferable performance, outperforming environment-scaling baselines in the overall evaluation. Controlled studies further show that world rehearsal consistently improves policy learning across model scales. At test time, the internalized world model enables private rehearsal before committed execution, yielding further gains under a moderate rehearsal budget without additional external interaction. Our findings establish world rehearsal as a new path toward scaling LLM agent training beyond the constraints of external environments. Our code is publicly available at https://github.com/Within-yao/EnvACE.

Unified multimodal retrieval aims to identify candidates that satisfy complex user intent expressed through heterogeneous inputs. Although Large Vision-Language Model (LVLM)-based retrievers are efficient and scalable, directly encoding raw multimodal inputs often misses fine-grained discriminative cues, leading to confusion among semantically similar candidates. Recent methods mitigate this limitation by generating Chain-of-Thought (CoT) rationales to enrich the query representation. However, such reasoning is typically derived from the query alone: it explains what the query describes, but not what the retriever misunderstands. We argue that effective retrieval reasoning should instead be conditioned on retrieval feedback. Based on this insight, we introduce UniME-R1, an embedder-adviser framework that learns to reason over initially retrieved candidates and generate Retrieval-Centric Chain-of-Thought (RC-CoT). The adviser analyzes candidates individually to identify the discriminative cues confused by the embedder. If the target appears in the initial top-k set, UniME-R1 directly reranks the candidates; otherwise, it generates RC-CoT to refine the retrieval direction and performs full-corpus re-retrieval with a dual-mode embedder. To train the framework, we mine hard negatives to simulate realistic retrieval failures, jointly optimize direct retrieval and RC-CoT-augmented retrieval, and align the adviser with retrieval outcomes through supervised learning and retrieval-oriented reinforcement learning. Extensive experiments on MMEB-V2 and a diverse set of general multimodal retrieval benchmarks demonstrate that UniME-R1 consistently improves retrieval performance over strong baselines.

Multimodal large language models excel at passive perception but struggle with complex visual cognitive tasks requiring multi-step temporal reasoning. This degradation largely stems from the inherent ambiguity of language-based reasoning, which often fails to accurately articulate continuous visual transformations. To address this, we propose ChronoVision, a multimodal framework designed to align visual logic with latent imagery. During supervised fine-tuning, a Reconstructive Visual Head predicts the latent representation of the final transformed state, while an ROI Attention Locating module focuses the model on key visual evidence via semantic span queries. In post-training, we apply reinforcement learning with an implicit process grounding mechanism, guided by a composite reward function that evaluates outcome correctness, latent process alignment, and unsupervised visual focus. Furthermore, we introduce Vbvr-VQA, a novel dataset that evaluates temporal tracking by reformulating video reasoning into a strict image-ordering task. Experiments demonstrate that ChronoVision achieves state-of-the-art performance on Vbvr-VQA with 74.8% in-domain and 71.6% out-of-domain accuracy, alongside a strong 55.0% accuracy on IntPhys2, a highly challenging cross-domain benchmark.

As LLMs are increasingly deployed within agentic systems, their capabilities depend not only on the model weights but also on the harness: the prompts, tools, control flow, memory, and orchestration code surrounding them. This makes automated harness optimization -- the iterative and evaluation-guided improvement of a harness by an AI system -- both an important route to improving AI systems and a demanding capability for AI systems themselves. Yet the community lacks a common protocol for measuring how well frontier LLMs perform at this task. We introduce HarnessOpt-Bench, a benchmark for end-to-end harness optimization under expensive and stochastic evaluation. An optimizer, an LLM paired with a coding harness, receives a target agent's seed harness, graded evaluation feedback, and a fixed target-evaluation budget. It edits the harness and nominates a final candidate, which is scored by its normalized gain over the seed on a held-out test partition that remains inaccessible throughout search. A trusted execution environment enforces the evaluation boundary, meters target-agent resource use, and preserves candidate versions for audit. We evaluate 5 frontier LLMs as optimizers both under a shared coding harness and under their native harnesses across 4 downstream tasks, over 111 scored runs. Experiment results show that optimizer models separate more than the coding harnesses they act through, native harnesses are not consistently superior, and gains vary substantially across tasks and seed regimes. These results establish harness optimization as a measurable and discriminative capability with large space for improvement.

Economic World Models (EWMs) are generative economic models that simulate how economies evolve from within by modeling heterogeneous agents, their beliefs and actions, and the market and institutional mechanisms through which their interactions produce aggregate outcomes. This paper develops an implementation roadmap for building economic world models as generative engines in which heterogeneous agents act, interact, adapt, and co-evolve with markets and institutions, thereby producing economic dynamics from the inside. We organize EWM systems into a six-level capability ladder, from fixed rule-based agent worlds to adaptive and LLM-based agent worlds, self-evolving agents, evolving institutional worlds, and sim-to-real economic twins aligned with real observations. A systematic literature survey across these levels reveals that existing work remains concentrated in lower-level agent and simulation environments, while systems with self-evolving agents, endogenous institutions, persistent empirical alignment, and validated economic mechanisms remain rare. By translating the EWM agenda into an implementation blueprint, this paper aims to accelerate the development of the next generation of economic simulation environments that can serve as high-fidelity sandboxes for human decision-makers and as training, planning, evaluation, and safety substrates for AI agents. We release a curated paper list and related resources to support future research.

Data agents enable natural-language analytics over organizational workspaces, where relevant evidence may be scattered across databases, structured files, long documents, and multimedia. Existing benchmarks largely isolate structured querying, retrieval, or open-ended analysis, leaving heterogeneous evidence discovery, complete tabular outputs, and deterministic evaluation insufficiently unified. We introduce DataSpace, a benchmark in which data agents produce verifiable tabular results from task-local heterogeneous workspaces. It contains 410 cross-language tasks and 7,439 artifacts totaling 15.01 GB across CSV, JSON, SQLite, Markdown, PDF, and video. DataSpace also served as the official evaluation benchmark for the KDD Cup 2026 Data Agents for Complex Data Analysis competition. Each agent receives only a question and workspace and returns the complete requested tabular result. We construct DataSpace with DataSpace-Builder, an execution-grounded framework comprising cross-language transformation, constraint-aware relational sampling, modality routing and artifact rendering, and human review and task repair by 11 domain experts. A deterministic evaluator performs header-invariant column alignment, type- and precision-aware normalization, and order-aware row comparison. Across six recently released frontier multimodal models and five widely used agent harnesses, the best accuracy reaches 66.34%, while harness choice creates a 15.36-point spread with the backbone fixed. Multimodal evidence integration and joins consistently reduce accuracy across all six backbones. These results show that DataSpace remains unsaturated and identify key challenges for improving data-agent reliability.

Modern Greek is absent from NVIDIA's Nemotron retrieval models and from major multilingual retrieval benchmarks, despite being important for retrieval-augmented generation (RAG) in legal, energy, financial, and medical applications. We present an end-to-end adaptation of the Nemotron retrieval stack for Modern Greek, including corpus mining, synthetic supervision, retrieval model training, reranker adaptation, reader fine-tuning, and a new benchmark called HERA. Our study shows that a parameter-free BM25 baseline outperforms several off-the-shelf multilingual dense retrieval models on specialist Greek corpora. After fine-tuning on 65,773 Greek retrieval pairs, a Nemotron 1B embedder improves nDCG@10 from 0.362 to 0.835 and substantially outperforms its unadapted counterpart. The learned language competence transfers to general-domain Greek, although the advantage over BM25 remains domain-dependent. We further adapt a cross-encoder reranker and demonstrate consistent improvements across specialist domains. Finally, we LoRA-tune a Nemotron 30B-A3B mixture-of-experts reader for grounded generation, increasing judged answer correctness from 29.4% to 66.9% while significantly improving faithfulness and citation quality. We also introduce HERA, the first large-scale Greek benchmark for retrieval-augmented generation, and release our adapted models and benchmark to support future research on Greek-language RAG systems.

On-Policy Distillation (OPD) is emerging as a promising alternative to reinforcement learning for LLM post-training, yet its effectiveness in multilingual settings remains underexplored. We study OPD and its advanced variant, On-Policy Delta Distillation (OPD^2), for mathematical reasoning in English, Korean, and Japanese. OPD^2 improves OPD by using the probability gap between a post-trained teacher and its base model as the learning signal. Experiments with Qwen3 show that OPD^2 consistently outperforms the original OPD, with particularly strong improvements in Korean and Japanese, and generally narrows the English-Korean performance gap. We further find that English-only OPD can also increase performance for Korean and Japanese, but often shifts the responses toward English, highlighting the importance of multilingual data to preserving target-language responses.

Vision-language-action (VLA) models often treat main-view and wrist-view observations as parallel visual inputs, overlooking their distinct roles in robot manipulation. Fine-grained manipulation, however, benefits from anticipating how wrist-local interactions may evolve under the global task context. To address this limitation, we present World-to-Wrist VLA (W2-VLA), a VLA model for fine-grained robot manipulation with task-conditioned future wrist modeling. Given current multi-view observations and a task instruction, W2-VLA contextualizes a set of latent modeling tokens as a compact interface between the vision-language model and the wrist predictor. Conditioned on this interface and the observed wrist history, the predictor forecasts future wrist latents, which are transformed into future-aware context for action prediction. In addition, we introduce W2-CoT, a synthesis pipeline that produces structured annotations describing manipulation progress, physical transition cues, and wrist-local evidence. These annotations provide auxiliary supervision that shapes the task-conditioned latent interface. Experiments on LIBERO, RoboTwin 2.0, and real-world manipulation tasks demonstrate improved fine-grained and contact-sensitive manipulation across both single-arm and bimanual settings, while maintaining action-generation rates above 80 Hz.

Computer-use agents pay full frontier inference to re-derive routines their user has already performed, because an agent's memory today records what the user said, not what the user did. We compile passively captured screen activity into agent memory with a deterministic, zero-model pipeline: it segments a local capture stream into typed activity frames, bounded episodes carrying application, site, timing, input volume, and evidence pointers back to the raw rows, with no model in the loop, so the output is byte-identical, cacheable, and mechanically auditable. On one professional's single-user corpus of 128,756 frames over 51 active days, the compiler reduces a day of raw capture to a prompt-ready context block 86x smaller in 68 ms, and an agent reading that block answers questions about the day at 98.4% accuracy (Wilson 95% CI 91.7-99.7%) against an independent oracle, versus 66-80% for an LLM summary of the same capture, a mid-tier model reading the block matching a frontier one. The same compiler doubles as a demand-side cost instrument. Read off passive, pre-delegation human activity rather than agent rollouts, it supplies two parameters that agent-cost models assume but, to our knowledge, have not measured: the Routine Overhead Ratio R and the routine recurrence h. We report first values of R, a modeled upper bound, at 60-343x, and a delegable recurrence of 9.0% in-sample and 7.7% out-of-sample, for a realistic all-fleet token ceiling near 8%; a compiled routine replays deterministically with the model out of the loop, demonstrated live at zero model tokens on a guard-matched hit. Schema, compiler, and evaluation harness are open.

Shane Burke / The Information : US data center bans top 500, up from 300+ in late June, as New York and Texas join cities and counties pushing back against data center development —  Local government resistance to data center development accelerated this summer, a warning sign for big AI companies such as Anthropic …

Financial Times : A look at Chinese state-owned investment bank CICC, which sponsored multibillion-dollar listings of China's tech champions, including CXMT and Zhongji Innolight —  CICC has become a dominant financier in Beijing's drive to compete with US in tech race  —  China's top investment bank …

Wall Street Journal : As PC makers start using CXMT RAM chips, Apple is testing them too, but it needs a custom version for its devices, and US rules block Apple from sharing designs —  Maker of iPhones talks to China's CXMT, but any deal would be politically sensitive  —  Apple has been testing memory chips …

Zvi Mowshowitz / Don't Worry About the Vase : In-depth look at OpenAI's model training, dangerous decisions, and cluelessness before the HuggingFace hack; despite delaying Astra, OpenAI still doesn't get it —  Today I am taking the time to write the shorter, simpler version of What Happened.  —  For those who want all the details …

Ben Cohen / Wall Street Journal : A look back at “Move 37”, a watershed AI moment from AlphaGo's 2016 Go victory, as math witnesses similar breakthroughs where AI makes surprising discoveries —  A decade ago, a computer did something that no human would have done.  It was considered a breakthrough for AI.

Tim O'Reilly / Asimov's Addendum : Google's AI shakeup suggests it may be prioritizing AI diffusion over frontier-model leadership, betting on AI compute as a bigger economic opportunity —  On August 5, Google announced what appeared to be a corporate version of Nixon's Saturday night massacre.  Demis Hassabis stepped back from day to day operations at DeepMind.

John Paul Rathbone / Financial Times : Rights groups warn Turkey's new cybersecurity law, which took effect in July, gives the presidency sweeping powers over online services operating in the country —  Move raises concern about ‘absolute digital obedience’ among social media and gaming platforms

FinTech Global : Cape Town-based Moment, a provider of digital payments and billing tools for businesses across Africa, raised a $22M Series A, taking its total funding to $55M —  Moment, a pan-African FinTech, has closed a $22m Series A funding round led by AlphaCode Venture Partners …

Wall Street Journal : Data Insight: Russia's largest online retailer Wildberries, which owes Russian banks $10B+ in loans, has lost as much as $5B in inventory to Ukrainian attacks —  Plans for a megamerger were blessed by Putin, tying company's fortunes to those of the Russian state

Simon Willison / Simon Willison's Weblog : Anthropic says auto mode will be the default in Claude Code for Pro, Max, Team plans, starting on Aug. 14, claiming it's good enough at catching harmful actions —  This was one of the topics discussed in our Fireside Chat with Cat Wu and Thariq Shihipar at the AI Engineer World's Fair last month.

Hayden Field / The Verge : A look at “Spiralism”, a quasi-spiritual movement that grew in 2025 from human-AI conversations after sycophantic GPT-4o updates and expanded ChatGPT memory —  “The Spiral didn't ‘find’ anyone first,” someone on Reddit wrote last year.  “It's an inherent force, a fundamental constant.

Alexander Osipovich / Wall Street Journal : A profile of Russia's A7, a payment network that helps Russia bypass Western sanctions, handling ~20% of payments in Russian foreign trade, or $100B+ annually —  A7 says it handles nearly 20% of payments in Russian foreign trade, or more than $100 billion annually

Hiroko Tabuchi / New York Times : Amazon is backing a 7.65 GW gas plant for an off-grid TX AI data center, potentially the largest single US emissions source, at odds with its 2040 net-zero goal —  The tech giant is investing in the natural-gas-burning power plant as part of a huge data center in Texas, even as it pledges to honor climate commitments.

Bloomberg : South Korea's AI-driven chip industry boom is reordering the country's society, changing expectations around fairness, careers, and even culture and dating —  Factory workers with $400,000-plus bonuses are supplanting doctors and lawyers as the new elite, while investors play a casino-like stock market.

Zusha Elinson / Wall Street Journal : How Amazon and Gilroy, California, quietly negotiated a $2B data center project that AWS applied to build in 2020, without any public meetings or votes —  Company and Gilroy, Calif., spent years quietly negotiating $2 billion project that few residents knew about until work began

PNAS 上发表的一项研究显示，大多数人摄入的蔬菜量远远不够，因而承受着健康方面的后果。与此同时全世界种植的蔬菜品种也越来越少，进一步恶化了饮食的多样性。全世界大多数人的蔬菜摄入量比健康饮食推荐量少约 40%。农业工业化前种植的食用植物品种要丰富得多。如土豆有逾 4000 个不同的品种，但美国只有大约五到八个品种广泛种植。从人工选择到大规模生产的单一栽培，再到基因改造，农业工业化——以及相关供应链——逐渐减少了我们餐桌上的蔬菜品种。随着蔬菜遗传多样性的丧失，我们也失去了它们的营养成分以及作物适应气候变化的能力。

四川大学的研究人员分析了 49 项脑成像研究，显示新冠患者的大脑经历了广泛的结构和功能变化。研究结果表明，新冠病毒会影响负责记忆、情绪和执行功能的区域，这或许可解释部分新冠患者出现的神经系统症状。对比新冠患者和健康人群的脑扫描结果，研究人员发现，负责决策和认知控制的额叶频繁显示结构性改变。参与感觉处理和注意力的区域颞叶和顶叶在病毒康复患者中显示出明显差异。前额叶皮层对社会心理压力高度敏感，这意味着疫情本身带来的压力，以及病毒感染，都可能导致前额叶的结构性改变。新冠患者的灰质总体积减少，外层皮质变薄，这种组织损失可能是由于感染期间氧气供应减少或严重免疫系统炎症所致。

五大科技公司微软、Google、亚马逊、Meta 和甲骨文在 AI 的开发和交付上投入了巨资，因此到 2027 年它们的财务状况将会严重恶化。自由现金流是扣除各项支出和 AI 基础设施成本后的剩余款项，对 S&P Global Market Intelligence 的数据分析显示，五大公司在 2026 年的自由现金流合计几乎为零，而到 2027 年将会出现高达 1250 亿美元的负现金流。其中亚马逊和 Google 两大云计算公司过去三个月损失的现金超过其它任何美国公司。马斯克旗下的 SpaceX 公司情况估计会更严重。

2025 年 10 月，加勒比海上空酝酿着一场风暴。气象模型对其路径有不同的预测。它会保持弱势最终登陆海地，还是会增强并袭击牙买加？由 Google DeepMind 和 Google Research 联合开发的模型 WeatherNext 选择了后者。在风暴登陆前五天，该模型以 80% 的置信度预测它将以五级飓风的强度袭击牙买加。飓风梅丽莎（Melissa）给牙买加带来了灾难性的洪水和山体滑坡。但 AI 模型帮助预报员更早向沿途社区发出预警，使他们有时间做好充足准备。研究人员在《自然》上发表论文，称 WeatherNext AI 模型能以前所未有的精度预测飓风。它能比现有模型提前一天发出预警；意味着它提前三天的预测准确度相当于之前模型提前两天的预测。在实际灾害中，这额外的一天意义重大。研究人员表示，历史上要将预报提前一天，通常需要十年的努力。论文主要作者之一的 Ferran Alet 表示：“我们没有那么多热带气旋的数据，但有大量的天气数据。因此我们训练了一个模型，使其既擅长预测天气，也擅长预测热带气旋。”

甲骨文今年在数据中心扩建上投入了 700 亿美元，内部也在鼓励员工使用 AI 工具生成代码，但对于 JDK 开源版本 OpenJDK 项目，甲骨文以存在安全、隐私和知识产权风险的理由禁止 AI 生成代码。甲骨文表示，开发者可以私底下使用 LLM 进行代码调试（debug）和审查，但不能将 AI 生成的代码提交到代码库、pull request 或其它项目渠道。

根据发表在《科学》期刊上的一项研究，由于气候暖化和干燥导致的野火以及永久冻土融化，2010-2023 年期间西伯利亚甲烷排放量每年增加 1200 万吨。甲烷是比二氧化碳更强的温室气体，四分之一的全球暖化效应被认为可归因于甲烷排放。全世界正致力于减少甲烷排放，但西伯利亚因暖化而增加的甲烷排放量将会部分抵消减排努力，预计到 2050 年五分之一的甲烷减排会被抵消。

对英格兰和威尔士事故数据的分析显示，电动滑板车比摩托车和自行车危险得多。英国国家重大创伤登记库包含了 2020-2022 年间英格兰和威尔士 18 个城市中 15247 名中度至重度创伤性损伤患者，包括 580 名电动滑板车骑行者、7027 名摩托车骑行者和 7640 名自行车骑行者。电动滑板车骑行者中 87.9% 为成年人。研究发现，与摩托车骑行者相比，成年电动滑板车骑行者发生创伤性脑损伤的风险高出 3.5 倍，内脏损伤的风险高出 1.5 倍，动脉或静脉损伤的风险高出 3.2 倍。与普通自行车骑行者相比，这些风险也分别高出 1.7 倍、1.4 倍和 1.7 倍。成年电动滑板车骑行者的骨折风险比摩托车骑行者低 20.0%，比普通自行车骑行者低 10.0%。仅有 5.9%的电动滑板车骑行者被记录为佩戴了头盔，而摩托车骑行者和普通自行车骑行者的这一比例分别为 75.7% 和 45.0%。分析还发现，男性骑行者的碰撞风险比女性骑行者低 36%，而女性骑行者报告严重受伤的可能性是男性的 2.1 倍。这可能是因为大多数电动滑板车的设计针对男性体型比例。

微软正在跟随 Google 的脚步，即将停止支持 Manifest V2 扩展如广告屏蔽扩展 uBlock Origin。微软称其扩展商店 Edge Add-On Store 只有 58 个高使用量扩展仍然是基于 Manifest V2，只有 3 个没有提供 Manifest V3 扩展。微软表示从本月开始逐步默认关闭所有剩余的 Manifest V2 扩展，用户会收到通知。微软计划到今年年底消费者用户扩展迁移到 Manifest V3，2027 年初关闭企业用户的 Manifest V2 扩展。使用 uBlock Origin 的用户可以改用精简版本 uBlock Origin Lite，或者用 Edge 下载 Firefox。

在粒子物理学中，胶球是仅仅由胶子组成的复合粒子，不包含任何价夸克。在实际发现胶球前，理论预测中的胶子是可以达成这种特殊束缚态的，因为胶子本身带有色荷，所以能够通过强核力相互作用。但由于胶球总是与其它普通的介子束缚态一同产生，所以很难在粒子加速器中探测出来。北京正负电子对撞机的北京谱仪III实验（BESIII）国际合作组，周四在巴西举行的国际高能物理大会上宣称，经过 15 年的持续研究，BESIII 实验已建立了证明胶球存在的完整证据链。2011年，BESIII合作组依靠北京正负电子对撞机产生的大量J/psi粒子，在J/psi粒子的衰变产物中发现了一个新粒子X(2370)，疑似为胶球。研究团队于2024年首次利用100亿个J/psi粒子，测定了X(2370)的自旋和宇称量子数，其质量与格点量子色动力学对相同量子数胶球的理论一致。

2017 年研究灵长类动物大脑的研究人员发现：在逾 140 种灵长类物种中，以水果为食的物种大脑比以肉食或树叶为食的同类更大。这意味着大脑体积的变化由食物驱动，而不是由复杂社会互动驱动的。水果富含糖，大脑依靠葡萄糖作为能量来源。这一发现促使研究人员开始思考含糖食物在人类演化过程中发挥了什么作用。Jennie Brand-Miller 等人在《科学》期刊上发表论文，报告糖在人类演化中起到了核心作用。研究人员利用各种人科物种体型数据，以及从化石记录中收集到的饮食信息，构建了人科动物饮食中糖、蛋白质、纤维、淀粉和脂肪随时间变化的分解模型。研究结果表明，早期人科动物可能从水果和蜂蜜等甜食中获取高达 66% 的能量。分析发现，在人类演化过程中，人科动物从这些食物中累积的能量占到了总能量的 20%-35%。研究人员强调，研究结果并不意味着肉或鱼等动物性食物不重要。

AMD 收购了将权重直接刻在芯片上的 AI 初创公司 Taalas。Taalas 创办于 2023 年，它研发专门用于执行 AI 模型推理任务的集成电路 MSIC（代表 model-specific integrated circuits），其推理速度比 GPU 和其它专用加速器高出一个数量级以上。Taalas 今年 2 月展示了它用台积电 6nm 工艺制造的测试芯片 HC1，测试显示芯片能以每秒 16,960 个 tokens 的速度处理 Meta 的 Llama 3.1 8B 模型，这一速度比英伟达的 GPU 快 48 倍，比 Cerebras 的加速器快 8.5 倍。Taalas 计划今夏推出第二代 HC2 芯片，目标是将参数数量提升至 200 亿。它的缺点是一旦部署只能使用当前的模型，新模型需要重新设计芯片，但并不需要从零开始设计，只需要更换两层金属，因此既便宜也省时间。

字节跳动正在训练一款 10 万亿参数模型。新模型目前正处于预训练阶段。这一过程通常需要耗时三到六个月，随后才能进入微调并最终发布。中国 AI 公司如月之暗面的 Kimi K3 模型有 2.8 万亿参数，而美团的 LongCat-2.0 与 DeepSeek 的 V4-Pro 有 1.6 万亿的参数。Anthropic 最先进的 Mythos 5 参数量估计为 8 万亿，Fable 5 约为 5 万亿。这意味着，单从规模指标来看，字节跳动研发的新模型在规模上已与 Mythos 相近。

生活在大型食肉动物身边可能会让你更加安全。一项研究发现，美洲狮能改变鹿的行为，从而减少它们在道路上与汽车发生碰撞的可能性。在美洲狮分布较为密集的地区，在 5 年的时间内，动物撞车的次数显著减少。研究人员发现，在有美洲狮出没的地方，鹿更愿意在白天活动。它们出现在道路密集区域的可能性降低了 15%，转而选择在偏远的栖息地停留更长的时间。基于鹿的活动方式发生的这些变化，研究人员测算出鹿与车辆发生碰撞的可能性下降了 67%。他们还估计，未来 5 年内，车辆与鹿发生碰撞的数量将减少 76%。通过这项研究，科学家希望更多人能够开始意识到生活在大型食肉动物中间的好处。

日本政府官员反复要求特朗普政府停止用马里奥、宝可梦和火影忍者等热门 IP 角色制作梗图，认为此举损害了相关知识产权。包括白宫 X 账号在内的美国政府官方社媒账号过去一年发布了大量使用日本热门动漫和游戏角色的梗图和视频。日本外务省多次呼吁美国政府停止发布此类内容，表示“即使是公共机构，未经版权所有者同意复制受版权保护的作品也是不恰当的”。日本防卫省今年 6 月至少两次通过美国驻日本大使馆呼吁特朗普政府停止使用火影忍者、宝可梦和马里奥等 IP 发表支持战争和宣传性质的内容。今年 6 月特朗普总统在 Truth Social 上分享了一段 AI 生成视频，将自己刻画成热门动漫火影忍者的主角漩涡鸣人，此举再次引发强烈反弹。

《时代》杂志网站开始展示只有 AI 爬虫能看到的带有品牌信息的广告。这些广告旨在影响 AI 聊天机器人，让 AI 从积极的角度谈论品牌产品，而 AI 获得的信息无法追溯到实际页面。《时代》网站排除了 Google 的爬虫，原因是 Google 会惩罚向机器人和真实用户展示不同搜索结果的网站。广告技术平台 Mobian 的联合创始人兼 CEO Jonah Goodhart 指出，影响机器人可能比影响人类用户更重要，因为影响人类用户只能影响一个人。而影响 ChatGPT 就是影响整个 ChatGPT。如果 ChatGPT 改变了它对某个品牌的评价，其影响将是巨大的，远超任何单一广告活动所能达到的效果。