OrangeBot.AI Digest — 2026-06-07

Daily AI stories personalised from your Google apps

View, monitor, and analyze 1.8M+ US jobs

Flight map from cabin Wi-Fi

Turn your voice into text — local or cloud, your choice

Simple file sharing on your iPad or iPhone

Profile for publishers/creators to highlight work on Search

My keyboard fell apart. Now it's your problem.

Generate and edit images with precise scene control

Build and manage Shopify stores from one chat

Track, plan, and release.

Add Tab Bar, Mega Menu & more to any website — no code

Level up the way you play with Steam, Dota 2, and more

The world’s most accurate AI for investors

Powers faster, efficient reasoning for long-running agents

Get real-world tasks done with autonomous AI agents

Code language models need repository-level context to resolve imports, APIs, and project conventions. Existing methods inject this knowledge as long inputs (retrieved through RAG or dependency analysis) or through per-repository fine-tuning and LoRA -- costly at repository scale and brittle to evolving codebases. We introduce Code2LoRA, a hypernetwork framework that generates repository-specific LoRA adapters, effectively injecting repository knowledge with zero inference-time token overhead. Code2LoRA supports two usage scenarios: Code2LoRA-Static converts a single repository snapshot into an adapter, suitable for comprehension of stable codebases; while Code2LoRA-Evo maintains an adapter backed by a GRU hidden state updated per code diff, suitable for active development of evolving codebases. To evaluate Code2LoRA against parameter-efficient fine-tuning baselines, we build RepoPeftBench, a benchmark of 604 Python repositories with two tracks: a static track with 40K training and 12K test assertion-completion tasks, and an evolution track with 215K commit-derived training and 87K commit-derived test tasks. On the static track, Code2LoRA-Static achieves 63.8% cross-repo and 66.2% in-repo exact match, matching the per-repository LoRA upper bound; on the evolution track, Code2LoRA-Evo achieves 60.3% cross-repo exact match (+5.2 pp over a single shared LoRA). Code2LoRA's code can be found at https://anonymous.4open.science/r/code2lora-6857; the model checkpoints and RepoPeftBench datasets can be found at https://huggingface.co/code2lora.

Role-playing language agents (RPLAs) should play characters whose values and behavior evolve as the story progresses, not maintain a fixed persona. Existing benchmarks measure factual recall at a given chapter, not whether responses align with the character's psychological trajectory, especially in scenarios the source text never explores. We introduce ArcANE (Arc-Aware Narrative Evaluation), an automatically constructed benchmark spanning 17 novels and 80 principal characters. A Character Arc segments the narrative into phases along a psychological axis, and each probe poses the same scenario across phases, spanning both situations within the source text and situations beyond it. Across six models and six context modes, conditioning on the Character Arc tops every other context strategy on every model, and the gap is largest on scenarios outside the source text where retrieval has nothing to find. We further fine-tune open-weight models on the same data to obtain ArcANE-8B/32B, which widen the Arc advantage even more on scenarios outside the source text.

Agents are widely deployed as assistants over documents, tools, and code. However, they typically act only on explicit user requests, which surface only the problems the user has noticed, while many other important problems coexist, hidden in plain sight, within the broader user context, with their total number unknown in advance. We frame this as the task of discovering multiple hidden problems from context, in which coexisting problems should be uncovered, grounded in supporting evidence, and paired with concrete actions. To this end, we introduce TIDE, a template-guided iterative framework with two complementary mechanisms. Specifically, motivated by the observation that single-pass prediction anchors on the most salient cases and yields generic claims, we propose iterative discovery, which surfaces a small batch of candidates per round while conditioning on what has already been found, so subsequent rounds extend coverage; and thought templates, reusable schemas distilled from previously solved cases that specify what contextual signals to attend to and how to connect them, anchoring each prediction in a recognizable problem class. We validate TIDE on two realistic settings, personal workspaces and software repositories, across four model backbones, showing substantial gains over single-shot and parallel multi-agent baselines on task coverage, identification, and resolution.

Planning for real-world problems by language models often involves both world and user constraints, which may not be fully specified upfront and are progressively disclosed through interaction. However, existing benchmarks still underexplore adaptive planning under such progressively revealed dual constraints. To address this gap, we introduce AdaPlanBench, a dynamic interactive benchmark for evaluating whether Large Language Model (LLM) agents can adaptively plan and re-plan under progressively revealed world and user constraints. AdaPlanBench is built on 307 household tasks, with a scalable constraint construction pipeline that augments each task with dual constraints. At runtime, agents interact with the environment in a multi-turn protocol where hidden constraints are revealed only when the agent proposes a plan that violates them, requiring iterative plan revision under accumulating feedback. This makes planning challenging, as agents must infer and track constraints from feedback while re-planning effectively. Experiments on ten leading LLMs show that adaptive planning under dual constraints remains challenging, with the best model reaching only 67.75% accuracy. We further observe that performance degrades as more constraints accumulate, with user constraints posing a particularly large challenge and failures often stemming from weaker physical grounding and reduced effectiveness. These results establish AdaPlanBench as a testbed for dual-constrained interactive planning and highlight the challenge of reliable adaptation to dynamically revealed constraints in LLM agents.

We introduce VideoKR, the first large-scale training corpus specifically designed to strengthen knowledge- and reasoning-intensive video understanding. It comprises 315K video reasoning examples over 145K newly collected, CC-licensed, expert-domain videos. We develop a human-in-the-loop, skill-oriented example generation pipeline that targets progressively deeper video reasoning capabilities while ensuring the difficulty, diversity, and reliability of both the examples and their CoT rationales. We also curate VideoKR-Eval, a new expert-annotated benchmark where questions require genuine video understanding and knowledge-intensive reasoning rather than textual shortcuts. Our experiments show that, under a standard SFTrightarrowGRPO pipeline, models post-trained on VideoKR outperform prior post-training approaches on knowledge-intensive video reasoning while remaining competitive on general video reasoning, highlighting data design as a key driver of progress in video reasoning. We further conduct comprehensive ablations to isolate the contributions of VideoKR, providing actionable insights for future work.

While household robots are often evaluated based on task completion, everyday domestic environments involve value-conflicting situations in which robots are expected to choose actions that prioritize other values than task success, such as human autonomy, efficiency, or social appropriateness. Yet, there are no benchmarks for evaluating robots' value preferences in such scenarios. We introduce RobotValues, a benchmark to evaluate household robot planners in 10K value-conflict scenarios. Each instance consists of a realistic household image with multiple plausible robot actions that prioritize different human values. We construct RobotValues through LLM-assisted scenario generation, stakeholder-grounded value extraction, image generation and automatic quality control. Using RobotValues we evaluate VLMs used in robotics and find that models exhibit default value preferences, including safety and accommodation, while underselecting privacy-prioritizing actions. When the models are instructed to prioritize specific values that conflict with their own preferences, they often fail to override their default actions, choosing incorrect actions for 80% of the time. These findings suggest that household robot evaluation should measure not only task completion or safety compliance, but also whether robots can choose among plausible actions when human values conflict.

Prior work has shown that large language models (LLMs) can translate unseen or low-resource languages by undergoing continued training or even by encoding a grammar book in their context. However, both methods typically overfit specific languages, with limited zero-shot transfer at test time. To translate extremely low-resource languages at scale, we argue that LLMs must acquire the meta-skill of utilizing in-context linguistic knowledge rather than memorizing specific languages. In this paper, we propose a reinforcement learning (RL) approach to unseen language translation given rich linguistic context, using a surface-level translation metric (chrF) as the reward. Empirically, despite the lightweight reward, our RL-trained models effectively extract and apply relevant linguistic information from the provided context, leading to better translations on completely unseen languages than in-context learning or supervised fine-tuning. Our analyses suggest that outcome-based RL can extend beyond conventional reasoning tasks like math and coding to serve as a recipe for language learning from context.

Developing unified video generation and editing models capable of interpreting interleaved multimodal inputs is a promising yet challenging frontier field. Existing unified frameworks predominantly rely on massive models (typically 13B parameters or more) and incorporate source video conditions for editing by concatenating sequence tokens. This concatenation inevitably doubles the sequence length, quadrupling the computational complexity of the self-attention mechanism and introducing prohibitive overhead. To address these bottlenecks, we present LoomVideo, a highly efficient 5B-parameter unified architecture for both video generation and editing. LoomVideo replaces the standard text encoder with a Multimodal Large Language Model (MLLM) and employs Deepstack injection mechanism to align multi-layer MLLM features with the Diffusion Transformer (DiT). Crucially, we introduce a zero-overhead Scale-and-Add conditioning approach for video editing. By scaling and directly adding the clean source video latent to the noised target latent, this elegant design eliminates the need for token concatenation, drastically reducing computational cost while maintaining robust capabilities for complex, non-rigid edits. Furthermore, a Negative Temporal RoPE strategy is seamlessly integrated to handle multiple reference images. Extensive experiments demonstrate that our compact 5B model achieves state-of-the-art or highly competitive performance across comprehensive benchmarks, exhibiting exceptional superiority in e-commerce and fashion generation scenarios. Benefiting from the zero-overhead conditioning mechanism, LoomVideo achieves at least a 5.41x acceleration in inference speed compared to models of similar capabilities, paving the way for highly practical and efficient video foundation models.

Experience internalization converts contextual experience from past interactions into reusable parametric capability, offering a promising path toward continual learning in large language models (LLMs). While prior work has predominantly focused on single-iteration transfer, we discover that under multi-iteration experience learning, existing methods suffer from a progressive capability collapse rather than compounding improvement. We systematically examine this failure through three vital dimensions of experience internalization: (1) Experience Granularity: We find that principle-level experience is more durable than instance-level experience, as it effectively abstracts transferable strategies away from trajectory-specific details. (2) Experience Injection Pattern: Our analysis reveals that step-wise injection significantly outperforms global injection by aligning experience with intermediate decision states, a property that is critical for long-horizon tool use. (3) Internalization Regime: We demonstrate that off-policy context-distillation on high-quality teacher trajectories provides a substantially more stable training signal than on-policy context-distillation, which is inherently limited by local corrections on student-induced flawed states. Together, these insights yield a simple yet robust recipe for stable and sustainable experience internalization, providing concrete guidance for engineering self-evolving and continually learning LLMs.

We study the personal camera roll visual question answering setting. In this setting, a conversational AI assistant can access a user's personal camera roll and retrieve relevant photos to answer queries, ranging from simple factual questions (e.g., ``Name of the food I tried yesterday?'') to more open-ended ones (e.g., ``Recommend some dishes I have never eaten before''). Given the vast nature of the personal camera roll (i.e., multiple years, hundreds to thousands of photos), a successful AI assistant needs to understand a long-horizon, highly personalized visual content stream in order to navigate and locate the correct and/or relevant information. To support this, we collect and manually annotate questions that mimic real-world usage. The final dataset, camroll, contains 50 users, 31,476 images, and 2,500 QA pairs. We further design camroll-agent, a conversational AI agent equipped with hierarchical memory and a minimal set of tools for efficient navigation over large, personalized visual memory. Experimental results show that camroll-agent outperforms numerous baselines and methods for long-context understanding AI agents system. Together, the camroll dataset and camroll-agent highlight the gap in AI agents' long-context reasoning: personalized visual memory requires different approaches from standard long-context textual memory, especially when consistency, visual details, and user-specific context are present.

Standard continuous-time generative models rely on monolithic architectures that must navigate vastly different signal regimes, from isotropic noise to intricate data distributions. While scaling model capacity improves performance, deploying a massive network uniformly across the entire generative timeline is inherently inefficient. In this work, we propose Complexity-Balanced Splitting (CBS), a principled framework for temporal capacity allocation that distributes the generative workload across multiple specialized sub-networks. Grounded in function approximation theory and de Boor's equidistribution principle, CBS partitions the diffusion timeline into segments of equal approximation burden, allocating more representational capacity to regions where the generative dynamics are more difficult to model. To estimate this local complexity, we introduce two complementary and tractable monitor functions: a spatial measure based on the flow's Dirichlet energy, and a geometric measure based on the acceleration of the sampling trajectories. Using a lightweight auxiliary model to estimate these complexity profiles, our approach eliminates the need for heuristic temporal splits or computationally expensive search procedures. Extensive evaluation across multiple architectures (SiT, JiT, and UNet) and datasets demonstrates that CBS consistently improves synthesis quality without increasing per-step inference cost. In particular, CBS improves FID by ~35% on SiT-XL with CFG relative to naive temporal partitioning. Project page is available at https://noamissachar.github.io/CBS/.

Existing autonomous driving datasets have enabled major progress, but fall short in sensor fidelity, map completeness, or geographic diversity. We present KITScenes Multimodal, a European dataset built around high-fidelity sensors and maps. Our fully synchronized sensor suite combines high-resolution global-shutter cameras, long-range lidar beyond 400m, 4D imaging radar, and redundant GNSS/INS localization. Our HD maps are, to our knowledge, the most complete of any sensor dataset, validated through autonomous driving trials on open-source software. For the first time in a public dataset, all driving-relevant traffic elements, such as traffic lights, are mapped in 3D to a reprojection-accurate level with full topological connectivity. Recorded in cities with irregular street layouts and mixed traffic modes, our dataset complements existing datasets by broadening the available geographic diversity. We also introduce four benchmarks, each advancing spatial learning for embodied AI: online HD map construction, long-range depth estimation, novel view synthesis, and end-to-end driving. Project page: https://kitscenes.com/

Video generation models have made impressive strides in synthesizing visually compelling content, yet their outputs remain confined to the virtual domain. A natural question follows: how well do these models reflect the physical world when their generated videos leave the screen and enter reality? We propose robotic manipulation as a concrete, measurable window onto this question: if a model has truly internalized physical laws, the motion it depicts should translate into executable robot behavior. We introduce Dream.exe, an evaluation framework that operationalizes this criterion through a video-to-execution pipeline. Given a scene image and a task description, Dream.exe synthesizes a manipulation video, converts the generated motion into robot trajectories, and executes them in a physics simulator, yielding a grounding signal that purely visual metrics cannot offer. Using this pipeline, we evaluate 8 models spanning frontier closed-source generators, open-source generators, and robot-specific models. Our benchmark covers 101 manually curated manipulation tasks at three levels of physical complexity, measured across visual quality, trajectory fidelity, and execution success. Encouragingly, several models achieve measurable execution success, suggesting that generative priors learned from internet-scale data already encode meaningful physical knowledge. Yet visual quality proves a poor predictor of executability, exposing a dimension of model capability that standard visual evaluations do not capture. Dream.exe will be open-sourced at https://github.com/showlab/Dream.exe.

Large language model (LLM) agents are increasingly applied to long-horizon tasks such as scientific discovery and machine learning engineering (MLE), where sustained self-evolution becomes a key capability. However, existing MLE agents suffer from inter-branch information isolation, memoryless search, and lack of hierarchical control, which together hinder long-horizon optimization. We present MLEvolve, an LLM-based self-evolving multi-agent framework for end-to-end machine learning algorithm discovery. By extending tree search to Progressive MCGS, MLEvolve enables cross-branch information flow through graph-based reference edges and gradually shifts the search from broad exploration to focused exploitation with an entropy-inspired progressive schedule. To allow the agent to evolve with accumulated experience, we introduce Retrospective Memory, which combines a cold-start domain knowledge base with a dynamic global memory for task-specific experience retrieval and reuse. For stable long-horizon iteration, we further decouple strategic planning from code generation with adaptive coding modes. Evaluation on MLE-Bench shows that MLEvolve achieves state-of-the-art performance across multiple dimensions including average medal rate and valid submission rate under a 12-hour budget (half the standard runtime). Moreover, MLEvolve also outperforms specialized algorithm discovery methods including AlphaEvolve on mathematical algorithm optimization tasks, demonstrating strong cross-domain generalization. Our code is available at https://github.com/InternScience/MLEvolve.

Inference-time skill augmentation provides a lightweight way to improve data-analytic agents by injecting reusable procedural knowledge without updating model parameters. However, discovering effective skills for data analysis remains challenging, as reliable supervision is expensive and success criteria vary across analytical formats. This raises the key question of how to discover reusable data-analysis skills from unlabeled exploration alone. We propose DataCOPE, an unsupervised verifier-guided skill discovery framework for data-analytic agents. DataCOPE derives verifier signals from the exploration trajectories and uses them to characterize relative quality or aggreement among trajectories. It iteratively coordinates a Data-Analytic Agent for trajectory generation, an Unsupervised Verifier for signal extraction, and a Skill Manager for contrastive skill distillation. For report-style analysis, we instantiate the verifier as an Adaptive Checklist Verifier that derives task-specific criteria, scores reports by verifiable coverage, and iteratively refines the checklist. For reasoning-style analysis, we instantiate it as an Answer Agreement Verifier that groups trajectories by answer agreement and uses self-consistency as an auxiliary signal. We evaluate DataCOPE on report-style analysis from Deep Data Research and reasoning-style analysis from DABStep. Across both settings, DataCOPE consistently improves held-out performance over baselines. Averaged across four model settings, DataCOPE improves the mean score by 9.71% and 32.30% on report-style and reasoning-style tasks respectively.

Bloomberg : As wealth managers confront an AI reckoning, the tech is, for now, easing their workloads by picking up routine tasks, freeing up more time to advise clients —  Wealth managers, who can make upwards of $500,000, are confronting a chatbot reckoning.  —  Down the Wealth Walkway on South Beach …

Wall Street Journal : KPMG survey: only 26% of companies have a comprehensive view of their AI costs, while 50% have some visibility and 22% have none or only see costs after billing —  Only 26% of companies say they have a comprehensive view of their AI costs, according to a new survey

Mark Gurman / Bloomberg : Inside the early 2025 senior exec meeting where Apple admitted its AI failures; Tim Cook became intimately involved in Apple's AI roadmap around the same time —  Also: What to expect at WWDC 2026.  —  A top secret meeting finally got Apple to take its AI crisis seriously.

Hassan Mujtaba / Wccftech : Memory chip vendors say CXMT's DDR5 prices match Samsung, SK Hynix, and Micron, and it has a supply advantage in client markets as it is not prioritizing HBM —  CXMT's DDR5 memory is apparently not as cheap to procure as reports claim, falling in the same range as DRAM from other manufacturers.

Dwarkesh Patel / Dwarkesh Podcast : Q&A with Google DeepMind's Director of AGI Economics Alex Imas and Epoch AI's Phil Trammell on what remains scarce after AGI, redistributing AI wealth, and more —  “One robot now turns into many robots next year, but the number of ballerinas is the same.”  —  Economics of AGI episode w Alex Imas and Phil Trammell.

Toby Sterling / Reuters : ASML CEO Christophe Fouquet applauds the EU's tech sovereignty plans but says substantial EU involvement in steering “strategic projects” is a cause for concern —  The CEO of top European tech firm ASML (ASML.AS) on Friday welcomed most proposals made by the EU Commission this week …

Tanya Pandey / The Economic Times : Industry data: India's six largest IT companies grew combined revenue from $95.9B in FY23 to $103.1B in FY26, while headcount remained broadly steady at ~1.9M —  India's top IT companies are achieving revenue growth with stable employee numbers.  This trend highlights a shift towards automation and improved productivity.

Reuters : French carriers Bouygues Telecom, Orange, and Free-iliad sign an MOU with Altice France to acquire SFR for €20.35B, including debt, expected to close in H2 2027 —  Bouygues Telecom (BOUY.PA), Orange (ORAN.PA) and Free-iliad Group said on Saturday they have signed a memorandum …

Cristina Criddle / Financial Times : OpenAI plans to overhaul ChatGPT in the coming weeks, turning it into a superapp with coding tools and AI agents to serve as a gateway to higher-margin products —  $850bn start-up to recast hit chatbot as a route to higher-margin products before a potential IPO.

Isabella Kwai / New York Times : How LinkedIn is transforming into a “post-cringe” social network as it courts high-profile influencers, who are building audiences on corporate-friendly topics —  Celebrities and paid influencers have transformed the career-focused platform, which for some users, has become a job unto itself.

Caleb Mutua / Bloomberg : Report: in May, supply of unsecured bonds from hyperscalers hit $155B YTD, 45%+ more than 2025's total issuance; some AI-infra bond sales are 4x oversubscribed —  Credit heavyweights like DoubleLine Capital LP and Oaktree Capital Management are buying debt now that can perform …

Robert Wright / Financial Times : Several UK police forces have been told to stop using AI to prepare court statements, citing concerns that inaccurate outputs could contaminate legal procedures —  Safeguards must be in place before forces automate justice tasks, says head of Police.AI  —  A number of UK police forces …

Leo Schwartz / The Information : White House AI advisor Sriram Krishnan says he will leave his role at the end of June; sources: Krishnan plans to start a pro-Trump AI policy institution —  Sriram Krishnan, a top AI policy advisor in the Trump administration, plans to leave his position at the end of June …

Financial Times : Raspberry Pi closed up 27%+ on June 5 after saying it expects adjusted EBITDA of at least $38M in H1, putting it on track to beat $42M est. for the full year —  UK maker of tiny low-cost computers expects ‘robust demand’ to push unit sales above 4mn in first half

Josh Eidelson / Bloomberg : FOIA docs reveal Amazon's extensive control over delivery drivers it insists are not employees, in a case the NLRB sought to settle on terms favorable to Amazon —  The feds were pushing a landmark case about Amazon's control of its contract drivers.  Then the president put Amazon's former lawyer in charge.

美国政府考虑持有 AI 公司股份。OpenAI CEO Sam Altman 正与白宫就政府可能入股这家 AI 公司进行持续磋商。双方的讨论已持续一年多，本周 Altman 在华盛顿会见了多位议员和官员，就监管和 AI 的最新发展进行了磋商。作为潜在协议的一部分，OpenAI 可能会向美国政府捐赠股权，用于建立某种公共财富基金。该基金可以“投资于多元化的长期资产”，让公民能获取 AI 发展的“收益”。在特朗普的第二个任期内，政府已入股了英特尔、IBM 以及量子和关键矿产公司。

在 1970 年代，Parul Gayen 生活在德里的贫民窟，那儿到处都是孩子。她的母亲有 6 个兄弟姐妹，她的祖父有 11 个兄弟姐妹。她的丈夫 Swapan 有 6 个兄弟姐妹——第 7 个夭折了，两人在 16 岁时结婚，有 3 个孩子。如今她已经 58 岁，但他们的孩子只有两个决定生育，而且只生 1 个。时代变了。她说，一个孩子会感到孤独。印度如今是世界人口最多的国家，但它正走在中国的人口开始减少的道路上——中国人口自 2021 年起开始减少。印度生育率下降的速度比预期的更快也更早。印度人口众多的贫困邦的生育率正向富裕邦看齐：人口 7700 万的泰米尔纳德邦和人口约 1 亿的西孟加拉邦的总和生育率均为 1.3，与芬兰相同。印度城市的平均总和生育率为 1.5。印度人口的峰值预计为 15.5 亿。

数据显示，2026 年春季加州伯克利 CS 10 课（The Beauty and Joy of Computing）的不及格率高达 35.3%，CS 61A 课（计算机程序的构造和解释）的不及格率达到了 10.6%。而在 2025 年和 2024 年春季，这两门课的不及格率均未超过 10%。教这两门课的教授 Dan Garcia 认为不及格率上升与学生使用大模型相关：学生被发现使用大模型如 Claude、ChatGPT 和 Google Gemini 考试作弊，或过于依赖大模型完成作业但对知识一知半解因此未能对考试做好准备。其它原因包括数学基础薄弱以及师资力量不足。

天文学家利用位于智利的 ALMA 大型毫米波/亚毫米波阵列长达五年的高解析度观测资料，首次清楚看见银河黑洞 Sgr A* 向外吹出的高温气流，解开困扰天文学界逾 50 年的谜团。根据理论，黑洞在吞噬周围气体时，部分物质也会以气流或喷流形式向外释放，但银河系中心黑洞的这种现象过去始终难以直接观测。研究团队观测距离 Sgr A* 约 3 光年范围内的一氧化碳分子讯号。一氧化碳是追踪冷分子气体的重要指标，能帮助天文学家描绘黑洞周围的气体分布。结果发现，在冷气体分布中存在一个巨大圆锥状空洞，方向正对着黑洞。研究人员结合 NASA Chandra X 射线天文台的观测资料，发现这个空洞内充满高温热气体。这代表 Sgr A* 正持续向外吹出高温高能气流，将周围冷气体扫开或加热，形成这个特殊结构。虽然这股气流不像某些活跃星系中的黑洞喷流那样剧烈，但研究团队估计，它至少已持续存在约 2 万年。

食品营养分析自由软件 GNUtrition 在时隔 14 年后释出了新版本 v0.33。GNUtrition 上一次更新是在 2012 年。GNUtrition 的食品营养信息使用的是美国农业部的数据库 Food and Nutrient Database for Dietary Studies(FNDDS)。v0.33 主要变化是：用 C 语言取代 Python 2 进行了重写，UI 从 GTK 2 升级到 GTK 3，旧营养数据库 Nutrient Database of Standard Reference 于 2018 年停止更新，因此改用了 FNDDS 数据库，等等。

微软最近的诸多行为（封了国际刑事法院首席检察官的账号，与美国众议院分享了荷兰监管机构员工的信息）给欧洲敲响了警钟，促使欧洲开始拥抱数字主权，减少对软件巨人的依赖。欧盟制定的数字主权标准包含了对开源授权的要求。微软创办的商业软件联盟（BSA）发表声明对此表达了强烈反对，称“真正保护欧洲的是治理、审计和降低风险的能力，而不是公司在哪里注册成立。此类标准会增加成本，降低获得一流安全解决方案的机会，可能与欧盟的国际贸易承诺相冲突。”

2025 年 10 月比特币创下了 123,742 美元的币值记录，8 个月之后它的币值已经跌去了一半，本周五比特币币值一度跌破 6 万美元。这一波跌势被认为与 Zcash 的漏洞相关。Shielded Labs 披露了 Zcash (ZEC) Orchard 隐私池的一个高危漏洞，可能威胁到代币供应的完整性。如果被利用，攻击者能在不被发现的情况下无限量伪造 ZEC 代币。这一消息导致 ZEC 在 24 小时内币值下跌了 42%，整个加密货币市场也受到牵连。Zcash Foundation 发表声明称未发现漏洞被利用的证据。

SpaceX/xAI 的聊天机器人 Grok 显然用户太少而导致马斯克（Elon Musk）耗巨资购买的英伟达 GPU 大量闲置，为了避免数据中心空转，SpaceX 近期先后与 Anthropic 和 Google 两大 AI 巨头达成了类似的算力出租协议：Anthropic 同意在 2029 年之前每月向 SpaceX 支付 12.5 亿美元租用 Colossus 1 数据中心的算力，Google 每月向 SpaceX 支付 9.2 亿美元租用 11 万个英伟达 GPU 及相关计算基础设施。SpaceX 未透露 Google 租用 Colossus 1 还是 Colossus 2 数据中心。与 Anthropic 的协议类似，与 Google 达成的协议也包含终止条款。SpaceX 和 Google 都可以在 2026 年 12 月 31 日之后提前 90 天通知对方终止交易。

由于国际空间站俄罗斯舱段的漏气过去几天从每天一磅空气增加到两磅（0.9 公斤），NASA 命令国际空间站上的宇航员待在飞船内，做好紧急撤离的准备。NASA Crew-12 任务的四名宇航员——两名美国宇航员、一名法国宇航员和一名俄罗斯宇航员——于美国东部时间周五 9.04am 接到 NASA 任务控制中心的命令，进入与空间站对接的 Crew Dragon 飞船，穿上宇航服，以防漏气情况需要紧急撤离。漏气的舱段位于 Progress（进步号）气闸舱和 Zvezda（星辰号）服务舱之间的 PrK 模块，漏气原因是微小的结构裂缝。最近几个月 NASA 和俄罗斯航天局一直在讨论漏气的原因和可能的修复方案。

Brave 浏览器过去几年积累了加密货币钱包、AI 助手、新闻流和奖励计划等不太欢迎的功能。为了回应用户对臃肿功能的不满，Brave 推出了精简版 Brave Origin 浏览器。Linux 平台免费，但其它平台则要付费，且价格不菲。Brave Origin 移除了 Brave Rewards、钱包、Leo AI、新闻流、Talk、VPN、Tor 等功能，保留了内置的广告和跟踪器屏蔽功能 Brave Shields，它的一次性授权费用为 59.99 美元，最多可用于 10 台设备。60 美元是否物有所值则取决于用户了。

越来越多的研究将超加工食品与心脏病、糖尿病、过早死亡等关联起来。但科学家仍在争论究竟是什么导致了健康风险：是食品本身的营养质量，还是生产过程中使用的工业加工和添加剂。根据《American Journal of Public Health》期刊上的一项研究，加工过程本身可能在其中发挥着重要作用。超加工食品的加工过程会改变食物细胞结构、流失有益化合物，引入添加剂以及包装的化合物。对美国长达 20 年的健康营养数据分析显示，超加工食品的热量每增加 10%，健康指标就会恶化。食用超加工食品的人体重更高、血糖控制更差、血压更高、胆固醇水平较差。他们更容易患上糖尿病、代谢综合征和癌症，在研究期间有更高的死亡风险。在考虑了超加工食品的营养质量，以及饱和脂肪、添加糖或钠的含量之后，这种关联仍然存在。

根据发表在《科学》期刊上的一项研究，大黄蜂能利用工具解决问题。昆虫加入到了能解决“盒子香蕉”问题的动物行列，展现出了基本智能。在盒子香蕉问题中，黑猩猩通过叠盒子够着了之前够不着的香蕉。在最新研究中，研究人员根据大黄蜂修改了盒子香蕉问题：它需要将聚苯乙烯球滚到特定位置，然后爬上去够到低天花板上的人造花。参与实验的大黄蜂只有几周大，研究人员训练它们将人造花与糖水奖励联系起来。在基础测试中 75% 的黄蜂成功够到了花朵；在更复杂的测试中，30 只黄蜂中有 23 只成功了。研究人员指出，即使昆虫的大脑非常小，它们也能灵活解决各种新问题。

根据 Cloudflare 的统计，基于 HTTP 请求的机器人流量已远超人类，由于数据混乱机器人流量超过人类的确切时间不太清楚。目前机器人流量占 57.5%，人类占 42.5%。Cloudflare 统计的是 AI 智能体，这些 AI 智能体能代表人类浏览网页，阅读产品页面、查看价格、执行比较航班等多步骤任务、抓取和索引网页内容——但用于 AI 大模型而非搜索引擎，以及充当私人助理去订餐比价和购物，处理客户服务等。就应用使用、流媒体播放和无限滚动信息流的总时长而言，人类用户仍然是主要群体。按国别/地区划分，直布罗陀岛的机器人流量比例最高（92.1%），其次是新加坡（76.4%）和伊朗（76.4%），伊朗可能是 VPN 用户比较多。

苹果宣布全球 App Store 生态系统在 2025 年促成了逾 1.4 万亿美元开发者营业额与销售额。在 App Store 生态系统促成的营业额和销售额中，超过 90% 完全归开发者所有，无需向 Apple 支付任何佣金。苹果未单独披露 App Store 收入，而是将其计入服务业务。服务业务在 2025 财年贡献了 1091 亿美元，占苹果总收入 4161 亿美元的近三分之一。iPhone 业务收入最高达到 2095 亿美元。根据 Analysis Group 的分析，1.4 万亿美元中 1490 亿美元来自数字商品和服务，1.1 万亿美元来自实体商品和服务。中国市场贡献了最大的销售额 5620 亿美元，其次是美国 4530 亿美元、欧洲 1840 亿美元和日本 520 亿美元。

Google 旗下企业 Debug 正寻求政府许可在加利福尼亚州和佛罗里达州释放 3200 万只雄蚊。这些雄蚊携带了沃尔巴克氏体细菌（Wolbachia），会导致细胞质不亲和性，意思是雄蚊的精子无法让野雌蚊的卵子受精。理论上这会导致蚊群数量逐代减少。雄蚊不会叮咬人，只有雌蚊才会，因此 Debug 并没有释放大量吸血昆虫。Debug 正在等待美国环保署的批准，公众意见征询截止日期 6 月 5 日。目前的公众意见显示很多人持有阴谋论观点，声称“人不是实验鼠”。