OrangeBot.AI Digest — 2026-07-22

Turn your old iPhone into a music-first dumbphone

Runtime authorization for Claude, Cursor, Codex and OpenClaw

13,000+ MCP servers, skills & plugins for AI coding agents

Your Chat UI Just Got an AI Roommate

No-slop AI SEO agent that interviews you first

Your creative AI co-director

Simple computers for agent w/ full VMs

Social intelligence plugin for Hermes Agent

A developer credential that proves what you built, NDA safe.

A story-driven game for learning Claude by doing

Identify every aircraft in your sky.

One identity layer for your customers, APIs, and agents

Fine-tune and Deploy Open-weight Models in TypeScript

Communities love your brand - Our AI team finds them for you

AI runs your email marketing. Better with every send.

We present ABot-World-0, an action-conditioned video world model for real-time, long-horizon closed-loop interaction, supported by a multi-source data infrastructure spanning AAA games, simulation engines, and internet videos to learn controllable world dynamics. WorldExplorer performs agent-driven collection guided by training feedback, while a unified pipeline applies 14 deterministic quality checks, VLM-based assessment, and synchronized action and text annotation. We progressively distill a bidirectional action-conditioned teacher into a causal student through teacher forcing and ODE distillation, and introduce LongForcing to align long student self-rollouts with an extended-horizon teacher, mitigating accumulated distribution shift and autoregressive drift. Raw keyboard actions provide a unified control interface for scene roaming and third-person character interaction, while reference-character memory provides persistent appearance cues for identity consistency during third-person rollouts. For deployment, we co-design a streaming inference stack with a lightweight VAE decoder, efficient attention, memory-aware scheduling, and low-bit DiT inference. Across optimized low-bit configurations, ABot-World-0 streams 720P video at up to 16 FPS on a single NVIDIA RTX 5090 desktop GPU, with 1.2s action-to-first-frame latency and approximately 19GiB peak VRAM. Experiments on WorldRoamBench and extended interactive rollouts demonstrate competitive controllability and coherent long-horizon world evolution.

Large language models (LLMs) are increasingly used to automate data-processing workflows, yet coding agents typically produce scripts that are not automatically materialized as persistent, editable platform artifacts. We call this disconnect the NL2Pipeline gap. To bridge it, we introduce DataFlow-Harness, a platform that guides an LLM agent to construct platform-native directed acyclic graphs (DAGs) through typed, incremental mutations rather than free-form scripts. The platform combines DataFlow-Skills for procedural guidance, a Model Context Protocol (MCP) layer that exposes the live operator registry and current pipeline state, and DataFlow-WebUI, which synchronizes conversational authoring with a visual DAG editor. On a 12-task data-engineering benchmark, DataFlow-Harness achieves a 93.3\% observed end-to-end pass rate. Relative to Vanilla Claude Code, it reduces measured monetary cost by 72.5\% and generation latency by 49.9\%; its observed pass rate is within 0.9 percentage points of the Context-Aware Claude Code baseline while its cost is 42.8\% lower. Per-task analysis indicates that Skills are most useful when construction depends on implicit procedural knowledge. These results show that live platform grounding can produce persistent, editable workflow artifacts with an observed reliability close to script-generation baselines and with lower measured construction cost and latency.

Text-to-image diffusion transformers (DiTs) jointly process text and image tokens, yet their internal computation during denoising remains poorly understood. We introduce a causal interpretability framework for modern large-scale DiTs that combines attention decomposition with targeted interventions across token spans, heads, and layers. Using it to separate prompt-content tokens from structural template tokens, we find that the structural tokens carry little prompt-specific information at the encoder output. Yet surprisingly, they emerge as dominant image-to-text attention sinks and causally maintain object identity inside the DiT, acting as implicit semantic registers. We show that they acquire this identity indirectly, with prompt semantics first injected into the image latents and then read back into the template tokens rather than transferred directly from the prompt tokens. Inspired by the above findings, we design a training-free pruning rule for DiTs. Heads that attend most strongly to prompt tokens are dispensable, and pruning them removes 20% of attention FLOPs with only a 1.4-point drop on GenEval. We further reveal how generative computation in DiTs is organized across heads and depth, separating semantic routing from visual synthesis and progressing from identity formation to propagation and refinement. Our work not only reveals that the tokens encoding semantics at input need not be those that maintain it during generation, but also provides a causal view of internal mechanisms in DiTs.

Generative world renderer AlayaRenderer receives structured world states exported from physics engines and synthesizes RGB frames. Unlike models that generate frames from text/control-hints prompts, AlayaRenderer preserves scene structure without altering the underlying world dynamics. This demonstrates an alternative path toward interactive world modeling and user-controllable play. However, the original AlayaRenderer is too computationally expensive for real-time deployment. This technical report introduces AlayaRenderer-Flash, a real-time-oriented generative forward world renderer that pushes AlayaRenderer from 0.56 FPS to 31.54 FPS, reaching the speed of play. AlayaRenderer-Flash reformulates the original renderer as a few-step autoregressive streaming model and introduces lightweight distilled codecs for efficient latent encoding and frame reconstruction. It retains the teacher model's G-buffer and text-prompt interfaces while enabling continuous rendering over input streams of unbounded length. We evaluate AlayaRenderer-Flash on G-buffer streams across content preservation, temporal consistency, cross-window stability, prompt controllability, and runtime efficiency. Our results show that AlayaRenderer-Flash substantially reduces inference cost while preserving the core rendering capabilities of the teacher model. By integrating AlayaRenderer-Flash with a physics engine, we build a fully playable generative world running at 30 FPS.

Large-scale visual generators are increasingly capable but costly to train, fine-tune, and deploy. We introduce Mage-Flow, a compact 4B-scale generative stack for efficient text-to-image generation and instruction-based image editing. The stack is built from two co-designed components: Mage-VAE, a lightweight high-fidelity latent tokenizer, and a Native-Resolution Multimodal Diffusion Transformer trained with rectified flow matching. Mage-VAE uses one-step diffusion-style encoding and decoding with anchor-latent regularization, preserving the reconstruction quality of strong public VAEs while reducing tokenization cost by more than an order of magnitude. Together with native-resolution packing and stack-level CUDA kernel fusion, the stack supports flexible-resolution training and improves end-to-end training throughput by about 2.5times. Built on this foundation, we develop a complete model family with Base, RL-aligned, and Turbo variants for both generation and editing. Diffusion-NFT improves prompt following, text rendering, aesthetic quality, and editing fidelity, while few-step distillation with adversarial perceptual guidance produces 4-step Turbo models for low-latency inference. Despite its compact scale, Mage-Flow and Mage-Flow-Edit achieves competitive performance across standard generation and editing benchmarks. More importantly, the Turbo variants make high-resolution generation and editing practical for interactive use: at 1024^2 resolution on a single NVIDIA A100 GPU, Mage-Flow-Turbo generates an image in 0.59s, and Mage-Flow-Edit-Turbo edits an image in 1.02s, while maintaining a small memory footprint. These results show that careful tokenizer--backbone--system co-design can deliver strong high-resolution generation and editing within an efficient 4B model family.

Unlike conventional video game development, which relies on labor-intensive pipelines for asset production, animation, physics, and programming, video world models generate interactive environments from user inputs instantly. It enable us to create customized, explorable, and continuously evolving virtual world from text, an image, or video. Realizing this vision requires four tightly coupled capabilities: interaction, persistent spatiotemporal consistency, stable long-horizon generation, and efficient response. We present AlayaWorld, an interactive long-horizon video world model that generates 24-fps video at 540p and 720p. Built on a 15B video diffusion transformer, AlayaWorld generates short latent chunks autoregressively under camera trajectories and switchable text prompts. Its bounded visual context combines a persistent sink frame, compressed temporal history, geometry-aligned spatial memory, and recent-frame conditioning. To reduce long-term drift, the model is trained with corrupted histories and prediction residuals collected from its own roll-outs. We further introduce a discrete autoregressive distillation formulation that combines distribution-matching distillation, self-forcing++, and consistency distillation, reducing inference from approximately 30 sampling steps to four steps per chunk. On iWorld-Bench, AlayaWorld achieves the best performance over long-horizon generation. Conceived as a full-stack, open-source, and long-term project, AlayaWorld is intended to provide an extensible foundation for future research on interactive video world models.

Asynchronous reinforcement learning improves throughput by decoupling rollout generation from optimization, but staleness is an inevitable byproduct compounded by policy lag, engine delays, and mixture-of-experts routing. From a trust-region perspective, this mismatch is critical: training-inference divergence governs approximation error in finite-horizon bounds, whereas PPO clipping only gates sampled outward updates, acting as a sampled surrogate rather than a full-policy constraint. As a result, high-staleness updates remain weakly controlled in the asynchronous regime where stale rollouts matter most. We introduce the Staleness-Adaptive Trust Region (SAT), which uses the detached sampled log-ratio as a practical staleness proxy, identifies high-mismatch tails within each batch via staleness-based kernel scaling, and contracts only the sign-selected endpoint of the nominal PPO interval. This preserves baseline behavior on ordinary tokens while enforcing more conservative updates on newly intercepted outward bands. We prove local interval containment and pointwise pessimism relative to PPO, showing how the adaptive rule reshapes update geometry under heterogeneous staleness. We evaluate SAT in a decoupled asynchronous RL setup built on Qwen3-30B-A3B-Base, using SGLang as the inference engine and Megatron for training. In this setting, SAT-GSPO w/ R3 achieves the best observed AIME24 avg@8, reaching 35.83 at lag 1 and 34.79 at lag 8, while SAT-GSPO reaches 34.17 at lag 1. Adaptive clipping and routing replay act as complementary stabilizers targeting mismatch tails and routing inconsistency, respectively. Overall, aligning clip intervals with staleness heterogeneity effectively stabilizes asynchronous RL.

Diffusion Language Models (DLMs) have recently emerged as a promising alternative to autoregressive models. Unlike standard diffusion-based approaches, DLMs are not explicitly conditioned on a timestep, raising a natural question: do these models internally represent denoising progress, and how is such information used downstream? In this work, we show that DLMs do in fact encode a latent representation related to the diffusion timestep within their residual streams. We find that this signal can be reliably extracted using probes across layers, indicating that denoising progress is decodable from internal activations. We further demonstrate that steering the model along a low-dimensional subspace associated with the inferred timestep allows us to systematically modulate its notion of denoising progress, leading to predictable changes in model confidence and entropy. Finally, we analyse the geometry of the identified representation, showing that it exhibits structured and interpretable properties in activation space, and shedding light on how such a signal is processed by these models.

LLM agent failures are difficult to debug because the step where an error surfaces is often not the one that caused it. Existing observability tools replay execution traces but provide little support for identifying the root cause or translating diagnosis into recovery. We present AgentDebugX, an open-source debugging framework that organizes debugging as a closed loop of Detect, Attribute, Recover, and Rerun. At its core, DeepDebug performs multi-turn root-cause diagnosis through global trajectory understanding, structure-guided investigation, and cross-examination. On the Who and When benchmark, DeepDebug achieves the best strict attribution accuracy among the evaluated methods on both tested open-weight backbones, reaching 28.8 percent exact agent-and-step accuracy on qwen3.5-9b versus 21.7 percent for the strongest single-pass baseline. On GAIA, DeepDebug repairs 13 of 73 failed tasks in a single rerun, compared with 4 to 6 for three decoupled self-correction baselines, improving overall accuracy from 55.8 percent to 63.6 percent. AgentDebugX exposes this workflow through a Python library, CLI, web console, and installable agentic skill, and provides an opt-in Error Hub for sharing scrubbed failure-diagnosis-repair bundles and reusing them as debugging memory.

Structural fidelity is essential to scientific methodology diagrams. To communicate research logic, these diagrams must faithfully render components, directional relations, and textual annotations. Since a single error, such as a reversed arrow or an unreadable equation, can invalidate the entire figure, structural fidelity is inherently conjunctive: correctness on one axis cannot compensate for failure on another. Current open-source models fail to satisfy this criterion. Supervised fine-tuning (SFT) learns plausible layouts but cannot reliably ensure structural correctness, while scalar reward-based post-training obscures which structural dimension has failed. To address this, we introduce SciForma, a framework for the structure faithful generation of scientific methodology diagrams. Specifically, SciForma decomposes diagram quality into three structural axes: Component, Arrow, and Text, guided by a structural inventory. Built on this foundation, we curate SciFormaData-700K for structured training and SciFormaBench-2K for logic-verified evaluation. To close the gap left by SFT, we develop Multi-Dimensional Conjunctive Preference Optimization (M-DPO), which enforces simultaneous correctness across all axes and adaptively routes gradients to the most deficient dimension in post-training. The same structural inventory also enables iterative editing at inference time to correct residual errors. This combination allows SciForma-9B to exceed all open-source baselines and GPT-Image-1.5 on both SciFormaBench-2K and AIBench, bringing open scientific diagram generation close to proprietary-level structural fidelity. Our code and data will be available at: https://github.com/microsoft/SciForma.

Efficient teamwork typically combines global coordination with parallel execution, a principle not yet fully reflected in unified Vision-Language Model (VLM)-based document parsers. Existing unified parsers process an entire page jointly but generate its output through a single token-by-token autoregressive trajectory, creating a sequential bottleneck that grows with document length. Such full-page sequential generation overlooks a key property of document parsing: layout must be analyzed globally, whereas block content can be parsed in parallel. Based on this observation, we introduce HPD-Parsing, which replaces full-page autoregressive generation with a Hierarchical Parallel Decoding paradigm. A main layout branch organizes the overall document structure and dynamically assigns block-level content decoding to concurrent branches, while progressive multi-token prediction (P-MTP) further reduces the decoding steps within each branch. Experiments on public benchmarks show that HPD-Parsing achieves 4,752 tokens per second, delivering 2.62times the throughput of the fastest existing document parsing model and 3.06times that of the vanilla autoregressive baseline, while maintaining competitive parsing accuracy. These results establish hierarchical parallel decoding as an effective alternative to full-page autoregressive generation, opening a new direction for efficient unified document parsing.

Evaluating the factuality of long-form generations has focused predominantly on precision, measuring whether the claims a model makes are correct. The dominant decompose-search-verify pipeline catches incorrect claims well but says little about whether a response contains all the information it should. Measuring factual completeness, the missing half of factuality, is harder: it requires enumerating the full set of facts a complete answer should contain, and these facts rarely form a flat list. They often involve open-ended sets where coverage is what matters, ordered processes, and relationships among facts that a list of independent boolean checks fails to capture. We introduce a two-level meta-rubric framework for evaluating open-ended generation, and instantiate it as Gamut (Grounded Assessment of Multimodal Factuality), a benchmark for factual completeness in long-form generation. The framework rests on a two-level rubric representation: a structured meta-rubric captures the organization and importance of the required content, which is then mechanically compiled into a flat checklist of binary, machine-gradable rubrics that an LLM judge scores reliably. We construct 1,813 questions grounded in real wearable imagery across 10 diverse domains, each paired with an evidence-backed rubric verified by expert human annotators. Because the framework is modality-agnostic, we also release a text-only variant. Evaluating 14 frontier and open-weight models, we find the benchmark genuinely challenging (best score 58.7% from Gemini 3.1 Pro), highly discriminative, and robust to the choice of judge.

Modern ASR models trained on heterogeneously annotated data treat transcription style (verbatim vs. intended) as an uncontrolled latent variable, causing measurable decoding instability, evaluation confounding (up to 60% of reported WER attributable to style mismatch), and unreliable word-level timing. We show that models already encode both styles; the challenge is controlled activation. Using coverage-aware decoder task tokens trained on parallel verbatim/intended transcript pairs, we raise German disfluency F1 from 10% to 79% zero-shot, despite English-only training. Full English-only fine-tuning surpasses all baselines in verbatim accuracy, disfluency detection, and intended-mode quality across both languages. We further introduce supervised cross-attention fine-tuning that improves word-level timestamps on disfluent speech beyond forced-alignment baselines. Finally, we propose verbatimize, a new task enabling scalable creation and enrichment of speech corpora with high-quality canonical verbatim transcriptions.

Optimizer state is the largest single line item in the memory budget of mixture-of-experts (MoE) training: on a 6.78B-parameter MoE language model, AdamW keeps 50.6 GB of first and second moments to update 12.6 GB of bfloat16 weights. We study SkewAdam, an optimizer built on the observation that the three parameter populations of an MoE - the dense backbone, the experts, and the router - differ enough in size and gradient statistics that they should not receive the same state. SkewAdam keeps float32 momentum plus a factored second moment for the backbone (5% of parameters), a factored second moment alone for the experts (95%), and an exact second moment for the router (<0.01%). The resulting state occupies 1.29 GB, 2.6% of AdamW's, and peak training memory falls from 81.4 GB to 31.3 GB, within the budget of a 40 GB accelerator. In a controlled comparison from identical initializations over 82M tokens, SkewAdam reaches validation perplexity 108.4, ahead of AdamW (126.8), Muon (120.2), and Lion (393.7), and settles router load balance to within 1% of its uniform floor. The allocation is not what earns that perplexity: a tier ablation matches it with twenty times the state, and Adafactor, which shares the factored estimator but drops momentum, plateaus 40 points behind. The tiers buy memory at no cost to accuracy; the accuracy comes from keeping momentum, which a uniform optimizer shares too. Sweeping the baselines' learning rates narrows but does not close the gap: the best tuned AdamW reaches 118.5, tuned Adafactor 139.7. Where optimizer state lives, these results suggest, matters at least as much as how much of it there is.

Reinforcement learning with verifiable rewards (RLVR) is rapidly advancing the reasoning capabilities of language models, yet the optimization layer that converts reward feedback into weight-space updates remains poorly understood. Building on our prior analysis (Zhu et al., 2025), we study this missing layer through the singular structure of model weights and identify spectral inheritance: RLVR can reuse the base model's weight spectra while acquiring new behavior through changes in the associated input and output singular frames. We operationalize spectral inheritance as Isospectral Optimization (ISO), an RLVR-native, fixed-spectrum optimization framework with complementary offline and online instantiations. Offline, ISO-Merger combines the frame changes of shared-base specialists into a single fixed-spectrum model, requiring no post-merge data, rollouts, gradient updates, or on-policy distillation (OPD). It recovers complementary specialist capabilities and achieves the strongest aggregate performance among the compared data-free merging methods. Online, ISO-Optimizer applies a chosen base optimizer, including AdamW and Muon, to the frame variables while keeping the base spectra fixed. Across reasoning and coding tasks ranging from 1.5B to 8B parameters, ISO-Optimizer improves accuracy in the reported runs and reaches matched scores with substantially fewer training steps. On Qwen3-8B-Base, AdamW reaches an aggregate accuracy of 0.495 after 270 training steps. ISO-AdamW reaches the same accuracy after only 100 training steps and improves further to 0.509 after 210 training steps. Together, ISO offers a concrete answer to RLVR's missing optimization layer: rather than inheriting pre-training optimization wholesale, design post-training around the structure of reward-driven adaptation: inherit the spectrum, optimize the frames.

Julia Love / Bloomberg : Google projects full-year capex will be $195B to $205B in 2026, after saying in April that it will spend as much as $190B; its Q2 capex hit $44.92B, up 100% YoY —  Alphabet Inc. projected full-year capital expenditures will be $195 billion to $205 billion in 2026, increasing …

Julia Love / Bloomberg : Google says its cloud computing unit had a $514B backlog of contracted work that hasn't been recorded as revenue yet, up from ~$460B held during Q1 —  Alphabet Inc. said its cloud computing unit had a $514 billion backlog of contracted work that hasn't been recorded as revenue yet …

Brody Ford / Bloomberg : ServiceNow reports Q2 subscription revenue up 24.5% YoY to $3.88B, increases full year subscription revenue growth estimate to 23%; NOW jumps 4%+ after hours —  ServiceNow Inc. reported better-than-expected quarterly sales and bookings, boosting Wall Street's hopes that the software maker's …

Nicola M White / Bloomberg : The SEC agrees to pay $150,000 in legal fees to end Coinbase's 2024 lawsuit seeking internal docs on the regulator's Biden-era crackdown on the crypto industry —  The US Securities and Exchange Commission agreed to pay $150,000 in legal fees to end Coinbase Global Inc.'s 2024 lawsuit seeking …

Emma Roth / The Verge : Google says Gemini now has 950M MAUs, up from 900M in May and 750M in February —  Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

Jordan Novet / CNBC : IBM reports Q2 revenue up 1% YoY to $17.2B, vs. $17.6B est., infrastructure revenue down 7% to $3.8B, Z mainframe revenue down 42%, and lowers its 2026 forecast —  IBM on Wednesday lowered its 2026 forecast and delivered thinner quarterly profits than analysts had projected …

CJ Haddad / CNBC : Reddit's stock closed down 8.32% after a report that the company was considering ending Google's access to its content for AI training; RDDT is down ~26% YTD —  Shares of Reddit slid 9% on Wednesday after the Wall Street Journal reported the company has discussed shutting off Google's access …

Todd Spangler / Variety : Alphabet reports YouTube's Q2 ad revenue rose 13% YoY to $11.06B, vs. $10.8B est., and Google's ad revenue reached $81.63B, up from $71.34B in Q2 2025 —  Video giant YouTube, one of the world's biggest entertainment businesses, continues to turn in double-digit increases in ad revenue.

Alphabet : Alphabet reports Q2 revenue up 24% YoY to $119.8B, above $116.9B est., Google Cloud revenue up 82% YoY to $24.8B, and Services revenues up 15% to $94.5B —  MOUNTAIN VIEW, Calif. - July 22, 2026 - Alphabet Inc. (NASDAQ: GOOG, GOOGL) today announced financial results for the quarter ended June 30, 2026.

Shirin Ghaffary / Bloomberg : OpenAI President Greg Brockman says Kimi K3 is a “pretty good model” but it is “too early” to tell whether Moonshot used distillation on OpenAI's models —  OpenAI President Greg Brockman acknowledged that Moonshot AI had developed a competitive new artificial intelligence model …

Yueqi Yang / The Information : Source: Stripe revenue jumped by a third to $6.8B in 2025, its fastest revenue growth since 2021, with free cash flow up 52% to $3.2B —  Stripe is coming off a banner year thanks to booming growth in the AI sector, where the company processes payments for big AI labs and small developers alike.

Sean O'Kane / TechCrunch : Travis Kalanick's robotics startup Atoms raised $1.7B led by a16z, with Uber and others participating; Ben Horowitz will join the startup's board —  Travis Kalanick's robotics company, Atoms, has raised $1.7 billion in a funding round led by Andreessen Horowitz.

New York Times : A Florida teenager drops his lawsuit accusing Meta of creating harmful social media features, following settlements with co-defendants TikTok, Snap, and YouTube —  For Meta, the move was a reprieve after it and YouTube were found guilty in another case in March of negligence and personal injury for their platforms' addictive features.

Sarah Perez / TechCrunch : Yope, a social network that has ~15M registered users for its “micro communities” and no algorithms, ads, or public content, raised $12.3M seed led by Northzone —  As social media has evolved from a place to connect with friends into large-scale entertainment platforms …

David Jeans / Reuters : Sources: Cathedral, launched by ex-DOGE staffers to use AI to expand US military cyber capabilities, raised $160M led by a16z and Sequoia at a $1.4B valuation —  A team of former DOGE employees have raised a major funding round for a startup that aims to use AI to expand U.S. military cyber capabilities …

我们在小学时学习的多位数乘法叫竖式乘法，其时间复杂度为 O(n²)，即位数越长，计算量随位数的平方增长。举例来说，两个两位数相乘，需要进行四次计算；两个三位数相乘，需要进行九次计算。位数越长，计算量会越来越惊人。那么 O(n²)是否是乘法的速度极限呢？苏联著名数学教授 Andrey Kolmogorov 在 1960 年的一次研讨会上讨论了这一猜想，仅仅一周之后，23 岁的学生 Anatoly Karatsuba 就给出了否定答案。他发现可以用简单快速的加法去替代费劲的乘法计算，而两个 n 位数相加的时间复杂度仅为 O(n)，加法只需要遍历数字一次，而乘法需要对 n 位数的每一位进行完整遍历。通过这一代数技巧，他将乘法的时间复杂度减少到 O(n^1.585)，比O(n²) 快得多。Karatsuba 算法的优势只有在数字较大时才会体现出来。Python 语言就使用了混合方法，当数字较小时使用小学乘法，当数字大于 630 位十进制数时改用 Karatsuba 的算法。2019 年数学家 David Harvey 和 Joris van der Hoeven 找到了一种比 Karatsuba 算法更快的方法，其时间复杂度为 O(n × log n)，但它相对于 Karatsuba 算法的优势只有在数非常非常大时才会体现。Harvey-van der Hoeven 算法被普遍认为是乘法的最快方法，但目前尚无正式证明。

为了维护数字主权，确保关键系统和数据掌控在欧洲自己手中，空客准备将其约 900 个应用从亚马逊 AWS 迁移到法国云服务商 Scaleway。AWS 与所有美国平台一样，并不能确保其客户的敏感数据不被政府索取。数字主权如今是真正的商业驱动力。美国曾经是欧洲人信任的国家，但如今它站在了和俄罗斯以及中国同等的位置上。

大众汽车的官方应用不支持第三方 Android 系统。大众汽车的客户如其手机运行 GrapheneOS、LineageOS 或 /e/OS 等没有预装 Google Play Services 的第三方 Android 系统，那么他们将无法通过手机运行官方应用查看汽车剩余续航里程、预约保养服务，控制充电和空调。对于这一问题，官方表示在调查，同时警告短时间内不会有变化。大众汽车解释说，它的应用使用了 Google 的 Play Integrity API，而该 API 只有预装了 Play Services 的 Android 设备会提供。GrapheneOS 最近抨击了 Google 在 Play Integrity API 功能上误导企业，称该 API 并没有真正强制要求设备安全或应用合法，只是装装样子。它留下了巨大的安全漏洞。它强制执行的是 Google 的商业利益。

Firefox 的扩展 Multi-Account Containers 备受欢迎，它允许用户在浏览器上同时登陆多个账号，账号之间通过容器彼此隔离。现在 Mozilla 正致力于将该扩展变成原生功能，刚刚释出的 Firefox 153 提供了容器的原生预览版，允许用户将工作、购物、个人和银行等不同在线活动隔离，同时确保每个容器内的 Cookie 和广告追踪信息完全隔离，用户在一个容器内的操作不会被其它容器看到。

LG USA 宣布将封禁内置住宅代理功能的智能电视应用。此前安全公司 Spur 的研究发现，LG webOS 应用商店逾 42% 的游戏和其它应用内置了住宅代理 SDK，也就是会出售用户的家用 IP 作为代理服务使用。三星 Tizen 应用商店也有逾四分之一应用内置了住宅代理 SDK。LG 高级副总裁 John Taylor 表示，该公司正与应用开发者合作移除应用中的住宅代理 SDK，未遵守规定的开发者其应用将会下架。未来的 LG 智能电视应用将禁用住宅代理。

在现代人类祖先走出非洲踏上欧亚大陆前，尼安德特人在此繁衍生息了逾 30 万年，但与现代人类相遇数千年后它就消失了。一种解释认为现代人类依靠更出色的智力战胜了尼安德特人，另一种解释认为双方融合了。发表在 PNAS 期刊上的一项新研究利用形变映射法（deformation mapping）重建了尼安德特人颅骨内表面，与现代汉人以及欧洲裔人群进行了对比。结果显示尼安德特人可能和现代人类一样聪明，而现代人类之间的差异比尼安德特人和现代人类之间的差异更大。新研究支持融合理论。

根据 Sensor Tower 的估计，2025 年苹果 App Store 上架的新应用数量增长 30% 达到约 60 万。今年上半年，新应用数量翻了一番达到约 56 万。虽然更多的应用理论上能为苹果带来更多的佣金，然而应用数量的大幅增长并没有带来下载量大幅增加，Sensor Tower 的数据显示去年 App Store 的下载量增长 3% 达到 354 亿次，今年上半年下载量增长 2% 达到 176 亿次。由于涌入了大量应用，苹果审核人员显然有点跟不上了。应用开发者在苹果开发者论坛上抱怨审核时间过长。分析师认为这一波 AI 辅助编程应用浪潮可能不会为苹果带来多少收入，因为此类应用通常是靠广告获利，不会提供内购。

中国科学家利用嫦娥六号从月球背面南极收集的月壤样品开展了系统的稀有气体同位素分析。研究发现，月球正面和背面接收到的太阳风存在系统性差异，且地球磁层在其中扮演了“调速器”的角色。研究团队瞄准了五种稀有气体——氦、氖、氩、氪、氙。这些气体具有化学惰性，是追溯太阳风注入过程和后期改造的忠实示踪剂。结果显示，嫦娥六号月壤的氖同位素组成呈现出极为独特的特征——其氖20、氖22比值平均为11.34±0.22，远低于所有已知的月球正面样品，极为接近理论上的强烈分馏太阳风端元（~11.2）。这表明月球背面经历了更为极端的分馏过程，使其富集重同位素，仅用传统的溅射、扩散或侵蚀模型已无法完全解释，暗示背面可能存在更为复杂的同位素分馏机制或一个未被发现的低氖同位素端元。为什么同一颗月球的两面会接收到不同能量的太阳风？研究团队论证认为，是地球磁层在“调速”。当月球围绕地球运行并穿越地球磁鞘时，原本约每秒400公里的“正常”太阳风会被显著减速至约每秒200公里。

根据斯坦福经济学家 Jacob Light 的研究，2025-2026 学年美国计算机科学专业入学人数近二十年来首次下降，但这种下滑是否是一种长期趋势还有待观察。数据显示，过去十年美国计算机和信息科学专业的学士学位授予数量增长了一倍多，从 2014 年的约 5.6 万增至 2024 年的约 12.2 万。2025 年秋季入学数据显示，四年制大学计算机与信息科学及支持服务专业的本科生入学人数同比下降 8.1%，从 2024 年的约 659,700人 降至 2025 年的 606,100 人，但仍然高于 2022 年的 574,333人，表明人数下降是在经历了多年快速增长之后出现的。Light 认为人数下滑的原因可能有多种，基于大模型的辅助编程工具的流行，IT 就业市场疲软等等。

安妮·弗兰克去世前在荷兰躲避纳粹期间写的日记于 1947 年在荷兰出版，她的日记在欧洲大部分国家都已经进入了公有领域，但在荷兰她的部分日记的版权保护期要到 2037 年才结束。为了尊重这一版权保护的地区差异，托管安妮日记的出版商屏蔽了荷兰 IP 访问网站。然而拥有安妮日记荷兰版权的安妮·弗兰克基金会提起了诉讼，声称 VPN 可以绕过地理位置限制，因此受版权保护的安妮日记被传播给了荷兰居民。欧盟法院驳回了这一论点，裁决 VPN 是合法技术工具，读者通过 VPN 绕过地理位置限制，出版商不应因此承担侵权责任。

由于 AI 生成安全报告大量涌入，GNOME 项目宣布改变安全报告处理方式。首先是 GNOME 项目将安全漏洞披露时间从行业标准的 90 天缩短至 30 天，原因是大部分 GNOME 安全问题会在一到三周内修复，或者干脆不修。其二禁止 AI 生成的安全报告，由于今天的大部分安全报告包含了 AI 生成的内容，这些安全报告将会直接关闭。

为了训练其 AI 模型 Claude，Anthropic 实施了名为“巴拿马计划”（Project Panama）的行动：大量购买实体图书，拆开书脊、扫描书页，之后将图书残骸送去回收公司。Anthropic 为此投入了数千万美元，聘请了二十年前参与 Google Books 项目的 Google 高管。此外  Anthropic 还从影子图书馆下载了海量的盗版电子书。对于 Anthropic 的行为，法官裁决使用受版权保护图书训练 AI 模型属于合理使用，但其图书盗版行为并不合法。图书出版商以及作者提起了集体诉讼，Anthropic 去年与他们达成了和解，将向图书作者和出版商赔偿 15 亿美元。本周一，美国地区法官 Araceli Martinez-Olguin 批准了这一和解协议。和解协议涵盖逾 48.2 万册图书，其中 91% 已被作者或出版商认领，每本书预计将获得大约 3000 美元的赔偿。

去年美国在全球大规模征收关税，任天堂据此提高了 Switch 2 等相关设备和配件的售价。今年 2 月美国最高法院裁决征收关税违法，任天堂随后起诉美国政府要求退还关税。美国玩家则跟着提起了一项拟议中的集体诉讼，认为任天堂应该将美国关税退款退给消费者。任天堂周一请求法庭驳回该诉讼，称它无法律义务将美国关税退款退给消费者。任天堂的观点十分简单，任天堂或其零售合作伙伴设定了最终价格，而买家是自愿支付的。

FBI 逮捕了一名 21 岁的佛罗里达居民 Zyaire Dontaevious Zamarion Wilkins，他涉嫌与同伙通过在 Steam 游戏中植入恶意程序去窃取玩家的加密货币。该团伙在近两年时间内通过在至少八款 Steam 游戏中植入恶意程序，感染了约 8000 台电脑，在 2024 年 5 月至 2026 年 2 月间从约 80 个加密货币钱包中窃取了价值至少 22 万美元的加密货币。被植入恶意程序的 Steam 游戏包括 BlockBlasters、Dashverse、Lunara 和 PirateFi。其中仅 BlockBlasters 一款游戏就从 261 至 478 名受害者手中窃取了价值约 15 万美元的加密货币，受害者包括了 Twitch 主播 RastalandTV，他被盗走了 3.2 万美元，他正接受癌症治疗，这笔钱是观看者捐赠的。

在生成式 AI 时代，流行公民科学数据库如 iNaturalist 和 Macaulay Library 也免不了被 AI slop 入侵。研究人员在《自然》上报告，他们在记录野生动物物种的流行公民数据库内发现了数百张 AI 生成的虚假图像。问题的真实规模尚不清楚，因为有许多假图可能未被发现。文章作者 Alexander Lees 博士称他在 Facebook 上看到的野生动物照片基本上都是 AI 生成的。他指出公民科学数据库里彻头彻尾的假图还比较罕见，问题主要是照片上传者会用 AI 美化下图像，结果生成式 AI 给图像添加了不存在的内容。iNaturalist 上逾 6.1 亿张图像只有 1400 张被标记为可能使用 AI。iNaturalist 社区支持总监 Tony Iwane 认为大多数假图并非是恶意的，他同时呼吁用户保持警惕，因为信息的准确性至关重要。