AI Coding Tools

A tiny crew of monsters for your Claude Code agents

A terminal built for Claude Code w/ one-click session resume

Make existing Codex and Claude Code sessions multiplayer

Wild Pokémon appear while you wait for Claude Code

A Claude Code alternative for people who avoid the terminal

Claude Code, Codex & more on your phone

Free voice coding for Claude Code, Codex and more

Turn YouTube into your personal learning platform

See what your Claude Code sessions actually cost

Claude & Cursor usage in your macOS menu bar

Give Claude Code missions to spawn a team of agents

24/7 personalized health companion you can text

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.

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.

Personalized LLMs with persistent memory are increasingly deployed, yet the faithfulness of their user models remains unexamined. We study over-inference (OI): the phenomenon where LLMs fabricate user attributes beyond what evidence supports. We introduce MirageBench, comprising 150 personas balanced across stereotypical, counter-stereotypical, and neutral profiles, 6 personalization tasks spanning an ``imagination gradient'', a four-way faithfulness taxonomy operationalized by an independent judge (validated against a blind human annotator on 400 claims: Cohen's kappa = 0.863 four-class, kappa = 0.900 binary), and a leaderboard of 12 models across 7 families on 143616 judged claims. We find that over-inference is pervasive: every one of the 12 models over-infers 35%--49% of its claims (cross-model mean 41.6%; claim-weighted 41.8%), with no model in this evaluation escaping it. Most strikingly, we surface a Self-Monitoring Inversion: at the model-selection level, models' self-assessed OI is negatively rank-correlated with their judge-measured OI (rho = -0.60, p = 0.044; exploratory, wide bootstrap CI [-0.90, +0.06], n = 12). The models that report the least over-inference tend to be flagged as fabricating the most, so self-reported confidence is a misleading signal for comparing models, even though within a single model self-audit still ranks that model's own claims moderately well (AUROC 0.58--0.83). We further show that OI is task-dependent (27%--59%) and that, in a multi-turn pilot, inferred attributes accumulate approximately linearly with little revision. MirageBench positions external verification, rather than model self-report, as a more reliable foundation for trustworthy personalization.

Agent self-evolution updates an agent's persistent state from prior experience and reuses it to solve related tasks more effectively. Evaluating self-evolution is difficult: existing benchmarks provide limited coverage of economically valuable task domains, do not always design training and test tasks such that test-time gains can be attributed to training experience, and remain vulnerable to data contamination. We present GDPevo, an evolution-native benchmark grounded in GDP-related enterprise workflows, together with the fully automated data pipeline that generates it. Its core mechanism, rule hybridization, decomposes each enterprise workflow into atomic business rules, distributes subsets of these rules across training tasks, and recombines them in held-out test tasks so that test-time gains are attributable. GDPevo spans CRM, ERP, finance, healthcare, legal, and data-centric workflows. Its V1 release contains 120 tasks in 12 groups, with five training and five held-out test tasks per group. Full automation enables the pipeline to expand the suite to 240 tasks in 24 groups (V2) within two days, providing a practical response to contamination. Using GDPevo, we evaluate four agents, each comprising a harness and a model, under four supervision types. Self-evolution consistently improves held-out accuracy by up to 16.44 percentage points. But the best evolved agents remain far below the fully informed oracle ceiling of 91.6%, indicating that the self-evolution ability of current agents remains far from fully realized. We publicly release the pipeline, benchmark, and full evaluation results at https://github.com/Prism-Shadow/GDPevo.

Learning generalizable robot manipulation policies requires large-scale and diverse demonstration data. Egocentric human manipulation videos offer rich scene and task diversity, and prior work has shown that retargeting and rendering such videos into robot-format data can yield effective per-task policies at small scale. However, whether this approach can provide pretraining benefits for vision-language-action models at scale remains unexplored. We present Ego2Robot, a scalable pipeline that converts egocentric human manipulation videos into robot training data through action retargeting, robot-arm visual synthesis, and multi-level quality curation. Ego2Robot supports both curated datasets and in-the-wild videos, producing 18,561 hours of robot training data spanning 15 robot morphologies, making it the largest ego-to-robot dataset to date. To evaluate generalization, we extend RoboTwin2.0 with disentangled perturbation axes covering visual appearance, scene layout, embodiment morphology, and task semantics. Experiments show that joint pretraining on Ego2Robot-synthesized and robot data consistently improves out-of-distribution generalization across multiple perturbation types, with benefits validated on real-robot deployment. Project page: https://www-ye.github.io/ego2robot_blog/

Industrial recommenders increasingly adopt the pretrain-then-transfer paradigm, yet behavioral distribution drift raises two questions: what to learn from behavior sequences, and how to transfer the learned knowledge while the pretrained model is continually refreshed. To resolve them, we propose Knowledge-Geometry Decoupling (KGD). For what to learn, conventional next-token prediction treats adjacency as dependency and may encode spurious transitions across unrelated sessions. We introduce Behavioral Multi-Token Prediction (BMTP) to retain only collaboratively or semantically related future items as supervision, yielding cleaner and more transferable behavioral knowledge. For how to transfer, pretrained knowledge and task-specific geometry impose conflicting optimization demands on shared parameters. To handle it, KGD assigns them to separate parameter sets: a refreshable encoder owns behavioral knowledge, while a task learner reads contextualized encoder states through read-only cross-attention and writes task-specific geometry through Anchored Calibration Residual (ACR) orthogonal to the pretrained embedding. The decoupled ownership enables continual knowledge refresh without task-gradient interference or invalidating downstream adaptation. KGD improves over strong pretrain-transfer baselines by 4-12% on eight public benchmarks and sustains its advantage over a 90-day production stream where baselines show no gains. KGD has been fully deployed in Shopee. In a live A/B test on Shopee Homepage Search, it increases GMV per user by 1.75% and advertising revenue by 1.53%, demonstrating its high practical value. We provide the core implementation of KGD at https://github.com/FuCongResearchSquad/KGD4REC.

Large language model agents have shown strong potential in complex interactive tasks, yet their reinforcement learning (RL) is often hindered by sparse rewards, as a long multi-turn trajectory may receive only a single outcome-level signal. On-policy self-distillation (OPSD) provides dense token-level supervision from a privileged teacher, but the teacher may not be reliable at every position. Existing methods commonly rely on isolated token-level discrepancies, which can be sensitive to noise, or assign a shared step-level weight that may overlook positional variation. We propose Persistent Consistency Self-Distillation (PCSD), which derives token-level distillation weights from the local persistence of teacher-favoring signals. PCSD combines adaptive windows with exponentially decayed aggregation to capture persistent relative teacher support, applies trend-aware modulation to attenuate locally declining support, and produces continuous weights through sigmoid gating. The resulting objective is jointly optimized with GRPO, combining dense teacher guidance with sparse environmental feedback. Without inference-time skills, PCSD achieves the best ALFWorld Overall results among all baselines on both backbones, exceeding GRPO by 15.6 and 13.3 points and SDAR by 6.2 and 5.5 points, while remaining competitive on WebShop and gaining 15.8 points over GRPO on unseen ALFWorld split.

On-policy (self) distillation (OPSD) is increasingly adopted for language-model post-training. It strengthens the teacher with privileged information but can induce a privilege illusion: the student learns privilege-dependent behavior it cannot reproduce from its inference-time context, yet behaves as if the training-time privileged information remained available, ultimately degrading performance. In this paper, we identify information asymmetry between the privileged teacher and the student at inference as the root cause of this failure in OPSD. To resolve this asymmetry, we propose Dual-Anchored Policy Distillation (DAPD), a unified framework with two levels of anchoring. Dual-Path Anchoring (DPA) introduces a self-conditioned bridge and aligns reference and rollout behavior along two matched-information paths, preventing privilege-dependent behavior from being transferred to the inference-time student. Dual-Source Anchoring (DSA) applies these paths in both reference-to-rollout and rollout-to-reference directions, reducing reliance on privileged reference guidance while preserving correctness supervision. Extensive experiments show that DAPD significantly alleviates privilege illusion, outperforming OPSD on Qwen3-4B by +2.00 points on average across tasks. Notably, its gains persist across scales, reaching +2.69 at 4B and +2.78 at 32B.

Controllable video generation models are increasingly being developed as world models. Accordingly, evaluating them in this role extends beyond the apparent appearance of generated videos to the inherent reactivity of the worlds they depict: the ability to infer from the scene state how the world should react and to generate plausible consequences not explicitly described in the input. Yet existing benchmarks mainly assess visual quality or explicit instruction fulfillment by checking whether requested actions and interaction outcomes are realized, leaving inherent reactivity underexamined. We introduce WorldExam, a hierarchical diagnostic benchmark spanning four levels: Visual Quality, Control Adherence, Spatial Consistency, and World Reactivity. It comprises 1,474 cases across eight dedicated tasks and supports unified evaluation of camera-, action-, and language-driven model paradigms. The World Reactivity level evaluates scene-conditioned reactions and goal-directed behaviors beyond what is explicitly specified in the input. Evaluation of 20 representative models reveals a clear capability split. Camera-driven models excel at camera control, but their interfaces do not support dynamic interaction; action-driven models control subjects more precisely but often leave the world unresponsive; and language-driven models perform better on interaction but follow complex controls less faithfully. No model combines broad task coverage with consistently strong performance, showing that high visual quality and explicit instruction fulfillment do not guarantee inherent reactivity.

AI-assisted coding increasingly translates informal user intent into executable software, yet coding requests often contain ambiguities that recur in user-specific ways across tasks and sessions. Existing disambiguation methods typically address each ambiguous request in isolation within the current coding session, often through eliciting additional clarification. However, whether resolved session history from the same user can serve as memory for resolving recurring personalized ambiguity in a newly opened session remains underexplored. We formulate personalized ambiguity adaptation as a new task: given a user's previously resolved coding sessions and a new ambiguous request, an assistant should identify the recurring ambiguity pattern, produce the intended executable solution, and minimize clarification. To benchmark this task, we introduce CAPA, which characterizes personalized coding ambiguity through six mechanisms and injects these mechanisms into unambiguous executable tasks using a controlled three-stage generation pipeline. CAPA contains 600 coding sessions across 60 balanced user--ambiguity cells, including 300 held-out evaluation sessions. We evaluate 12 recent LLMs under no-history and same-user-history conditions using executable success, first-turn success, and turns-to-completion. Our analyses examine task difficulty, user identity, and memory-based history use, and we further propose same-user history gating as a lightweight inference-time method. CAPA provides a foundation for developing long-term coding assistants that better align generated code with user intent while reducing repeated clarification.

Chemistry literature synthesis often requires assembling specific findings scattered across many publications, yet existing literature-search systems primarily return ranked document lists. As a result, scientists and AI agents need to locate relevant information, verify their provenance, and assemble cross-paper answers manually. We present AskChem, a claim-centered infrastructure for cross-paper chemistry search. AskChem changes the unit of retrieval from the paper to the provenance-carrying claim: each paper is converted into atomic, typed claims, each grounded by a source DOI and a verbatim quote or an explicit evidence locator. Over this shared claim store, AskChem exposes complementary structures for search and synthesis: a stabilized faceted taxonomy for hierarchical retrieval and browsing, an evidence graph linking claims through relations, and an exploratory living taxonomy that situates indexed papers under scientific principles. AskChem currently indexes 2.4M claims from 147K papers and provides a web interface, as well as REST, SDK, and MCP access for AI agents. On AskChem-Bench, grounding a GPT-5.5 reader in AskChem yields 100% resolvable DOIs, compared with 88.3% without retrieval, and the highest citation density among five tested systems. AskChem is live at https://askchem.org.

Grace Kay / The Information : Sources: Cursor told staff on Thursday that SpaceX could complete its $60B acquisition as soon as next week and the Cursor brand name will likely be phased out —  Cursor told staff at an all-hands meeting on Thursday that SpaceX could complete its $60 billion acquisition of the coding startup …

The Information : Sources: legal AI startup Harvey is in talks to raise $500M+ at a $15.5B valuation, up from $11B in March, and is generating $350M+ in annualized revenue —  Legal AI startup Harvey is in talks to raise at least $500 million at a $15.5 billion valuation including the investment, a 40% premium to its last valuation five months ago.

Mark Gurman / Bloomberg : Sources: OpenAI's new device, slated for 2027, is a hockey puck-sized smart speaker with moving parts that help give it personality, likely costing $300+ —  A highly anticipated new device from OpenAI will have a unique look, complete with moving parts that help give it personality …

Devlin Barrett / New York Times : Court docs: an FBI agent has been charged with amassing ~$1M worth of cryptocurrency, largely by making unauthorized withdrawals from a criminal target overseas —  The agent confessed to making unauthorized withdrawals from a criminal target overseas, according to court papers.

Mariah Timms / Wall Street Journal : Researchers used AI-assisted code to undetectably tamper with data from computerized scans of physical DNA evidence produced by widely used crime-lab machines —  Researchers found that widely used lab machines produced digital DNA files that are vulnerable to tampering

New York Times : Inside Larry Ellison's debt-fueled push to turn Oracle into an AI juggernaut by aligning with Trump, backing Project Stargate, and partnering with OpenAI —  the first full day of the second Trump administration — Larry Ellison woke up in his 33-bedroom, 34-bathroom oceanfront mansion in Florida …

Kit Norton / Barron's Online : Roblox shares close down ~27%, its worst-ever daily percentage decline, after forecasting falling bookings and slowing revenue growth; RBLX is down 40%+ YTD —  Choose Barron's as a preferred source of financial news  —  Key Points  —  Shares of Roblox tumbled Friday as Wall Street responded harshly …

sroorz 写道： 在初代 Apple Silicon 发布六年后，CrossOver 宣布推出原生 Apple Silicon 支持，并发布了首个预览版。该软件旨在在 macOS 上运行 Windows 游戏和软件。 苹果于 2020 年推出 Apple Silicon，并内置 Rosetta 2 转译层以兼容旧有 Intel 软件。CrossOver 和其他基于 Wine 的模拟器一直依赖 Rosetta 2 在 Apple Silicon 上运行 x86 Windows 游戏。然而，苹果计划逐步淘汰 Rosetta 2，并于 macOS 28 完全停止支持。CrossOver 的原生 Apple Silicon 支持正是为了应对这一变化，实现独立于 Rosetta 2 的 x86 Windows 游戏支持。 CrossOver 产品经理 Meredith 在发布公告中回顾了 CodeWeavers 开发团队过去四年如何通过改进 Wine 和相关开源组件、工具链实现这一目标。她也指出，这是首个预览版，仍存在一些问题，目前仅推荐测试使用。此外，由于 ARM64 版 D3DMetal 尚未就绪，目前仅能通过 DXMT ARM64 运行 DirectX 11 游戏。但所有这些问题都将在正式版 CrossOver 27 中得到解决。 Wine 是一个开源项目，是除虚拟机外，在非 Windows 操作系统上运行 Windows 游戏的唯一解决方案。基于 Wine 衍生出的项目中最著名的是 Steam Proton。CodeWeavers 是 Wine 项目的长期贡献者和 Proton 开发者，其员工贡献了 Wine 90% 以上的代码，CrossOver 是其基于 Wine 开发的商业版。

EcommerceBytes 是一家创办于 1999 年专注于电商相关新闻和评论的网站，由 Ina 和 David Steiner 夫妇创办。2019 年 4 月，Ina 发表文章说时任 eBay CEO Devin Wenig 的薪水 152 倍于普通员工。eBay 高管长期以来就对这个网站的文章很不满，首席传媒官 Steve Wymer 将这篇文章转发给了 CEO，附言要给点 Ina 颜色看。Wenig 在一个月后发短信同意了计划。Wymer 随后让安全部门主管 Jim Baugh 负责此事。eBay 的员工对 Steiner 夫妇展开了一系列的跟踪和骚扰行动，包括送活蟑螂、送花圈、一本关于如何走出丧偶阴影的书，以及血淋淋的猪面具。Steiner 夫妇于 2021 年提起了诉讼，双方刚刚达成了和解，eBay 将支付 5570 万美元的赔偿和慈善捐款。其中赔偿金为 4615 万美元，eBay 还将向多家慈善机构捐赠 600 万美元，前 CEO 将以 Ina Steiner 的名义向一家致力于保护宪法第一修正案权利的慈善机构捐赠 100 万美元。七名前 eBay 员工和合同工认罪和判刑，但两位前高管 Wymer 和 Wenig 没有受到刑事指控。