AI 情报文章归档

Video generation models can now depict remarkably realistic motion. Yet it is unclear how to use this motion beyond pixels: to animate a specific 3D asset and replay the animation in its 3D scene. For instance, a video model can show a mon…

Agentic artificial intelligence (AI) systems hold immense potential for automating complex workflows in materials discovery and design, but extending them to domain-intensive mechanistic investigations remains challenging. Here, we present…

Dexterous multi-finger hands hold promise to unlock diverse skills like functional grasps, in-hand reorientation, and dynamic handover that the human-designed world assumes, yet existing learning recipes force a trade-off. Goal-reaching re…

Motivated by the human ability to integrate experience across language, perception, and action, we introduce UniWALM, a unified world-action-language model that learns video, action, and language in a single shared representation, where ea…

Warehouse robot fleets continuously observe operational conditions, but preserving these observations as reusable semantic memory incurs substantial computation, communication, and storage costs. For example, explaining recurring aisle blo…

Planning under uncertain control execution noise requires reliable feedback policies across diverse environments. We introduce Planning Under Noise (PUN), a neural operator that learns noise-aware value fields for planning in unseen geomet…

Spatial intelligence has recently been assessed through generative tasks such as spatially grounded image editing, yet language-guided generation of spatially consistent views from partial multi-view observations remains underexplored. We …

Amazon 与独立非营利机构 AVERI 合作，对其内部安全评估数据集开展第三方独立审查，重点检查测试本身的质量与严谨性。审查确认了主题覆盖全面、测试方式多样、污染风险低等优势，同时提出扩大语言与风格多样性、增加多步对抗性提示词、保持评估新鲜度等改进建议。Amazon 据此更新测试集、引入分级难度框架、建立季度审查机制并扩大多语言覆盖。

UniqueHuman 发布健康智能体 Niq，面向偏头痛、纤维肌痛和 IBS 等慢性病患者，首发聚焦偏头痛。Niq 整合可穿戴设备数据、病历、化验结果和症状记录，在偏头痛发作前预警，并自行完成预约医生、处理处方和保险等后续事务。

OpenAI 宣布 API 现在可选择 12 小时提示词缓存。共享的指令、工具和上下文可以缓存更长时间，即使请求间隔数小时也能按更低的缓存输入价格计费，适合像 dots 这类全天复用上下文的持久化 Agent。

“电力不足”的说法过于简化了。要以极快速度上线大规模算力，每个环节都很难。 如果有两倍的时间，这会相对容易，但并没有。

Multi-token prediction (MTP) can accelerate language-model decoding by drafting several tokens and verifying them in a single target-model pass. However, the best speculation horizon changes with context: longer drafts offer greater amorti…

Anthropic 的 Claude Opus 5.5 和 Claude Sonnet 5.5 已在 Google Antigravity 2.0 桌面端、CLI 及 VS Code、JetBrains、Visual Studio 的 IDE 扩展中正式可用。

Graph generative models commonly use graph-specific Transformer architectures to jointly update node and edge features. However, these architectures do not directly benefit from recent hardware-aware attention implementations, limiting the…

Picture-description dementia classification uses a shared visual scene, but a transcript-level prediction alone does not identify the scene evidence underlying its inputs. NarraSpace links words and speech times to named picture elements a…

据 Bloomberg 报道，前 Google CEO Eric Schmidt 持有 Anthropic 约 1% 股份，源于 2021 年向其 Series A 轮投入数千万美元，当时公司估值为 4.61 亿美元，该持股约为 CEO Dario Amodei 2% 持股的一半。

Modern distributed microservice deployments generate massive volumes of heterogeneous, multi-format observability logs across physically and logically partitioned infrastructure. Isolating root causes during cascading failures remains an a…

Perplexity Decider 在 Vals AI 的 Financial ClaimBench 上达到帕累托最优

Mathematical reasoning remains underexplored for low-resource languages, including Dialectal Arabic. To bridge this gap, we introduce the DAR-MGSM (Dialectal Arabic Multilingual Grade School Math) Benchmark, the first manually curated benc…

Current evaluation of LLM-generated debate largely focuses on argumentative quality, leaving the value structure of generated arguments underexamined. We introduce a diagnostic metric framework for comparing how values are expressed and or…

TS2Vec is one of the strongest self-supervised methods for time series classification, introduced in 2022. It is designed to be trained anew on every dataset: self-supervised pre-training followed by linear probing. Since then, the field h…

Alphabet CEO Sundar Pichai 在 Big Technology Podcast 访谈中驳斥 Google 放弃前沿 AI、只做轻量 Flash 模型的说法，称前沿研发是公司的 ground zero。

Computer-use agents are limited by the interfaces humans give them. An agent may type a spreadsheet cell by cell when a single line of code would do, or print all two thousand lines of a file through a shell to find one function. Both the …

Neural implicit representations provide memory-efficient signed distance field (SDF) parameterizations for large-scale LiDAR scenes, but the geometry recovered from sparse and often noisy inputs carries jitter and artifacts. Classical geom…

RTL repair benchmarks should measure whether a model can fix broken RTL given a specification. To this end, we construct and hash-bind RTLRepairDataset, 5,784 (broken, fixed) training pairs; RTLRepairBench, 459 held-out cases with 374 synt…

My Downloads folder has a talent for becoming a digital storage room. PDFs, screenshots, source...

fal 发布与 Black Forest Labs 合作后训练的视频模型 FLUX 3 Max，在第三方独立人类偏好评测中获 1139 Elo，与 Seedance 2.5 并列图生视频质量第一梯队。该模型约 12 秒生成一段 10 秒 720/768p 视频，未来三周降价 75%，后续还将继续后训练改进。

Neuro-spectral architectures (NeuSA) are a class of physics-informed neural networks (PINNs) that have proven effective for solving time-dependent PDEs in the frequency domain. However, training and inference often rely on costly numerical…

Arena 更新 Agent Arena Pareto 前沿，OpenAI 的 GPT-6 Luna 和 GPT-6.1 Sol 在不同推理档位下占据 11 个前沿位置中的 6 个，加上 GPT-6 Astra (Max) 共 7 席。

Camera-trap imagery presents substantial variation in animal appearance, illumination, background clutter, and visual degradation. Masked autoencoders (MAEs) offer a label-free framework for modeling such data through reconstruction of ran…

We prove that every $\operatorname{ReLU}$ network can be exactly simulated by a $\operatorname{ReLU}$ network with two hidden layers. We present two proofs: one that is based on fillings in homological algebra and a direct combinatorial pr…

法院命令 Meta 披露批准为训练 Llama 1 和 2 而通过种子下载盗版图书的高管姓名。Meta 辩称早期数据获取无正式决策、属于集体讨论且没有单一决策者，法官 Hixon 驳回这些说法，要求 Meta 说明所有决策者以及未用于训练的盗版作品的去向。

Shadeform 发布 Flex VMs 和 Shadeform Spot，把闲置的裸金属 GPU 服务器转化为可租用的云算力，数小时内即可上架。作者认为关键在于无需构建 API，只需一个 IP 和 SSH 密钥即可完成接入，这消除了闲置算力难以触达买家的集成障碍。

看起来 AI 的网络风险可能更多是生态层面的问题，而非开放权重模型特有的问题。如果我们能在这里找到合适的平衡点，我认为有很多人愿意在 RSI 和即将到来的 AI 动态面前支持更大的开放性。

Pragmatic Engineer 访谈 27 家公司的工程负责人，并结合 CodeRabbit 和 Gitar 的汇总数据，调查软件工程师的 AI 工具支出与 ROI。

Perplexity CEO Aravind Srinivas 转引 Engadget 报道称，Perplexity AI 工作流现在可以在 Mac 上完全本地运行，要求配备至少 36GB 内存。

ChatGPT 推出 Locked Chats 功能，把敏感对话从常规历史记录和搜索中移入受保护空间。用户在网页端和桌面端可用 PIN 解锁，支持的移动设备上可用 PIN、Face ID、Touch ID 或设备解锁；锁定会话中的新消息不会更新常规对话使用的记忆。该功能即日起向 Free、Go、Plus、Pro、Business、Enterprise 和 Edu 套餐推出。

Jim Fan 宣布开源 Flora，一组用于解析物理世界、把原始视频转成训练信号的开放模型，称其以 125 倍更低的成本达到 Astra 级性能。Flora 已用于其内部数据引擎，在 500 块 GPU 上每天标注 20 万小时视频，提供带时间戳的高质量字幕、用于质量控制和 RL 后训练的密集进度奖励，以及用于现实基准测试的 rubrics 评分。

Databricks 在 Agent Bricks 平台推出 Agent Memory Beta，通过存储关键事实、偏好和工作流提升智能体质量并降低成本和延迟，底层由 Lakebase 提供、基于 Databricks 的 MemCurrent 记忆研究。

OpenRouter 宣布 Mistral 的 Voxtral 3 已上线其平台。Mistral 此次发布了两款新的语音转文本模型和一款新的文本转语音模型，支持 50+ 种语言，将用于驱动 Mistral 新的语音智能体 API。