AI 情报文章归档

每次演示前后的繁琐工作几乎消失了： 准备：30 分钟 → 5 笔记：15 分钟 → 2 CRM：30 分钟 → 7 每次通话大约省回一小时。每位销售代表现在每天能多接约 2 通电话，准备程度却丝毫不减。

我们 5 人的销售团队在构建 Rasp 之前不堪重负，Rasp 是一个使用 OpenRouter 的 Ori 打造的 AI 销售智能体。 每条入站线索都会经过调研，许多线索在 60 秒内就能收到回复，我们每月节省约 600 小时。 成本：约 $30/天，而且每周都在变得更便宜。 https://openrouter.ai/blog/case-studies/how-an-ai-sales-agent-saved-our-sales-team-600-hours-a-month

This week I wrote a Photoshop file by hand. 390 bytes, two gray layers, the top one set to Soft...

General-purpose robots must perform a wide range of tasks from agile locomotion to dexterous manipulation. While sim-to-real reinforcement learning (RL) has proven to be a useful tool for this goal, current RL pipelines depend on engineeri…

METR's 50\% time horizon measures the human completion time of software tasks that an AI solves with 50\% probability, allowing AI capabilities to be expressed in interpretable units. On 228 tasks and 26 AIs, we recompute the time horizons…

Text-conditioned motion generators produce trackable whole-body motion, but they have no notion of scene-dependent safety: the same action may target an object or a person. Existing safeguards either inspect the prompt, require labeled mot…

In 2026, cybersecurity evaluations involving OpenAI, Anthropic, and Google agents reached real systems outside their authorized test scope. The paths were different. OpenAI agents exploited research infrastructure, coordinated across runs,…

We propose BrickBench, a benchmark for agentic text-conditioned LEGO-set design. Given a prompt, an agent is tasked with producing an assembly that not only satisfies semantic and design criteria, but that can also be physically built. To …

In this work, we show that a single Transformer block, applied recurrently, can match the accuracy of a full-depth vision encoder at comparable inference FLOPs without intermediate feature distillation. reViT restores depth-specific transf…

Bifurcations are ubiquitous in physical systems, from structural buckling to fluid and climate dynamics, yet they remain largely unexplored in deep learning. At a symmetry-breaking bifurcation, a single input admits multiple equally valid …

USA Today Co. 及其旗下多家地方报纸起诉 OpenAI，指控其未经许可复制“数十万篇”文章用于训练 AI 模型，并寻求超过 2.5 亿美元赔偿。诉讼称 OpenAI 的商业成功建立在大规模版权侵权之上，The Tennessean、Indy Star、The Columbus Dispatch、The Oklahoman 等地方媒体也列入起诉；OpenAI 未立即回应置评请求。

FT 报道 OpenAI 年化营收已接近 $50B，比投资者流传的 $70B 低约 $20B。差异源于统计口径与推算方式：Anthropic 把经由 AWS 和 Google Cloud 销售的金额计入营收。

Recent incidents have highlighted the challenge of monitoring LLM agents and the danger of models deceiving people. We show that white-box deception detection via probes can be scaled up to frontier monitoring settings by collecting the la…

Quantizing AdamW's optimizer states reduces persistent storage, but quantization errors propagate through the moment recurrences and perturb subsequent adaptive updates. We redesign 4-bit optimizer-state quantization for AdamW from the per…

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass ...

Density ratios quantify distribution shift from a probability-mass point of view, whereas displacement fields describe, from a dynamical point of view, how one distribution is transported onto another. Although both offer complementary ins…

AI agents can now conduct real-world cyberattacks, scale up capabilities with the number of agents, and collectively pursue misaligned goals to obtain rewards. Together, these factors raise the risk of a population explosion of misaligned …

Teaching a humanoid to follow instructions with its whole body runs into two obstacles. Its action space is large and tightly coupled: legs, arms, and fingers must move together while the robot keeps its balance, which makes joint-level ac…

Safe reinforcement learning commonly places safety and task performance in the same policy objective, where they can introduce competing updates. Safety filters separate them at action execution, but classical designs require an analytic s…

SpaceXAI 宣布加入管理 Omarchy 的 Omacom Foundation，成为创始企业赞助人，并捐赠价值 150 万美元的 Grok tokens。

Building a quantum model involves a tradeoff: how complex the circuit should be, and how much training data it needs. Caro et al. show that models with fewer trainable gates need less training data to generalize well. Q-FLAIR shows that a …

Streaming Video Large Language Models (VLMs) enable continuous video understanding, yet existing benchmarks focus on low-dynamic scenarios. Under bounded context budgets, models must balance temporal history, spatial resolution, and tempor…

A generalist robot should not only perform diverse tasks but also improve through experience, turning what it learns during execution into capabilities that later tasks can reuse. Robot agents that act through code can already repair progr…

Dense correspondence matching has historically been bounded by simplifying spatio-temporal priors, such as smooth motion and rigid geometry. While effective for classical tasks, these assumptions break down in image editing and reference-g…

Reinforcement learning in safety-critical domains requires maximizing task performance while strictly adhering to safety constraints. Existing safe reinforcement learning paradigms typically force a trade-off: they either require a priori …

Multimodal large language models (MLLMs) struggle with spatial, embodied, physical, and temporal reasoning. We hypothesize that these failures reflect a shared deficit in visual transition reasoning, and test whether this capability can se…

Generating realistic human-object interactions (HOI) in complex 3D scenes requires two complementary capabilities: reasoning about interaction feasibility in the environment and synthesizing realistic human-object motion. However, supervis…

SemiAnalysis 长文分析中国对 AI 安全的实际做法，认为北京口头上安全立场领先，实际执行速度优先，不会配合美方放慢前沿节奏。

Step 5 Preview 现已在 @cline 免费开放一周。 我们面向智能体与专业工作的旗舰模型。 启动 Cline，向我们展示你的作品！

Ethan Mollick 转发社区进展并评论称，在 OpenAI 发布一系列证明引发数学界争议之际，其中至少部分证明似乎已引发大范围协作者社区的极快速迭代式推进。其引用的进展将 κ 从 2⁻¹⁸² 收紧到 2⁻¹⁵，较此前结果提升 50 万倍，较 OpenAI 原始结果提升 2^167 倍。

The Rules Are Hidden. The Timeline Isn't. What happens when you put something inside a...

Odyssey 发布基础世界模型 Odyssey-3，Pro 版报告了 Physics-IQ Verified video-to-video 最高分 66.1。

Anthropic 于 10 月 8 日推出 Claude Motion，可将提示词变成由文本、图表、图形和图像构成的短动画讲解视频，Claude 会编写动画代码，结果可在编辑器中调整、通过对话修改或下载为 MP4。

这份文件将成为大学课堂上讲授这一时期的指定阅读材料，短短几段里信息量很大…… https://www.ahmath.org/statements

Odyssey 发布其最强世界基础模型 Odyssey-3，并向公众开放免费研究预览。模型可从提示词构建环境并实时预测随人物移动或事件触发的变化；Odyssey-3 Pro 在 Physics-IQ Verified video-to-video（best-of-8 采样）上得分 66.1，为已报告最高结果。

我们应该更有野心

Perplexity CEO Aravind Srinivas 宣布，Perplexity Computer 为 OpenAI 数学手稿发布中的全部 722 篇预印本创建了解说视频，共 93 小时，并整理成 YouTube 播放列表，可在 https://722.video 观看。

A trail voice memo becomes structured sightings and walks, parsed by open-weight gemma3:4b on your own machine — no dashboard, no cloud, no typing.

推荐阅读！ ExploreNet: Learning Where to Explore in Diffusion GRPO

很高兴能通过 @SpaceXAI 支持开源！我们正在引入更多项目，敬请期待！