AI 情报文章归档

Context.dev 提供一个 API 让 AI 智能体读取实时网页：传入 URL 返回干净文本，也可搜索、爬取站点、查询公司信息或监控页面变化。它把抓取、搜索、公司数据和页面监控整合到一个 key 和一张账单下，避免智能体管线分别付费给 3-4 家供应商并写胶水代码。该 API 已获 800+ 客户使用，公司完成由 General Catalyst 领投的 900 万美元融资。

哇！

Agentic AI is everywhere right now, and for a student just getting started it can look impossible....

就在你的灵感时刻随便聊聊 💭

智能体战争对所有国家都将是大事。AI将有能力执行震网级别的网络行动，下一场军备竞赛将在自我改进的智能体之间展开。然而目前我们几乎没有针对这一点的评测。我认为没有任何国家为即将到来的规模和速度做好了准备

Grok Bot 又近了一步。 把它设为 iPhone 的操作按钮，按下即可与 Bot 对话。

专注推理的芯片公司 Tenstorrent 以 62 亿美元估值完成 15 亿美元融资，估值较 2024 年底的 26 亿美元翻倍以上，本轮由 Trace Capital 和 Innovation X 领投，Cadence Ventures、Eclipse、Fidelity、SanDisk 和 T. Rowe Price 参投。

Claude 宣布 Dashboards 和 Motion 进入 beta。Dashboards 可连接数据平台或 CRM，生成随数据变化实时更新的仪表盘，面向所有付费计划；Motion 把报告或图表变成动画讲解短片，以代码方式生成、无视频模型参与，可修改文字、数字和节奏后导出 MP4，面向 Team 和 Enterprise 计划。

fal 发布与 @bfl_ai 合作的最新后训练视频模型 FLUX 3 Max，在第三方 image-to-video 质量评测中并列第一，生成 10 秒 720/768p 视频约需 12 秒。未来三周该模型 75% 折扣，标注为同质量级别最便宜的前沿视频模型。

Every new chat with my coding agent started the same way. Ten minutes of me typing some version of...

Google has rolled out Gemini 4 Argon, a model positioned for complex, multi-step work across the...

Artificial Analysis 推出 AA-PersonalityBench，从 30 项人格特质和 21 项沟通特征两个维度衡量语言模型差异，基于 900 道精选问题、由 LLM 评审团两两对比打分。

Gartner 预测 AI 提供商 2026 年将投入近 1 万亿美元基建，2027 至 2030 年再投 6 万亿美元，并警告若维持当前资本开支节奏，到 2030 年底超 3.5 万亿美元支出或因收入无法覆盖成本而减值。

MiniMax 发布推理模型 M3.1 Flash，在 Artificial Analysis Intelligence Index 得 42 分，比 M3 高 13 分，单任务成本约 $0.22，API 定价 $0.15/百万输入 token、$0.60/百万输出 token，支持 100 万 token 上下文和图文输入，计划 10 月 23 日开放权重。

新的前沿即将到来。

MIT、波士顿儿童医院与 Inria 的研究者开发并训练了深度学习模型 E(3)-Pose，可在胎儿头部运动时辅助成像，生成供放射科医生查看的高分辨率三维图像。该模型把胎儿头部的左右对称性等物理原理嵌入网络架构，而非仅从数据中学习物理规律，从而更准确地识别大脑朝向。相关论文已在 2026 年欧洲计算机视觉会议（ECCV）上展示，目前模型主要基于第三孕期胎儿训练。

Arena 宣布 DeepSeek-V4.1-Flash-20261017 (Max) 成为 Agent Arena 第一开源模型，总排名第 11，基于超 7.5K 真实会话净提升 +5.1%。其中位任务成本 $0.20，约为 GPT-6.1 Sol 提升幅度的 37%、成本的 27%；Confirmed Success 提升 +11.5%，且无 Tool Hallucination 问题。

上周你用 @bot 版 Grok 完成了哪些事？

Microsoft AI’s MAI-Transcribe-2 achieved the lowest transcription error rate in 18 of 21 languages in...

Meta 介绍其 Muse 基础模型（含 Muse Wonder）在俄亥俄州 New Albany 的吉瓦级超级集群 Prometheus 上的训练情况。该集群拥有数十万 GPU，预训练 goodput 达 85% 以上，强化学习达 90% 以上，Muse Wonder 训练期间平均约 47K GPU 小时才发生一次中断，并附技术介绍链接。

A user opens an email assistant and asks it to summarise their unread messages. Somewhere in their...

Open the network tab while you use most AI writing tools. Watch what happens as you type an email, a...

这是我们正在用 Codex 打造的东西： https://x.com/i/broadcasts/1RKZzBELBbXKB

COOKED：第 1 期 🧑‍🍳 https://x.com/i/broadcasts/1RKZzBELBbXKB

NVIDIA 在德国 Jülich 超算中心的 JUPITER 上完成 104 量子比特容错量子计算模拟，这是迄今规模最大的容错量子处理器完整模拟。该模拟验证了 magic state cultivation 方案，所需物理量子比特比此前方法少 90%，并首次在真实硬件噪声下评估其表现。模拟基于 GH200 超芯片与 cuQuantum、CUDA-Q 运行，可让研究者在造硬件前预测设计性能。

Modern text-to-image (T2I) systems use internet-scale language pretraining for text-conditioned image generation. We ask whether language pretraining is necessary for strong T2I generation. We introduce Halcyon, a family of early-fusion mu…

Multimodal large language models (MLLMs) have made rapid progress on visual reasoning, yet many tasks require evidence that is difficult to obtain through internal reasoning alone: our controlled study on 22 benchmarks shows that simply in…

While LLM agents can generate visually plausible CAD assemblies, these designs often fail to fulfill their intended physical functions. We study functional CAD design, a setting where agents interpret underspecified requests to construct a…

Multimodal reinforcement learning with verifiable rewards (RLVR) provides reliable task-level supervision, yet offers limited guidance for localizing errors in visual reasoning. On-policy self-distillation (OPSD) complements this supervisi…

Reinforcement learning enables mobile robots to make fast and scalable navigation decisions in constrained human-shared environments. However, directly extending single-robot approaches to multi-robot settings often fails to ensure safety …

In this paper, we systematically investigate the mechanisms, scaling behavior, and generalization of physical Recursive self-improvement. We introduce PhysRSI-eval to quantify physical RSI through a five-axis metric: First-Attempt Performa…

Long-horizon agents must learn heterogeneous capabilities from interleaved trajectories. Updating shared parameters to learn new capabilities can degrade previously acquired ones, resulting in catastrophic forgetting. As a trajectory unfol…

A seemingly reasonable push in Sokoban can make the puzzle unsolvable. To avoid such mistakes, an agent has to anticipate the consequences of its actions before execution, especially when feedback from the real environment is sparse or act…

The goal of a world model is to capture the dynamics of an environment. This has the great appeal that it allows agents to plan and simulate a course of actions with the help of a world model. For these uses to be reliable, however, the le…

Systems that combine several large language models (LLMs) often use one context for many questions. When the context is passed between models as text, the receiving model processes all of its tokens again for every question. Avoiding this …

Bilevel optimization provides a framework for machine learning tasks such as hyperparameter optimization, data reweighting, and meta-learning. Existing Newton-based bilevel methods typically rely on the classical hypergradient formula, whi…

LLM agents that compute the right answer can still answer wrongly. Agent harnesses increasingly keep long tool outputs out of the context window, saving them to files that the agent searches or reads on demand. We ask where this costs accu…

We propose KronLSTM, a linear attention architecture based on the Kronecker Delta Rule, a novel update rule for associative matrix memories. KronLSTM extends the xLSTM family with a memory-revision mechanism that corrects targeted associat…

Machine learning is widely regarded as a promising application in which quantum computing could offer a computational advantage. However, general quantum machine learning models with provable guarantees on performance, trainability, and ef…

I built Sorcerer, a free, open-source macOS app (SwiftUI, macOS 15+, MIT) that simulates a community...