AI 情报文章归档

OpenAI 推出 Rosalind Academic Plan，面向美国高校研究实验室免费提供最多 30 个 Business Standard 席位，最长 12 个月，不含 API 访问。

腾讯混元团队开源 Hy Memory 2.0，包含记忆写入模型 Hy-Memory-30B-A3B（30B 总参数、约 3B 激活的 MoE）、可编程的 hy-memory SDK 和面向终端用户的桌面工具 MemHub，用小型专用模型替代通用大模型做记忆写入。

你的 logo，以高级动态呈现。 高端品牌视觉，无需制作成本 👇 https://x.com/PixVerse4Biz/status/2108549484812902712

Anthropic 宣布让非营利机构 METR 的六名员工获得公司硬件、办公空间和内部工具访问权，以审计其 AI 模型与安全实践，权限大致与公司内部风险团队相当。

Traffic Creator is one of my companies. Customers use it to send configurable, automated website...

受成龙启发，动作指导兼表演负责人 Themo 来到 @DreamLabLA，学习如何在部分 AI 动作捕捉特技中加强控制。 我们以一次策划会议开启前期制作，随后立即投入工作，捕捉他的动作，再通过视频到视频转换，将其变成全新的风格和世界。 由 Luma 制作。

Replit CEO Amjad Masad 在 TechCrunch Disrupt 上提出「自动驾驶公司」愿景：AI 智能体承担公司大部分基础工作，人类保留判断力与品味。他以 Replit 客服流程为例，工单先由智能体结合 GitHub issues、代码库和 Linear 工单完成初步分诊，人工决定修复后一键部署另一个智能体执行。

This article is submitted for the **Hacktoberfest Open-Source AI Challenge: Week 2 (Side...

英伟达RTX Spark电脑今晚8点开售，线下首发体验会人山人海，128G+1T的华硕ProArt创16酷卖37999元在线上瞬间被抢光。作者实测Qwen3.8-27B可跑到30 Token/秒，Qwen image 2.1约1分钟一张图，本地2B翻译模型可6路并发实现近0延迟实时翻译。

Researchers at Chalmers University of Technology have demonstrated an AI-enabled laboratory system...

American Forge 现已开放申请。 我们很高兴与 @ndstudio 合作，共同推进其让美国政府软件现代化的使命。https://x.com/ndstudio/status/2111075526399889572

10月16日，百度智能云与西安高新区管委会在百度智能云超级智能体大会·西安站完成战略签约，规划“1+8+N”合作框架，建设1个百度智能云西北总部基地、8个人工智能产业中心并拓展N个产业应用场景。首期落地OPC产业中心、具身智能创新中心、产教融合中心三大项目，并联合9所西安高校发起西北高校人工智能产业联盟。

百度智能云在西安超级智能体大会上宣布百度一见 6.0 开放内测，主打用自然语言和 Golden-Set 样本让人人都能造视觉 AI，并同场推出 AI 原生具身智能相机。百度搭子用户规模增长 9 倍，交付物需求增长 4.3 倍，专家套件使用量增长 7.3 倍。2026 年前三季度百度智能云大模型中标金额位列云厂商第一。

10月16日，百度智能云与远元集团在超级智能体大会 BAIDU AGENT SUMMIT（西安站）签署“远元集团品质服务智能巡检项目”，双方将结合大模型与视觉 AI 技术，用百度一见对门店接待、环境、操作、养护等环节做智能识别与标准校验。远元集团已拥有近万家门店，年服务客户超 1.12亿人次。

I’ve been looking at how AI coding agents are changing the way developers work, and one thing is...

Coding agents are sold per seat and metered by the token. Sub-agents, long contexts and retries hide the cost. Measure one task before you delegate ten.

它运行在与其他所有 OpenRouter 图像模型相同的 Images API 上，所以把你的编辑迁移到 Krea 3 只需改一下模型 ID。 详见：https://openrouter.ai/krea/k3-edit

Krea 3 不会改动你没有要求它触碰的部分：围绕产品构建新场景，标签依然清晰；或者给它角色参考，就能在整套图中保持同一批角色。

定价：1K 每张 $0.03，2K 每张 $0.06 参考图 1K 每张 $0.02，2K 每张 $0.07 thinking_mode：fast 为默认 当编辑需要对提示词进行更多推理时，设为 high（+$0.02）

OpenRouter 宣布 Krea 3 图像编辑模型上线，入口为 https://openrouter.ai/krea/k3-edit。该模型来自 @krea_ai，接受 1–4 张参考图加提示词并返回原生 2K 分辨率的编辑结果，多数编辑端到端耗时不足 8 秒。

Modal 工程博客介绍其为 VM Sandboxes 打造的 ImageFS-blk 块设备文件系统后端。

Dify 发文讲解 AI Agent 与 Workflow 的区别：Agent 由模型在运行时从获批工具中选择下一步，Workflow 则由搭建者预先定义步骤与分支。文章建议流程固定可重复时用 Workflow，需逐次判断时用 Agent，并可在 Workflow 中嵌入 Agent 节点处理单个灵活环节。

NVIDIA 宣布 RTX Spark 笔记本正式开售，来自 ASUS、Dell、HP、Lenovo、Microsoft 和 MSI，起价 $2,599。

Large language models (LLMs) are increasingly used to automate AI research, including language-model (LM) development, offering a potential path toward recursive self-improvement. However, turning experimental feedback into a learned abili…

Low-rank adaptation (LoRA) has become the standard method for parameter-efficient fine-tuning of large pretrained models, yet theoretical explanations for why low-rank updates suffice remain incomplete. Most existing analyses rely on the N…

Tool-augmented language-model agents rely on interface text: tool descriptions, argument schemas, and the candidate tool list shown at inference time. These interfaces often change after deployment even when the underlying tool semantics r…

Modern deep learning systems increasingly rely on high-dimensional vector operations, in which large input and parameter vectors are repeatedly stored, transferred, and processed. Although substantial research has focused on reducing neura…

Pretrained models are often reused across downstream tasks, motivating the question of whether this transfer learning procedure can provide a substantial performance improvement, namely, in terms of the scaling law. We address this questio…

The High Resolution Imaging Science Experiment (HiRISE) onboard the Mars Reconnaissance Orbiter (MRO) acquires imagery of Mars at 25--30 cm/pixel, supporting studies of martian geology, climate history, and landing-site characterization. I…

Real-time inference on small devices needs low activation memory without extra latency, which makes operator order the most valuable memory lever: it is free in time. We show that the usual way of deploying a better order, handing it back …

原来 le Chonk 还懂火箭科学。 Mistral Large 4 在 Ulam 的私有基准 STEMBench 的工程和天文学项目中得分不错：https://www.ulam.ai/stembench

Often, \emph{eg} in healthcare and predictive maintenance, what matters is which of several competing causes will strike first, and with what probability: survival analysis with competing risks. Corresponding deep learning methods are deve…

Enterprise data analysis requires agents to connect analytical requests with knowledge distributed across databases, business documents, and existing example tasks. Without an explicit representation of this knowledge, an agent may produce…

The joint reconstruction of the initial pressure and spatially varying sound speed in photoacoustic tomography from boundary measurements is considered. A regularized data-driven framework is proposed in which the unknown fields are repres…

Long-term agent memory requires recovering answer-relevant evidence dispersed across turns and sessions. Existing systems increasingly use structured memories and adaptive retrieval, yet often pass retrieved or expanded units directly to t…

Large language models (LLMs) can generate fluent outputs that lack source support or conflict with local operational constraints. We introduce Self-Reflective Chain-of-Thought (SRCoT), a modular architecture that organizes generation as a …

Native language identification (NLI) infers a non-native speaker's first language (L1) from their second language (L2) production. We investigate a token probability approach that scores each text sample as a continuation of a prompt prefi…

Liquid AI 发布决策模型 d1-ultra，以 65 分登顶 Artificial Analysis Decision Index v1.0，领先 Jev 1.14 的 57 分。

Multimodal large language model (MLLM)-based mobile agents are rapidly evolving from conversational assistants into autonomous systems that perceive, reason over, and act on smartphone environments. This capability also exposes a new secur…

A key challenge in on-policy distillation (OPD) is extracting effective supervision from teacher knowledge. Supervision from token probabilities may underutilize internal knowledge, while direct hidden-state matching forces the student to …