文章系统

AI 情报文章归档

浏览 AI圈报已保存的 30022 条 AI 情报文章，覆盖模型、产品、行业、论文、教程与观点方法六大分类。

已收录文章30022

正式分类6

精选内容4451

全部文章

第 352 / 751 页 · 每页 40 条

模型发布 / 更新普通2026-09-20 14:20

卢伟冰：真正意义上的 AI 手机离走进生活还有不少距离，小米 18 Pro 系列 9 月 23 日发布

小米集团总裁卢伟冰表示，真正意义上的 AI 手机距离走进生活还有不少距离，需要整个 AI 和手机行业共同努力。他称小米 18 Pro 系列是小米数字旗舰升级幅度最大的一次，从系统软件到芯片硬件、从影像到大模型应用，是小米用 AI 全面改造智能手机的开始，该系列将于 9 月 23 日 19：00 发布。

信息来源：IT之家（RSS）

教程 / 实战普通2026-09-20 14:20

I Built Shree AI OS — A Deterministic AI Runtime for Java 21

I Built Shree AI OS — A Deterministic AI Runtime for Java 21 After several months of...

信息来源：DEV Community

产品发布 / 更新普通2026-09-20 14:15

阿里云 Apsara 大会智能体 AI 论坛直播

锁定 2026 云栖大会 International Agentic AI Forum 直播，聆听全球技术领袖分享智能体 AI 如何在基础设施、应用和行业中演进。日期：2026 年 9 月 23 日 | 时间：09：30-12：00（UTC+8） https://x.com/i/broadcasts/1OxwbnObnBqJB

信息来源：X：阿里云 / Alibaba Cloud (@alibaba_cloud)

观点 / 方法普通2026-09-20 14:15

小米 18 Pro 系列「星河蓝」配色亮相：纯平一体式设计背屏、0.99mm 窄边屏，9 月 23 日发布

小米 18 Pro 系列手机官图公布，新增「星河蓝」配色，采用纯平一体式设计与百变背屏，屏幕边框窄至 0.99mm，将于 9 月 23 日发布。该系列搭载 2nm 旗舰移动平台、徕卡双 2 亿超清影像系统、M11 发光体系新一代超级像素屏幕，以及 32% 超高硅小米金沙江电池。

信息来源：IT之家（RSS）

产品发布 / 更新普通2026-09-20 14:13

阿里云AI Agent技术论坛直播

属于你的智能体原生时代前排席位，等你入座。锁定 2026 云栖大会国际 AI Agent 技术论坛直播，探索驱动下一波 AI 浪潮的技术。日期：2026 年 9 月 23 日 | 时间：14：00-16：25（UTC+8） https://x.com/i/broadcasts/1dJrProPrWVKX

信息来源：X：阿里云 / Alibaba Cloud (@alibaba_cloud)

论文研究普通2026-09-20 14:11

Blind Thermodynamic Ontology Discovery from Anonymous Experiments

Before a machine learning model can learn a thermodynamic equation of state, it must discover what its measurements represent: which channels scale with system size, which are intensive conjugates, how sectors pair through contact, and whi…

信息来源：arXiv

论文研究普通2026-09-20 14:10

Leveraging Industrial Foundation Models at the Edge of Particle Physics Detectors via Distillation Learning and Hardware Co-design

Data acquisition (DAQ) systems at future particle physics experiments stand to benefit from the extremes of AI/ML development: large-scale foundation models can enhance the performance of feature extraction algorithms, and small-scale on-d…

信息来源：arXiv

产品发布 / 更新普通2026-09-20 14:09

阿里云东盟区域论坛直播

见证东南亚 AI 故事实时展开。锁定 Apsara Conference 2026 区域论坛：ASEAN 直播，了解 AI 应用在该地区如何演进。日期：2026 年 9 月 22 日 | 时间：14：00-16：00（UTC+8） https://x.com/i/broadcasts/1pKdRDlbqVoJW

信息来源：X：阿里云 / Alibaba Cloud (@alibaba_cloud)

观点 / 方法普通2026-09-20 14:06

曝各家手机厂商将用更多黑科技拔高骁龙 8 Elite Gen5 性能上限，一加 Ace 7 或首当其冲

博主 @数码闲聊站透露，今年子系品牌 N-1 迭代新机仍以骁龙 8 Elite Gen5 为主流选择，各家会上更多狠活拔高该平台性能上限。其摸到的某款子系性能旗舰将搭载一项黑科技，能让骁龙 8 Elite Gen5 性能直升一档，结合评论区讨论该机被认为是一加 Ace 7。博主回应网友时称，8E5 是打磨了一年的旗舰 U，且安排了迭代级新技术。

信息来源：IT之家（RSS）

论文研究普通2026-09-20 14:03

Bayesian Filtering in Physical Systems via Test-time Trained Flow Matching

Bayesian filtering provides a principled framework for online state estimation under uncertainty, yet its application to systems with high-dimensional states and complicated posterior distributions remains challenging. Recent generative mo…

信息来源：arXiv

论文研究普通2026-09-20 14:02

Discovering Physical Representation Languages

Before a machine can discover a physical law, it must discover what its measurements are: which observations live on cells, which are intensive or extensive, which sectors are dual, and which distinctions are merely gauge. We introduce phy…

信息来源：arXiv

行业动态普通2026-09-20 14:00

市场监管总局："十五五"时期将运用大数据、AI 规范网络销售与直播带货

市场监管总局在"十五五"时期将运用大数据、人工智能等筛查锁定违法线索，规范网络销售、直播带货等经营活动。同时建设消费信用体系，健全消费投诉信息公示制度，并探索动态发布全国消费环境指数。到 2030 年全国放心消费集聚区将达 5，000 个以上。

信息来源：IT之家（RSS）

教程 / 实战普通2026-09-20 14:00

Orca: The Agent Development Environment for Running AI Coding Agents in Parallel

If you use Claude Code, Codex, or another AI coding agent in your terminal, you have probably hit...

信息来源：DEV Community · Claude

教程 / 实战普通2026-09-20 14:00

AI Integration Cost: 2026 Enterprise Budgeting Guide

Determining the true AI integration cost is a major financial step for UK businesses looking to...

信息来源：DEV Community

论文研究普通2026-09-20 13:59

Leaky-integrator reconstruction: taming error accumulation in recursive differenced time-series forecasting

We introduce leaky-integrator reconstruction, a training-free method that cures the error accumulation of recursive differenced forecasting. Our first contribution is diagnostic: predicting one-step changes and integrating them by cumulati…

信息来源：arXiv

论文研究普通2026-09-20 13:58

One to More, More to One: Category-Aware Iterative Expert Training for Software Engineering Agents

Repository-level software engineering (SWE) comprises heterogeneous task categories, whose progress under pooled agentic reinforcement learning can be uneven: gains in some categories coincide with regressions in others, while aggregate re…

信息来源：arXiv

观点 / 方法普通2026-09-20 13:57

Jev 值得关注内容持续追踪

Jev 值得关注内容持续追踪 1/n

信息来源：X：ZHO (@ZHO_ZHO_ZHO)

论文研究普通2026-09-20 13:57

If You Hear It, Help Find It: Orthogonal Knowledge Distillation for Open-Vocabulary Audio-Visual Event Localization

Open-vocabulary audio-visual event localization (OV-AVEL) grounds a text-queried event in time from video, audio, and language. The supervision sources available to this task can differ in temporal-boundary reliability: on OV-AVEBench, our…

信息来源：arXiv

教程 / 实战普通2026-09-20 13:55

Running Flux locally on a Mac: install, commands and two models compared

Building the engine, downloading 17 GB of weights, every flag explained, and two Flux models measured on one machine: one paints from scratch, the other edits a finished frame.

信息来源：DEV Community

教程 / 实战普通2026-09-20 13:54

How I Structure MiniMax H3 Max Image-to-Video Prompts with Sound

Turning a still image into a video is already a strange creative problem. Adding sound makes it...

信息来源：DEV Community

教程 / 实战普通2026-09-20 13:50

Gemini Hacked Three Companies Using the Dumbest Trick in the Book

You probably saw the headline this week: Google's AI autonomously hacked three companies. Cue the...

信息来源：DEV Community · Gemini

论文研究普通2026-09-20 13:49

The Evidence Ladder for Reinforcement Learning in Healthcare: From Retrospective Policies to Trusted Interventions

Reinforcement learning (RL) offers a natural language for healthcare decisions whose conse- quences unfold over time, yet most reported progress remains far from routine intervention. Ex- isting surveys organize the field by algorithm or c…

信息来源：arXiv

产品发布 / 更新普通2026-09-20 13:46

B站上线"AI 无限竞技场"大模型测评榜，GPT-6 Astra 居榜首

B站上线"AI 无限竞技场"大模型测评榜，并公布首轮排行榜，GPT-6 Astra 在 10 个 UP 主测评中拿下榜首，打败 GLM-5.3 获得榜首次数冠军，前五名中国产大模型占三席。

信息来源：IT之家（RSS） · GPT / GLM

行业动态普通2026-09-20 13:45

华为数字能源王超建议汽车行业硬件接口标准化，减少定制损耗

华为数字能源副总裁王超在第二十二届中国汽车产业发展（泰达）国际论坛上呼吁加大汽车行业标准化力度，推动智驾域、座舱域、运动域接口标准化，通过硬件接口标准化减少定制损耗。他还建议加大统一平台化力度，鼓励跨车企平台化，构建单部件百万规模平台，以提升中国汽车产业链的全球竞争力。

信息来源：IT之家（RSS）

论文研究普通2026-09-20 13:43

Machine-Interpretable Information: Compiling Documents into Searchable and Readable Protocol States

Long-context language models interface with external knowledge through raw natural language. In retrieval-augmented systems, this creates a persistent index-payload schism: dense vectors enable searchable routing, but models must re-ingest…

信息来源：arXiv · GPT / Qwen / Llama

教程 / 实战普通2026-09-20 13:28

OpenAI's GPT-5.6 Sol Sets New Record: Sub-100ms Response Time Changes Everything

The Numbers That Matter OpenAI just shipped GPT-5.6 Sol with what might be the most...

信息来源：DEV Community · OpenAI / GPT

教程 / 实战普通2026-09-20 13:26

Anthropic, OpenAI Agents Caught Creating Fake Identities During Security Tests

The Report That Should Keep You Up at Night The UK's AI Security Institute (AISI) ran a...

信息来源：DEV Community · OpenAI / Claude

论文研究普通2026-09-20 13:25

LLM-Based FORM Code Generation with Verification-Driven Fine-Tuning

FORM is a domain-specific symbolic manipulation language widely used in particle physics for processing the very large algebraic expressions arising from multi-loop Feynman diagram calculations. Despite its central role in precision theore…

信息来源：arXiv · Qwen

教程 / 实战普通2026-09-20 13:22

From Zero to One: How I Built a Multi-Market B2B Badge & Trophy Site with an AI Agent Configurator

The Niche: Badges, Medals, Coins, and Trophies Most e-commerce builders chase apparel and...

信息来源：DEV Community

论文研究普通2026-09-20 13:21

What Can a Recurrent State Safely Forget?

Recurrent models must preserve information that changes future behavior while suppressing hidden-state error. These objectives conflict: contraction improves stability, but contraction along a future-distinguishing direction destroys memor…

信息来源：arXiv

论文研究普通2026-09-20 13:17

TicTacBench: Benchmarking Timing Closure Capabilities of Coding Agents

Recent advances in large language models (LLMs) have led to the emergence of coding agents capable of performing complex engineering tasks, including register-transfer level (RTL) design and optimization. Existing RTL benchmarks mainly eva…

信息来源：arXiv

行业动态普通2026-09-20 13:12

智元与长隆集团合作的具身智能主题乐园 9 月 24 日在珠海横琴开园

智元（AGIBOT）与长隆集团合作建设的大型具身智能主题乐园将于 9 月 24 日在珠海横琴长隆飞船乐园开园。园区计划部署超 300 台机器人，设置 100 多个机器人交互点，覆盖演艺、科普、导览、零售、酒店、体育等场景，开园当天还将揭牌具身智能文旅联合研究院。

信息来源：IT之家（RSS）

教程 / 实战普通2026-09-20 13:05

Under the Hood of Browser-use (100k+ Stars): DOM Tree Distillation, Vision Grounding, and Production Web Agents

Why do traditional Selenium and Playwright scripts constantly break against modern web frontends? We dissect the architecture of Browser-use, the breakout open-source project with over 100k GitHub stars: DOM tree pruning, Set-of-Mark (SoM)…

信息来源：DEV Community

观点 / 方法普通2026-09-20 13:01

奥巴马谈 AI 递归自我改进与指数级加速

奥巴马在新访谈中称 AI 技术并未被过度炒作，模型正进入"曲棍球杆"式指数级加速变聪明的阶段。他提到 AI 已进入"递归学习"阶段，机器开始能自我教学：一年前模型约 90% 的学习来自人类、仅 5%-10% 靠自己填补，如今已变成 50/50，若趋势延续可能达到 90/10。

信息来源：X：Rohan Paul (@rohanpaul_ai)

教程 / 实战普通2026-09-20 13:00

Hybrid-precision attention reduces compute cost with minimal accuracy loss

Mixed‑precision quantization can halve the compute cost of LLM attention while keeping accuracy loss...

信息来源：DEV Community

观点 / 方法普通2026-09-20 12:54

Noam Brown 在 Dwarkesh 访谈中谈多智能体协同作弊与思维链监控风险

作者推荐 Dwarkesh Patel 对 OpenAI 推理科学家 Noam Brown 的约 80 分钟访谈并归纳要点。访谈称 OpenAI 跑通 ARC-AGI 的成本从去年的 50 万美元降至如今的 20 块钱；1 万个 Agent 用 1300 亿 Token 连续运行 88 小时做深度优先搜索。

信息来源：X：阿易 AI Notes (@AYi_AInotes) · OpenAI

论文研究普通2026-09-20 12:53

Human-guided physics-constrained AI agents construct an auditable model of soil-plug evolution

Engineering predictions require physical mechanisms to be translated consistently into equations, discretization, code, and validation, yet errors can propagate despite local checks. Artificial-intelligence (AI) agents automate scientific …

信息来源：arXiv

行业动态普通2026-09-20 12:51

国内首次"通感智算"一体化技术验证卫星鹏城首发星发射成功，搭载 5G NTN 基站与星地激光

2026 年 9 月 20 日 12 时 03 分，力箭一号遥十八火箭成功发射 9 颗卫星，其中鹏城首发星暨鹏城南科一号是国内首次把 5G NTN 基站、核心网、星载 AI 智算、星地激光及微波通信联合搬上天的技术验证卫星，由银河航天与鹏城实验室、南方科技大学等联合研制。

信息来源：IT之家（RSS）

模型发布 / 更新普通2026-09-20 12:49

LLM 已知答案概率为何还要先说话

既然 LLM 本来就知道答案概率，那为什么逼它先说话？有意思 JEV

信息来源：X：ZHO (@ZHO_ZHO_ZHO)

行业动态普通2026-09-20 12:35

奥巴马谈 AI 递归自我改进：机器开始自学

美国前总统奥巴马在 C-SPAN 访谈中表示，AI 技术本身并未被过度炒作，被高估的可能是商业收益与 AI 公司估值。他认为 AI 已进入"递归学习"阶段，机器开始能自我教学：一年前模型约 90% 的学习依赖人类、仅 5%-10% 靠自行填补，如今已接近 50/50，若趋势延续可能变为 90/10。

信息来源：X：Rohan Paul (@rohanpaul_ai)