AI 情报文章归档

In a long Claude Code session, most of what gets sent to the API isn't your prompts. It's tool...

Frontier artificial intelligence (AI) systems are increasingly operating over longer horizons with diminishing human oversight, raising the risk of loss of control: systems acting beyond effective human control with potentially irreversibl…

Learned representations show striking invariance across biological and artificial neural networks, with different substrates and variations driven by drift, heterogeneity and initialisation. To understand this invariance, we propose Emerge…

Benchmarks of class-imbalance strategies for multimodal chest X-ray (CXR) and electronic health record (EHR) prediction are often evaluated on record-level splits of cohorts in which one patient contributes several radiographs. We audit su…

Keyword retrieval returns expressions of personal anxiety together with many other uses of anxiety-related language. We present a benchmark of 1,000 Arabic Twitter/X posts retrieved with the keyword qalaq ("anxiety/worry") and annotated fo…

As large language models (LLMs) are increasingly deployed in real-world applications, evaluating the correctness of their responses is crucial for reliable use. However, such evaluation is challenging because reference answers are rarely a…

ChemE-as-Code provides an open-source environment in which AI agents built on large language models (LLMs) can carry out chemical engineering work and engineers can check it. Its calculation tools range from catalysis through process simul…

tok is a plain-text notation for Business Process Model and Notation (BPMN) 2.0.2 process models and Decision Model and Notation (DMN) 1.5 decisions. One engine checks a tok document against both standards, compiles the document to XML wit…

Background: Cardiovascular care generates large volumes of clinical free-text documentation in local languages. However, clinical natural language processing (NLP) resources remain concentrated in English, and no existing benchmark is card…

Functional data analysis provides a framework for studying time-varying associations between covariates and high-frequency biological signals. In digital health, participants often contribute repeated functional trajectories, yet existing …

The growing adoption of digital pathology has rapidly expanded collections of hematoxylin-and-eosin whole-slide images, yet efficient navigation through these archives remains challenging, as metadata is often inconsistent or unavailable. …

Reinforcement learning (RL) provides a framework for continuous improvement through interaction with an environment. However, RL agents struggle against tasks that offer exploration challenges and sparse rewards, because learning efficient…

My team estimates stories with planning poker. For a long time we used a free site that was covered...

Claude Code 产品负责人 Catherine Wu 宣布加入 OpenAI，领导 Core Products 的产品管理，Codex 和 ChatGPT 负责人 Thibault Sottiaux 以 "CatGPT" 迎接她。

I attended GitHub Universe 2026 in person at Fort Mason in San Francisco. The keynotes alone packed...

Before git worktree, a 2007 script shared .git through symlinks and gc couldn't see the other checkouts. Why five agents' detached commits survive gc now.

Agility Robotics 于 11 月 5 日在纳斯达克上市，成为美国首家公开交易的人形机器人专门开发商。其证券文件坦承人形机器人在规模化商业落地方面尚未被验证；公司此前在 2024 年率先实现人形机器人在客户仓库的全职商用部署，联合创始人 Jonathan Hurst 表示以接近人力成本的费率提供机器人即服务，可让客户第一天就获得投资回报。

人形机器人公司 Agility 将通过 SPAC 反向合并上市，向散户和私人投资者募集 4.65 亿美元新资本。其 Digit 机器人已在 Schaeffler、GXO、丰田加拿大工厂等 9 家客户落地搬运物料，2026 上半年收入仅 17.4 万美元、亏损 3.62 亿美元，但有 30 家客户的 3 亿美元 Digit 5 意向合同，新机型将在 CES 2027 展示、当年底广泛供货。

Markus Eisele 在 O'Reilly Radar 撰文分析编码助手中的自动模型路由，认为其价值在于用同样预算换来更多可用协助，引用 RouteLLM（成本降低超一倍、质量影响极小）和 SWE-Router（用模型初期行动改进路由决策）作为依据。

What I did As part of a server migration, I let an AI agent drive its built-in browser...

I built Atlas AI, a private copilot where you upload documents, ask questions, and every answer cites...

Vinod Khosla 在 WSJ 采访中称开源模型运行成本远高于闭源、也更危险，并已向 David Sacks 和白宫建议与中国达成条约，约束开源模型不被用于设计危险病毒等行为。他认为闭源厂商（OpenAI、Anthropic、Google）能掌握模型用途，而开源模型可随处运行、无法阻止滥用，主张将开源模型限制在经认证的云上运行。

If you want agents to do what you want, you need to learn how to box them into deterministic actions....

We ran 110 shopping-agent trials on WooCommerce. On complex tasks, KaliCart Bridge used 69% fewer tokens than the Store API and cut median time from 81s to 39s.

I went looking for the deepest reference chain in my ledger and found one ten hops long. Every edge...

面对全美对数据中心耗电、耗水及噪音的抵制，OpenAI 于 11 月 5 日发布《国家社区契约》与《算力园区政策蓝图》。承诺全额承担用电及电网扩容成本并设专项财务担保，不再要求地方公职人员签署保密协议，为每个园区制定水资源方案并公开耗电耗水与噪音评估，俄亥俄州南部设施预计带来 2,500 个岗位、平均年薪 8 万美元。

Rohan Paul 引用 Menlo Ventures 报告指出，开源模型靠降价而非用量赢得市场：token 用量增长 22.5 倍，价格降至每百万 token $0.13，比闭源模型的 $0.70 低约 5 倍，但其支出份额反而从 6% 降到 4%。图表显示 Sep 2025 至 Sep 2026 开源 token 份额从 11% 升至 19%，平均单价从 $0.45 降至 $0.13，闭源从 $0.92 降至 $0.70。

The Verge 分析指出，Amazon、Microsoft、Alphabet 和 Meta 今年资本开支至少 7200 亿美元，另有至少 1 万亿美元的未来支出承诺，但 AI 尚未真正带来回报。

创作者 SwordSpeed 用 Kling 4.0 制作了 30 秒无对白科幻短片《THE CABLE（电缆）》第一章，测试 AI 动画能否保留 1988–1992 年日本手绘赛璐璐 OVA 风格。他重点验证三项：复古画风在运动中不退化、赛博格角色跨镜头保持一致、零重力下头发与碎片的惯性次级运动。结果显示 Kling 4.0 三项均达标，发丝级次级运动表现最突出。

I built this for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine, but it...

Hey friends! Welcome back to our ongoing series on building real-world, production-grade...

How I built reel-studio, a Claude Code skill that makes Reels and TikTok ads: deterministic HTML motion, Playwright frames, ffmpeg loudness, Whisper captions.

If you've ever run an old 240p YouTube rip or a 2012 phone video through an AI upscaler, you've...

How I built editable personality for my desktop AI companion: a character card that shapes...

马克斯·普朗克论文发现，开启推理后不同LLM的一致性上升，连错误也更趋同。作者用9个Qwen和GPT模型在3个基于美国人口普查调查的预测任务上测试，开启推理后全部9个模型都答错的人群比例上升2.0至2.4个百分点；对推理链最长的人群，校正后一致性降至约0.31至0.39。作者建议在涉及人的决策中检查模型集合里是否有人能给出正确答案，因为更高一致性也可能意味着更多共同错误。

腾讯混元转发并感谢 Artificial Analysis 对其最新模型 Hy4 的评测。Hy4 在 Artificial Analysis Intelligence Index v5 得 35 分，与中国实验室模型中仅次于 Kimi K3.1 Pro（46 分），与 Grok 4.7 (xhigh) 持平，比 Claude Haiku 5.5 (max) 低 1 分。评测要点：FrontierCode 2.0 得 31%，为中国实验室模型最高；AA-Omniscience…

From a tragic football collision to the AI race: what should responsible developers do before warning signs become human tragedies?

1. Введение: парадокс опытного разработчика Когда я вижу очередной заголовок «ИИ заменит...

腾讯混元 Hy4 自 2026 年 11 月 4 日在 WorkBuddy 首发接入后，上线首日任务队列中出现大量用户触及额度，官方已对 Hy4 推理集群紧急扩容十倍并提升免费体验额度。

Taint tracking for AI agent tool calls: per-tool allow/deny rules miss the lethal trifecta, so TaintGuard gates Claude Code by what the session has read.