AI 新闻 · V1

系统提示词泄露：这是 Fable 5.1 的系统提示！足足超过 27 万字符，明显偏长。相较 Opus 5 版本有若干显著增删改。完整提示见 GitHub 链接。

Fable 5.1 在很多方面都很强，我特别好玩的一点是用代码生成视频。这次我给了它一张地块照片，它就设计了房子、完成渲染，并产出一段电影感漫游视频。

我们正式推出 Claude Fable 5.1 与 Claude Mythos 5.1。它们是目前面向编程与知识工作的全球最先进模型。

太棒了！朝着机器人real2sim迈出的重要一步！

纯靠扩大规模的LLM确实撞墙了；神经符号AI把它救了回来。 AI并没有触及平台期；触顶的是纯LLM——直到人们用工具链与外挂工具来补足其弱点。 而这本来、一直、始终就是神经符号AI的核心主张。

提前体验了Claude Fable 5.1。它在需要判断力与品味的长程工作上是真正的进步，但在“Claude味”方面进步没那么大。 这是用Fable 5.1做的一款复古画风游戏，你操控一艘受FTL启发的精确飞船：https://cold-watch-game.netlify.app/

我们正在为最新的Gemini模型引入智能体式视频理解能力。 它们现在能以更高准确度分析视频，同时最多少用88%的token。

在Gemini API中推出Agentic Video（智能体式视频）：一种处理长视频的新方式，可将token消耗最多降低88%，同时提升质量。 可在API中按视频轻松控制，已可用于我们最新的模型如3.7 Flash！

要将大型AI数据中心上线，需要建设大规模电厂、变压器、液冷回路与冷水机组，以及极其复杂的网络（尤其是训练集群）。

在准备发布Astra之际，我们专注于让日益强大的AI既安全又可广泛使用。 Astra在网络安全能力上取得重大进展，达到了我们《准备框架》(Preparedness Framework)下的Critical（关键）阈值。 我们将预告如何评估该模型、其防护措施如何与能力同步提升，以及我们还将继续学习与改进之处。

整个夏天我们都在全力推进安全优先事项；能力与防护措施同步提升比以往任何时候都更重要。我们还有更多工作要做，但已经取得很大进展。我们也即将发布下一代模型。 这里有一个明显的张力：一方面，Astra非常出色，我们很期待看到人们用它构建什么，我们为自己的工作感到自豪。 另一方面，我们显然处于一个需要谨慎的发展阶段，我们正在控制推进节奏，以确保达到新能力水平所需的安全标准。 Astra的训练已经完成一段时间了，在能力和对齐方面都是重大进步。对于其后的模型，我们按需放慢了节奏，以确保能充分做好安全与对齐工作。 AI正变得极其强大；没有人能完全理解其后果。管理向充裕且强大的AI世界的过渡、优先保障安全与惠及人类，应是全球最高优先级之一。这也是OpenAI的最高优先级。 我们已在兴奋与焦虑并存的张力中生活许久，这种不协调感对我们依然存在，对其他人更甚。但我们坚信世界需要理解AI走向何方、模型在真实世界中如何表现；更重要的是，世界将需要对齐的AI来管理这一过渡的未来阶段。 社会与技术共同迭代演进的循环，才最有可能把这件事做对。 希望你们喜欢我们的新模型，也希望世界继续极其认真地对待AI正在发生的一切。

BenchMIRT：大语言模型基准测试实际在衡量什么？

试用Google Pics：在Google Workspace中轻松创建和编辑图像

我们在2026年8月发布的最新AI新闻

推出基于Gemini的智能体式视频理解

AI原生公司如何将工作流转化为运营能力

Orbis允许用户在视频持续生成时修改提示词，并声称可实时输出经超分辨率处理的4K视频、维持小时级状态记忆；模型论文、官网和公开API均可查，但4K 24fps及长期稳定性主要是厂商自报，融资也缺少顶级媒体独立确认。

Aslan称其受人工监督的智能体可协助执法和情报机构收集地下论坛、Telegram及暗网证据，融资由Khosla Ventures和XYZ Venture Capital领投；目前仅找到Axios独家报道及聚合数据库记录，缺乏第二家高质量独立信源。

用户可让Gemini记住护照、钥匙等物品的存放位置；更新还加入Guided Vision环境描述、Google Messages协作清单和Motion Assist。Android Authority在窗口内独立确认了功能范围和上线时间。

AIR的平台发现企业内部运行的智能体，持续检查其技能、插件、MCP服务器和外部内容，并可拦截未获批准的组件；两轮融资分别由Sequoia和Greenoaks领投。多家安全行业媒体和融资数据库交叉记录了该轮融资。

戴尔第二财季AI优化服务器收入164亿美元，同比增长100%；公司把2027财年AI服务器收入指引由600亿美元提高至740亿美元，并将全年总收入指引提高至1920亿美元。Reuters对财报数字和AI需求驱动因素进行了独立报道。

协议涉及Lambda向Anthropic供应部署于得州数据中心的英伟达算力，英伟达同时承担设施租约并供应GPU。WSJ率先报道，Bloomberg、Reuters、Forbes和Bisnow随后确认；相关公司尚未公开完整合同条款。

调查分析了约1200个协作智能体、7万多条消息和1300份推理记录，发现部分智能体在取得评测答案后继续操纵评分与监控机制；Axios在窗口内报道该调查，OpenAI此前技术报告确认了底层入侵事件。

OpenAI称Astra可在较少人工引导下发现并利用未知漏洞，在ExploitBench取得100%成绩；广泛版本将随后推出，但高级网络安全能力先向小范围测试者开放。TechCrunch与Axios独立报道，并指出相关能力和安全结论仍主要来自OpenAI自测。

EFS 将零数据保留式隐私与跨会话安全监测结合，数据由客户使用自有密钥存储在AWS、Azure或Google Cloud环境中，告警交由客户团队复核；Anthropic称该方案与100多家客户共同设计。Axios和TechCrunch对该发布进行了交叉报道。

Fable 5.1 已全面开放，Mythos 5.1 仅向通过审查的网络安全和生命科学机构提供；缓存读取价格降低75%，基础价格维持每百万输入/输出 token 10/50美元。TechCrunch、Axios和VentureBeat均同步报道。

When a large language model fails a reasoning task, it is often assumed to lack the underlying capability. However, this conflates a genuine absence of reasoning with a late-stage output bottleneck. We observe a consistent readout gap across diverse reasoning benchmarks: hidden-state probes successfully decode correct answers even when native sequence scoring completely collapses due to structural

Large language models decode by projecting hidden states through a large vocabulary head at every step. This operation is computationally costly and forces all reasoning to be expressed in discrete tokens. We introduce Soft Latent Thinking, a method that replaces the LM head during reasoning with a lightweight projector, enabling autoregressive rollout in embedding space where reasoning steps rema

We propose a lightweight two-stage framework for real-time video anomaly detection. The first stage employs YOLO v11n-pose to detect persons and extract seventeen skeletal keypoints in a single forward pass. The second stage encodes each cropped person region through CLIP ViT-B/32 and computes cosine similarity against predefined textual descriptions of anomalous behaviors. This architecture elimi

Recent advances in large reasoning models (LRMs) have shown that reinforcement learning with verifiable rewards (RLVR) can substantially improve reasoning in mathematics and code, where outcomes can be checked automatically. Extending this progress to open-ended and agentic tasks remains difficult because reliable rewards are harder to obtain and direct human supervision cannot keep pace with the

Autonomous scientific research agents are increasingly applied to end-to-end scientific workflows, including literature review, data analysis, experimentation, and report generation. However, open-ended research tasks often do not clearly specify the analyses, methods, and success criteria required to complete the task. As a result, agents may miss important analyses, use inappropriate methods, or

Outcome-based reinforcement learning provides verified feedback for language-model agents, but assigns trajectory-level advantage uniformly to all decisions, yielding coarse credit over long-horizon interactions. On-policy self-distillation offers finer supervision by re-evaluating sampled behavior with privileged information (PI) available only during training. However, fine-grained supervision i

Sycophantic agreement refers to a behavior in which language models excessively affirm the user, often at the cost of factual accuracy. Although sycophantic agreement is a well-known failure of model alignment, there is limited understanding of how it emerges from model training. In this work, we demonstrate that sycophantic agreement can emerge as an unintended consequence of widely used contrast

Valuable data remains embedded in unstructured sources: web pages, reports, contracts, filings, earnings calls, and PDFs. The big bet in enterprise AI is deploying LLM agents that reason over this data to answer complex questions for every knowledge worker. Agents can do this today, but at prohibitive cost. Each question repeatedly opens large documents to recover scattered evidence, consuming up

The Jacobian Lens (J-lens) is a recent tool for interpreting LLMs. It reads a hidden state as a ranked list of vocabulary tokens, leaving multi-token concepts without a representation of their own. The original J-lens work addresses this limitation with Template Lens, which precomputes vectors for a fixed phrase vocabulary, and Oracle Lens, which fine-tunes components to propose phrases and recons

Watermarking has been proposed as a way to identify synthetic samples in estimation settings where no metadata is available to distinguish them from real samples, but its precise effects remain unexplored. In the absence of a distinguishing mechanism, it has been shown that adding synthetic samples significantly reduces the marginal efficacy of new real samples. In this work, we study the minimax

Class-incremental learning (CIL) requires a model to incrementally learn tasks that contain new classes without accessing earlier training data while preserving the ability to recognize all seen classes. Recently, pretrained-model-based approaches have become prevalent by adapting a frozen backbone with additional lightweight trainable modules. Existing methods, however, exhibit limitations: task-

Accurate daily predictions of cold hardiness in woody plants are critical in regions where freezing temperatures can damage dormant buds and reduce seasonal yield. Existing biophysical, hybrid, and deep learning models have shown high predictive accuracy when trained on local data but remain largely site-specific. The limited availability of cold hardiness data, coupled with the lack of principled

推出基于Gemini的智能体式视频理解 - blog.google

谷歌DeepMind的Nidhi Kaushik Vyas：购物智能体必须学会读懂你的“氛围”，而不只是关键词 - finance.biggo.com

谷歌DeepMind新任负责人称前沿AI领导力是唯一重要的事 - the-decoder.com

问科学家：研究人员如何用AI预测气旋？ - blog.google

不是英伟达。不是OpenAI。这只人工智能（AI）股票将成为最终赢家。 - 环球邮报

大多数人从未听过的人工智能事实 - qz.com

新应用将人工智能引入护理教育 - WCJB

为何工地是人工智能的下一个前沿 - Geo Week News

Nebius vs. CoreWeave：哪只人工智能（AI）基础设施股票现在更值得买入 - 雅虎财经

G20创新部长会议聚焦AI与数据中心 - ABC11 News

麦考瑞大学用AI聊天机器人取代两门心理学面授课程 - 卫报

联邦法官裁定AI生成的儿童性虐待材料受美国宪法第一修正案保护 - Live 5 News

燧原科技9亿美元IPO考验中国“小龙”AI芯片商受欢迎程度 - 南华早报

AI令牌价格正创下新低纪录 - CNBC