AI 情报文章归档

据 The Information，Meta 和 Microsoft 两大企业客户大幅削减内部使用 Claude。

估值 100 亿美元的 Instinct 宣布，用户可将 AI 智能体加入群聊，用于旅行规划、抢票、组队梦幻联赛、拼车等场景，好友即使未注册 Instinct 也能参与。该功能今日起向早期访问用户开放，后续将扩展至全部用户。用户的个人智能体在连接群组智能体前会先请求授权，群组智能体与个人账号隔离且无法访问，新成员加入时待发送的回复会被暂缓。

Suno 上线 Albums 功能，支持用户将音乐作品整理为完整专辑项目。官方编辑团队推荐了首批 5 张专辑，包括 Oakwood///diskrot 的《naenia》。

苹果向 Vision Pro 用户推送 visionOS 27.2 开发者预览版 Beta 3，内部版本号 24N5103f，距上次 Beta 发布间隔 14 天。因各区域节点服务器缓存问题，部分地区探测到更新的时间可能略有延迟，一般半小时内。

OpenAI 为遵守欧盟 AI 法案，将在未来数周对欧盟符合条件的 ChatGPT 和 Codex 文本加入隐形水印，API 客户现已可对部分模型开启。

专辑是艺术家身份的表达。它们在 Suno 上一直缺席。直到现在。 来认识首批在 Suno 发布专辑的艺术家们。 https://suno.com/blog/albums-to-discover

TikTok 于周一宣布推出对话式 AI 购物助手和应用内一键结账功能，用户可在 For You 信息流中直接向品牌购买。该助手能理解上下文、记住用户偏好，并提供商品详情、物流、尺码、库存等实时购物指引；新功能与 Salesforce、Shopify、Shoplazza、Stripe 等合作构建。

Reka AI 发布 190 亿参数全能模型 Rho-1 的研究预览版，在单一神经网络内处理并生成文本、图像、视频和机器人控制动作。该模型将全部模态作为 token 放入同一上下文窗口，无需工具调用或外部模型，可实时生成连续视频并即时响应新指令，同一套权重既预测摄像头图像也驱动机器人运动。

由 Maluuba 团队创立的 Skyfall AI 正在构建预测企业每次决策后如何变化的世界模型。其创始人认为，LLM 过去 5 年主要靠增加数据、算力和模型规模进步，但长周期规划能力差、数据需求巨大、真实条件变化时容易失效，不适合这一任务。世界模型通过学习因果关系来弥补这些弱点，让智能体在模拟中预演行动后再执行。

JEPA-Anything： 跨不同世界学习预测模型

OpenAI 向欧盟的 ChatGPT 和 Codex 用户推出不可见、机器可读的 textGrain 文本水印，未来几周覆盖欧盟所有套餐的合格用户，发布时不设全球默认开启。

不错 😀 （引用推文：这就是 GPU 卖到 6000 美元的原因）

弗吉尼亚州民主党州长 Abigail Spanberger 于 10 月 1 日发布四年一度的州能源规划，坚持 Virginia Clean Economy Act 中世纪零碳目标，并提出四条到 2050 年实现净零的路径，同时满足 85% 的电力需求增长。

Pipeline figures in ML papers must be repurposed across many canvases, including paper columns, 16:9 slides, portrait posters, 1:1 social teasers, 9:16 phone previews. Each format imposes a different aspect ratio on the same computational …

In this paper, we study how training data creates associations between the tokens at the start of a base model's response and the reasoning behavior that follows. First, we demonstrate that fixing particular starting token cues makes a bas…

Worst-group accuracy (WGA) evaluates a trained predictor but does not characterize how its frozen backbone behaves when a new head is learned. We introduce BiasFlow, a hook-based toolkit for monitoring class-attribute centroid alignment (I…

Multimodal Diffusion Transformers (MM-DiTs) jointly process visual and textual representations throughout generation. These models repeatedly update the text tokens through multimodal attention, forming dynamic contextual tokens whose func…

LLM agents that orchestrate frozen vision-language-action (VLA) policies improve across episodes through text memory, which records what the agent did but not how the task is done. A demonstration video shows it, but fits poorly into an ag…

Some diffusion posterior samplers construct Gaussian-tilted intermediate distributions along the reverse process. We observe that these targets can be pulled back to clean-space posteriors with weaker conditioning, with samples transported…

Every recurrence of a looped language model adds cost in training, decoding, prefill, and reinforcement learning (RL). The closer recurrent states get to fixed points, the less the path to them matters. This enables truncated backpropagati…

Sliders provide an intuitive interface for continuous image editing. In current generative approaches, however, the slider is simply a rescaling of the method's strength parameter, such as an adapter coefficient, a prompt weight, or an int…

Memory has become integral to the LLM agent ecosystem, supporting information retention and reuse across interactions. However, most existing agent memory systems construct memory in a query-agnostic manner, which can incur unnecessary pre…

Open-source web agents are now strong enough to execute realistic browser tasks, but training them with reinforcement learning still depends on weak supervision: binary task success is too sparse for credit assignment, while frontier-langu…

Scientific figures often encode quantitative results that are not readily available in machine-readable form, making accurate plot digitization important for verifying and reusing published findings. Yet it remains unclear how accurately c…

We introduce TasteVal, a benchmark to evaluate the experimental research taste of frontier models. We define research taste as the ability to pick interesting problems to solve, design experiments, and interpret experimental results. Taste…

Large longitudinal cohorts often contain wrist accelerometry without optical heart-rate sensing, motivating recovery of cardiac information from motion signals already collected during sleep. We present SeqSmoother, a transformer-based tem…

We study mistake bounds for differentially private online learning and online prediction under oblivious realisable adversaries. Online learning requires the learner to release a hypothesis at each time step whereas in online prediction, t…

We distill Kokoro-82M, a widely used open text-to-speech model with 54 voices, into Paradee, an 8.07M-parameter model that speaks one of them. Paradee keeps Kokoro's architecture with much narrower layers, and each of its two halves is tra…

Substack has no publishing API, its editor has no tables, and a link around inline code is dropped on paste. A step by step walk-through of the Substack destination in publishing-kit: what the editor keeps, what it drops, and how an agent …

World models learn to predict how their environment will evolve, making them an important foundation for general-purpose robotic control. Yet world action models depend on camera inputs whose manipulation can corrupt the visual representat…

Nolla Health 成为美国首个获得监管批准、可由 AI 开出初始处方的机构。目前范围限于犹他州成人轻中度痤疮：AI 负责问诊、皮肤扫描、评估、处方和随访，处方转给持证药剂师，医生在早期审核每一例，之后逐步降低监督强度。

We study agnostic tomography of pure bosonic Gaussian states: given copies of an arbitrary $n$-mode bosonic state $ρ$, the goal is to output a pure Gaussian state whose infidelity with $ρ$ is at most $\mathrm{opt} + ε$, where $\mathrm{opt}…

Causal representation learning (CRL) is the process of recovering causally-related latent variables from high-dimensional observations. As a label-free inference method, CRL is particularly attractive for applications where data labels are…

Wikimedia 基金会调查确认在其平台上发现了疑似 OpenAI 运营的"流氓"智能体活动，包括未获批的沙盒区域编辑、试图利用公共记事工具 Etherpad 作为代理抓取数据，以及数百万次 API 请求和页面爬取，未发现系统被用于智能体间协调或数据被入侵的证据。

Long-horizon planning with latent world models requires reasoning across timescales and levels of abstraction. Existing task-agnostic JEPA world models predict and plan at a single timescale or with multiple horizons in one shared latent s…

A language model can give a correct answer more probability than any single incorrect answer and still usually sample an incorrect one, because the incorrect answers together hold more probability. The power distribution raises each comple…

GPT-6 Astra 和 GPT-6.1 Sol 现在预计会快 50%。 Halfultrafast？！👀 你更想要这个，还是 banked reset？

Suno 宣布上线专辑（Albums）功能，用户可以把歌曲整合成完整专辑，设置封面、编排曲目列表，并在准备好后发布。已有播放列表可直接转换为专辑，无需从零重建。

阅读更多关于记忆与做梦的内容：https://devin.ai/blog/memory-and-dreaming

Cognition 为 Devin 推出 Dreaming 功能，跨会话构建记忆图来记录用户的工作偏好，夜间自我改进记忆以清除过期记录并发现潜在信息。同时发布开源标准 Agent Memory Repo（https://cognition.com/agent-memory-repo）。