US corporate AI spending hits a new high in October, but growth slows to 1.65 percent
Ramp AI Index显示,美国企业2026年10月AI支出达到新高,环比增长1.65%,OpenAI与Anthropic的客户支出分别增长2.1%和0.5%。
Ramp AI Index显示,美国企业2026年10月AI支出达到新高,环比增长1.65%,OpenAI与Anthropic的客户支出分别增长2.1%和0.5%。
英国 AI 芯片初创公司 Fractile 完成 $720 million C 轮融资,英国 Sovereign AI Fund 参与支持,计划明年交付第一批芯片与系统。
CoreWeave正增加客户预付款、按特定合同抵押借款及短期交易,以支持AI基础设施扩张并减少对无担保债务的依赖。公司第三季度收入为$3.66 billion,同比增长168%,净亏损为$709 million,季度利息支出为$846 million,预计未来3个月获得约$9 billion客户预付款和保证金。
Pomerado 推出网站转 API/MCP 工具,宣称可在几分钟内为没有现有 API 的网站构建智能体集成。用户可通过 Codex、Claude 或 Cursor 发起集成创建,由 Pomerado 构建、测试和托管,并处理登录、2FA 及网站变化后的修复。
AWS Emulator 是用于开发和测试的本地 AWS API 模拟器,无需 AWS 账号,请求发送至模拟器而非 AWS。它支持服务端模式和 Rust 测试内嵌模式,也可通过 aws-emulator 技能让 AI 编码智能体协助配置,后者要求 AWS CLI 2.35.0 或更高版本。该工具不适用于生产负载或安全测试,默认开启使用遥测,可通过命令行、环境变量或配置关闭。
AWS released AWS Emulator, a free, offline AWS test environment that runs on a local machine for coding agents. Its code was written and tested entirely by agents, built on more than 400 API models covering over 200 AWS services, and checked against real AWS for behavioral consistency. The code is published under CC0, and the GitHub repo shows the full commit history plus the known gaps where behavior still differs from real AWS. Engineers design the specs, tests and the merge gate, including security review.
Why it matters: AWS Emulator checks agent-written code against real AWS and gates merges, putting quality control in specs and acceptance tests engineers design.
Lenore Palladino指出,退休储蓄通过私募信贷和股权基金流入AI数据中心建设,使劳动者承担相关债务与需求风险。文章援引Americans for Financial Reform报告称,私募股权公司参与了25-30%的数据中心融资,并引用S&P对超大规模云服务商信用质量恶化的警告。她主张养老金减少持有与数据中心热潮相关、缺乏监管的金融资产,并将退休资金转向可负担住房和可再生能源投资。
Bo Lau 在 Hacker News 发布《Term Sheet, Please》,把《Papers Please》的证件审查玩法改成 Demo Day 后筛选 YC 创业公司并决定是否出资。玩家在 16:12 开始、Demo Day 次日 17:00 前核对材料事实并做出 Pass、Fund、Hold 或 Place 决定。
JioStar 用 OpenAI API 在 JioHotstar 中构建了对话式视频与板球数据发现助手,支持超过 15 种语言,由小团队约八周完成上线。该平台服务超 5 亿用户,曾支撑 7250 万并发直播体育观众,超 60% 用户偏好语音而非文字输入。团队将准确性、延迟与成本视为三方权衡,并强调评估体系与领域专家在投产中的关键作用。
Simon Willison 为 tools.simonwillison.net 上线标签浏览功能,可按 JavaScript API、库和 Web API host 分类检索其 244 个 vibe-coded 工具。
Google DeepMind 推出 AlphaProof Code,在 4 周内完成了通常需要数年的 Signal 密码核心代码形式化验证。专家定义安全要求后,基于 Gemini Carbon 的 AI 智能体编写了超过 6000000 行 Lean 证明代码,并将约 1% 的代码库标记为需要人工审查。
TSMC's board approved a plan to invest up to $1 billion total in OpenAI and Anthropic, taking direct equity stakes in both AI developers. Per Reuters, TSMC will put up to $500 million into OpenAI preferred shares for large investors, and another $500 million into Anthropic common stock at its expected IPO price.
Why it matters: TSMC plans to take stakes in Anthropic and OpenAI, and this circular funding between supplier and customer makes real AI hardware and service demand harder for the market to judge.
Goldman Sachs barred most traders and investment bankers from using Claude Code and GitHub coding tools in October 2026, after a three-month trial. Quantitative strategists keep access. The article says prototypes built by business staff and handed to engineers for deployment still carry code verification, security audit and incident-handling costs, and locally run agents raise credential exposure and audit gaps.
Why it matters: Goldman's split, cutting most business staff while keeping quant strategists, ties tool access to code review, security audit and failure ownership.
House House, the studio behind Big Walk, disclosed that it used large language models for a small share of technical fixes and bug work, well under 0.1% of the game's code. One Unity shadow bug fix was heavily reworked into a final 10 lines. Steam does not require disclosure of this kind of assisted coding, while IGF awards draw no such distinction.
Why it matters: The disclosure drew pushback in a scene with little tolerance for AI, and Steam and IGF treat assisted coding differently.
Yogurt 是免费、开源且注重隐私的 AI 会议笔记工具,定位为 Granola 的替代品。它无需向会议添加机器人即可录制通话,提供 Mac、Windows 和 Linux 下载,采用 MIT 许可证。项目地址为 https://github.com/FrigadeHQ/yogurt。
Pomerado 推出自助平台,将网站转为确定性、可自动修复的 API 和 MCP,支持接入用户并不拥有的网站。用户描述任务并回答工作流问题后,可在不到一小时内获得 MCP,后续调用通常在 5-30 秒内返回;平台通过 HTTP 请求或 Playwright 脚本完成示例任务,再生成可重复调用的接口,并支持认证、2FA 和失败后的维护修复。
Google 今天起向全球 iOS Gemini 用户推出 Apple CarPlay 支持,覆盖 Gemini Live 支持的所有语言。用户可在车载屏幕开启或继续语音对话,通过 Gmail 和 Google Calendar 查询日程、了解未读邮件及总结邮件长线程,对话记录会自动同步至 iPhone 上的 Gemini 应用。
klad 发布早期公开版本,用户与 Claude Code、Codex 等编码智能体可在同一画布上协作绘画,双方笔触实时同步。应用不含图像模型,保留可编辑的笔触、形状和填充数据,支持动画及导出至 Godot;源码见 https://github.com/klad-app/klad,指南见 klad.studio。
OpenAI's quarterly revenue reportedly passed $10 billion for the first time, up nearly 50% from $6.7 billion in the second quarter. Bloomberg cited anonymous sources who attributed the growth to strong demand for recent AI models.
Why it matters: OpenAI and Anthropic count annualized revenue differently: Anthropic includes cloud partners' sales, while OpenAI counts only its share of some partner sales.
投资情报公司 Heron Intelligence CEO Nate Singer 表示,Nvidia CEO 黄仁勋的"仙人跳"丑闻并非孤例,半导体行业至少还有三起类似的高管行为问题报告,且可能更多未被曝光。他称这些高管"越过了不该越过的界限",相关问题在业内已成公开的秘密,许多资深人士知情却默许。
Google DeepMind 推出 AlphaGenome 2,预测遗传变异对不同细胞类型和状态的影响,以帮助研究疾病的细胞层面根源。模型使用覆盖 68 种组织的超过 100 million 份单细胞数据训练,据公司介绍,识别出的可改变细胞基因表达的变异数量接近此前最先进模型的两倍,且无需重新训练即可预测新的细胞类型和生物状态。
Google 推出名为 Tabletop 的实验性 AI 工具,可通过描述需求生成储蓄追踪、家务表、行程规划等自定义协作空间。用户添加链接、文件和想法后提示 @tabletop 或选用模板即可生成互动工具,再分享邀请链接让群组成员像 Google Doc 一样同时使用。Tabletop beta 通过候补名单开放,目前仅限美国 18 岁以上用户。
德国防务科技公司Helsing发布智能体平台Clarity,将多源情报转为目标建议,经人工审核后向武器系统交付数据。公司称,数百个专用AI智能体处理传感器、开源情报和报告,系统部署在客户自身基础设施上,每次交付须由人授权并记录责任人姓名。文章援引同类系统在伊朗、加沙及一次涉及中国船只的误判案例,指出自动化提速、多源数据融合和人工审核仍可能伴随错误与盲目信任风险。
Databricks 周三发布更新版 Genie AI 智能体,公司称其可利用客户平台数据的本体提供上下文,无人监督地完成耗时的多步骤任务。CEO Ali Ghodsi 向记者展示了实际使用记录,包括分析员工满意度低分项、评估改进方案及总结失败的销售沟通,以获取管理层不易听到的负面信息。他表示,企业可通过 Unity Gateway 设置使用限额并切换到更便宜的开放权重模型来控制成本。
Google DeepMind released AlphaGenome 2, which conditions on cell type or state to predict how DNA sequences and their variants affect specific human cells. The model was trained on more than 107 million single-cell profiles. On the TenK10K dataset it identified causal variants 30% of the time, against 16% for the original model fine-tuned on single-cell data and for other specialized tools. It can also analyze unseen cell states without retraining.
Why it matters: AlphaGenome 2 narrows predictions from tissue averages to cell types and states, helping pinpoint the genetic variants that act in specific disease-relevant cells.
Anthropic began its IPO roadshow yesterday, and the roadshow video covers how Claude is trained and deployed across multiple chips, plus methods for getting more out of compute.
Why it matters: The roadshow video compares a multi-chip adaptation that took dozens of engineers months with the same job done by Mythos in three days.
一位 Independent Games Festival(IGF)评委主张明确 AI 使用政策,并支持评委自行决定是否考虑使用 AI 的参赛游戏。House House 声明,Big Walk 在开发后期使用 LLM 诊断和修复错误,涉及 less than 0.1% of the code;IGF 要求申报所有 AI 使用,但尚无统一评奖政策。
Atuin 18.24+ ships Threads, which searches, syncs and resumes agent sessions across machines and can fork a session into a different agent tool. It supports Claude Code, Codex, Pi, Omp and OpenCode. Session capture must be turned on, sync is end-to-end encrypted, and you can use the hosted server, self-host or keep everything local.
Why it matters: It pairs cross-machine agent session sync with history search, so later tasks can reuse earlier investigation and decision context.
Anthropic says its own model, Mythos, got Claude running on AMD GPUs in 3 days with almost no human intervention. The model ported or wrote nearly 100% of the kernels, measured on prompt processing on AMD MI355X. Adapting a new accelerator by conventional means takes 6 to 12 months.
Why it matters: Anthropic claims Mythos cut new-hardware adaptation from 6 to 12 months to 3 days, comparing model-written kernels with conventional engineering.
Plurai 完成 1300 万美元种子轮融资,由 Team8 领投,U&I、Velocity 和 Mercer 参投。该公司用基于企业自有数字系统生成的合成数据构建仿真环境,为 Claude Code、Codex、Copilot、Cursor 等智能体工具做评测与修复,并利用收集的数据后训练小型开源权重模型。
Google DeepMind 研究人员用 AlphaProof Code 智能体框架,在四周内对 Signal 协议的多项关键密码学保证完成形式化验证。研究未在所检查部分发现新漏洞,并在含有已知且早已修复漏洞的旧版本中正确识别问题,判定其无法通过形式化验证。验证范围不涵盖整个 Signal 移动应用,研究人员正公开此次分析的规范与证明,以支持其他形式化验证工作。
Adobe 在年度 Max 大会上公布多项实验性 AI 工具:Project Sight Span 可将任意平面图像转为可导航的 3D 场景并从新视角重新取景。
DeepL 的新模型 Aure 将原说话者的语气、情绪、重音和语速传递到目标语言,让 DeepL Voice 的实时翻译更自然。旧方案只在对话开始时采集静态声音样本,Aure 则随对话持续分析译文和原始音频,训练时分别处理声音身份、语调和内容,并通过强化学习避免输出带有口音。
ChinaTalk 调查发现,至少9家美国数据公司向中国 AI 实验室出售训练数据,交易规模约5亿美元,其中包括强化学习环境。中国实验室购买此前售予 Anthropic 或 OpenAI 的同类产品,以减少内部开发和无效训练,并结合 GPT、Claude 输出进行知识蒸馏。此类环境目前基本不受出口管制,调查作者主张将具有军民两用风险的编码、生物学环境纳入现有管制。
A Google engineer used 48 AI agents to scan NASA lidar data and flagged a possible Maya site in Mexico in under 72 hours, then got an official reply. The agents checked terrain imagery image by image, picked out plazas and a pyramid at least 8–12 m tall, searched literature and archives, and checked the raw point cloud. INAH said no archaeological site is registered at that location. The site is not formally confirmed, and fieldwork needs INAH authorization.
Why it matters: Agents screened lidar imagery, then cross-checked literature, the raw point cloud and official records, moving a visual lead toward an archaeological candidate.
Thomas Neumann 质疑 GenDB-V 相对 Umbra 和 DuckDB 的 22× 至 61× 加速,指出其性能收益依赖针对 TPC-H 的过度特化。
Nathan Barry 用线性规划编辑 Qwen3.5-0.8B 权重,缓解推理循环,以少 8× 的权重改动取得接近 GPT-6 Astra 编辑方案的基准表现。
OpenAI's Q3 revenue reached $10 billion, with enterprise business revenue doubling to become the main source. Revenue still trails Anthropic. The Wall Street Journal and The Information report OpenAI's adjusted operating loss widened from $10.1 billion to $10.9 billion, plus $2.6 billion in stock-based compensation not counted in that figure.
Why it matters: OpenAI's growth is now pulled by enterprise sales, while Anthropic's enterprise-and-tools focus still leads on both revenue scale and operating efficiency.
Security firm Claroc checked 37,571 Grok-generated apps that are still online. Of those, 35,578 could be opened without logging in, exposing patient records, plaintext passwords and SpaceXSI internal sales data. Claroc traces the leaks to default sharing settings and broken permission checks.
Why it matters: Claroc traced the exposure to default sharing settings and permission-check bugs, and outside accounts could read, change and delete team data after logging in.