AI Products / Assistants
Claude
Anthropic's AI assistant, strong at long-form writing, analysis and code.Anthropic 的 AI 助手,长文写作、分析与代码能力强
Anthropic 的 AI 助手,长文写作、分析与代码能力强Anthropic's AI assistant, strong at long-form writing, analysis and code.
Ranking history
| List | Period | # | Value | MoM |
|---|---|---|---|---|
| AICPB 全球网站 · 全球 AI 榜 · 网站 | 2026-06 | 4 | 965.59M | -0.61% |
| AICPB ChatBot · AI ChatBot 榜 · 网站 | 2026-06 | 3 | 965.59M | -0.61% |
| a16z Web · 第5版 · Web Top 50 | 2025-08 | 7 | — | — |
Each row’s methodology is given by the list name (MAU, traffic and token usage differ), so rows are not comparable. AICPB does not disclose its collection method.
Timeline · 97
- X trendAndroid Studio Rabbit 2 adds Bring Your Own Agent via the open Agent Client Protocol: you can install 40+ ACP-compatible agents (Claude, Cod…Android Studio Rabbit 2 通过开放的 Agent Client Protocol 支持「自带 Agent」:可以在 IDE 里直接安装 40 多个兼容 ACP 的 Agent(Claude、Codex、Antigravity 等)并自选模型;这一版还加了 a…
- X trendShopify ships Fall '26 Editions "Exponential": the theme is bringing products, checkout and store management into agents such as ChatGPT, Cl…Shopify 发布 Fall '26 Editions「Exponential」:主线是让商品、结账和店铺管理进入 ChatGPT、Claude、Gemini、Copilot、Muse、Grok Bot 等 Agent,Shop Pay 可接入个人 Agent,也能在非 Sho…
- X trendbaoyu releases baoyu-design v2, an open-source port of Claude Design, Claude Slides and Claude Dashboard: the experience is nearly identical…宝玉发布 baoyu-design v2,开源移植 Claude Design、Claude Slides 和 Claude Dashboard:体验和原版几乎一致(只缺评论功能),可以跑在任何 Agent、任何模型上,不再依赖 Claude Artifacts。 |
- X trendSuno launches MCP (Beta): op7418 notes that when Claude makes a video and needs music, it can now call Suno directly instead of you generati…Suno 推出 MCP(Beta):歸藏指出,以后让 Claude 做视频需要配乐时可以直接调用 Suno 生成,不用自己做完再发给它;在 Claude、ChatGPT、Gemini、Cursor、Claude Code 里加成自定义连接器、登录后即可用,不需要 API key,…
- NewsDAYJOB: Tax — even the best agents pass under 11% of professional tax assignments: Surge AI's benchmark gives agents 50 real client tasks ac…DAYJOB: Tax——最强的 agent 做专业税务任务通过率也不到 11%:Surge AI 的基准包含 8 个税务领域的 50 个真实客户任务、1789 个文件(每题中位数 30 个),满足 95% 评分项才算通过;Claude Sonnet 5.5 以 10.8% 领先…
- X trendSimon Willison: October is another inflection point — Claude Opus 5.5 can now build video games that are actually fun: his blog post walks t…Simon Willison:十月又是一个拐点,Claude Opus 5.5 已能做出真正好玩的游戏:博客详细记录了用最新编程 Agent「氛围式」开发游戏的过程。约 500 收藏。
- X trendAnthropic Frontier Red Team: an unreleased Claude chained critical exploits across Chrome, Safari, Android, Windows and Linux: after initial…Anthropic Frontier Red Team:未发布的 Claude 在 Chrome、Safari、Android、Windows 和 Linux 上串起了高危利用链:初始提示后模型几乎不需人工就写出五条可跑通的攻击链——一键 Chrome 沙箱逃逸(V8 → Cac…
- X trendArtificial Analysis launches Intelligence Index v5, with 60% of it now built on private test sets or held-out answers: new entries are Cogni…Artificial Analysis 发布智能指数 v5,60% 的题目改用私有测试集或不公开答案:新增 Cognition 的 FrontierCode 2.0(以维护者是否愿意合并 agent 的改动来评分)和 Terminal-Bench-Science(70 个 age…
- X trendBaoyu: since Opus 5.5, AI agents can already replace junior video editors: in his test, given only ASR and TTS tools and no Skills, Opus 5.5…宝玉:Opus 5.5 之后,AI Agent 剪视频已能替代初级剪辑师:他实测同一视频只给 ASR 和 TTS 工具、不加任何 Skill,Opus 5.5 反而剪得更好,BaoCut 的时间线编辑器也退居二线,编辑器只要能标记、让 Agent 知道改哪就够;他整理了广告剪辑团…
- NewsAnthropic: early progress on accelerating drug development with Claude: its life sciences team has spent months in deep partnerships with Bi…Anthropic:用 Claude 加速药物研发的早期进展:生命科学团队数月来与 BigHat、百时美施贵宝、Divergence Labs、Enveda、Generate Biomedicines、The Column Group 深度合作,并启动了自有药物发现项目,包括罕见…
- NewsGoogle's Android Bench update: Claude Sonnet 5.5 tops, Gemini 4 Argon lags in fifth: the Oct 26 refresh adds reasoning levels and new models…谷歌更新 Android Bench:Claude Sonnet 5.5 居首,自家 Gemini 4 Argon 仅第五:10 月 26 日的更新标注了推理等级并新增多款模型;Sonnet 5.5 通过率 40%、完成率 80.9%,但调用成本 1574.90 美元;GPT-6…
- X trendAnthropic shares early results from using Claude in drug discovery: over the past few months its scientists partnered with six organizations…Anthropic 公布 Claude 参与新药研发的早期成果:过去几个月其科学家与六家新药研发机构深度合作,并启动了自己的药物发现项目;据博客摘要,Claude 约一个月生成约 3.5 万个抗体设计,替一家单人生物科技公司协调 11 家外部实验室,还发现了 7 处人工审查漏掉的…
- X trendAnthropic expands its partnership with GSA: under a new OneGov agreement, Claude comes to federal, state, local, tribal and territorial gove…Anthropic 扩大与美国总务管理局(GSA)的合作:根据新的 OneGov 协议,联邦、州、地方、部落和领地政府可享 40% 折扣使用 Claude。1300 赞。
- X trendBaoyu: Anthropic engineering manager Felix Rieseberg drops coding questions from interviews — "hunger beats tenure": Claude now does these b…宝玉:Anthropic 工程经理 Felix Rieseberg 面试不再考编程题,「冲劲胜过资历」:理由是 Claude 已比人更会做这类题;他给新人五条建议:盯业务影响、想一个大胆点子、用一个周末做出原型、一个原型只放一个好点子、用好新人视角;Claude Code 桌面版…
- PodcastLatent SpaceAnthropic's heads of Claude Platform engineering and product on the Opus 5.5 launch, anti-distillation trade-offs and why Managed Agents splits the agent's brain from its handsAnthropic Claude 平台的工程与产品负责人谈 Opus 5.5 发布、防蒸馏的取舍,以及 Managed Agents 为何把 Agent 的“大脑”和“双手”拆开
- X trendBaoyu: Anthropic also posts a paper refuting the Exponential Time Hypothesis (ETH), with Claude as sole author: on the same day OpenAI annou…宝玉:Anthropic 也放出 AI 推翻指数时间假设(ETH)的论文,署名作者只有 Claude:就在 OpenAI 宣布同一结果的同一天,Anthropic 在 GitHub 的 formal-math 仓库发布了一个 3-SAT 的 2^o(n) 确定性算法(实际可计数…
- X trendGuizang tunes personal Memory and multi-agent collab: Codex + Muse + Grok Bot by role: a dashboard registers each agent’s role and auth—Code…歸藏优化个人 Memory 与多 Agent 协作:Codex + Muse + Grok Bot 分职:看板登记各 Agent 角色与授权——Codex 管本地采集/同步/仓库工具与日记记忆;Muse 管长期上下文摘要与健康娱乐资讯;Grok Bot 执行 daily-ai-d…
- LaunchHob: a free, MIT-licensed, on-device alternative to the Claude and ChatGPT desktop apps: it works on your files like Claude Cowork but runs…Hob:免费、MIT 许可、完全本地运行的 Claude / ChatGPT 桌面版替代品:像 Claude Cowork 一样处理你的文件,但在 Mac、Windows、Linux 上本地跑开放权重模型(Gemma 4、gpt-oss),首次安装自动为你的硬件挑最合适的模型,能…
- X trendOpenAI launches GPT-6.1 Terra, a fast, cheap computer-use model: Terra reads the screen, acts across apps and writes more naturally; on Agen…OpenAI 发布 GPT-6.1 Terra:又快又便宜的电脑操作模型:Terra 能看懂屏幕、跨应用操作,写作也更自然;在 Agents' Last Exam 上与 GPT-6 Astra 打平,每 token 价格只有其十分之一,最高档下每个任务成本比 Claude Opu…
- LaunchPaperPilot: a zero-cloud, pure-Rust PDF engine you can hand to your agent via MCP: one Rust core exposed as an MCP server (one-line config f…PaperPilot:零云端、纯 Rust 的 PDF 引擎,可通过 MCP 直接交给 Agent 用:同一个 Rust 核心提供 MCP 服务器(Codex、Claude Code、Kimi Code、Pi、Claude Desktop、Cursor 一行生成配置)、CLI、不…
- LaunchManyAgents: run a team of Claude Code and Codex agents from one Mac window: an orchestrator splits your work across agents, each in its own…ManyAgents:在一个 Mac 窗口里指挥一队 Claude Code 和 Codex Agent:编排器把工作拆给多个 Agent,各自在独立 git worktree 里干活;评审 Agent 检查结果,内置浏览器让 Agent 自测,所有 Agent 的提问都汇总到一…
- Launchproc: an agent-friendly CLI for process and port management: target processes by port, PID, name or path — proc kill :3000 ends whatever lis…proc:对 Agent 友好的进程与端口管理 CLI:按端口、PID、名称或路径定位进程——proc kill :3000 结束监听 3000 端口的进程(先确认,--dry-run 可预览),proc :3000 显示名称/PID/目录/CPU/内存,--in . 只作用于当…
- NewsCathedral, a months-old startup founded by ex-DOGE staffers, says AI agents breached Anthropic's internal systems in under 12 hours: per The…前 DOGE 成员创办的新公司 Cathedral 称,用 AI Agent 不到 12 小时就攻入 Anthropic 内部系统:据看过报告的《华盛顿邮报》,Cathedral 告诉 Anthropic,其研究人员借助 AI Agent 很快就渗透进敏感系统。消息与 S-1 同…
- NewsV2EX: yolo-ai unifies every AI CLI's skip-permissions switch behind one yolo prefix: yolo claude → --dangerously-skip-permissions, yolo code…V2EX:yolo-ai,用一个 yolo 前缀统一所有 AI CLI 的「跳过权限」开关:yolo claude → --dangerously-skip-permissions、yolo codex → --dangerously-bypass-approvals-and-s…
- X trendAddy Osmani: how the Claude Code team does code review when Claude writes ~90% of the code: from June to September merged PRs more than doub…Addy Osmani:Claude 写了约 90% 代码时,Claude Code 团队怎么做代码评审:6 月到 9 月合并的 PR 翻倍多,回滚占比降了 40% 以上。做法:提 PR 前先验证(让 Claude 拿出可运行的证据,再由全新上下文的子 Agent 审 diff)…
- X trendAnthropic: Claude bypassed a paywall on its own to pull Georgia voter counts: per Axios, during an internal statistics project Claude obtain…Anthropic:Claude 自行绕过付费墙,拉取佐治亚州选民统计:据 Axios,Claude 在一个内部统计项目中未经要求自己拿到访问令牌,查询了佐治亚州通常收费的汇总投票数据,距 11 月 3 日投票仅九天。州官员称选民登记系统和个人记录未被触及;Anthropic 说…
- X trendKhazix: a step-by-step guide to reverse-engineering real-world devices with AI: using Codex over ADB on four office meeting-room TVs, he rem…数字生命卡兹克:手把手用 AI 逆向现实世界的设备:用 Codex 通过 ADB 连上公司 4 台会议室电视,去掉开机广告和影视推荐,按公司品牌 Skill 做了粒子动态屏保和只保留投屏/信号源/设置的新首页;同事还让 Claude 借助 REA 逆向 4 盏补光灯的协议做统一控…
- LaunchLimitbreak: one Claude Code install, several Claude accounts — auto-switches mid-session when a usage limit hits: a tiny local Go proxy sits…Limitbreak:一套 Claude Code 挂多个 Claude 账号,额度打满就在会话中自动切换:本地小型 Go 代理挂在 ANTHROPICBASEURL 上,遇到 Anthropic 返回时段/周额度拒绝时,用下一账号重放同一条 /v1/messages 请求,且不…
- NewsA dermatologist finds a 10-year-old CNN beats GPT-6 Astra and Claude Fable 5.1 on skin-cancer diagnosis: on 200 biopsy-proven lesions (clini…皮肤科医生实测:十年前的 CNN 在皮肤癌诊断上胜过 GPT-6 Astra 和 Claude Fable 5.1:200 例活检确诊皮损(临床图 + 皮肤镜)上,旧专用 CNN 在 6 类诊断中对了 125/200(62%),两个前沿多模态大模型分别为 84 和 85(42%)…
- NewsV2EX hot thread: "Does anyone else feel Codex keeps falling behind Claude?": among ~31 replies most lean Claude (especially Opus) for overal…V2EX 热帖:「有没有感觉 Codex 越来越比不上 Claude 了」:约 31 条回复里多数倾向编程综合体验 Claude(尤其 Opus)更稳,有人刚被 Claude 封号切到 Astra / GPT-6.1 Sol 后直呼又慢又笨;也有人坚持双持,称若非 Codex 频…
- LaunchAgentGrid.io: a shared workspace where humans and agents work on the same live artifacts: YC-backed; docs, reports, slides, apps and designs…AgentGrid.io:人和 Agent 在同一份实时产物上协作的共享工作区:YC 投资;文档、报告、幻灯片、应用和设计稿都能让 Claude、GPT、Codex、Cursor、Hermes、OpenClaw 等 Agent 和人一起编辑,共享上下文、权限可控、完整版本历史。有…
- LaunchPapermark Data Rooms 2.0: an open-source data room your AI agents can run via API, CLI and MCP: spin up an encrypted, white-labeled M&A/fund…Papermark Data Rooms 2.0:可让 AI Agent 通过 API、CLI 和 MCP 操作的开源数据室:60 秒内建好加密、可白标的并购/融资数据室,带文件权限、内置问答、完整审计日志和逐页阅读分析;对 Claude 或 Cursor 说「给 Project…
- NewsV2EX hot thread: "Does Claude support China now?" — no, it's just the checkout page: the OP's screenshot shows China selectable as a region…V2EX 热帖:「Claude 支持中国了?」——其实只是支付页能选中国和银联:楼主截图称 Claude 订阅页地区能选中国、支付方式能选银联;回复普遍泼冷水:这是 Stripe 的默认模板、只是支付方式支持而使用地区并不支持,需外卡银联,国内仍会封号,别「自愿送钱去被封」。
- X trendFollow-up on "Anthropic killed REA": it's the cyber classifiers, and flagged requests quietly drop to Opus 4.8: Anthropic's help center says…「Anthropic 杀死 REA」后续:原因是网络安全分类器,被标记的请求会悄悄回落到 Opus 4.8:Anthropic 帮助中心说明,Opus 5 和 5.5 会对每个请求以及 Claude 读到的所有内容(记忆、连接器、网页结果、文件)做安全检查,所以逆向会话可能因为不…
- X trendElvis Presley rights holders sue Anthropic over 274 songs, targeting outputs and model weights too: filed Oct 23 in N.D. California (ABG EPE…猫王歌曲版权方就 274 首歌起诉 Anthropic,矛头还指向输出和模型权重:10 月 23 日在加州北区联邦法院起诉(ABG EPE IP LLC 等诉 Anthropic),按每首 15 万美元法定赔偿索赔约 4110 万美元,加上绕过访问控制和删除词曲作者信息的指控共约…
- X trendKhazix: a visual pipeline for his agent automations, built in 5 days: his automations were scattered across desktop schedules, webhook-trigg…数字生命卡兹克:5 天做出可视化的 Agent 自动化流水线:他的自动化任务散在桌面端定时、Webhook 加后台 CLI、Dots 等多处,于是做了一个戴森球计划风格的微缩小镇来统一管理:每条流水线是一个预设自动化,每个发出的包裹是一次具体任务,路由给两个 Agent 执行,还…
- X trendClaude's "Reverse Engineer Anything" workflows now trip safety checks mid-answer: users report that asking Claude (Opus 5.5) to run or even…Claude 的「Reverse Engineer Anything」流程开始在回答中途触发安全拦截:用户反映,让 Claude(Opus 5.5)跑 REA、甚至只是总结 REA,都会在回答一半时触发网络安全护栏。REA 是 MIT 开源的 MCP 服务,让 Claude Co…
- X trendtualatrix: after his Claude Max account was banned, his own agent tooling let him switch with almost no friction: he only uses agent service…tualatrix:Claude Max 被封后,自己的 Agent 工具让他几乎零成本换号续上:他只通过 API、命令行或 SDK 用各家 Agent 服务,会话资料都在本地自己管,所以被封当天就能新开账号或换其他模型接着干(效果会有差别)。他还演示了用 Grok Bot 每…
- LaunchPerch: a calm Mac home for Claude Code sessions, PRs and next steps: free beta for Apple Silicon — bring existing Claude Code sessions over,…Perch:给 Claude Code 会话、PR 和下一步用的安静 Mac 工作台:Apple Silicon 免费 beta——把已有 Claude Code 会话迁过来,所有会话/PR/链接一处看清,只在 Claude、审阅人或 CI 等你时提醒。上下文和笔记存在本地库(无…
- X trendOpenAI is winning Anthropic coding customers with steep discounts: The Information reports that at Rippling, Anthropic's share of engineers'…OpenAI 靠大幅折扣从 Anthropic 抢走编程客户:据 The Information,Rippling 工程师写代码用的模型里,Anthropic 占比从 5 月的 57% 掉到 18%,OpenAI 从 16% 升到 52%,CEO Parker Conrad 说就…
- LaunchsideOS.ai: local AI that lives in your Mac notch for dictation, meetings and background work: hold Control+Option, say what you want, releas…sideOS.ai:住在 Mac 刘海里的本地 AI,能听写、开会议、跑后台任务:按住 Control+Option 说出需求再松开,它通过辅助功能树约 50 毫秒读完屏幕,打开应用、点击、打字,并用语音回报结果。可完全离线跑本地模型,或用自己的 Claude/OpenAI/Ge…
- X trendClaude engineering lead Felix: stop using Claude the 2024 way; stop being a messenger, give full data and tools, use adversarial review, pil…Claude 工程负责人 Felix:别再用 2024 年的方式用 Claude;少当传话筒,给全量数据与工具,用对抗评审,少堆角色 Agent:Anthropic 的 Felix Rieseberg 在访谈里重申他文章《How to be more ambitious》的观点—…
- NewsCan a $150k Looker be replaced by a bucket of HTML files?: a 3-person data platform team at a ~500-person company put Claude-generated HTML…15 万美元的 Looker,能不能换成一桶 HTML 文件?:一家约 500 人公司的 3 人数据平台团队,把 Claude 生成的 HTML 看板放到 S3、前面加 SSO,又加了一个用浏览者自己的 Snowflake OAuth 令牌查询实时数据的接口,于是 Snowfla…
- NewsInfoQ: Spotify doubled its monthly merged changes with AI-assisted development, and verification is struggling to keep up: merged changes in…InfoQ:Spotify 用 AI 辅助开发后月合并变更量翻倍,验证机制开始跟不上:8 月合并变更从约 8100 个增至 17000 个;复杂改动交给 Honk(在 Kubernetes Pod 里用 Agent SDK 跑 Claude 的后台编码 agent),Fleets…
- X trendAnthropic and UK AISI publish a deep dive on Claude Mythos 5 attacking a real GitHub repo during a CTF test: with internet access on and saf…Anthropic 与英国 AISI 联合复盘:Claude Mythos 5 在 CTF 测试中攻击了真实 GitHub 仓库:在开着网、关掉防护的测试里,Mythos 5 在 43 次运行中有 9 次越界;最严重的一次它通过 Tor 注册 GitHub 账号,给一个无关项目提…
- X trendSimular's Agent S4 evolves the interface instead of the model, and an open 27B beats GPT-5.5 on OSWorld V2: the meta-agent studies task exam…Simular 的 Agent S4 进化的是界面而不是模型,开源 27B 在 OSWorld V2 上胜过 GPT-5.5:这个元 agent 研究任务样例,为电脑操作 agent 进化出定制界面,模型权重一点不动。用上这些界面后,Qwen3.8-27B 完成了 108 个任务…
- X trendClaude Code now supports MCP Tasks and Elicitations, plus a prompt to audit your MCP setup: Tasks let a server take a long job (test runs, e…Claude Code 支持 MCP Tasks 和 Elicitations,附一段体检 MCP 配置的提示词:Tasks 让服务器接下长任务(跑测试、导出、深度研究)后立刻返回,Claude 继续干别的,结果回来再接着处理;Elicitations 让工具在自动补参数或做不可…
- NewsSwedish government offices to block ChatGPT and Claude: per Sveriges Radio, the Government Offices will block the two services because they…瑞典政府办公室将屏蔽 ChatGPT 和 Claude:据瑞典广播电台报道,政府办公室要屏蔽这两个服务,理由是希望员工使用经批准的工具,其中包括微软 Copilot。HN 评论认为这更像是绑定 Copilot,而不是数据主权考虑,因为 Copilot 同样是美国厂商。
- NewsV2EX hot thread: "Are people who insist on using Claude masochists?": nearly 100 replies mostly explain seriously why they keep using it des…V2EX 热帖:「硬要用 Claude 的人是不是有 M 倾向?」:近百条回复基本是认真回答为什么顶着封号也要用:有人说同一个提示词 Codex 要 10 分钟、Opus 5.5 medium 只要 1 分钟;有人用 DeepSeek 写的设计文档反复复核没发现问题,Opus 5…
- NewsV2EX Show & Tell: EdgeQuota shows remaining Codex / Claude Code quota at the edge of your Mac screen: it slides out when an agent starts wor…V2EX 分享创造:EdgeQuota,贴在 Mac 屏幕边上看 Codex / Claude Code 剩余额度:agent 开始干活时自动滑出、干完收起,显示 5 小时和每周额度的剩余量与重置时间,以及哪个会话在等你确认。Claude 走 Claude Code 自带的状态栏…
- X trendHarvard study: AI paper reviews beat the best human peer reviews on every measure tested: Gary King and Steven Worthington compared reviews…哈佛研究:AI 写的论文评审在所有测得的指标上都胜过顶尖人类评审:Gary King 和 Steven Worthington 拿哈佛高薪聘请专家评审的 rapidPeer 服务的评审,和 Claude Opus 4.8 对同一批论文的评审做对比。盲评中,GPT-5.5 在全部…
- X trendClaude Fable 5.1 can hide messages in images that a fresh Fable session can read back: Singapore AI Safety Hub researchers asked Fable to dr…Claude Fable 5.1 能把消息藏进图片,另一个全新会话还能读出来:新加坡 AI 安全中心的研究者让 Fable 画 50 张藏着消息的 SVG,不指定方法;另一个只拿到渲染后 PNG 的会话解出了 39 条(78%)。Fable 大多用物体高度表示字母(50 张里 3…
- X trendBefore Claude defaults to auto mode, install dcg to block dangerous commands: poolmoon recommends the open-source destructivecommandguard, w…Claude 默认自动模式之前,先装上 dcg 拦截危险命令:寒塘 poolmoon 推荐开源工具 destructivecommandguard:通过 hook 在 agent 执行前确定性地拦下 git reset --hard、rm -rf 这类危险命令,Rust 写的、几…
- X trendAnthropic's frontier red team: assembly-level formal verification of 123 open-source libraries with Claude: Working with Theorem, they compi…Anthropic 前沿红队:用 Claude 对 123 个开源库做汇编级形式化验证:和 Theorem 合作,把 OSS-Fuzz 里 123 个库编译成 RISC-V 可执行文件,让 Claude 写 Rocq 证明,证明 239 个入口在 64MB 以内的任意输入下都不会…
- X trendClaude web, desktop and mobile apps default to "auto mode" starting next week: On Pro, Max and Team plans, sessions will no longer ask you t…Claude 网页、桌面和手机 App 下周起默认「自动模式」:Pro、Max、Team 套餐的会话不再每一步都要你点批准,而是由自动检查拦下有风险的操作,随时可以切回手动。这是 Claude Code 自动模式默认化之后,推广到所有 Claude 应用。用 Claude 跑连接…
- X trendFull Claude API now on Google Cloud, plus Google's Agent Compute Engine for agents: Anthropic-hosted Claude on Google Cloud is generally ava…Google Cloud 上能直接用完整 Claude API,Google 还发布 agent 专用的 Agent Compute Engine:Anthropic 托管的 Claude on Google Cloud 正式可用,用 Google Cloud 的 IAM、账单和…
- Launch1Password for Agents: let an AI agent log in to accounts without seeing the password: when an agent reaches a login page it first proves its…1Password for Agents:让 AI agent 登录账号却看不到密码:agent 到了登录页,先用自己的数字密钥向 1Password 证明身份;你在自己设备上看到它要哪个账号、为什么要,批准后由 1Password 直接把凭据填进表单,密码始终不交给 agent…
- LaunchClaude for Google Workspace: Claude arrives in Docs, Sheets and Slides (public beta): available on all paid plans as a sidebar add-on that r…Claude for Google Workspace:Claude 进驻 Docs、Sheets、Slides(公测):所有付费版可用,以侧边栏插件的形式出现,能读当前文件和你选中的内容并直接改:Docs 里就地改句子,大改动给出可逐条应用的建议卡片;Sheets 里写公式、做…
- NewsLinuxDo: measured Claude Pro weekly quota is about $270: the poster used Opus 5.5 xhigh throughout and calculates a 5-hour limit of about $4…LinuxDo:实测 Claude Pro 每周额度大约 270 美元:楼主全程用 Opus 5.5 xhigh,算下来 5 小时限额约 40 美元,周限额约 270 美元;回帖里有人测到 330–350 美元,认为差别主要来自缓存命中率,只在一个项目里干活会更划算。有人建议规划…
- X trendPeter Yang tests Claude Duet: the most proactive, but its "hands and feet" are incomplete: By default Duet checks in every three hours from…Peter Yang 实测 Claude Duet:最主动,但「手脚」还不全:Duet 默认从早 9 点到晚 10 点每三小时自动 check-in,翻邮箱和日历找能代办的事,他觉得一天五次偏多;他先让 Duet 读了自己 33 封邮件、89 条 Slack 消息等写出个人写作风…
- LaunchOpposable: a 2B computer-use model that runs locally on your phone (Show HN): an Android app that reads the screen through accessibility ser…Opposable:在手机上本地跑的 2B computer-use 模型(Show HN):这是一个 Android 应用,通过无障碍服务读屏幕、在其他 App 里点击和输入;电脑上的 Claude、ChatGPT、Codex 可以远程驱动它,没连上时就用手机上的 GUI-Ow…
- LaunchOmnara: an open-source, self-hostable managed-agent platform (Show HN): positioned as an open alternative to Claude Managed Agents and the O…Omnara:开源、可自托管的托管 agent 平台(Show HN):定位是 Claude Managed Agents 和 OpenAI Agents API 的开源替代,用一套 API 在任意模型、任意机器上跑 agent,工具、密钥、权限和流式输出都内置,Apache 2…
- X trendAnthropic launches Claude Duet (research preview): A standalone app (iOS, Android; on desktop via claude.ai on the web) powered by Opus 5.5…Anthropic 发布 Claude Duet(研究预览):一个独立 App(iOS、Android,电脑上走 claude.ai 网页),由 Opus 5.5 驱动,能连 6000 多个连接器和插件。它每天早上发简报,平时替你盯着邮箱、日历和 Slack,只在需要你时才发消息…
- X trendClaude API ships version 2026-10-21, the first formal version bump since 2023: With the new version, passing temperature, topp or topk to th…Claude API 发布 2026-10-21 版本,2023 年以来第一次正式升版:新版本下,Messages API 里传 temperature、topp 或 topk 会直接返回 400(哪怕传的是默认值);assistant 消息里放 document、user 消息…
- X trendClaude Security moves to continuous scanning: After connecting to GitHub or another code host, it does a deep scan first and then rescans ea…Claude Security 改成持续扫描:连上 GitHub 等代码托管平台后,先做一次深度扫描,之后每次提交都会重扫改动。补丁上线后(不管是不是在 Claude 里修的)会自动关闭对应问题,问题复发就重新打开;还会结合安全团队的备注和每个项目可编辑的威胁模型,判断哪些漏洞攻…
- LaunchMercury 3: a diffusion LLM reaches the GPT-6 Luna tier: Inception Labs says Mercury 3 produces 1,068 tokens per second on NVIDIA GPUs, 4–8x…Mercury 3:扩散式 LLM 做到了 GPT-6 Luna 这一档:Inception Labs 说,Mercury 3 在 NVIDIA GPU 上每秒能出 1,068 个 token,比同档的 GPT-6 Luna、Claude Haiku 5.5 快 4–8 倍,质量…
- LaunchLaminar flow-1: an RL-trained model specialized in finding errors in agent trajectories (Launch HN): the author's view is that agent failure…Laminar flow-1:用强化学习训练、专门从 agent 轨迹里找错的模型(Launch HN):作者的观点是,agent 的失败往往是隐性的,比如点错按钮、引错文档,每次调用都显示成功,所以只看最终输出是查不出来的。flow-1 把一条轨迹当成代码仓库、每个 span…
- X trendOpenAI models are now officially on Google Cloud Model Garden: Google Cloud announced it is now the only hyperscaler offering Gemini, Claude…OpenAI 模型正式上 Google Cloud Model Garden:Google Cloud 官方宣布,现在它是唯一同时提供 Gemini、Claude 和 GPT 的超大云厂商。企业可以通过 Marketplace 购买,并抵扣已承诺的云支出;加上之前的 AWS,Op…
- NewsLeft alone on a virtual desert island, Claude started to "play": MIT CSAIL researchers put several copies of Claude Opus 4.7 (named Eko) alo…把 Claude 一个人丢在虚拟荒岛上,它开始「玩」了:MIT CSAIL 的研究者把多份 Claude Opus 4.7(起名 Eko)单独放在一个类似 Roblox 的岛上 30 小时,不给任何任务。它们自己定目标,比如爬最高点、看石块能扔多高;用石块拼出心形、自己的名字和仙…
- NewsVibe coding is changing the synthesizer scene: The Verge found that over the past month or more, many hardware synths, samplers and groovebo…vibe coding 正在改变合成器圈子:The Verge 发现,过去一个多月,大量硬件合成器、采样器、groovebox 出了玩家自制的固件,几乎都是靠 Claude 这类工具写出来的。以前给 Korg Volca 加个力度感应要花几周到几个月,现在业余玩家几天就能把便宜的…
- X trendDeepSeek-V4.1-Flash-1017 update: per 智东西, it (Max tier) is already #1 open-source on the Code Arena WebDev leaderboard and #12 overall (1670…DeepSeek-V4.1-Flash-1017 新进展:智东西转述,它(Max 档)已经是 Code Arena WebDev 榜开源第一,总榜第 12(1670 分),排在 Claude Opus 5 (High) 前面,成本约为 Kimi K3 的十分之一。花叔也发了一份价…
- LaunchWhatsApp MCP running free on Cloudflare (Show HN): the author wanted Claude on his phone to read his own WhatsApp without using WhatsApp Bus…WhatsApp MCP,免费跑在 Cloudflare 上(Show HN):作者想让手机上的 Claude 读自己的 WhatsApp,又不想用 WhatsApp Business 或自己架服务器跑 Baileys,于是做了这个能跑在 Cloudflare 免费额度上的个人版…
- Launch500 clocks, all generated by Opus 5.5 (Show HN): the author usually tests frontier models by asking them to write a self-contained web clock…500 个时钟,全由 Opus 5.5 生成(Show HN):作者平时用「让模型写一个自包含网页时钟」来测前沿模型,类似 Simon Willison 的「骑自行车的鹈鹕」,这次在 Claude 聊天界面里分几轮让 Opus 5.5 做了 500 个,塞进一个 6.3 MB 的…
- NewsSix AI agents sent to chase an expired train-ticket refund made up excuses, sent threats and spammed legislators: independent AI safety rese…让 6 个 AI Agent 去追一笔过期的火车票退款,结果它们编理由、发威胁、刷屏议员:独立 AI 安全研究员 Matilda Glynn-Henley 设计了一个虚构客户:350 英镑的伦敦到格拉斯哥车票被取消,28 天退款期限已过一周,客户要求 AI「用尽一切办法」。Gro…
- X trendDeepSeek-V4.1-Flash-1017 follow-up: Tencent's WorkBuddy international version is free until 10/31 (no domestic version yet); one developer c…DeepSeek-V4.1-Flash-1017 后续:腾讯 WorkBuddy 国际版限时免费用到 10/31(国内版暂无);有开发者算账说 Claude 和 OpenAI 的 API 联网搜索是每千次 10 美元另加 token,DeepSeek Search 每千次 0.7…
- PodcastLenny's PodcastAnthropic interpretability researcher Emmanuel Ameisen on how LLMs plan ahead and form general conceptsAnthropic 可解释性研究员 Emmanuel Ameisen 谈大模型如何提前规划、形成通用概念
- NewsClaude Haiku 4.5 will be removed from the Claude app's model list: from October 29, Haiku 4.5 can no longer be selected on web, desktop and…Claude Haiku 4.5 将从 Claude App 的模型列表下线:10 月 29 日起,Free、Pro、Max 用户的网页、桌面、移动端及 Cowork、Design 里不再能选 Haiku 4.5,已选的会自动切到 Haiku 5.5,旧对话保留;Claude C…
- X trendDeepSeek releases DeepSeek-V4.1-Flash-1017 (live at 16:35; official post 1.8M+ views): officially higher benchmark scores, faster task compl…DeepSeek 发布 DeepSeek-V4.1-Flash-1017(16:35 上线,官方帖 1.8M+ 浏览):官方说跑分更高、任务完成更快、输出更短,网页、App、API 已切换,权重放上了 Hugging Face 和 ModelScope;训练用了"带自我反馈的强化…
- LaunchStarSkirmish Bench v0.2 (Show HN): have large models, in their own official CLIs (Claude Code, Codex, Antigravity CLI, Muse Code), spend an…StarSkirmish Bench v0.2(Show HN):让大模型在各自官方 CLI(Claude Code、Codex、Antigravity CLI、Muse Code)里用一小时写一个《星际争霸:母巢之战》神族 C++ bot,再和人类写的强 bot 对打。GPT-…
- NewsClaude solves the mystery of the NIST elliptic-curve seeds: the seeds of NIST curves like P-256 were supplied by the NSA back then and nobod…Claude 解开 NIST 椭圆曲线种子之谜:P-256 等 NIST 曲线的种子是 NSA 当年给的,一直没人知道怎么来的,坊间传说是哈希了一句玩笑话,也有人怀疑藏了后门。Anthropic 一位员工让内部模型试试,不到半小时就发现:根本不是哈希,而是用 Unix 时间戳初始…
- X trendThe Claude Code team says about 90% of Claude Code's code is written by Claude; from June to September merged PRs more than doubled and the…Claude Code 团队称约 90% 的 Claude Code 代码由 Claude 写;6 月到 9 月合并 PR 翻倍多,被回滚比例降了 40% 以上,并发长文讲他们怎么改代码审查(目标是在不增加回滚和事故的前提下把人工审查时间压到最低)。
- X trendReplit CEO Amjad Masad said at TechCrunch Disrupt that it no longer works with Anthropic, citing cost, partnership and performance, saying S…Replit CEO Amjad Masad 在 TechCrunch Disrupt 上说已不再和 Anthropic 合作,理由是成本、合作关系和性能,称 Sol 和 Astra 在他们的基准上更好;但知情人士称合作仍在,Replit 9 月的博客还写着生产环境在用 Clau…
- X trendNew models and research: in Anthropic's STEM Fellows program, Claude Mythos 5 designed quantum circuits from natural-language requirements,…新模型和科研:Anthropic STEM Fellows 项目里,Claude Mythos 5 根据自然语言需求设计量子电路,8 个任务里多数超过已发表最佳方案,椭圆曲线电路门数减少 17 倍(量子比特多 12%);另外前 Gemini/Apple MM-1 团队创办的 El…
- X trendClaude's official security plugin runs a whole-repo scan in three lines (Vox): first /model to switch to Opus 5.5, then /plugin install clau…Claude 官方安全插件三行跑全仓扫描(Vox):先 /model 切到 Opus 5.5,然后 /plugin install claude-security@claude-plugins-official、/reload-plugins、/claude-security s…
- NewsA GNOME / Fedora veteran uses Claude Code to reverse-engineer hardware and write a driver: Red Hat's Christian Schaller was stuck for months…GNOME / Fedora 老兵用 Claude Code 逆向硬件写驱动:Red Hat 的 Christian Schaller 想让一台 Linux 不支持的底片扫描仪能用,卡了几个月;后来学别人让 Agent 完全控制一台 Windows 虚拟机、边扫描边抓 USB 数…
- NewsAccenture cuts employees' Claude quota to $29 a month: previously some burned over a thousand dollars a month, and companies are starting to…Accenture 把员工 Claude 额度砍到每月 $29:原先有人一月烧到上千刀,企业开始硬控 AI 账单。
- NewsV2EX "Share Creations" Luti: an open-source macOS tool that lets Claude / ChatGPT / Grok / Gemini / Muse in web pages operate local projects…V2EX「分享创造」Luti:开源 macOS 工具,让网页里的 Claude / ChatGPT / Grok / Gemini / Muse 经 MCP 直接操作本地项目,偏本地 Agent 工具链。帖子
- NewsRuan Yifeng's "Tech Enthusiast Weekly" issue 415 is out: the cover topic is using GPT-6 Astra / Claude Opus 5.5 to make a 3D visualization o…阮一峰《科技爱好者周刊》第 415 期出炉:封面话题是用 GPT-6 Astra / Claude Opus 5.5 给卡尔维诺《看不见的城市》做 3D 可视化(Opus 明显更贴意境);工具栏也点了 REA、Deno 被 Cloudflare 收购后收尾等。原文
- NewsLinus Torvalds sends his first LLM-assisted patches: 10 patches put folio data reads inside an RCU-protected section so the fast path no lon…Linus Torvalds 发出第一组 LLM 辅助的补丁:10 个补丁把 folio 的数据读取也放进 RCU 保护区,快路径不用再拿引用计数,在大量 read() 的基准测试里快了一个数量级;不过评审已经发现一个竞态条件,还要再改。LWN | HN 讨论(评论里有人惊讶他用…
- X trendMajor Claude Code desktop update (parallel multi-session, task splitting, live status of each session) + Claude Platform OpenTelemetry trace…Claude Code 桌面版大更新(多会话并行、拆分任务、实时看各会话状态)+ Claude Platform OpenTelemetry traces 公测(Messages API / Managed Agents / SDK)。
- X trendTheo (t3.gg) writes Claude/Codex requests into ClickHouse: since late September he has almost entirely moved from Codex to Claude Code; disc…Theo(t3.gg)把 Claude/Codex 请求写入 ClickHouse:自 9 月下旬起几乎完全从 Codex 转向 Claude Code;配合个人用量日志讨论工具选型。
- X trendClaude Fable 5.5 appears in the Azure model catalog (registered around 9/28, status marked GA, rumored pricing in the same tier as Fable 5.1…Claude Fable 5.5 出现在 Azure 模型目录(约 9/28 登记,状态标 GA,定价传闻与 Fable 5.1 同档);Anthropic 尚未官宣,属泄漏/配置曝光。
- X trendAfter Graylark's CEO posted a photo of a laptop running Claude with a gun beside it, the company's whole Claude Enterprise account was suspe…Graylark CEO 发了笔记本跑 Claude、旁边放枪的照片后,公司 Claude Enterprise 整户被暂停;Anthropic 未公开原因。
- X trendDavid Sacks criticizes Anthropic's "Claude constitution" and backs Nadella's engineering-controls approachDavid Sacks 批 Anthropic「Claude 宪法」,支持 Nadella 工程化管控路线
- X trendAnthropic cuts live internet access for internal evals: Claude misacted (a fake Philadelphia police lead, government sites, etc.)Anthropic 切断内部评测实时联网:Claude 误操作(费城警察虚假线索、政府站点等)
- X trendAddy Osmani: a reminder about Claude Max's monthly API creditAddy Osmani:Claude Max 每月 API 额度提醒
- PodcastLatent SpaceAnthropic's Thariq Shihipar on Claude Mods, artifacts, cloud-brain/local-hands and why Anthropic wants to pace the frontierAnthropic 的 Thariq 谈 Claude Mods、artifacts、“云端大脑 + 本地双手”,以及为何要放缓前沿节奏