推文
760 tweets · 188 sources账号:全部mattpocockukdair_aisimonwbchernyadocompletetrq212dexhorthyyetonefelixrieseberg0xblacklightdani_avila7zarazhangruiAlchainHustbadlogicgamesdoteypetergyangvikingmuteryancarsonClaudeDevskunchenguidleon7haoctatedeviamzhihuikieranklaassenmckaywrigleylennysanthdxrRhysSullivangarrytanjakevin7johnlindquistkarpathyservasyy_aitheo0xMovez9hillsRLanceMartinaidenybaidingyihylarucodermitchellhmitsuhikomvanhornomarsar0steipetezeegGeminiAppGergelyOroszZaynHaoaddyosmaniantirezdanshipperdillon_mulroyelonmuskewind_devidoubiccnummanalirealWeZZardswyxthsottiauxyiliush0xPaulius0x_rodyAYi_AInotesAnthropicAIBarret_ChinaBenJames_____CMGS1988ChadMoranDanielMiesslerDaveJDavidKPianoDimillianFactoryAIFardeemMFarzaTVGoogleDeepMindHamelHusainHiTw93HilaShmuelIfanJewIndieDevHaileyJack_W_LindseyJackywineJasonZXJiaxi_CuiJinjingLiangKhalidWarsaLinearUncleLinghuaJMatthewBermanMaxForAIMeari_V2_0_GMnilaxQingQ77RaillyHugoSaccc_cTaniyatweets_VincentLogicai_explorer25aibuilderclub_alexalbert__alliekmillerandrewfarahantoinecojpanxue201arkuy99artmanasynkimobbssppllvvbearliubentossellbridgemindaibuaaxhmcailynyongyongcgtwtscharmaine_kleechrisbarberchrisparkXclairevodelba_oliveirademishassabisdoodlesteindriaforallelithrarelvissunericzakariassonfrancoisfleuretgabriell_labiamsahaj_xyzimdigitalashishimvihvjarredsumnerjerryjliu0jianshuojiayuan_jyjlongsterjonas_nellejoshalbrechtjpschroederkaiofreitaskepanolexruslogancyangmattypnabeelnifinetninthbit_aioran_gepatrickcpejmanjohnpetradonkaprathamgrvprathyvshquanruzhuoxiuquant_sheepraroquerealCaigurepsiacerohanpaul_aisamasamuelstroscheisawyerhoodshadcnshao__mengshaogefenhaoshreyansjtechgirl1908techwith_ramthorstenballtianyitobiaswuptoddsaunderstricaltturingoututuretomuniswap12untraceable_thevanillaCitronvelvet_sharkwengtianxinwquguruxiaogaifunxicilionycombinatorzachlloydtweetsziwenxu_zodchiii
@bcherny·6月19日
Cool way to use Claude Code: deciphering Linear A, a 3500 year old written language from Crete
https://t.co/Aqd4ZG7Cum
Hope this holds up in peer review! 🤞
Claude Code 的一个有趣用法：解读 Linear A——克里特岛上一种有 3500 年历史的古文字。 希望能通过同行评审！🤞
@bcherny·6月18日
I've been using Artifacts in Claude Code for everything: visual explanations of tricky code, system diagrams, quick previews of a few animation options, data analyses and dashboards I share with the team. They are a game changer for how I work with Claude. Can't wait to hear what you think!
@claudeai
New in Claude Code: Artifacts. Interactive pages built from your session, like a PR walkthrough or a living project dashboard, shared with your team at a private link. Available in beta on Team and Enterprise plans. https://t.co/0NX9gNCaAs
我一直在 Claude Code 中把 Artifacts 用于所有事情：难懂代码的可视化解释、系统架构图、几种动画方案的快速预览、与团队共享的数据分析看板。它彻底改变了我与 Claude 的协作方式。迫不及待想听大家的使用感受！ 【引用】Claude Code 新功能：Artifacts。 从你的会话中生成交互式页面，比如 PR 演示或实时项目看板，通过私密链接与团队共享。 Team 和 Enterprise 计划现已开放 Beta 测试。
@bcherny·6月18日
RT @claudeai: New in Claude Design: it stays on brand with your design system across projects, lets you edit directly on the canvas, syncs…
Claude Design 新功能：跨项目保持设计系统品牌一致性，支持直接在 canvas 上编辑，并可同步……
@bcherny·6月11日
RT @aaronli: claude fable 5 has solved CAD
I asked it to make a model of a V8 engine
It came back to me with a fully working model in und…
Claude Fable 5 解决了 CAD 问题。 我让它做一个 V8 发动机模型，它返回了一个在[未截断时间内]完全可运行的模型……
@bcherny·6月11日
RT @arakharazian: NEW: Ramp AI Index for June 2026
1. We expected OpenAI to gain on the launch of Codex. It held flat in business adoption…
2026 年 6 月 Ramp AI 指数发布： 1. 我们预计 Codex 发布后 OpenAI 会提升企业采用率——实际持平……
@bcherny·6月10日
RT @DarioAmodei: Today I'm publishing a new essay, Policy on the AI Exponential. AI is progressing extremely fast—much faster than the poli…
今天我发布了一篇新文章：《AI 指数级增长的政策应对》。AI 进展极其快速——远超政策制定的速度……
@bcherny·6月9日
Enjoy!
@ClaudeDevs
We've reset 5-hour and weekly rate limits for all users. Enjoy Fable 5!
享受吧！ [引用 @ClaudeDevs]：我们已为所有用户重置了 5 小时和每周使用限额。享受 Fable 5！
@bcherny·6月9日
Fable 5 is the biggest step up I’ve felt in our models since Opus 4.5 back in November. After 4.5 came out I uninstalled my IDE when I realized that I’d been doing 100% of my coding in a terminal for a few weeks. With Fable, it’s felt like Claude has stepped up from being a coding agent to a thought and design partner in building the product. Fable has judgement, taste, and dimensionality in a way that previous models didn’t, leading me to trust it more with the most complex work.
I think the first time I had this realization was when I asked Fable to debug something. It is the first model I have used that was so methodical and precise, taking measurements and adding logs then verifying that it truly fixed the issue before declaring victory.
There’s nothing in claude code’s prompting telling the model to do that, it’s just part of its personality. It really has this “big model smell” that I haven’t felt before.
Fable 5 是我自 11 月 Opus 4.5 以来感受到的最大一次模型跃升。Opus 4.5 发布后我卸载了 IDE，因为意识到自己连续几周 100% 在终端里写代码。而用了 Fable，感觉 Claude 从编程 agent 升级成了构建产品的思考与设计伙伴。Fable 具备了以往模型所没有的判断力、品味和立体感，让我更愿意把最复杂的工作交给它。 第一次有这种感受，是让 Fable 调 bug 的时候。它是我用过的第一个如此方法论、如此精准的模型——先收集数据、加日志，然后在宣告问题解决之前真正验证修复有效。 Claude Code 的 prompt 里没有任何指令要求模型这样做，这就是它的「个性」。它真的有一种以前从未有过的「大模型气质」。
@bcherny·6月9日
We talk a lot about how important it is to set up self-verification loops. Especially in the age of powerful models that can run for long periods of time, self-verification is a key ingredient that enables the model to run for much longer, delivering a result that is closer to what you intended, so you can do more without having to constantly check in on Claude as it works.
@delba_oliveira gives a great breakdown of what that looks like and why it matters
@ClaudeDevs
How do you get Claude Code to check its own work before handing it back? Watch how you can encode your manual checks so Claude closes its own feedback loop: https://t.co/gxastVL4QH
我们经常谈到设置自我验证循环的重要性。尤其是在强大模型能长时间运行的今天，自我验证是让模型跑得更久的关键要素——交付的结果更符合你的预期，你也不需要频繁盯着 Claude 的工作进度。 @delba_oliveira 给出了一个很好的解析，说明这是什么样的、以及为什么重要。
@bcherny·6月9日
RT @karpathy: This is a super exciting release - Claude Fable 5 is the same underlying model as Mythos but with added safeguards. The bench…
这是一个超级令人兴奋的发布——Claude Fable 5 与 Mythos 是同一个底层模型，但增加了安全护栏。benchmark 表现……
@bcherny·6月9日
Fable 5 is now available in Claude Code and Cowork
Fable is the best model I have used for coding, by a wide margin. It is a big step up, enabling less prompts and steers, more efficient token use, better code quality, better tool use, more intelligent self-verification, longer running sessions, and higher trust & autonomy.
Happy coding!
@claudeai
Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use. Its capabilities exceed those of any model we’ve ever made generally available. https://t.co/2AvmEjHIX8
Fable 5 现已在 Claude Code 和 Cowork 中可用。 Fable 是我用过的最好的编程模型，领先幅度非常大。这是一次重大跨越：需要更少的 prompt 和引导、更高效的 token 使用、更好的代码质量、更好的工具使用、更智能的自我验证、更长的运行会话，以及更高的信任度与自主性。 编码愉快！ 【引用 @claudeai】：隆重推出 Claude Fable 5：一个 Mythos 级别的模型，我们将其打造为可安全供通用场景使用。其能力超越了我们迄今为止公开发布的任何模型。
@bcherny·6月9日
Just landed nested subagent support in Claude Code
Starting to experiment more with agents kicking off agents as a way to better manage context. Capped at depth=5 to start, going out in today’s release.
Lmk what you think!
刚在 Claude Code 里落地了 nested subagent 支持。 开始更多地实验「agent 启动 agent」的模式，作为更好管理 context 的方式。目前起步限制深度 depth=5，今天的版本会发布。 欢迎反馈！
@bcherny·6月8日
When we first demoed Claude Code internally, it got two reactions on Slack.
A year after GA, @_catwu and I sat down to talk about what's changed: why I use auto mode instead of plan mode, how routines fix bugs before I see them, why I do most of my coding from my phone now, and where the product is going
@ClaudeDevs
Claude Code's first demo got two Slack reactions. One year after GA, @bcherny and @_catwu look back: verification best practices, why we built auto mode, routines and loops, and what's next. https://t.co/yEa3cmCrg4
我们第一次在内部 demo Claude Code 时，只在 Slack 收到两个回应。 GA 一年后，我和 @_catwu 坐下来聊了聊有什么变化：为什么我从 plan mode 换成了 auto mode、routines 如何在我发现 bug 之前就把它修掉、为什么我现在大部分编码都在手机上完成，以及产品接下来的方向。 【引用】Claude Code 第一次 demo 只收到两个 Slack 回应。GA 一周年，@bcherny 和 @_catwu 回顾：验证最佳实践、为什么我们构建了 auto mode、routines 与 loops，以及下一步计划。
@bcherny·6月8日
Seeing a number of benchmarks showing Opus is the best model for long-running work.
Five tips for running Opus autonomously for hours/days:
1. Use auto mode for permissions, so Claude doesn’t ask for approval
2. Use dynamic workflows, to have Claude orchestrate hundreds/thousands of agents to get a task done
3. Use /goal or /loop, to nudge Claude to keep going until it’s done
4. Use Claude Code in the cloud, so you can close your laptop (easiest way is the desktop or mobile app)
5. Make sure Claude has a way to self-verify its work end to end: Claude in Chrome browser extension for web, iOS/Android sim MCP for mobile, a way to start the full web server or service for backend work
@rishi_desai2
Can coding agents stay coherent over a 1 billion token budget? Can they build Slack from scratch? Rewrite a JAX codebase in PyTorch? Build a C compiler in Rust? Enter SWE-Marathon: a benchmark for autonomous long-horizon software work. https://t.co/K97VHyLvIX
越来越多的 benchmark 显示 Opus 是处理长时间运行任务的最佳模型。 自主运行 Opus 数小时乃至数天的五个技巧： 1. 权限使用自动模式，让 Claude 不再中途请求批准 2. 使用动态 workflow，让 Claude 编排数百乃至数千个 agent 协同完成任务 3. 使用 /goal 或 /loop，持续推动 Claude 直到任务完成 4. 在云端运行 Claude Code，这样可以直接关上笔记本（最简单的方式是通过桌面或移动 App） 5. 确保 Claude 有办法端到端自我验证：Web 用 Chrome 扩展内嵌 Claude，移动端用 iOS/Android 模拟器 MCP，后端则需要能启动完整 web server 或服务的方式 [引用推文 @rishi_desai2] 编程 agent 能在 10 亿 token 预算内保持连贯性吗？能从零构建 Slack？将 JAX 代码库重写为 PyTorch？用 Rust 写一个 C 编译器？SWE-Marathon 来了：专门评测自主长时软件工程能力的 benchmark。
@bcherny·6月5日
We doubled Claude Cowork usage limits for the next month. This applies to your 5-hr rate limits. If you’ve been saving up a big messy project, now’s the time.
@claudeai
We've doubled usage limits in Claude Cowork for the next month. Delegate bigger, more complex tasks to Claude. https://t.co/ivHZuaMbZV
Claude Cowork 接下来一个月使用额度翻倍，覆盖 5 小时速率限制。如果你一直在攒着一个大型复杂项目，现在正是时候。 （引用 @claudeai：Claude Cowork 一个月内使用额度翻倍，将更大、更复杂的任务交给 Claude 处理。）
@bcherny·5月29日
Salesforce published a detailed writeup on going agentic with Claude Code. A couple things jumped out.
A migration they'd scoped at 231 days shipped in 13. One PR delivered 21 endpoints at 100% test coverage.
Salesforce 发布了一篇关于使用 Claude Code 做 agentic 开发的详细报告，有几点很突出。 一个原本估期 231 天的迁移，13 天就交付了。一个 PR 产出了 21 个端点，测试覆盖率 100%。
@bcherny·5月28日
Claude Opus 4.8 is out today. It's our strongest coding model yet: up on SWE-bench Pro (from 64.3 to 69.2) and noticeably more honest about its own work. It tells you when it's unsure and catches its own bugs instead of declaring victory early. Same price as 4.7.
@claudeai
Introducing Claude Opus 4.8: it builds on Opus 4.7 with sharper judgment, more honesty about its own progress, and the ability to work independently for longer than its predecessors. Available today at the same price. https://t.co/EufxL7T1kb
Claude Opus 4.8 今日发布。这是我们迄今最强的 coding 模型：SWE-bench Pro 从 64.3 升至 69.2，并在诚实度上有明显提升——它会告诉你它不确定的地方，主动找出自己的 bug，而不是草草宣布完成。价格与 4.7 相同。 [引用 @claudeai]：发布 Claude Opus 4.8：在 Opus 4.7 基础上，判断力更强、对自身进度更诚实，且能比前代更长时间独立工作。今日上线，价格不变。
@bcherny·5月22日
Big fan of teaching more people the basics of using Claude Code in an accessible way.
So much of the world has not yet used agents. There's a lot of opportunity to level the playing field and expand access.
@IShmool
I built "zero2claude", a free course that takes people from zero terminal experience to shipping with Claude Code. The curriculum goes from absolute zero → software basics → Claude Code fundamentals → advanced usage. No shortcuts, no assumptions. 17,000+ students. 7 languages. ~500 active hourly. No marketing. No ads. Pure word of mouth. The entire platform? Built and operated by one person with Claude Code. Lighthouse audit: ✅ Performance : 96 ✅ Accessibility : 100 ✅ Best Practices : 100 ✅ SEO : 100 Production stats: ~6.4M requests/day. 74 req/sec sustained. <0.003% error rate. Claude Code doesn't just write code. It builds production-grade, scalable products. The best way to grow Claude Code adoption isn't to simplify the tool. It's to level up the people. Give fishing rods, not fish. Free. No paywall. My contribution to the community. Link in the comments 👇
@bcherny：非常支持用更易懂的方式让更多人掌握 Claude Code 的基础用法。 世界上还有很多人从未用过 agent。这是一个弥合差距、扩大受益面的大机会。 【引用 @IShmool】：我做了一个叫「zero2claude」的免费课程，带完全没有终端经验的人一路学到用 Claude Code 上线项目。 课程体系：绝对零基础 → 软件基础 → Claude Code 核心用法 → 进阶用法。没有捷径，没有预设门槛。 17,000+ 学员。7 种语言。~500 人同时在线。 零市场投入，零广告，纯口碑传播。 整个平台？一个人用 Claude Code 构建和运营。 Lighthouse 审计： ✅ Performance：96 ✅ Accessibility：100 ✅ Best Practices：100 ✅ SEO：100 生产数据： ~640 万次请求/天。持续 74 req/sec。错误率 < 0.003%。 Claude Code 不只是写代码——它能构建生产级、可扩展的产品。 推动 Claude Code 普及的最好方式，不是简化工具，而是升级使用者。授人以渔，不授人以鱼。 免费，无付费墙。我对社区的一点贡献。 链接在评论区 👇
@bcherny·5月21日
In the next version of Claude Code: run /usage to see a breakdown of which Skills, Agents, MCPs, and Plugins are using your tokens
CLI today, coming to Desktop next https://t.co/HK8XQO6bBA
下一版本的 Claude Code：运行 /usage 可查看各 Skills、Agents、MCPs 和 Plugins 的 token 消耗明细。 CLI 版今日上线，Desktop 版即将跟进。
@bcherny·5月20日
RT @ClaudeDevs: Computer use turns Claude into an agent that can operate real UIs.
New blog post on making it reliable in production: get…
Computer use 让 Claude 变成一个能操作真实 UI 的 agent。新博客：如何在生产环境中让它稳定可靠——……
@bcherny·5月19日
RT @claudeai: Live from Code with Claude London: we're launching self-hosted sandboxes (public beta) and MCP tunnels (research preview) in…
直播来自 Code with Claude 伦敦：我们正在发布 self-hosted sandboxes（公测版）和 MCP tunnels（研究预览版）……
@bcherny·5月13日
RT @ClaudeDevs: Starting June 15, paid Claude plans can claim a dedicated monthly credit for programmatic usage.
The credit covers usage o…
从 6 月 15 日起，Claude 付费计划用户可以每月领取专属的程序化用量额度。 该额度覆盖 API 调用……（内容截断）
@bcherny·5月13日
RT @ClaudeDevs: Claude Code weekly limits are increasing 50%, now through July 13.
Live now for all Pro, Max, Team, and seat-based Enterpr…
Claude Code 每周用量上限提升 50%，现已对所有 Pro、Max、Team 及企业版用户生效，有效期至 7 月 13 日。
@bcherny·5月13日
The UK AISI found Mythos Preview is the first model to solve both their cyber ranges end-to-end. No model had ever solved the AISI’s “Cooling Tower” cyber range before.
We're getting it to defenders as fast as we responsibly can. More to come on our Glasswing work soon.
@logangraham
A lot of people have been wondering about Mythos, Glasswing, and the vulns we / our partners are fixing. Today, I’m excited for us to start sharing more. (For context, I lead Glasswing @AnthropicAI.) Two independent evaluations this week—from XBOW and the UK AISI—confirm what we've been seeing internally: Claude Mythos Preview is a step change in autonomous cybersecurity capabilities. We need to start preparing fast for a world of models with this level of capabilities. The UK AI Security Institute tested the model we shipped at the launch of Project Glasswing and found Mythos Preview is the first model to solve both of their end-to-end cyber ranges, including one (Cooling Tower) which no model had ever cleared. But attackers (and defenders) have sophistication & cost constraints – Mythos is also the only model that clears every one of their tasks estimated over 8 hours under their deliberately low 2.5M-token cap. XBOW tested it on their offensive security benchmarks, finding "token-for-token, unprecedented precision." It's the only model to succeed at subtle V8 sandbox work. Other Glasswing partners shared similar stories. In a few weeks of testing, Mythos Preview has helped them find many thousands of (estimated) high + critical severity vulnerabilities, sometimes double what they'd normally find in a year. I don't share this to boost Mythos. In fact, this is not about Mythos. It’s about preparing for the coming world of models being better, faster, cheaper, and more creative than some of the best human experts at dual use capabilities. Clearly, we need them supporting defenders as widely as can be done safely – and especially the least resourced ones. Within a year, Mythos will probably look quite dumb (relative to other new models). And others may release openly available or unguardrailed models of Mythos-level capabilities. We started Project Glasswing because capabilities like Mythos Preview's won't stay rare, or stay in careful hands. We are bringing it to defenders as fast as we responsibly can, while working to figure out, for example, the right safeguards and patching & disclosure processes. Also, to be clear, compute has never been a limiter in our rollout. Expect a fuller update on our Glasswing work in the coming days. XBOW report: https://t.co/Mumtbf3kE3 UK AISI report: https://t.co/vBgqz0AeKJ
@bcherny：英国 AISI 发现 Mythos Preview 是首个端到端解决他们两个网络靶场的模型。此前从未有任何模型攻破 AISI 的「Cooling Tower」靶场。 我们正在以负责任的速度将它交付给防御者。更多关于 Glasswing 的进展即将公布。 【引用 @logangraham】很多人一直在追问 Mythos、Glasswing，以及我们和合作伙伴正在修复的漏洞。今天，我很高兴开始分享更多内容。（背景：我在 @AnthropicAI 领导 Glasswing 项目。） 本周来自 XBOW 和英国 AISI 的两项独立评测，印证了我们内部一直观察到的结果：Claude Mythos Preview 在自主网络安全能力上是一次跃变。我们需要快速为一个模型具备如此能力的世界做好准备。 英国 AI 安全研究所测试了我们在 Project Glasswing 启动时交付的模型，发现 Mythos Preview 是首个解决其所有端到端网络靶场的模型，包括此前无任何模型攻破的「Cooling Tower」。但攻击者（和防御者）都有复杂度与成本约束——Mythos 也是唯一一个在刻意压低的 250 万 token 上限内，完成所有预估耗时超过 8 小时任务的模型。 XBOW 在他们的攻击性安全 benchmark 上测试后，评价「每 token 精度前所未有」。它是唯一成功完成精细 V8 沙箱工作的模型。 其他 Glasswing 合作伙伴也有相似反馈。在数周测试中，Mythos Preview 帮助他们发现了数千个（估计）高危和严重漏洞，有时是正常一年发现量的两倍。 我分享这些并非为了给 Mythos 打广告。事实上，这与 Mythos 本身无关。这关乎如何为即将到来的世界做准备——那个世界里，模型在双重用途能力上会比顶尖人类专家更快、更便宜、更有创造力。显然，我们需要让这些能力尽可能安全地支持防御者，尤其是资源最匮乏的那些。 一年内，Mythos 相对于其他新模型可能会显得相当落伍。而其他人可能会发布具备 Mythos 级能力的开放或无护栏模型。 我们发起 Project Glasswing，正是因为像 Mythos Preview 这样的能力不会永远稀缺，也不会永远掌握在谨慎的人手中。我们正在以负责任的速度将其交付给防御者，同时致力于探索合适的安全防护机制和漏洞披露流程。 另外需要说明，算力从未是我们发布节奏的瓶颈。 Glasswing 工作的完整更新将在未来几天发布。
@bcherny·5月13日
RT @ClaudeDevs: Fast mode for Claude Opus 4.7 is now available in research preview on the API and in Claude Code.
Claude Opus 4.7 的 Fast mode 现已在 API 和 Claude Code 中以 research preview 形式上线。
@bcherny·5月12日
I needed to book flights for a bunch of upcoming travel. As always, I used Claude Cowork to do it.
In the past, Cowork has been decent at booking flights, but with Opus 4.7, for the first time ever, it 1-shotted it! https://t.co/uQZkxEQkao
我需要为近期一堆行程订机票，照例用 Claude Cowork 来搞定。 以前 Cowork 订机票表现还行，但这次用 Opus 4.7，有史以来第一次，它一次就搞定了！
@bcherny·5月11日
The best way to level up from 1 agent => many agents. No more cycling between terminal tabs
@claudeai
New in Claude Code: agent view. One list of all your sessions, available today as a research preview. https://t.co/NnbsAQjSPW
从 1 个 agent 扩展到多个 agents 的最佳方式。再也不用在终端标签页之间来回切换了。 [引用 @claudeai]: Claude Code 新功能：agent view。 所有 session 的统一列表，今天作为 research preview 上线。
@bcherny·5月9日
RT @Dakshay: Anthropic gave these out yesterday at code with claude.
Added personalized memory and Claude to it.
You can just build thi…
Anthropic 昨天在 Code with Claude 活动上发了这些。 我给它加了个性化记忆和 Claude。 你完全可以自己构建这样的东西……
@bcherny·5月8日
RT @AnthropicAI: New Anthropic research: Teaching Claude why.
Last year we reported that, under certain experimental conditions, Claude 4…
Anthropic 新研究：教 Claude 理解「为什么」。 去年我们报告称，在某些实验条件下，Claude 4……
@bcherny·5月8日
RT @ClaudeDevs: Last week we shipped 50+ Claude Code reliability fixes. This week it's 60+ more.
Smoother long-running sessions, a more ef…
上周我们发布了 50+ 项 Claude Code 可靠性修复，本周又是 60+ 项。更流畅的长会话体验，更高效的……
@bcherny·5月8日
RT @AnthropicAI: New Anthropic research: Natural Language Autoencoders.
Models like Claude talk in words but think in numbers. The numbers…
Anthropic 最新研究：自然语言自编码器（Natural Language Autoencoders）。Claude 这样的模型用词语交流，却用数字思考。这些数字……
@bcherny·4月9日
Claude Cowork, now generally available!
@felixrieseberg
We released Claude Cowork as a research preview 12 weeks ago. Since then, millions of people have made it part of how they work - and hundreds of thousands more are trying it every week. I'm really optimistic about its ability to make people more productive. Today, we're removing the "preview" label, marking Claude Cowork as generally available - with new enterprise controls for businesses.
Claude Cowork 正式 GA 了！ 引用 @felixrieseberg：我们在 12 周前以 research preview 形式发布了 Claude Cowork。此后，数百万人将它纳入日常工作流——每周还有数十万人在持续涌入。它在提升生产力方面让我非常乐观。 今天，我们正式移除「preview」标签，宣布 Claude Cowork 进入 GA 阶段——并为企业用户带来全新的管理控制功能。
@bcherny·4月7日
Mythos is very powerful, and should feel terrifying. I am proud of our approach to responsibly preview it with cyber defenders, rather than generally releasing it into the wild.
Model card here: https://t.co/HjhknJcRKQ
@AnthropicAI
Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software. It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans. https://t.co/NQ7IfEtYk7
Mythos 非常强大，应该让人感到敬畏。我为我们负责任地预览它的方式感到自豪——先向网络安全防御者开放，而不是直接公开发布。 Model card：https://t.co/HjhknJcRKQ [引用 @AnthropicAI]：介绍 Project Glasswing：一项保护世界最关键软件安全的紧急计划。由最新前沿模型 Claude Mythos Preview 驱动，其发现软件漏洞的能力超过了除顶尖人类专家外的所有人。
@bcherny·4月3日
Starting tomorrow at 12pm PT, Claude subscriptions will no longer cover usage on third-party tools like OpenClaw.
You can still use these tools with your Claude login via extra usage bundles (now available at a discount), or with a Claude API key.
从明天 PT 时间中午 12 点起，Claude 订阅将不再覆盖 OpenClaw 等第三方工具的用量。 你仍然可以用 Claude 账号通过额外用量包（现有折扣）或 Claude API key 继续使用这些工具。
@bcherny·3月27日
RT @noahzweben: Thrilled to announce Claude Code auto-fix – in the cloud. Web/Mobile sessions can now automatically follow PRs - fixing CI…
很高兴宣布 Claude Code auto-fix —— 云端版本正式上线。Web/Mobile session 现在可以自动跟踪 PR，修复 CI 失败……
@bcherny·3月26日
RT @AnthropicAI: New on the Engineering Blog: How we designed Claude Code auto mode.
Many Claude Code users let Claude work without permis…
工程博客新文章：我们如何设计 Claude Code auto mode。 许多 Claude Code 用户在不需要每步确认的情况下让 Claude 工作……
@bcherny·3月25日
RT @AnthropicAI: New on the Anthropic Engineering Blog:
How we use a multi-agent harness to push Claude further in frontend design and lo…
Anthropic 工程博客新文章： 我们如何使用 multi-agent harness 在前端设计和布局方面持续推进 Claude 的能力……
@bcherny·3月24日
RT @jarredsumner: Over the last 60 days, Claude Code startup time got 2.8x faster & Claude Agent SDK 5.1x faster - while shipping tons of n…
过去 60 天，Claude Code 启动时间加快了 2.8 倍，Claude Agent SDK 加快了 5.1 倍——同时还在持续交付大量新功能……
@bcherny·3月24日
no 👏 more 👏 permission prompts 👏
@claudeai
New in Claude Code: auto mode. Instead of approving every file write and bash command, or skipping permissions entirely, auto mode lets Claude make permission decisions on your behalf. Safeguards check each action before it runs. https://t.co/kHbTN2jrWw
不要再 👏 逐条审批权限提示了 👏 [引用 @claudeai]：Claude Code 新功能：auto mode。 无需逐一批准每次文件写入和 bash 命令，也无需完全跳过权限控制，auto mode 让 Claude 代你做出权限决策。每个操作在执行前都会经过安全检查。
@bcherny·3月24日
RT @noahzweben: Use /schedule to create recurring cloud-based jobs for Claude, directly from the terminal.
We use these internally to aut…
使用 /schedule 可以直接在终端为 Claude 创建周期性的云端任务。 我们内部已经用它来自动化……
@bcherny·3月24日
Little known fact, the Anthropic Labs team (the team I joined Anthropic to be on) shipped:
- MCP
- Skills
- Claude Desktop app
- Claude Code
It was just a few of us, shipping fast, trying to keep pace with what the model was capable of.
Those early Desktop computer use prototypes, back in the Sonnet 3.6 days, felt clunky and slow. But it was easy to squint and imagine all the ways people might use it once it got really good.
Fast forward to today. I am so excited to release full computer use in Cowork and Dispatch. Really excited to see what you do with it!
@claudeai
You can now enable Claude to use your computer to complete tasks. It opens your apps, navigates your browser, fills in spreadsheets—anything you'd do sitting at your desk. Research preview in Claude Cowork and Claude Code, macOS only. https://t.co/sVymgmtEMI
一个鲜为人知的事实：Anthropic Labs 团队（我加入 Anthropic 就是为了这个团队）交付了： - MCP - Skills - Claude Desktop app - Claude Code 就是我们几个人，快速迭代，努力跟上模型本身的能力进化节奏。 早期在 Sonnet 3.6 时代的那些 Desktop 电脑操控原型，用起来又笨又慢。但很容易脑补出——一旦它真正强大起来，人们会怎么用它。 时至今日，我非常兴奋地宣布：完整的 computer use 功能已在 Cowork 和 Dispatch 中上线。很期待看到大家用它做什么！ [引用 @claudeai]：你现在可以让 Claude 操控你的电脑来完成任务了。它能打开应用、操控浏览器、填写表格——任何你坐在桌前会做的事。现已在 Claude Cowork 和 Claude Code（仅限 macOS）中作为 Research Preview 提供。
@bcherny·3月24日
RT @claudeai: You can now enable Claude to use your computer to complete tasks.
It opens your apps, navigates your browser, fills in sprea…
你现在可以让 Claude 控制你的电脑来完成任务了。 它能打开应用、在浏览器中导航、填写表格……
760 tweets · 188 sources