The capability is already here.

"Good morning, everyone. It is such a pleasure to see you all. Thank you for joining us."

"早上好。很高兴见到大家。谢谢你们今天来。"

"When I think about why I'm here today, I go back to the first time I wrote a computer program and it worked."

"每次想到我今天为什么会站在这里，我都会回到那一刻——我第一次写程序、它居然跑通了。"

"I didn't grow up coding. I grew up in the foothills of the Appalachians. I never built my own computer. I didn't even play video games."

"我不是从小写代码长大的。我在 Appalachian 山脚下长大，从没自己装过电脑，连游戏都不打。"

"The first time I actually tried to build anything complicated was in my college computer science classes."

"我第一次真正动手写点稍微复杂的东西，是在大学的计算机课上。"

"Now, this was so long ago, we had to wait in line to log directly into the servers because they were the only thing powerful enough to handle our race tracers."

"那年代久远到——我们得排队登服务器，只有它跑得动我们写的 race tracer。"

"This might be familiar to some of you — the hum of servers, the smell of old pizza, coffee, and that very specific aroma of a windowless basement computer lab."

"这场面有人可能熟——服务器的嗡鸣、隔夜披萨和咖啡的味道，再加上没窗的地下机房特有的那种气味。"

"I can still remember that feeling of hitting compile and waiting to see if my program worked — that feeling of joy, discovery, a little relief, and the excitement that I had made something that had never existed in the world before."

"我现在还记得那种感觉——按下 compile，然后等着看它跑不跑得通；那种喜悦、那种发现感、那种"终于跑通了"的松一口气，还有"我做出了世界上从没存在过的东西"的兴奋。"

"That's what hooked me, and that's why I'm here today."

"我就是被这种感觉勾住的，这也是我今天在这里的原因。"

"But so much has changed."

"但今天的世界已经完全不一样了。"

"What I could only get by waiting in line at a college computer science lab — that's available to anyone, any day of the week, anywhere in the world."

"那种当年只能靠排队进大学机房才换得到的东西，今天任何人、任何时候、世界任何地方都拿得到。"

"No line, no weird smells, and no barriers — just that same feeling of excitement, joy and relief."

"不用排队，没有怪味，没有门槛——只剩下一样的那种兴奋、喜悦、松一口气。"

"And I know a lot of you feel the same way."

"我知道在座很多人有同样的感受。"

"People say to me all the time, 'I feel like Claude has given me superpowers.' It's one of my favorite things to hear."

"经常有人跟我说："Claude 像是给了我超能力。" 这是我最爱听到的话之一。"

"And we're seeing people use those powers."

"我们也确实看到大家在用这份能力。"

"For instance, Scott McVicker runs developer infra at Stripe."

"举个例子。Scott McVicker 在 Stripe 负责 developer infra。"

"One of his teams had 50,000 lines of Scala that needed to become Java before they could upgrade their JDK."

"他的一个小组手上有 5 万行 Scala 代码，要先全部改成 Java，才能升 JDK。"

"The initial estimate was 10 engineering weeks. They used Claude and finished in four days."

"最初估的是 10 个工程师周。他们用 Claude，4 天搞完了。"

"And sometimes speed isn't just about efficiency. It's about what's waiting on the other side."

"而且有时候，速度不只是效率问题——它关乎"那一头到底有谁在等着"。"

"Felicia Kakuru is the co-founder and CEO of Binti."

"Felicia Kakuru 是 Binti 的联合创始人兼 CEO。"

"Her software runs the systems that caseworkers use to place kids in foster care — the paperwork, the home visits, the licensing process."

"她的软件支撑的是 caseworker 把孩子送进寄养家庭的整套流程——文书、家访、资格审批。"

"This year, her team used the Claude API to give caseworkers back hours they used to spend on paperwork."

"今年她的团队用 Claude API，把 caseworker 原来填表用掉的时间，一小时一小时还回去了。"

"That took 20 days off the process of licensing a foster family."

"寄养家庭的资格审批整整缩短了 20 天。"

"20 days. It's not just an efficiency metric. That's a kid connecting with a family."

"20 天。这不是一个效率指标，这是一个孩子早 20 天进到一个家。"

"And that excitement, joy, relief, that feeling of discovery — it's something I hear from everyone."

"那种兴奋、那种喜悦、那种松一口气、那种发现感——我在每一个人身上都听到过。"

"I'm gonna guess though that everyone here experiences it differently."

"不过我猜，在座的每个人体验它的方式都不一样。"

"Some of you are living on the frontier every day, some of you are bringing along the people around you, and some of you came here because, like me, you can feel the ground shifting under us and you want a view of what's to come."

"有人每天活在前沿；有人在带着身边的人一起往前走；有人来这里，是因为像我一样，能感觉到脚下的地正在移动，想抢一眼接下来会发生什么。"

"Trust me, I feel all of those things often in the same morning."

"相信我，这三种感觉我经常在同一个上午全部经历一遍。"

"I come to work with a plan, and then I have to tear it up by lunchtime because something new has happened. Does that sound familiar?"

"早上带着一个计划来上班，到中午之前就得撕掉重写，因为又出新东西了。是不是听着耳熟？"

"And, you know, that makes sense when we step back and look at how fast these models are getting better."

"其实你只要往后退一步，看一眼这些模型在以多快的速度在变好，这种感觉就完全合理了。"

"At Anthropic, we talk a lot about the exponential, and I think that's what we're all feeling right now."

"在 Anthropic，我们经常聊"指数级"。我觉得现在所有人感受到的就是这个东西。"

"Remember a couple years ago, the frontier of model development was something so good it could write a decent email — and we were pretty happy about that."

"还记得吗，几年前模型领域的"前沿"是——能写一封像样的邮件，我们就已经挺满足了。"

"A year ago, we were standing on this stage — Opus 4.0 was the headline, and the idea that an agent could run for an hour without a human checking in felt like a stretch goal."

"一年前我们也站在这个台上，那次的头条是 Opus 4.0，而"一个 agent 能连续跑一小时不用人盯"——当时听起来都像一个挺激进的目标。"

"But then six months ago, agents were running end-to-end overnight, so we'd wake up to finished work."

"然后到 6 个月前，agent 已经能整宿端到端跑下来了——我们一觉醒来就看到活儿做完了。"

"And then last month, Mythos read the entire OpenBSD source tree and found a 27-year-old vulnerability that had survived every human reviewer, every fuzzer, every static analyzer thrown at it for almost three decades."

"然后上个月，Mythos 把整棵 OpenBSD 源码树读了一遍，挖出来一个躺了 27 年的漏洞——这玩意儿在过去差不多 30 年里，扛住过所有人工评审、所有 fuzzer、所有 static analyzer。"

"The jumps keep getting bigger and the intervals keep getting shorter."

"每一跳越来越大，间隔越来越短。"

"But even though model capabilities are improving on an exponential, most organizations are still adopting AI on a linear path, and that means there's a gap between what AI can do and what it's actually doing for people."

"但是——模型能力在指数级跃升，绝大多数公司采用 AI 还是线性的；这中间就出现了一个 gap：AI 能做的事 和 它真正在帮人解决的事 之间的落差。"

"Closing that gap, translating model capability into something real people use to solve their problems — that's what developers do, that's what you all are doing, and we're seeing it happen."

"把这个落差填上、把模型能力翻译成"真人能拿来解决真问题"的东西——这就是开发者在做的事；就是你们在做的事。我们也确实看到它在发生。"

"Year over year, API volume is up nearly 17X on the Claude platform."

"同比看，Claude 平台的 API 调用量涨了近 17 倍。"

"And on Claude Code, the average developer is now spending 20 hours per week running Claude."

"在 Claude Code 这边，一个开发者平均每周已经在用 20 小时跑 Claude。"

"Now, like you, we've been shipping a lot lately, and we want you to walk away from today with a clear picture of where we're headed, so you can plan for it and ride the exponential with us."

"我们最近也跟你们一样在猛发版本。我们希望今天结束的时候，你能带着一张清晰的路线图走出去——知道我们要往哪儿去，好做你自己的规划，跟我们一起搭上这个指数级。"

"Let me say upfront, we don't have a new model to unveil."

"先把丑话说在前头：今天没有新模型要发布。"

"Today is about how we're making our products work better for you, so you can close the gap for the rest of the world."

"今天讲的是——我们怎么把产品做得更趁手，好让你们去帮全世界把那个 gap 填上。"

"And this morning, we'll show you what that looks like."

"接下来这个上午，我们就来给你看这事到底长什么样。"

"First, Diane will talk about our foundation — the model layer. She'll share more about our frontier models and what's coming."

"首先 Diane 会讲我们的地基——模型层；她会说前沿模型的现状和接下来的方向。"

"On the Claude platform, we're shipping updates to Claude-managed agents — outcomes, dreaming, multi-agent orchestration."

"在 Claude 平台上，我们要给 Claude-managed agents 上一批更新——outcomes、dreaming、multi-agent orchestration。"

"And Angela and Caitlyn will walk you through how the platform handles the infrastructure so you don't have to."

"Angela 和 Caitlyn 会带你看：怎么让平台帮你扛掉那些基础设施的活儿——你不用自己再扛。"

"And on Claude Code, Cat and Boris will walk you through how you can use new primitives like routines, to let Claude Code prompt itself even when you're away from your computer."

"在 Claude Code 这边，Cat 和 Boris 会带你看 routines 这类新原语——怎么让 Claude Code 自己 prompt 自己，哪怕你人都不在电脑前。"

"But all this comes back to you and what you're going to build, 'cause most people will never call the Claude API, they'll never open a terminal and type Claude."

"但所有这些最终都会回到你——回到你接下来要造的东西。因为绝大多数普通人永远不会去调 Claude API，永远不会打开终端敲一个 Claude。"

"They'll experience AI through something one of you built on the Claude platform — whether that's a designer exploring new directions with Canva, or a lawyer getting a brief out the door faster with Legora, or a developer using any one of the world's best coding agents."

"他们体验 AI 的方式，是通过你们在 Claude 平台上做出来的某个产品——可能是一位设计师在 Canva 里试新方向，可能是一位律师用 Legora 更快出一份 brief，可能是一位开发者在用世界上最好的某一个 coding agent。"

"So thank you. You all shape what AI feels like for everyone else."

"所以——谢谢你们。是你们在替所有其他人，决定 AI 用起来是什么感觉。"

"We'd never be able to build everything that people need to solve their problems. That's something only you all can do."

"所有人解决问题需要的东西，我们一家公司不可能全都做出来。那只能靠你们。"

"And one way we wanna show our gratitude is by sharing a little exciting news."

"我们想表达感谢的方式之一，就是放个让人兴奋的小消息。"

"As of today, we are increasing rate limits for developers on Claude Code and the Claude platform, to help you keep building and closing that gap for the world."

"从今天起，我们给 Claude Code 和 Claude 平台上的开发者上调 rate limit——帮你们继续造、继续帮世界把那个落差填上。"

"More specifically, we are doubling Claude Code's five-hour rate limits for pro max, team, and seat-based enterprise plans, and we're raising our API limits considerably for Claude Opus."

"具体说：Claude Code 的 5 小时 rate limit，对 Pro Max、Team、按席位计费的 Enterprise 这几档全部翻倍；Claude Opus 的 API 限额也大幅上调。"

"We're making this possible by expanding our compute partnerships."

"我们能这么做，是因为又扩了一批算力合作。"

"We're partnering with SpaceX to use all the capacity of their Colossus 1 data center, and we're investing this directly into individual developers and small teams."

"我们和 SpaceX 合作，用下他们 Colossus 1 数据中心的全部容量——而且我们要把这部分算力，直接砸给个人开发者和小团队。"

"Over time, we'll continue to explore every way to help you get the best out of Claude, including our existing compute efforts and even bolder bets."

"接下来我们会继续找各种办法，让 Claude 给你最大的产出——既包括现有的算力布局，也包括一些更激进的押注。"

"So thanks for being here today. Thanks for partnering with us to shape what AI looks like for the world. Thanks for giving people superpowers."

"再次谢谢你们今天来。谢谢你们和我们一起，去定义 AI 在这个世界上长什么样。谢谢你们在给所有人发超能力。"

"Up next, Diane, who leads our research PM team."

"下一位是 Diane，她带的是我们的 research PM team。"

"Thanks, Ami. I'm Diane, and I joined Anthropic back in 2023, and I've been a part of every model since Claude 2.0."

"谢谢 Ami。我是 Diane，2023 年加入 Anthropic，从 Claude 2.0 开始的每一个模型我都参与过。"

"For those of you who are counting, that's bringing 18 versions of Claude across Haiku, Sonnet, Opus, and now Mythos, to users and developers like you."

"对认真数过的同学：这总共是把 18 个版本的 Claude——横跨 Haiku、Sonnet、Opus，现在又有了 Mythos——交到了像你们这样的用户和开发者手上。"

"We wrestled with making Opus 3.0 great at adhering to JSON, and also making it the best at writing long-form code."

"为了让 Opus 3.0 老老实实遵守 JSON 格式、同时还能成为写长代码最好的那个，我们没少较劲。"

"With Sonnet 3.5 new — or as we all now finally know it, Sonnet 3.6 — we taught Claude to make and use a computer safely."

"到了 Sonnet 3.5 New——也就是大家现在终于知道的 Sonnet 3.6——我们教会了 Claude 安全地操作和使用一台电脑。"

"And with Sonnet 3.7, which had a tendency to be slightly too over-eager, we figured out the right way to expose that to users and developers so you could get the most out of Claude."

"再到 Sonnet 3.7——它有一点偏"过分主动"的毛病——我们找到了合适的方式把这一点暴露给用户和开发者，让你们能从 Claude 身上压榨出最多的东西。"

"This time last year, we used Claude 4.0 to be able to use thinking dials in a way that worked well, and to address test time compute. And we haven't slowed down."

"去年这个时候是 Claude 4.0——我们终于把"思考强度旋钮"这件事做得能用，也开始正经处理 test-time compute。从那以后我们就没慢下来过。"

"In the last 12 months, we shipped eight frontier models to developers and users, each one building upon the last, allowing you to write better code and the products you build go further than previously before."

"过去 12 个月，我们一共把 8 个前沿模型推到了开发者和用户手上——每一个都是前一个的延续，让你能写出更好的代码，让你做的产品比以前能走得更远。"

"The model layer underpins everything else you'll hear about today, and that's the bottom line."

"模型层托住的是今天接下来要讲的所有东西——这是底线，也是前提。"

"As model intelligence increases, your starting line moves forward, and you can do more than ever before."

"模型智力越往上走，你的起跑线就被往前推一段，你能做的事比过去任何一刻都多。"

"We talk about the exponential a lot at Anthropic. You heard it a little bit from Ami as well."

"我们在 Anthropic 经常聊"指数级"。你刚才也从 Ami 那里听到一些。"

"For me, the exponential means that as model intelligence increases, the use cases you can build and deliver to your users increase exponentially."

"对我来说，"指数级"的意思是：模型智力一往上涨，你能搭出来、交付给用户的 use case 跟着指数级地涨。"

"For example, agentic coding is far more impactful than code autocomplete."

"举个例子。Agentic coding 这件事，影响力远远不是 code autocomplete 能比的。"

"And in this way, new products and new experiences create new markets and grow the pie for everyone."

"这种方式下，新产品和新体验会带出新的市场——把蛋糕做大，所有人受益。"

"In research, we don't think about the exponential as SWE-bench numbers going up."

"做研究的时候，我们看"指数级"，不是 SWE-bench 数字往上爬。"

"It's also about creating and tracking capabilities that previously didn't exist until we designed and created them."

"它更多是关于：去定义、去追踪那些以前根本不存在、直到我们设计出来才存在的能力。"

"Tool use, computer use, thinking that adapts to the problem, agentic loops that hold a plan over hundreds or thousands of steps, and long context windows that teach Claude knowledge that it previously didn't have."

"工具使用、电脑使用、根据问题自己调节强度的思考、能撑住上百乃至上千步的 agentic loop、能让 Claude 学到原来不会的东西的超长 context window——这些都是。"

"And these capabilities don't just stop at code."

"而且这些能力不只停在写代码上。"

"Today, Claude can generate and iterate on visual designs, analyze and create complex work deliverables, and also navigate business domains you might be a part of, in open-ended, ambiguous fashions."

"今天的 Claude 能做视觉设计的生成和迭代，能分析和产出复杂的工作交付物，也能在你所在的业务领域里——在那些开放式、模糊不清的场景里——把事情走通。"

"That's because the model intelligence — the core foundation — has gotten smart enough and strong enough to support all of this."

"这是因为模型智力本身——也就是底层的地基——已经聪明到、强壮到撑得起这一切了。"

"When you're building on Claude, you're building on the model family that created these capabilities first and has had the most time to make them reliable."

"你基于 Claude 来做产品，意味着你基于的是——最早做出这些能力、并且有最长时间把它们打磨可靠的那一个模型家族。"

"Let me make that concrete with our latest model, Opus 4.7."

"我用我们最新的 Opus 4.7 把这件事说具体一点。"

"Amp, the coding agent, moved their entire smart mode onto Opus 4.7 because it scored the highest on their benchmarks, and they were able to simplify their tooling, change their scaffold, because the model no longer needed the help."

"Coding agent Amp 把他们整个 smart mode 都迁到了 Opus 4.7 上——因为这个模型在他们 benchmark 上跑得最高；而且他们顺手把工具链精简了、把 scaffold 也改了，因为模型已经不需要那些拐杖了。"

"Rakuten ran it on their benchmarks and resolved three times the number of production engineering tasks than they previously could."

"Rakuten 在他们自己的 benchmark 上跑过——能解决的生产工程任务数量直接是过去的 3 倍。"

"And finally, Intuit saw Opus 4.7 identify its own logical faults during the planning stage, figure out what was wrong, backtrack, resolve it, ultimately leading to faster and cleaner execution."

"最后是 Intuit。他们看到 Opus 4.7 能在 planning 阶段就识别出自己的逻辑漏洞——搞清楚错在哪、回滚、修掉——最后整个执行又快又干净。"

"The day after we launched Opus 4.7, we launched Claude Design by Anthropic Labs — one of my favorite launches this year."

"Opus 4.7 发布的第二天，我们发了 Claude Design by Anthropic Labs——这是我今年最喜欢的发布之一。"

"Already, people are building production interfaces with a combination of Claude Design and Claude Code."

"已经有人在用 Claude Design 加 Claude Code 的组合，搭出生产环境用的产品界面了。"

"This is because Opus 4.7 has a real taste for visual design — the right nuances to show, while adhering to your design principles."

"原因是 Opus 4.7 对视觉设计真的"有审美"——它知道在遵守你定的设计原则前提下，还能拿捏哪些微妙的分寸。"

"And we also hear from everyday users that people like to use Claude because it understands a full assignment, and can figure out when to push back and question assumptions."

"我们也从日常用户那里听到——大家喜欢用 Claude，是因为它能完整地理解你交给它的任务，并且知道什么时候该顶回来、该质疑你的前提。"

"At the same time, as each of us already know having built on these systems, the models are unfinished and they're works in progress."

"与此同时——所有亲手在这些系统上做过东西的人都知道——模型还远没做完，它们都还在路上。"

"They can still be stumped sometimes by very basic things, and also lose the thread when you introduce a lot of context."

"它们有时候会被一些非常基础的东西卡住，也会在你塞一大堆上下文之后丢掉主线。"

"That's what makes this exciting, and thanks for being on this journey with us."

"但这恰恰是让这件事变得有意思的地方——也谢谢你们一路在这条路上和我们一起走。"

"Here's a little about what we're working on and what's ahead."

"接下来说一点我们在做什么、下一步是什么。"

"First, higher judgment and better quality code taste. This means versions of Claude that you can trust with complex autonomous engineering work."

"第一，更高的判断力，更好的代码品味。意思是——你可以放心把复杂的、自主的工程任务交给它的那一代 Claude。"

"Second, context windows that feel infinite when combined with high quality memory, so it feels like you could do long running task while getting better results."

"第二，配合高质量的 memory，context window 用起来会"感觉是无限的"——你能跑长任务，结果反而越跑越好。"

"And finally, multi-agent coordination, powering teams of agents in instances of Claude that collaborate on big goals that are far too big for any single instance ever could."

"最后是多 agent 协同——一群 Claude 实例组成的 agent 团队，去合力扛那些"任何单个实例都不可能扛下来"的大目标。"

"The way I think about progress and model intelligence is task horizon, which is a measure of how long a version of Claude or a model can work autonomously while improving on its deliverables and the quality of its work."

"我衡量"模型智力进展"的方式叫 task horizon——它衡量的是：一个版本的 Claude 或一个模型，能自主工作多长时间，同时还在改进它产出物的质量。"

"Last time this year, models can work for minutes."

"去年这个时候，模型能自主工作的时间是以分钟计的。"

"Now, most of you and I probably have had agents that are running for hours on end, and tomorrow we'll have agents that are proactive, always on, and know what to work on without losing the thread."

"今天，你和我大多数人手上都有过能连续跑好几个小时的 agent；明天我们会有"始终在线、主动的、自己知道该做什么、还不会跑偏"的 agent。"

"So, what do we as developers make of all of this?"

"那么我们这群开发者，该怎么看这一切？"

"The exponential will keep improving, and you need to build for emerging capabilities, not just for today's versions of Claude."

"指数级会继续往上走。你写代码要为正在涌现的能力做准备，不要只盯着今天这一版 Claude。"

"This is because new models will be far more capable than the ones we have access to today."

"因为下一代模型会比我们今天手上能用到的强得多。"

"It used to be that we have to build scaffolding to keep every version of Claude up, and now scaffolding is there to actually amplify model intelligence."

"以前我们写 scaffold，是为了把每一代 Claude 撑起来；现在 scaffold 的作用是——真正去放大模型已有的智力。"

"Used to have to design complex iterative loops, give it the right tools, figure out how to do retries — and now all of that can be folded into the right thinking and the right execution right within the model."

"以前我们得自己设计复杂的迭代循环、配工具、想怎么重试——现在这些都可以折回模型内部，靠它正确的思考和执行来完成。"

"You are already seeing where this can go. Opus preview, Mythos, is the next point of that exponential, and it's not a small step."

"你已经能看到这件事会往哪儿去。Opus preview——Mythos——就是这条指数曲线上的下一个点，而且这一步不小。"

"Therefore, how we all work with model and Claude need to change."

"所以——我们和模型、和 Claude 协作的方式，是必须要变的。"

"Here are some things we think about at Anthropic."

"分享几条我们在 Anthropic 内部经常想的事情。"

"First, you need to design for the next version of Claude, not just the current one."

"第一，你要为下一版的 Claude 写代码，而不是只为当前这一版。"

"We've seen countless times that the developers who win are the ones who optimize their architectures to absorb the next intelligence jump, not just today's incremental accuracy."

"我们见过太多次了——最后赢的开发者，是那些把架构调整成"能吸收下一跳智力"的人，而不是只死磕今天这点点滴滴的精度提升。"

"This means maintaining and creating harder evals, building ambitious prototypes that you don't think might work today — because that's how you'll notice when the exponential is improving and moving under you, that something that previously didn't work all of a sudden starts passing."

"具体说，就是去维护并造更难的 eval；去做那些今天你觉得不太可能跑起来的"激进原型"——因为只有这样，当指数级真的往前推进、悄悄从你脚底下挪过去时，你才会及时发现："咦，原来跑不过的，突然过了。""

"That's a sign that you probably have something magical to give to your users that didn't work before."

"这就是一个信号——你手上很可能有了一个之前根本做不到、但现在能交给用户的"魔法时刻"。"

"And here's what the teams that are getting the most out of Claude have figured out."

"那些把 Claude 用得最透的团队，都已经悟到一件事。"

"Model upgrades are a business opportunity. The teams that are getting the most out of Claude models are the ones who make upgrade cheap."

"模型升级本身就是一个商业机会。把 Claude 模型用得最透的，是那些"让升级这件事变得很便宜"的团队。"

"This is automated evals, simple scaffolding, and ambitious prototypes and uses of capabilities that others haven't imagined yet."

"具体长什么样：自动化的 eval、尽可能简单的 scaffold、激进的原型，以及别人还没想过的能力用法。"

"We believe that the exponential on slide one is gonna keep looking like that."

"我们相信第一张幻灯片上那条指数曲线，会继续是那个样子。"

"As model intelligence increases, US developers have the chance to make a head start to experiment with building new use cases, creating exciting new products for your users, and ultimately creating new designs and markets, and ultimately, growing the pie."

"模型智力一直往上走，开发者就有机会抢一个先手——去试新 use case，给用户做出让人眼前一亮的新产品，最后是开出新的设计、新的市场，把蛋糕真的做大。"

"Everything Caitlyn and Angela are about to show you will give you the tooling to make all of this possible and come to life."

"接下来 Caitlyn 和 Angela 要给你看的所有东西，目标就是把上面这一切——从可能、变成能跑。"

"Thank you so much for being here."

"非常感谢你们今天来。"

"So model capabilities are on the exponential, but businesses are still operating on the linear."

"模型能力跑在指数级上，但企业还在线性运转。"

"And so as a business, it's never been more important to make sure that you're really able to harness the power of that exponential."

"对一家企业来说，"能不能把这条指数级的能量真正用起来"——从来没有像今天这么重要。"

"But what's stopping businesses from really snapping to that?"

"那么，是什么挡在企业和"指数级"之间？"

"Well, it really boils down to two key problems."

"说到底就是两个核心问题。"

"The first one is getting the right outcomes. Getting the right outcomes is still too difficult, right?"

"第一，是"出对结果"。让模型出对结果，到今天还是太难了。"

"You have to be able to do a lot of prompt optimization, tool construction, harness engineering."

"你要做大量的 prompt 调优、工具构造、harness 工程。"

"There's still a lot of work that it takes to really steer the model to exactly where you need it to go."

"要把模型精确地导向你想要的地方，中间这些活儿一点都没省。"

"That's right. And the second problem is you wanna ship fast, but you wanna ship scalably at the same time."

"没错。第二个问题是——你想发得快，但又得发得稳、能上规模。"

"Everyone in the tech space is moving insanely fast right now, and you've gotta keep up. But to win, you need quality too."

"整个科技圈现在都跑得疯一样快，你不跟上不行；但要赢，光快还不够，质量得跟上。"

"It's really easy to ship prototypes, but it's really hard to scale in production."

"做 prototype 太容易了；把它推到生产环境、跑稳上规模，是另一回事。"

"So we built the Claude platform to give you everything you need to get those great outcomes and to ship with speed and scale at the same time."

"所以我们做了 Claude 平台——把你需要的一切都给你，好同时拿到"出对结果 + 又快又稳上规模"这两件事。"

"The platform comes with API primitives that are tuned to Claude models."

"平台自带一套针对 Claude 模型调过的 API primitives。"

"It gives you infrastructure to build and scale agentic systems, and it gives you controls to operate those systems."

"它给你搭和扩 agentic system 用的基础设施，也给你运营这些系统的控制面。"

"So if we go back to our problem statements that we hear all the time from different businesses, one of the most common ones that Caitlyn and I hear all the time is how much businesses need high intelligence — but of course, at lower cost."

"回到我们整天从各家企业那里听到的痛点——Caitlyn 和我听得最多的一条就是：企业要"高智力"，当然同时还要"更低成本"。"

"So one of the ways that we're solving this is with the advisor strategy."

"我们解决这件事的一个办法，是 advisor 策略。"

"This is really easy to implement. All you have to do is update your tools array inside the messages API."

"接入特别简单——只要在 messages API 里更新一下你的 tools 数组就行。"

"And what we're doing is that we're actually providing an agent architecture that splits execution from advising."

"我们提供的，其实是一种"把执行和咨询拆开"的 agent 架构。"

"So in execution, you can pick a smaller model — it'll be a little cheaper. But then that small model, when it needs advice on what to do next, it can actually reach out to a bigger model for help."

"执行的时候你可以挑一个小一点的模型——便宜些；这个小模型在不知道下一步该干嘛的时候，可以反手去找一个大模型问意见。"

"Yeah, so in practice, this means you could use a Haiku or a Sonnet class model to execute, and Opus as an advisor."

"对，落地下来就是——用 Haiku 或者 Sonnet 这一档来跑执行，用 Opus 当顾问。"

"And when we ran this with Sonnet executing and Opus advising, we saw Sonnet perform way better than Sonnet alone."

"我们用 Sonnet 跑执行 + Opus 当顾问跑了一遍——Sonnet 的表现远比单独跑 Sonnet 要好。"

"But more importantly, we saw Sonnet perform even more cheaply than Sonnet alone, because Opus was able to advise it to get its work done better."

"更重要的是——它比"单独跑 Sonnet"还更便宜，因为 Opus 能在关键时候给它指路，让它更省事地把活儿干完。"

"A great example of this was Eve Legal."

"一个很好的例子是 Eve Legal。"

"Eve Legal used the advisor strategy, and they told us they got frontier model quality at five times lower cost."

"Eve Legal 用了 advisor 策略，他们告诉我们——拿到了前沿模型的质量，但成本只要 1/5。"

"Yeah, and that's awesome. We love stuff like this because you can go ahead and use this in things like a freemium model, right?"

"对，这个我们特别喜欢——因为你可以把它用在 freemium 这种模式里。"

"When you offer these kinds of freemium experiences to your users, you have to be conscious of the costs that you're going to incur, but obviously you wanna make sure you have a good experience for them."

"做 freemium 体验的时候你要盯成本，但你又得保证用户体验过得去——这个策略两头都顾到了。"

"It's also really great for areas where you have an extremely high volume amount of workloads — and of course, you'd have to be a little conscious of your ROI."

"它也特别适合那种调用量极大、必须算 ROI 的场景。"

"So that's all great, but what about speed and scale? Those two things that Caitlyn mentioned to be really difficult to achieve at the same time."

"这套搞定了"出对结果 + 降本"。那 Caitlyn 提到的另一头——又快又稳上规模——怎么办？"

"Well, most recently we introduced Claude Managed Agents."

"我们最近上了 Claude Managed Agents。"

"Claude Managed Agents is an agentic harness that's paired with production-grade infrastructure."

"Claude Managed Agents 本质是一套 agentic harness，再绑上一套生产级别的基础设施。"

"Teams are able to go from prototype to production in literally a matter of days."

"团队从 prototype 到生产环境，真的是几天的事。"

"The teams that we've worked with have been able to ship literally 10 times faster with managed agents."

"我们合作过的团队，靠 managed agents 发版速度真的是过去的 10 倍。"

"Another great thing with managed agents that we love is that it bundles in a lot of the best practices out of the box."

"另一个我们很喜欢的点——它把很多最佳实践开箱就内置好了。"

"So for example, when you build an agent, one of the best practices that you wanna make sure that you do is of course give it memory."

"举个例子。建一个 agent，最重要的最佳实践之一，当然是给它配 memory。"

"That way the agent persists user preferences, and it kind of remembers closer to what you want it to do every single session."

"有了 memory，agent 才能记住用户偏好，每个 session 都更接近你想让它做的事。"

"It's a little difficult to build memory, and so this is an example of a best practice that we actually just bundle in out of the box, and it's automatically tuned for Claude."

"自己造 memory 是有点麻烦的事——所以这就是一个我们直接内置的最佳实践，而且专门为 Claude 调过参。"

"And we want to make sure that everyone hears this. When we do give you memory, that memory is ultimately yours."

"这一点我们想让所有人都听清楚——我们给你的 memory，最终归属是你的。"

"So you can take that, import it wherever you'd like."

"你可以拿走，导到任何你想用的地方。"

"So one of our favorite examples of someone who built on managed agents was Notion."

"我们最喜欢的一个 managed agents 客户案例是 Notion。"

"Notion wanted to build for speed and scale at the same time, so they chose to build on managed agents, and they built the ability for you to fire off Claude agents directly within their product experience for long-running, complex, autonomous tasks."

"Notion 想同时做到"又快又能上规模"，所以选了 managed agents——他们直接在产品里做了一个能力：你可以在 Notion 里就地起一个 Claude agent，去跑那些长时间、复杂、需要自主完成的任务。"

"Yeah. And we love that feature. That's super cool."

"对，这个功能我们也特别喜欢，真的酷。"

"All right. Well, today we are upgrading, uh, Claude Managed Agents with three really powerful features."

"好。今天，我们要给 Claude Managed Agents 升级 3 个特别能打的能力。"

"We're going to be introducing multi-agent orchestration so that you can actually create fleets of agents to solve really complex tasks."

"第一是 multi-agent orchestration——多 agent 编排——让你能搭出一支 agent 编队，去啃真正复杂的任务。"

"We're going to introduce outcomes which allow you to specify exactly what success looks like, and then Claude will literally just iterate until it gets it done."

"第二是 outcomes——你描述清楚"成功长什么样"，然后 Claude 就会一直迭代，直到把它做出来。"

"And we're going to introduce dreaming. And this one we're really excited about."

"第三是 dreaming——这个我们自己最兴奋。"

"But with dreaming, Claude is actually able to self-learn. It's able to inspect over its previous sessions, figure out skills that it missed, lessons it should have learned, and actually apply those directly to memory on its own."

"有了 dreaming，Claude 能自我学习——它会回头看自己过去跑过的那些 session，找出它当时漏掉的技能、本来应该学到的教训，然后自己把这些写进 memory。"

"But instead of just talking about all these with you, we're actually gonna show you live what these things look like. So, Caitlyn, let's do that. Let's do it."

"光讲没意思——我们现场演给你看。Caitlyn，开搞。开搞。"

"So Caitlyn and I have been inspired by some of the announcements earlier today. There's greater API rate limits for Opus, and we've, uh, you know, been hanging out with a particular space company most recently."

"我和 Caitlyn 被今天早上的几个发布激发了灵感——Opus 的 API 限额更高了，而我们最近也跟某家航天公司走得比较近。"

"So we've been inspired to create a little startup of our own — a fictional one to be clear — uh, called Lumara."

"于是我们顺手"创业"了——明确说，是虚构的——一家叫 Lumara 的小公司。"

"And with Lumara, we decided why don't we build agentic software that helps us autonomously land drones on the moon? And we really care about speed and scale at the same time."

"Lumara 干嘛呢——我们决定造一套 agentic software，自动把无人机降到月球上。我们对速度和规模这两件事同时很在意。"

"So obviously we are going to build this on Claude Managed Agents."

"那毫无悬念——我们要拿 Claude Managed Agents 来搭。"

"Exactly. So let's say we line up our first customer, and this hypothetical first customer wants to land drones on the moon to mine for hypothetical materials."

"没错。我们假设拿下了第一个客户——这位（虚构的）首单客户想在月球上降无人机，去开采（虚构的）矿物。"

"Um, and this is a big, ambitious job. And despite all of our dreams, we're not actually aerospace engineers, so we're gonna need really awesome agents to get this work done for us."

"这是一个又大又野的活儿。但理想归理想，我们俩不是航天工程师——所以我们得靠一群很能打的 agent 替我们把活儿干了。"

"So we're gonna integrate all three of the new features that we just talked about. Um, and we did this for our first customer. And I'm gonna show you using the Claude API CLI how we actually set that up."

"我们要把刚才那 3 个新能力全用上。我们已经为这位首单客户做了一遍——我下面用 Claude API CLI 给你演示我们是怎么搭的。"

"So first things, first things first. This is a big job. So we actually want multiple agents to help us get this work done."

"先说最重要的——这是个大活儿，我们得用多个 agent 一起干。"

"Um, so I'm gonna go ahead and show you the agents that we actually have set up for our customer."

"我先把我们给客户配好的几个 agent 给你看。"

"The first one, we have a commander agent. And really the commander agent's job is to make sure that this whole mission goes well."

"第一个是 commander agent。它的工作是保整个任务跑成。"

"Then we've got a detector agent. And the detector agent's job is to make sure that we're actually finding the sites to land on that will have high quality mining materials."

"然后是 detector agent——它负责帮我们找到真正含高质量矿料的着陆点。"

"And then we've got our navigator agent. And the navigator is making sure that we're landing our drones safely and flying them to their destination."

"再然后是 navigator agent——它管两件事：把无人机安全降下去，再把它飞到目的地。"

"So I'm gonna go ahead and actually set up our commander to be a coordinator of the other two agents."

"接下来我把 commander 配置成另外两个 agent 的协调者。"

"And when this is running, what's actually happening is that the commander is spinning up a session. And then for each of these sub-agents, they have their own independent threads so that they have independent context windows."

"跑起来之后实际发生的事是——commander 起一个 session；它下面每一个 sub-agent 都有自己独立的线程，也就有独立的 context window。"

"This is a very intentional design, and we found that by doing this all together, and then merging in all the results, we get better performance."

"这是一个有意为之的设计——我们发现：让它们各跑各的，最后再把结果合回去，整体表现更好。"

"Exactly. So that's multi-agent. Let's go ahead and integrate outcomes."

"对。这就是 multi-agent 那一块。下一步我们把 outcomes 也接上。"

"So the way outcomes work is — we wanna make sure that our customer, which has very specific criteria for what they want to accomplish, can define that criteria. And then we can provision a grader agent that actually makes sure we get that outcome."

"outcomes 的工作方式是——我们要让有非常具体目标的客户，能把"什么算成功"写下来。然后我们就给他配一个 grader agent，由它来盯着结果是否真的达成。"

"And so outcomes actually just start with a pretty simple markdown file."

"outcomes 其实就是从一个非常简单的 markdown 文件开始。"

"So here you can see — in a markdown file. Again, really, really simple. And it just kind of outlines this criteria that shows us whether a run is successful."

"你看这里——就是一个 markdown 文件，真的非常非常简单。它就是把"一次跑算不算成功"的判断标准列出来。"

"So we want our drones to touch down softly. We want them to land on clear ground. Um, and pretty importantly actually, we wanna have enough fuel in reserve so that we can get our drones safely back to earth."

"我们希望无人机落得稳；落在没有障碍物的地面上；还有非常重要的一条——剩余燃料要够把无人机安全带回地球。"

"Um, so in order to actually set out this rubric for our outcomes, I'm gonna go ahead and send an event, um, to our session that defines this rubric as our outcomes."

"为了把这套评分标准注册成 outcomes，我下面会向 session 发一个 event，把它登记进去。"

"Yeah. And when this is running, like Caitlyn mentioned, we actually create a separate grader. And this grader agent is evaluating across the session as to whether or not in each run we've actually met the rubric that was specified."

"对。跑起来之后——刚才 Caitlyn 也说了——我们会另外起一个 grader agent，它跨整个 session 评分，看每一次跑有没有达到刚才登记的评分标准。"

"Now, of course, it could one-shot this — but more likely than not, it probably needs to iterate over a couple sessions to do it. And as Caitlyn has highlighted here, you can specify the max number of iterations you want to allow this to do."

"当然它可以一次成功；但更常见的情况是，它会要迭代好几个 session。Caitlyn 这里也标了——你可以指定最多允许它迭代多少次。"

"So we've integrated multi-agent, we've integrated outcomes, and it's time to test."

"multi-agent 接好了，outcomes 也接好了——是时候测试了。"

"Our customer gave us some data on six hypothetical sites that they might wanna land their drones on, for us to run some simulation sessions and see what happens."

"客户给了我们 6 个候选着陆点的数据，让我们跑模拟、看结果。"

"So I'm gonna pop over into Lumara's dashboard, where you can see that I ran, um, a simulation against these six sites."

"我切到 Lumara 的 dashboard——你可以看到，我已经针对这 6 个站点跑了一轮模拟。"

"Yeah. And this is pretty good. I mean, this is a one-shot with our whole system. It's got the multi-agent architecture, it's got the outcomes feature integrated."

"结果还不错。这是我们整套系统第一次跑——multi-agent 架构 + outcomes 都接上了。"

"And you can see here that it's solved four out of six sites correctly, but clearly it could have done a bit better on sites three and sites four."

"你能看到——6 个里它做对了 4 个，但 3 号点和 4 号点本来还能更好。"

"And like any good founders, we obviously want to hill climb on this system."

"那作为创始人，我们当然想 hill climb——把它继续往上推。"

"And normally hill climbing is a pretty difficult process. You have to put a lot of work together to go and do that. But we're gonna show you how we are hill climbing on this with just dreaming."

"通常 hill climb 是个很费劲的活儿——你得攒一堆工作。我们下面演示——光靠 dreaming，就能把这件事做了。"

"Yeah. So we ran the simulation yesterday. We weren't quite happy with our results, and we came into the Claude developer console — into our dreaming interface."

"对。我们昨天跑了那一轮模拟，对结果不太满意——就进了 Claude developer console，打开 dreaming 界面。"

"And you can see I can actually just hit this little button that says dream — choose a memory store where a dreaming agent can go and look over all of those past simulation sessions and write its learnings to memory, so that all of our new sessions can actually reference those learnings in memory to do a better job."

"你看——我直接点这个写着 dream 的按钮，挑一个 memory store。这个 dreaming agent 就会去回顾过去所有的模拟 session，把学到的东西写进 memory；之后我们再跑的所有 session 都能引用这块 memory，把活儿干得更好。"

"So I did this last night. Um, and this is our dream that ran. Um, and you can see that, uh, we wrote a bunch of stuff to memory which is awesome."

"我昨晚就这么做了——这就是昨晚跑的那次 dream。你可以看到——它写了一大堆东西到 memory，相当顶。"

"Um, and crucially and most importantly, the agent actually chose to write a descent playbook — so that all of our additional sessions that we run going forward will have this playbook to reference, um, including all sorts of heuristics from the different missions that it ran previously."

"而且——非常关键的一点——这个 agent 自己决定写了一份"下降 playbook"。以后我们再跑的所有 session 都能参考这份 playbook，里面包含了它从过去那些任务里总结出的各种经验法则。"

"This is, like, a really robust playbook of a whole bunch of information that we can go ahead and use."

"这是一份相当扎实的 playbook，里面有一大堆我们可以直接用上的信息。"

"So this ran overnight last night, um, and I popped back into Lumara's dashboard this morning and ran a new simulation now that our system is upgraded."

"它昨晚跑了一整夜——今天早上我回到 Lumara 的 dashboard，系统已经升级，我重新跑了一轮模拟。"

"And that's awesome. Right, we were able to actually hill climb and not regress on any of the ones that we cared about, but the two sites that could've improved have actually improved."

"结果很赞。我们真的 hill climb 上去了——之前做对的那几个一个都没退化，而本来还能更好的两个站点，现在确实变好了。"

"And to hill climb on this, all we had to do was just have Caitlyn press a button in the console called dream."

"我们做的所有事，就是让 Caitlyn 在 console 里点一下那个叫 dream 的按钮。"

"All right, awesome. Let's wrap up."

"好，挺顺。我们收一下尾。"

"And everything that we showed you here today live in this demo is available on the Claude platform for you all to build on."

"今天 demo 里给你看的所有东西，今天就已经在 Claude 平台上，你们可以直接用。"

"Multi-agent orchestration, outcomes, and dreaming are now making the Claude Managed Agents primitive so much more powerful, so that you can use it to construct deep, powerful agentic systems that scale."

"multi-agent orchestration、outcomes、dreaming 这三件套，让 Claude Managed Agents 这个原语强大了好几个量级——你可以用它去搭那种"深度强大、还能上规模"的 agentic 系统。"

"Whether you're landing drones autonomously on the moon or you're building the next big business, Claude Managed Agents is here to help you construct the agentic systems that you need and solve the problems that you're looking to solve."

"不管你是在月球上自动着陆无人机，还是在做下一个大生意——Claude Managed Agents 都能帮你把你想搭的 agentic system 搭出来、把你想解决的问题解决掉。"

"Now, we would love to pass it on to Cat and Boris, who will show you how Claude Code is making it even more fun to build as a developer."

"接下来——把舞台交给 Cat 和 Boris。他们会带你看，Claude Code 怎么让"做开发"这件事变得更好玩。"

"Thank you. Angela and Caitlyn just showed you how the Claude platform closes the gap between what models can do and what agents businesses ship."

"谢谢。Angela 和 Caitlyn 刚才给你看的是——Claude 平台怎么把"模型能做的事"和"企业最终发出去的 agent"之间的落差填上。"

"We have a related challenge on Claude Code. We also want to close the gap between model capabilities and what every developer can actually do with them."

"我们在 Claude Code 这边面对的是一个相邻的难题——我们也想把"模型能力"和"每一个开发者实际能用它做什么"之间的落差填上。"

"First, I just want to thank all of the developers here in this room."

"开始之前，我先想谢一谢在场所有的开发者。"

"Thank you for trusting Claude Code on your production databases back when Sonnet 3.7 was our frontier model and when our product was a bit rough around the edges."

"谢谢你们——在 Sonnet 3.7 还是我们前沿模型、产品还有点毛糙的时候，就敢把 Claude Code 放到生产数据库上用。"

"Your support is what makes our team so excited to come in every day and make Claude Code even better."

"是你们的支持，让我们团队每天都很兴奋地来上班、把 Claude Code 做得更好。"

"Let's start with why Claude Code exists."

"我们先从 Claude Code 为什么存在说起。"

"Software development is being reinvented in real time."

"软件开发这件事，正在被实时重写。"

"The mission of Claude Code is to close this gap between the great ideas that you all have and shipping a product to the market."

"Claude Code 的使命，是把"你脑子里那个好点子"和"产品真正发到市场上"之间的距离压到最短。"

"And the way that we enable this is — we build tools that elicit the frontier intelligence from our models, and we make them accessible to every builder."

"我们做这件事的方式是——造能把模型的前沿智力榨出来的工具，再让每一个 builder 都用得到。"

"And we don't think of ourselves as having a finished roadmap for you."

"我们也不觉得自己手上有一份"做好了的 roadmap"递给你。"

"We think of ourselves more like mountaineers — kind of like climbing alongside you into terrain that none of us has explored before, learning together what works as we go."

"我们更像是登山者——和你一起，往一片谁都没去过的地形里爬，边爬边一起摸清楚什么走得通。"

"And we're growing with you. We're growing with increasing AI capabilities, and we're navigating these new challenges together."

"我们和你一起成长——和涨上来的 AI 能力一起成长，也一起摸索那些新出现的挑战。"

"I still remember a year ago when I would give Claude Code a task, and I would carefully review every single edit it was trying to make, every single permission prompt, giving it really detailed feedback on what I liked and what I didn't, and just, like, holding its hand every step of the way until the result was good."

"我现在还记得——一年前我交给 Claude Code 一个任务的样子：每一次它要改的文件我都仔细看，每一个权限弹窗我都过一遍，给它非常细的反馈说我喜欢什么、不喜欢什么，几乎是手把手扶着它一步一步走，直到最后那个结果让我满意。"

"I remember some of these tasks would be like 100, 200 permission prompts until I got the final result."

"我记得有的任务，要点 100、200 个权限弹窗才能到最终结果。"

"And now most of you are running in auto mode."

"现在你们大多数人都跑在 auto mode 上了。"

"You're delegating permissions to Claude, um, and you're checking in after Claude has done a lot of its work and has a PR for you to review."

"你们把权限授权给 Claude，等它把大量工作做完、提了一个 PR 给你 review 的时候，你才回头看一眼。"

"Over the last year, we've expanded the number of ways that you can use Claude. We started with the terminal, then we launched the IDE, and now we have desktop."

"过去一年，我们把你能用 Claude 的方式扩开了——一开始只有终端，后来是 IDE，现在又有 desktop。"

"We started with the CLI. This is still the interface for power users who want a minimal text interface, who want all the latest customizations and the most control."

"我们从 CLI 开始。它今天仍然是 power user 的入口——你想要极简的文字界面，想第一时间用上所有最新的自定义能力，想要最大化的控制权，就用 CLI。"

"Then we added IDE, because a lot of you actually want the same powerful agents but you want to follow along with all the code changes that it's making."

"然后我们加了 IDE 形态——因为很多人想要一样强的 agent，但同时希望能跟着看它每一笔改动。"

"And then based on all of your feedback for something a bit more visual, we knew where we had to go next."

"再后来——综合你们想要"更可视化一点"的反馈——我们知道下一步要往哪儿走。"

"We launched our newest surface, Claude Code on Desktop."

"我们发了最新的形态：Claude Code on Desktop。"

"It's a surface that's designed for people who want this, like, full-screen graphical interface, built-in preview so you can watch as Claude develops your app, a sidebar control plane for all your agents, and the ability to render images and rich outputs."

"它是给那些想要"全屏图形界面、内置 preview 能看 Claude 实时把你的 app 写出来、侧边一个管理所有 agent 的控制面板、能渲染图片和富文本输出"的人准备的。"

"We've built desktop to be a control plane not only for your local sessions but for your remote ones as well, with visual indicators for which agents are stuck and which ones are ready to go."

"我们把 desktop 做成了一个控制面——不仅管你本地的 session，也管远程的；哪个 agent 卡住了、哪个已经跑完待你确认，都有视觉提示。"

"The IDE and the desktop app are built on the Claude Agent SDK — the same SDK that many of you are already building on."

"IDE 和 desktop 都是基于 Claude Agent SDK 搭出来的——和你们很多人正在用的那一套 SDK 是同一个。"

"Many enterprises have adopted Claude Code tools wall-to-wall."

"已经有很多企业把 Claude Code 这套工具铺到全员了。"

"At Anthropic, this has driven a 200% increase in the number of PRs per engineer while keeping the same code quality bar, even as our engineering team has scaled substantially."

"在 Anthropic 自己——这件事让我们每位工程师的 PR 数量涨了 200%，代码质量门槛不变，与此同时工程团队规模也大幅扩张。"

"Together with you all, we're discovering and redefining the future of what software engineering looks like."

"和你们一起，我们正在发现、并重新定义软件工程接下来长什么样。"

"We're embracing these new challenges by embracing automations powered by Claude to overcome each."

"我们应对这些新挑战的方式，是用 Claude 驱动的一系列自动化——一个一个 challenge 砍下来。"

"I'm going to walk through a few of them right now. Here's the feedback that we heard from our users and what we've built with the help of this community."

"我下面挑几个跟你过一过——这是我们从用户那里听到的反馈，以及借这个社区的力做出来的东西。"

"We heard from you that you want to spend less time on code review."

"你们告诉我们——希望花在 code review 上的时间能少一点。"

"So we shipped code review that deploys a team of agents to catch critical bugs on your behalf."

"于是我们发了 Code Review——它会替你部署一支 agent 小队，去抓那些关键 bug。"

"Thousands of companies use this every day, including all internal Anthropic teams."

"现在每天有几千家公司在用这个能力——也包括 Anthropic 内部所有团队。"

"We heard from you that you really want to code on the go. So we launched remote control, and we added Claude Code to the iOS and Android Claude apps so that you can fire off a task from anywhere."

"你们告诉我们——想随时随地写代码。所以我们做了 remote control，把 Claude Code 加进了 iOS 和 Android 的 Claude app 里——任何地方都能起一个任务。"

"You're no longer walking around with an open laptop balancing it trying not to fall, um, and you're no longer stuck at your desk."

"你不用再端着一台开着的笔记本边走边小心别摔了，也不用被钉在工位上。"

"You can now go to a park, touch grass, and still code."

"你现在可以去公园里 touch grass 一下，照样写代码。"

"Um, we heard that you're spending a lot of time babysitting PRs — fixing flaky CI tests, addressing code review comments, resolving all the merge conflicts. So we added Autofix."

"我们也听到——你们在 PR 上耗很多时间：修 flaky 的 CI 测试、回 code review 的评论、解 merge conflict。所以我们加了 Autofix。"

"It just listens to all these events, proactively puts up fixes so that your PRs are always green."

"它就盯着所有这些事件，主动把修复贴上去——你的 PR 永远是绿的。"

"We heard from you that you're kicking off Claude Code tasks on new tickets and new customer bug reports, so we thought we should build routines."

"你们告诉我们——你们一直在拿新工单、新客户 bug report 去手动起 Claude Code 任务。我们想：那应该有 routines。"

"So we have routines, and it lets you configure once, listen for web hooks, API events, or run on a schedule, and it'll just kick off Claude Code automatically for you."

"所以现在有了 routines——你配置一次，让它监听 webhook、API 事件或者按计划跑，它就会自动替你起 Claude Code。"

"So instead of you having to manually kick things off, Claude will handle it."

"你不用再手动按按钮——Claude 来扛。"

"And last, we heard from you that you're launching so many features that your security teams are having a hard time keeping up. So we built Claude Security."

"最后一个——你们告诉我们：你们发新功能太快，安全团队跟不上了。于是我们做了 Claude Security。"

"It scans your whole code base overnight, and it can kick off Claude Code to address the vulnerabilities that it finds."

"它整宿扫你的整个代码库，扫出来的漏洞还能直接起一个 Claude Code 去修。"

"All these primitives compose together, and this helps all of us together adapt to the future of what engineering looks like."

"所有这些原语可以组合起来用——这样我们才能一起跟上"工程接下来长什么样"。"

"Everything I've covered is something that you can pick up today."

"我刚才讲的所有这些，今天就能用。"

"It's been especially exciting to see how a range of companies have taken these tools and adopted them at the scale of entire orgs."

"特别让人兴奋的是——一批公司已经把这些工具推到了整个组织的规模。"

"First, I wanted to share about Shopify."

"先说 Shopify。"

"They power e-commerce for millions of merchants worldwide, and they've imbued AI across the entire engineering org and changed their culture."

"Shopify 在为全球几百万商家撑起整个电商业务——他们把 AI 渗进整个工程组织，把文化都改了。"

"They're using Claude Code across the company, both on engineering teams but also non-engineering — so design, product, data science."

"他们全公司都在用 Claude Code——不只工程团队，还有非工程的：设计、产品、数据科学。"

"They're building it directly into their platform and standing up tools at scale."

"他们把它直接接进自己的平台，规模化地搭工具。"

"Andrew McNamara is the director of applied AI at Shopify, and in his words, 'The speed is just crazy.'"

"Andrew McNamara 是 Shopify 的 applied AI director——他原话是："这速度简直离谱（The speed is just crazy）。""

"Claude Code has completely transformed how they build internal tools."

"Claude Code 彻底改变了他们做内部工具的方式。"

"Another example is Mercado Libre. They're Latin America's most popular e-commerce platform. They serve over 100 million buyers."

"另一个例子是 Mercado Libre——拉美最大的电商平台，服务超过 1 亿买家。"

"Their org is 23,000 engineers and everyone runs on Claude Code."

"他们整个组织是 2.3 万名工程师，每一个人都在跑 Claude Code。"

"When that happens across an org, the work itself changes shape."

"当这件事在一个组织里铺开之后，工作本身的形态都会变。"

"Engineers are pointing agents at tech debt that people haven't touched in a long time and people don't have time for."

"工程师在把 agent 指向那些"很久没人动、也没人有时间动"的技术债。"

"It's reviewed more than 500,000 PRs with human oversight, and modernized more than 9,000 apps."

"在人工监督下，他们已经 review 了超过 50 万个 PR，把 9,000 多个 app 现代化重构了。"

"Oscar Mullen, who leads technology, is aiming for 90% autonomous coding and a fully agent-driven PR loop by Q3 of this year."

"他们的技术负责人 Oscar Mullen 给出的目标是——今年 Q3 之前，编码 90% 自动化、PR 流程完全 agent 驱动。"

"And we hear this from many others across the industry."

"我们在业内还从很多其他人嘴里听到类似的话。"

"The detail I love the most here actually isn't this number. It's that a lot of the managers and VPs we talked to are like, getting their hands dirty in the code base again."

"我最喜欢的细节其实不是这个数字。是——我们聊过的很多 manager 和 VP，又开始亲手碰代码库了。"

"Claude Code is putting coding back in the hands of people who've spent the last decades on roadmaps and reviews, and now they're back building."

"Claude Code 把"写代码"这件事，重新塞回那些过去十几年只在做 roadmap 和 review 的人手里——他们又开始 build 东西了。"

"We see this across the industry. Millions of developers are getting more products shipped at a higher quality than before."

"这种事在整个业界都在发生——几百万开发者，正在以更高的质量、发出更多的产品。"

"Now, let's see what this actually looks like in practice."

"接下来，我们看看这件事真正落到一台机器上、是什么样的。"

"To take you through it, please welcome the head of Claude Code, Boris Cherny."

"带你过这一段的——欢迎 Claude Code 的负责人，Boris Cherny。"

"Whoo. Thanks, Cat. Can we do a quick selfie? Phew."

"哇。谢谢 Cat。我们能先来张自拍吗？呼。"

"All right. Before I jump into this demo, I just wanna mention something. Everything that we're showing today still feels magical to me, and I work on Claude Code every day."

"好。开始 demo 之前我想先说一句——今天给你们看的这些东西，对我来说还是有点像魔法——而我每天都在做 Claude Code。"

"Even in Anthropic, we share screenshots back and forth of the cool things that people are building with Claude and things that people are doing in the wild, and honestly, I just feel excited to be on this journey together, um, discovering all of this."

"在 Anthropic 内部，我们也整天互发截图——看大家用 Claude 在搭些什么、外面在玩出什么。说实话，能和你们一起在这条路上一点一点发现这些东西，我自己觉得挺兴奋的。"

"So, uh, today I'm excited to share a few more examples of what this looks like."

"今天我想再多分享几个具体长什么样的例子。"

"Unfortunately, we can't all work in the lunar drone business, so for this demo, let's imagine that we're an engineer at Acme Pay, and it's a payments infrastructure company."

"很遗憾我们不可能都在做登月无人机的生意——所以这个 demo 里，我们假装自己是 Acme Pay 的一个工程师。Acme Pay 是一家做支付基础设施的公司。"

"We're gonna start at the Claude desktop app, and we're gonna start by working on a single task."

"我们从 Claude desktop app 开始，先跑一个任务。"

"In this session, Claude's working on adding refunds to Acme's merchant dashboard."

"这个 session 里，Claude 在给 Acme 的商户后台加一个退款功能。"

"It's building a full implementation. Uh, idempotency, so a duplicate webhook doesn't double refund a merchant."

"它要做的是一个完整实现——幂等性，重复 webhook 不会让一家商户被退两次款。"

"There is multi-currency handling across all the regions Acme serves, and audit logging for the compliance team."

"还有 Acme 服务的所有地区的多币种处理，加上给合规团队用的 audit log。"

"It writes the implementation, and it's gonna verify its own work."

"它先把代码写出来，然后会自己验证自己写的对不对。"

"Claude pulls up the merchant dashboard, it triggers a refund and... ah. There's no success toast. That's a real edge case."

"Claude 打开商户后台，触发了一笔退款——啊，没有"成功"的 toast 提示。这是一个真实存在的 edge case。"

"Claude sees the failure, it traces it back to a race condition in the optimistic update, it fixes it. And it's gonna verify that it actually works in a browser before it calls the task done."

"Claude 看到这次失败，把它追到 optimistic update 里的一个竞态条件上，然后把它修了。它会再到浏览器里实际验一遍，才敢说任务完成。"

"Now, uh, let's zoom out. This session wasn't running alone."

"我们镜头拉远一点——这个 session 不是一个人在跑。"

"It's actually one of many sessions that were all running in parallel and being managed in parallel."

"它其实是同时有很多 session 在并行跑、在被并行管理的其中一个。"

"In the Claude desktop app, you can now see all your Claude Code sessions — which ones are running, which ones need your input, and which have PRs that have already been merged and closed."

"在 Claude desktop app 里，你能看到自己所有 Claude Code session——哪些在跑、哪些等你输入、哪些已经把 PR merge 关掉了。"

"Synchronous coding is now just a slice of what's happening at any given moment."

"同步写代码（synchronous coding）今天只是任意时刻发生的事情中的一小块。"

"And we think that going forward, a lot more code is gonna start to be written in an async way."

"我们认为，往后会有越来越多的代码以异步的方式写出来。"

"And this is why we keep talking about verification."

"这也是我们为什么一直在反复说 verification。"

"If Claude can check its work, you can just let it run while you work on something else, and you come back to a fully working result."

"如果 Claude 能自己检查自己的活儿，那你就可以让它在那儿跑——你去干别的——回头你拿到的是一个完全跑得通的结果。"

"For me personally, a lot of my code nowadays is written by routines. I'm not the one doing the prompting. I'm the one creating a routine that does the prompting."

"我个人来说——今天我大量的代码都是由 routines 写的。不是我在 prompt，是我在写一个会去 prompt 的 routine。"

"For engineers in the room, think of it like a higher order function. Routines are a higher order prompt."

"对在场的工程师来说，把它想成 higher-order function——routines 就是 higher-order prompt。"

"For example, the refund session that we just looked at — a teammate filed a GitHub issue overnight."

"举个例子。我们刚才看的那个退款 session——其实是一个同事昨晚开了一个 GitHub issue。"

"Our routine watching the repo picked it up async, and then kicked off the work in Claude."

"我们一个盯着那个 repo 的 routine 异步抓到了它，然后就把活儿在 Claude 这边起起来了。"

"With routines, developers can set up async automations and wake up to PRs that are ready to merge."

"有了 routines，开发者可以配一套异步自动化——一觉醒来，PR 已经待 merge 了。"

"Here's our routines view."

"这是我们 routines 的视图界面。"

"Routines can be run on a schedule, they can be kicked off by web hooks, or they can even be kicked off by arbitrary API calls."

"routines 可以按计划跑、可以被 webhook 触发，甚至可以被任意 API 调用触发。"

"You can run them locally on your machine or on remote cloud compute."

"你可以让它在你本地机器上跑，也可以放在远程云算力上跑。"

"Let's look at one more feature. This is CI Autofix that Cat talked about earlier."

"再看一个能力——这就是 Cat 之前提到的 CI Autofix。"

"And what it's doing is it's watching the PR the prior session just opened. Its job is to babysit the PR to get it all the way to production."

"它在干的事——盯着上一个 session 刚提的那个 PR。它的工作就是把这个 PR 一路守到上线。"

"It's gonna autofix any comments from code review and security review."

"代码评审和安全评审里的所有评论，它会自动改掉。"

"It's gonna autofix CI, and it's gonna auto rebase if there's merge conflicts."

"CI 挂了它会自动修；遇到 merge conflict 它会自动 rebase。"

"And look at what just happened. The CI flaked on the network timeout, the routine woke up, it diagnosed it as a known infra issue, it retried the job, and now it's green."

"看刚才发生了什么——CI 因为网络超时挂了，routine 醒过来，把它诊断成一个已知的基础设施问题，重试了一次——现在绿了。"

"And actually, in the Claude Code code base, we have it not just retry — we have it fix the root cause every time."

"顺便说一下——在 Claude Code 自己的代码库里，我们不只让它重试，我们让它每次都直接修根因。"

"The engineer who owns the PR is never gonna see a red X, and that work is off their plate."

"那个 PR 的负责人永远看不到那个红色叉——这块活儿就从他的工位上拿走了。"

"And that's the shift. The default isn't, 'I'm gonna prompt Claude Code.' The default is now, 'I will have Claude prompt Claude Code.'"

"这就是那个范式 shift。默认动作不再是"我去 prompt Claude Code"——默认动作变成了："我让 Claude 去 prompt Claude Code"。"

"Everything you just saw is available today, including routines, the latest updates to the Claude desktop app."

"你刚才看到的所有东西今天就能用——包括 routines，包括 Claude desktop app 这次的最新更新。"

"We're excited for you to try it out and let us know what you think."

"快上手玩一玩——我们很想听你的反馈。"

"We hope these features continue to close the gap between your ideas and shipping products."

"我们希望这些能力继续把"你脑子里的点子"和"产品发出去"之间的距离压短。"

"And, uh, that's really what every talk today was pointing out."

"其实今天每一场演讲指的都是同一件事。"

"Diane's capability curve, Angela and Caitlyn's agents that grade and improve themselves, what Cat and I just showed you — these are three layers of one story."

"Diane 的能力曲线，Angela 和 Caitlyn 那些会自己打分、自己改进的 agent，Cat 和我刚才给你看的——这是同一个故事的三层。"

"The capability is already here. The gap left is how fast we put it to work."

"能力已经就位。剩下的差距，是我们多快把它用起来。"

"I encourage you to spend the rest of today exploring these layers."

"今天剩下的时间，建议你把这三层都摸一遍。"

"Research talks if you're evaluating the models, Claude Platform sessions if you're building for your users, uh, or Claude Code workshops if you wanna learn more ways to bring Claude into your day-to-day development work."

"如果你在做模型评测——去看 research talk；如果你在为你的用户做产品——去 Claude Platform 的分会场；如果你想学怎么把 Claude 揉进日常开发——去 Claude Code workshop。"

"Dive in, go deep, and start building with us. Thank you."

"扎进去，往深里走，跟我们一起开始 build。谢谢大家。"