Keep thinking it’d be cool to write a short sci-fi story called “The last prompt.” But then I remember Isaac Asimov did it already and it’s wonderful. In 1956.
我总觉得，写一个叫作“The last prompt”的短篇科幻故事会很酷。但随后我就想起，Isaac Asimov 早就写过了，而且写得非常精彩。那还是在 1956 年。
GR
每条推文 hover 显示单独 ▶
Fable just found a 15-30% memory efficiency improvement in Turbopack / Next.js, nearly autonomously. @tobi asked me today: what have been your “holy s***” moments with AI? My answer was: it’s every single week. And it’s accelerating. In fact, “WTFs/day” might just be my favorite metric for AI progress. It’s one thing to read benchmarks or stories online. It’s another to watch these machines pull engineering feats every day. 3 days ago? Sol helped us find novel vulnerabilities in some of the most audited code in the world. Today? Fable helps us ship this large optimization of a very complex Rust codebase. I just saw some results of work we’ve done to shrink binaries by 10-20x. List goes on.
Fable 刚刚在 Turbopack / Next.js 中发现了 15-30% 的内存效率提升，而且几乎是自主完成的。@tobi 今天问我：你和 AI 相处时，有哪些“holy s***”时刻？我的回答是：每一周都有。而且还在加速。事实上，“WTFs/day” 可能就是我最喜欢用来衡量 AI 进展的指标。读 benchmark（基准测试）或网上的故事是一回事；每天亲眼看着这些机器完成工程壮举，又是另一回事。3 天前？Sol 帮我们在世界上一些经过最严密审计的代码里找到了新的漏洞。今天？Fable 又帮助我们在一个非常复杂的 Rust 代码库中交付了这项大规模优化。我刚刚还看到一些我们在将二进制文件体积缩小 10-20 倍方面所做工作的结果。这样的例子还可以一直列下去。
Shopify + Vercel are two sides of the same coin. A congregation: entrepreneur-first, developer-led, web-obsessed. So excited to bring our infrastructure for apps and agents to millions of businesses together
Shopify + Vercel 是同一枚硬币的两面。它们汇聚在一起：以创业者为先，由开发者驱动，对 web 无比痴迷。非常兴奋能把我们面向 apps 和 agents 的基础设施，一起带给数百万家企业。
I've been dreaming about this ship for a while. Lots of tedious infra work behind the scenes that gave us beautiful results… Up to ⏱︎ 30% faster deployments, ⚡︎ 60% better time-to-first-byte, ▽ less data transfer usage, and ⛁ wildly more efficient underlying storage!
我已经期待这次发布有一阵子了。幕后做了大量枯燥但必要的 infra（基础设施）工作，最终带来了很漂亮的结果……部署速度最高提升 ⏱︎ 30%，time-to-first-byte（首字节时间）改善 ⚡︎ 60%，data transfer（数据传输）使用量更低 ▽，而且底层存储效率也提升了 ⛁ 非常多！
One line of code can save you a lot of 💰 on tokens!
一行代码就能帮你在 token 上省下不少 💰！
If you’re using an AI model router or gateway other than @vercel AI Gateway - tell me why! Reply or DM
如果你用的 AI model router（AI 模型路由器）或 gateway（网关）不是 @vercel AI Gateway——告诉我为什么！欢迎回复或 DM
The big lesson from AI is that everything is code. A slide deck is code. Design is code. That cool promo video? Code. Excel automation? Code. The universe? Probably made of code too.
AI 带来的最大启示是：一切都是 code。幻灯片 deck 是 code。设计是 code。那个很酷的宣传视频？也是 code。Excel 自动化？是 code。宇宙？大概也是由 code 构成的。
Thankful to 🇦🇷 Argentina for once again giving it all. Grit, passion, talent, faith. To Messi for proving he’s the best ever. And to America 🇺🇸 for hosting such a wonderful event. Enhorabuena tíos 🇪🇸!
感谢 🇦🇷 Argentina 再一次倾尽全力。坚韧、激情、天赋、信念。感谢 Messi 证明了他是有史以来最好的。也感谢 America 🇺🇸 主办了这样一场精彩的赛事。Enhorabuena tíos 🇪🇸！
Incidentally, I’m very bullish on cybersecurity as one of the best benchmarks for superintelligence. The “IQ test” of software engineering. The best engineers I’ve worked with in my career have usually had a deep background or interest in security. It’s actually easy for a model to “one-shot an XYZ clone” and impress people on X. But that’s not a good test. Finding, patching, reversing, and exploiting require a cognitive skill that transcends programming languages, runtimes, frameworks… It demands true reasoning power from the model and “corner thinking”. Very, very few humans excel at this, let alone in ordinary day-to-day software writing. Seeing Kimi K3 do so well here bodes well for open models.
顺便说一句，我非常看好 cybersecurity，把它视为衡量 superintelligence 的最佳标尺之一。它可以说是 software engineering 的“IQ test（智商测试）”。在我的职业生涯中，与我共事过的最优秀工程师，通常都有深厚的 security 背景，或至少对它有浓厚兴趣。实际上，让一个模型“一次生成一个 XYZ clone”并在 X 上惊艳众人，是很容易的。但那不是一个好的测试。发现、修补、逆向和利用漏洞，要求一种超越 programming languages、runtimes、frameworks 的认知能力……它要求模型具备真正的推理能力，以及“corner thinking（角落式思考）”。真正擅长这件事的人类都非常、非常少，更不用说在普通日常的软件编写中也是如此。看到 Kimi K3 在这方面表现这么好，对 open models 来说是个好兆头。
Based on internal evals: ▪️ Kimi K3 is top-tier at cybersecurity There is chatter on X that Moonshot benchmark-overfit. These are stealth evals. Model has raw IQ. ▪️ Sol is a leap ahead in cyber capability At a significantly higher cost, but quite remarkable still. ▪️ Fable refuses everything We couldn’t get it to complete the run at all. What’s interesting is that Sol in comparison was much more open to helping with defensive cyber hardening TL;DR: frontier, open-weight cybersecurity capability is here. Try it on for defensive purposes.
根据内部 eval（评测）：▪️ Kimi K3 在 cybersecurity（网络安全）方面属于第一梯队。X 上有人在议论 Moonshot 是 benchmark-overfit（对基准测试过拟合）。但这些是 stealth evals（隐藏评测）。这个模型有真材实料的 raw IQ。▪️ Sol 在网络安全能力上领先了一个量级。成本也显著更高，但依然相当惊人。▪️ Fable 则什么都拒绝。我们完全没法让它跑完整个流程。有意思的是，相比之下，Sol 更愿意帮助做防御性的网络安全加固。TL;DR：前沿级、open-weight（开放权重）的网络安全能力已经来了。出于防御目的，你可以试试看。
The term "AGI" has aged very poorly. AI can't possibly me more different from human intelligence. It's *far better* than human intelligence, for most economically-relevant tasks. But that doesn't make AIs better than humanity. There are wonderful things humans can do that these superintelligences (the better term) can't. For one, we care-for and care-about other humans in a powerful and inherent way, that we then teach machines to emulate. We must keep that as the top priority in everything we do. We have to be as pro-human-life as it gets. AIs can replace tasks you do, but they can't replace the proverbial you. This is why they horribly suck at writing! Even when given a corpus of your writing and 200 Skills and 50 subagents running in gRaPh LoOps, they produce robotic prose… because they are robots. I know one definitive way that people can become irrelevant. They stop being themselves: lose their identity, delegate all their writing, their unique thoughts, the creative ways in which they can use the machines. Opting out of 'weighing in'. I think we're safe though. There's nothing more distasteful right now than AI replies in social media or (🟢 T H I S T H I N G ) in landing pages. Quality and humanity will prevail.
“AGI” 这个词已经老化得很糟糕了。AI 不可能和 human intelligence（人类智能）更不同了。对于大多数具有经济相关性的任务来说，它*远远强于*人类智能。但这并不意味着 AIs 比 humanity（人类整体）更好。人类能做到许多美妙的事，而这些 superintelligences（我认为更准确的说法）做不到。首先，我们会以一种强大而内在的方式 care-for 和 care-about 其他人类，而机器只是后来被我们教着去模仿这一点。我们必须把这件事作为我们做任何事情时的最高优先级。我们必须尽可能坚定地站在人类生命这一边。AIs 可以替代你做的任务，但它们不能替代那个抽象意义上的“你”。这就是为什么它们写作会糟糕得离谱！哪怕给它们一整套你的写作 corpus（语料），再配上 200 个 Skills 和 50 个在 gRaPh LoOps 中运行的 subagents（子 agent），它们产出的依然是机械化的文字……因为它们本来就是机器人。我知道一种人真的会变得无关紧要的明确方式：那就是他们不再做自己——失去自己的身份，把所有写作都委托出去，把自己独特的想法，以及自己能以创造性方式使用机器的能力，都一并交出去。选择不再“发表看法”。不过我觉得我们还是安全的。眼下没有什么比社交媒体上的 AI 回复，或者落地页上的（🟢 T H I S T H I N G）更令人反感了。质量和人性终将胜出。
🆓 Sandbox data for downloads. Happy Friday! Time to ship more agents
🆓 可供下载的 Sandbox 数据。周五快乐！是时候发布更多 agent 了
Kimi K3 is the best performing model on ahead of Fable, reaching a comparable success rate in less time. This is the first time that an open model is ahead of all proprietary ones for this comprehensive web engineering benchmark. Notes: ▪️ Benchmarks don’t always tell the full story, although this is important signal, adding to mounting evidence that this could be a breakthrough moment for open models ▪️ No model as of yet has reached 100% completion on this set of evals. The top performer peaks at 92% and 96% “with help”
Kimi K3 在 ahead of Fable 上是表现最好的 model，以更短时间达到了可比的成功率。这是 open model 首次在这个综合性的 web engineering 基准测试中领先于所有 proprietary model。备注：▪️ benchmark 并不总能讲完整个故事，不过这仍是一个重要信号，也进一步印证了越来越多的证据：这可能是 open model 的一个突破性时刻 ▪️ 目前还没有任何 model 在这组 evals（评测）上达到 100% 完成率。当前最佳表现者的峰值是 92%，而“with help”时可达 96%
I’m excited to welcome two legends of developer tools, Pete Hunt (@floydophone) and Nick Schrock (@schrockn), to Vercel. Pete was one of the pioneers of @reactjs at Meta. He made an early bet to power Instagram Web with ⚛️ React, evangelizing it internally and externally. He will be running Frameworks and leading @nextjs. I couldn’t imagine a better person to lead React’s most popular framework to even greater heights. Nick co-invented @graphql, solving some of the gnarliest data infrastructure and access issues at Facebook scale, with a delightful developer experience. He will be working on Agentic Developer Experience, solving the problem of enabling the next billion agents and leading the way to a future of self-improving software. It’s a dream-come-true for a founder of a startup to welcome engineering minds of this caliber who are also wonderful humans. You probably want to work with them, and they’re hiring 😁. Their DMs are open, from job applications to bug reports!
我很高兴欢迎 developer tools 领域的两位传奇人物，Pete Hunt (@floydophone) 和 Nick Schrock (@schrockn)，加入 Vercel。Pete 是 Meta 的 @reactjs 先驱之一。他很早就押注用 ⚛️ React 来驱动 Instagram Web，并在公司内外积极推广它。他将负责 Frameworks，并领导 @nextjs。我想不出还有谁比他更适合带领 React 最受欢迎的 framework 迈向更高的高度。Nick 是 @graphql 的共同发明者之一，在 Facebook 这样的规模下，他解决了一些最棘手的数据基础设施和访问问题，同时还带来了令人愉悦的 developer experience（开发者体验）。他将投身于 Agentic Developer Experience，解决如何赋能下一个十亿 agent 的问题，并引领我们走向一个软件能够自我改进的未来。对于一位 startup 创始人来说，能欢迎这样级别的工程头脑加入，而且他们还都是非常棒的人，简直是梦想成真。你大概会想和他们一起工作，而他们也正在招人 😁。他们的私信是开放的，无论是投递工作申请还是提交 bug report（缺陷报告）都可以！
Vercel Sandbox: ◾ Growing DAUs at 100% m/o/m ◾ 3.5M+ sandboxes created per day ◾ Best-in-class Active CPU pricing model ◾ Powering @notion, @airtable, @meta, @zapier, @coderabbitai, @interaction, @conductor_build, @blackboxai… 🐐 farm My DMs are open for migration help or if missing anything with:
Vercel Sandbox：◾ DAU（月活跃用户数）按月环比以 100% 增长 ◾ 每天创建 350 万个以上的 sandbox ◾ 同类最佳的 Active CPU 定价模型 ◾ 为 @notion、@airtable、@meta、@zapier、@coderabbitai、@interaction、@conductor_build、@blackboxai… 提供支持 🐐 farm 如果你需要迁移帮助，或者觉得还缺少任何相关能力，欢迎私信我：
Some really cool usecases of Web Analytics API: ▪️ Ask your agent to correlate visitors, custom events (“purchase”, “checkout”), with the evolution of your deployments and performance ▪️ Build custom frontends, and plot this data alongside e.g.: Stripe’s and Resend’s
Web Analytics API 的一些非常酷的 use case（使用场景）：▪️ 让你的 agent 把访客、custom event（自定义事件，“purchase”“checkout”）与 deployment（部署）演进和性能变化关联起来 ▪️ 构建自定义 frontend（前端），并将这些数据与例如 Stripe 和 Resend 的数据一起绘制出来
We're opening up the dataset of AI token flows on @vercel AI Gateway. Fascinating insights contained within!
我们正在开放 @vercel AI Gateway 上 AI token（令牌）流动的数据集。其中包含了非常有意思的洞察！
The @agentmail team has been cooking. Tell your agent to 𝚟𝚎𝚛𝚌𝚎𝚕 𝚒𝚗𝚜𝚝𝚊𝚕𝚕 𝚊𝚐𝚎𝚗𝚝𝚖𝚊𝚒𝚕 – no signup, automatic setup and unified billing.
@agentmail 团队最近很能打。告诉你的 agent 去执行 𝚟𝚎𝚛𝚌𝚎𝚕 𝚒𝚗𝚜𝚝𝚊𝚕𝚕 𝚊𝚐𝚎𝚗𝚝𝚖𝚊𝚒𝚕——无需注册、自动设置、统一计费。
The two most popular features so far have been: 1️⃣ The ease of use / filesystem API 2️⃣ Observability We’re doubling down on both!
到目前为止，最受欢迎的两个功能是：1️⃣ 易用性 / filesystem API 2️⃣ 可观测性（Observability）我们会在这两方面继续加倍投入！
Powerful building block for autonomous, self-optimizing websites and applications. Give agents the ability to set up and tune experiments with feature flags.
这是构建 autonomous（自主式）、self-optimizing（自优化）网站和应用的强大基础模块。让 agents（智能体）能够用 feature flags（功能开关）来设置并调优实验。
“Open-weight models ran 29% of gateway tokens, up from 11% in April”
“Open-weight models 占 gateway tokens 的 29%，高于 4 月份的 11%”
Make the model a cog in a machine you own. ◾ AI SDK → open model API ◾ → open Agent API ◾ AI Gateway → open ZDR inference Startups and enterprises must own their data, evals, model choices, software layer. Don't outsource your brain.
让模型成为你所拥有的机器中的一个齿轮。◾ AI SDK → 开放的 model API（模型 API） ◾ → 开放的 Agent API ◾ AI Gateway → 开放的 ZDR inference（零数据保留推理） 初创公司和企业必须掌控自己的数据、evals（评估）、模型选择以及软件层。不要把你的大脑外包出去。
PS: open models are about to get exorbitantly fast. Watch this space
附注：open models 很快就会变得快得惊人。请关注后续动态
It’s model release week. I suspect Meta Spark 1.1, Grok 4.5, and GLM 5.2 will significantly displace token market share. Most agentic tasks require reasonably high intelligence at fast speeds. It’s a good time to be using ▲ AI Gateway!
这是 model 发布周。我怀疑 Meta Spark 1.1、Grok 4.5 和 GLM 5.2 会显著挤占 token 市场份额。大多数 agentic tasks（agent 式任务）都需要在高速下具备相当高的智能水平。现在正是使用 ▲ AI Gateway 的好时机！
You don't need a LinkedIn, you need a page on your website describing and linking to what you shipped
你不需要一个 LinkedIn，你需要的是你网站上的一个页面，用来说明并链接到你已经发布的东西
I now own Excited to give my children clean email handles. Domains are awesome. The internet is awesome.
我现在拥有了 Excited，可以给我的孩子们分配干净简洁的 email 地址。Domains 很棒。互联网很棒。
Mythos / Sol cybersecurity capabilities are equally useful in an offensive as well a defensive capacity. If adversaries get ahold of an equivalent offensive capability, it poses a serious threat to US companies that remain unaware of latent vulnerabilities. In the meantime, I strongly recommend running deepsec[1] or similar harnesses with the available frontier models. [1]
Mythos / Sol 的 cybersecurity（网络安全）能力无论用于 offensive（进攻性）还是 defensive（防御性）场景，都同样有用。如果对手掌握了同等级的 offensive capability（进攻能力），这将对那些仍未察觉潜在漏洞的 US 公司构成严重威胁。与此同时，我强烈建议使用现有的 frontier models（前沿模型）运行 deepsec[1] 或类似的 harnesses（测试框架）。[1]
Agents are particularly hard-to-debug software. For one, and by design, AI models behave in non-deterministic ways. Even two identical prompts don't always yield the same output. But agents are also complex distributed systems. They involve multiple steps of computation across functions and sandboxes, touching dozens of API services that can go down, rate limit you, etc. Nailing down observability out of the box for on Vercel was a key priority for the team, and the feedback so far has been 🔥↓
agent 尤其是很难调试的软件。首先，而且这是设计使然，AI 模型会以非确定性的方式运行。即使是两个完全相同的 prompt，也不总会产生相同的输出。但 agent 也是复杂的分布式系统。它们涉及跨函数和 sandbox 的多步计算，会触达几十个 API 服务，而这些服务可能会宕机、对你进行 rate limit 等等。为 Vercel 开箱即用地做好 observability（可观测性），是团队的一个关键优先事项，而到目前为止收到的反馈非常 🔥↓
I made a @hyperframes_ video of how I found out this morning
我做了一个 @hyperframes_ 视频，讲我今天早上是怎么发现这件事的
Next.js "Ways to fix this" with "Copy prompt" buttons is a work of (agentic) art
Next.js 带有“Copy prompt”按钮的“Ways to fix this”堪称（agentic）艺术之作
How we imbue coding agents with our design standards
我们如何让 coding agents 融入我们的设计标准
Grok Imagine Video has become the clear leader in @vercel AI Gateway, now amounting to ~50% of videos generated by developers 🤯
Grok Imagine Video 已成为 @vercel AI Gateway 中当之无愧的领跑者，如今约占开发者生成视频总量的 ~50% 🤯
AI will bring forth an unprecedented surge in entrepreneurship. From 'solopreneurs,' to the revitalization of the small & medium business segment, to the emergence of the largest companies of our times… providing the foundation of it all.
AI 将带来一波前所未有的创业浪潮。从“solopreneurs（单人创业者）”，到 small & medium business（中小企业）板块的复兴，再到我们这个时代最大公司的涌现……而这一切的基础正在被提供出来。
The data of tokens and uptime recovered by @vercel AI Gateway is truly astonishing
@vercel AI Gateway 恢复的 token 和 uptime（在线时长）数据，确实令人惊叹
𝚆𝚎𝚋𝚂𝚘𝚌𝚔𝚎𝚝 (and 𝚜𝚘𝚌𝚔𝚎𝚝.𝚒𝚘) now supported on Vercel, from CDN to Fluid. Special full circle moment for @vercel
现在 Vercel 已支持 𝚆𝚎𝚋𝚂𝚘𝚌𝚔𝚎𝚝（以及 𝚜𝚘𝚌𝚔𝚎𝚝.𝚒𝚘），覆盖从 CDN 到 Fluid。对 @vercel 来说，这真是一个特别的圆满时刻
The team cooked on performance. 'Everything the light touches' was optimized Simba. Painting, layout, WebGPU shaders, blocking scripts. Every frame scrutinized. The best part is that we'll be updating our with the lessons learned!
团队在性能方面下了狠功夫。凡是“阳光照到的每一处”，Simba 都做了优化。绘制、布局、WebGPU shader、阻塞脚本，全都涵盖。每一帧都经过仔细审视。最棒的是，我们还会用这些经验教训来更新我们的——！
Coding agents will squeeze every ounce of IKEA effect out of you, if you let them.
如果你放任 coding agent，它们会把你身上的每一分 IKEA effect 都榨得干干净净。
Happy Father's Day. Thankful to my dad who spent every last penny on buying our family 'the bicycle of the mind'
父亲节快乐。感谢我的父亲，把家里最后一分钱都花在给我们家买来“the bicycle of the mind”上。
Genuinely impressed, almost shocked, at how good GLM-5.2 by @zai_org is at coding. This changes things.
我真的对 @zai_org 的 GLM-5.2 在 coding（编程）方面有多厉害感到印象深刻，几乎可以说是震惊。这会改变很多事情。
The next hot programming language is… markdown. A minimal eve agent: 📂 𝚊𝚐𝚎𝚗𝚝/ 📄 𝚒𝚗𝚜𝚝𝚛𝚞𝚌𝚝𝚒𝚘𝚗𝚜.𝚖𝚍 📂 𝚜𝚔𝚒𝚕𝚕𝚜/ 📄 𝚢𝚘𝚞𝚛-𝚎𝚡𝚙𝚎𝚛𝚝𝚒𝚜𝚎.𝚖𝚍 Deployable in one command: 𝚟𝚎𝚛𝚌𝚎𝚕. It’s the most accessible programming has ever been. And likely will ever be, at least for the generation of software fully defined and controlled by us humans. (As a fun fact, one of the initial prototypes for eve was codenamed 𝚕𝚊𝚜𝚝 by @timolins, both in homage to ‘@nextjs for agents’ but also in recognition of how enduring eve’s design feels to us.)
下一个热门编程语言是……markdown。一个最小化的 eve agent：📂 𝚊𝚐𝚎𝚗𝚝/ 📄 𝚒𝚗𝚜𝚝𝚛𝚞𝚌𝚝𝚒𝚘𝚗𝚜.𝚖𝚍 📂 𝚜𝚔𝚒𝚕𝚕𝚜/ 📄 𝚢𝚘𝚞𝚛-𝚎𝚡𝚙𝚎𝚛𝚝𝚒𝚜𝚎.𝚖𝚍 只需一条命令即可部署：𝚟𝚎𝚛𝚌𝚎𝚕。这是编程有史以来最易于接触的时刻。并且很可能也将一直如此，至少对于那一代仍由我们人类完整定义并控制的软件而言是这样。（顺带一提，eve 的一个早期原型由 @timolins 命名为 𝚕𝚊𝚜𝚝，这既是在致敬“@nextjs for agents”，也体现了我们觉得 eve 的设计具有一种持久生命力。）
Agents are motivating so many healthy software habits. Open APIs, documentation (skills), tests (evals), Unix (CLIs), payment & commerce protocols, even wide 𝙰𝚌𝚌𝚎𝚙𝚝 use (markdown/json/html). The original vision of the WWW coming to life before our eyes.
Agents 正在推动许多健康的软件习惯。开放 API（应用程序编程接口）、文档（skills）、测试（evals）、Unix（CLIs）、支付与商业协议，甚至广泛使用 𝙰𝚌𝚌𝚎𝚙𝚝（markdown/json/html）。WWW 最初的愿景，正在我们眼前变成现实。
[1]
[1]
(Part of) the wonderful team that made Vercel SHIP London possible 🫶
让 Vercel SHIP London 得以成为现实的优秀团队中的一部分 🫶
React → Next.js → @aisdk is more relevant than ever, given the intense model competition landscape. Just today, GLM 5.2, an open model, surpassed Opus 4.8 in our Next.js Evals ( 🤯 But the world needs a practical solution for how to build and deploy agents. Just like React needed Next.js to solve the task of building an actual web application. And that's eve.
鉴于当下激烈的 model（模型）竞争格局，React → Next.js → @aisdk 这条演进路径比以往任何时候都更相关。就在今天，GLM 5.2 这个 open model（开放模型）在我们的 Next.js Evals 中超过了 Opus 4.8（🤯 但这个世界需要一个关于如何构建和部署 agent（智能体）的实用解决方案。就像 React 需要 Next.js 来解决构建一个真正的 web application（Web 应用）的任务一样。而这正是 eve。
Quick one to set the stage: ☑️ 30 minute function invocations 🆕 24 hour sandbox lifetimes
先快速交代一下背景：☑️ 30 分钟的 function 调用时长 🆕 24 小时的 sandbox 生命周期
v0 commits to shipping the best skills by default. Our goal is to give you the equivalent of a @vercel product engineer like @shuding and @shadcn on each prompt. But you can also now grab any skill from or add your team’s private set.
v0 承诺默认交付最佳的 skills。我们的目标是：在你的每一次 prompt 中，都相当于配备一位像 @shuding 和 @shadcn 这样的 @vercel 产品工程师。但现在，你也可以从任意 skill 集中获取 skill，或者添加你团队的私有 skill 集。
A sandbox A function A server A build Are you getting it!? These are all expressions of the same underlying compute infrastructure. With tweaks to load balancing, concurrency, persistence, overcommit… 2026 is the year serverless and servers finally converge. With no gotchas
一个 sandbox（沙箱） 一个 function（函数） 一个 server（服务器） 一个 build（构建） 你明白了吗！？这些其实都是同一底层计算基础设施的不同表现形式。只要对 load balancing（负载均衡）、concurrency（并发）、persistence（持久性）、overcommit（超额分配）等做一些调整……2026 将是 serverless（无服务器）与 servers（服务器）最终走向融合的一年。而且没有任何 gotchas（坑）。
One of our most requested features, longer Vercel function runtime, is here. What looks like an innocent tweak of a constant… is actually the conclusion a multi-year compute platform investment. Builds, Sandbox, and now Functions run on our homegrown microVM-based Fluid compute infrastructure. This investment has enabled innovations like function multi-concurrency, Active CPU pricing, and Secure Compute for private connectivity to existing cloud workloads… with lots more to come, in fact, in the coming days and weeks 😎
我们最常被请求的功能之一——更长的 Vercel function runtime（函数运行时长）——现在已经上线。看起来这只是对某个常量做了一个不起眼的调整……实际上，它是我们多年 compute platform（计算平台）投入的成果总结。Builds、Sandbox，以及现在的 Functions，都运行在我们自研的、基于 microVM 的 Fluid 计算基础设施之上。这项投入已经促成了诸如 function multi-concurrency（函数多并发）、Active CPU pricing（Active CPU 计费）以及用于私有连接到现有云工作负载的 Secure Compute 等创新……而且接下来还会有更多，实际上，就在未来几天和几周内 😎
My flight to London is Starlink-enabled 😭 The greatest advancement to air travel since the Wright brothers. God bless America
我飞往 London 的航班配备了 Starlink 😭 这是自 Wright brothers 以来航空旅行最伟大的进步。God bless America
has passed 700,000 skills. wild! all organic and community-driven. the open⎵ai ecosystem!
已经超过 700,000 个 skills 了。太夸张了！全部都是自然增长、由社区驱动的。这个开放的 ai 生态系统！
If you don’t love her at her foggiest, you don’t deserve at her sunniest
如果你不能在她最雾气蒙蒙的时候爱她，你也不配在她最阳光灿烂的时候拥有她
[1]
We just shipped 𝙷𝚊𝚛𝚗𝚎𝚜𝚜𝙰𝚐𝚎𝚗𝚝, a unified abstraction to orchestrate and integrate any agent’s “brain” into your app. @aisdk now frees you from both model and agent lock-in. (And it doesn’t just get you portability, it’s also delightful to use ofc!)
我们刚刚发布了 𝙷𝚊𝚛𝚗𝚎𝚜𝚜𝙰𝚐𝚎𝚗𝚝：这是一个统一的 abstraction（抽象层），可用来编排并集成任何 agent（智能体）的“brain”到你的 app 中。@aisdk 现在让你同时摆脱 model 和 agent lock-in（锁定）。（而且它带来的不只是 portability（可移植性），用起来也真的很顺手，ofc！）
Vercel + Shopify is too good… by @foda: ◾ 500+ orders processed in *2 minutes* ◾ Built with @v0 + @cursor_ai ◾ Fully custom @nextjs storefront on headless So long on the web. Anyone can now dream → build → ship → sell
Vercel + Shopify 实在太强了……作者 @foda：◾ 在 *2 分钟* 内处理了 500+ 笔订单 ◾ 使用 @v0 + @cursor_ai 构建 ◾ 在 headless（无头架构）上打造了完全自定义的 @nextjs storefront（店面前端） Web 的漫长周期结束了。现在任何人都可以从梦想 → 构建 → 发布 → 销售
🇬🇧 London calling Excited for Vercel Ship next week Some special announcements…
🇬🇧 London 在召唤，期待下周的 Vercel Ship，会有一些特别的公告……
What I love about Silicon Valley is that the future is up for grabs, ready for anyone to build. I get intros for angel investing to all kinds of people. I take everyone equally seriously. 2 lads & a dog, or a 5-time award-winning entrepreneur. No place more meritocratic.
我喜欢 Silicon Valley 的一点在于，未来仍有待争取，任何人都可以去创造。我会收到各种人的 angel investing（天使投资）引荐。我对每个人都一视同仁地认真看待。无论是 2 个小伙子和一条狗，还是一位 5 次获奖的创业者。没有哪个地方比这里更 meritocratic（任人唯贤）。
Vercel CLI now allows you to: ◾ create AI Gateway API keys ◾ pass a --𝚋𝚞𝚍𝚐𝚎𝚝 to cap their spend ◾ set a --𝚛𝚎𝚏𝚛𝚎𝚜𝚑-𝚙𝚎𝚛𝚒𝚘𝚍 for the quota Think of it as virtual credit cards for AI tokens 🤖💳
Vercel CLI 现在允许你：◾ 创建 AI Gateway API key ◾ 传入 --𝚋𝚞𝚍𝚐𝚎𝚝 来为其支出设置上限 ◾ 设置 --𝚛𝚎𝚏𝚛𝚎𝚜𝚑-𝚙𝚎𝚛𝚒𝚘𝚍 作为配额刷新周期 可以把它理解为 AI token（令牌）的虚拟信用卡 🤖💳
Opus wrote us a VM and then Mythos verified it
Opus 为我们写了一个 VM，然后 Mythos 对它进行了验证
Vercel AI Gateway recovers on average over 1T tokens a month 🤯 Much like Stripe recovers revenue with smart retries on failed payments or credit card updates. And we do it with 0️⃣ zero markup over the labs; adding redundancy, zero-data retention enforcement, observability, usage APIs, caps, …
Vercel AI Gateway 平均每月可挽回超过 1T tokens 🤯 很像 Stripe 通过对失败支付进行智能重试或更新信用卡信息来挽回收入。而且我们是在 labs 原价之上 0️⃣ 零加价的情况下做到这一点；同时还提供冗余、zero-data retention（零数据留存）强制执行、observability（可观测性）、usage APIs（用量 API）、上限控制，……
Agent filesystem state can now be read, written and mounted independently of Sandbox lifecycle. We’ve been developing a novel virtual storage infrastructure solution for all our compute products. Storage that’s decoupled, but attachable to Builds, Functions, Sandboxes & more.
现在，agent 的文件系统状态可以独立于 Sandbox 生命周期进行读取、写入和挂载。我们一直在为所有计算产品开发一种新颖的虚拟存储基础设施解决方案。这种存储与运行环境解耦，但可以挂载到 Builds、Functions、Sandboxes 等之上。
Shoutout to @andrewqu founder and CEO of 😁 you can just ship things
特别致敬 @andrewqu，😁 you can just ship things 的 founder 和 CEO。
Use the Skills API to make all your agents and platforms smarter. Think of it as the npm registry for agent capabilities and extensibility. Free and open.
使用 Skills API，让你的所有 agent 和平台都更智能。可以把它看作是用于 agent 能力与可扩展性的 npm registry。免费且开放。
Congrats Void team! We @vercel reaffirm our collaboration on an open platform for the web, with our investment in @nitrojsdev, open runtimes, and native support for Vite-based frameworks like Nuxt, Svelte, and TanStack Start 🫡
祝贺 Void 团队！我们 @vercel 通过投资 @nitrojsdev、推动开放 runtime，以及为基于 Vite 的框架（如 Nuxt、Svelte 和 TanStack Start）提供原生支持，重申我们在构建面向 web 的开放平台上的合作 🫡
Grok Imagine Video on @vercel AI Gateway – the top image-to-video model on
@vercel AI Gateway 上的 Grok Imagine Video——当前顶级的 image-to-video 模型，运行于
▲ + ❄️ Generating frontends on top of your business data is one of the killer apps of coding AI. The genie is out of the bottle. Never going back to clunky and rigid dashboards. @vercel already ran on Snowflake. But with @v0 and @nextjs, we’re now getting 1000x the value.
▲ + ❄️ 在你的业务数据之上生成 frontend，是 coding AI 最具杀手级应用潜力的场景之一。精灵已经出瓶了。我们再也不会回到那些笨重又僵化的 dashboard。@vercel 本来就已经跑在 Snowflake 上了。但有了 @v0 和 @nextjs，我们现在获得的价值高出了 1000 倍。
YES-CODE An entire category of software, "no-code", was built under the presumption that code is expensive, difficult, and scarce. Coding agents have forever changed the equation. Code is now cheap, easy, and abundant. I remember @cramforce being asked by an analyst long ago: "soo, is @vercel like a no-code platform?" Without hesitation he goes "no, it's the absolute opposite. It's a yes-code platform." 😁 A key thing we set out to do was to be uncompromising in quality and sophistication of the things you could host. No-code solutions took shortcuts and set hard ceilings, typically on performance and sophistication. Our mission is to create the easiest cloud for agents that you never graduate from. PS: welcome to Vercel!
YES-CODE 有整整一类软件——“no-code”——是建立在这样一个前提之上的：code 很昂贵、很困难、也很稀缺。coding agents 永远地改变了这个等式。现在，code 变得便宜、容易，而且充沛。我记得很久以前，有位分析师问 @cramforce：“所以，@vercel 算是一个 no-code platform 吗？” 他毫不犹豫地回答：“不，它恰恰相反。它是一个 yes-code platform。” 😁 我们从一开始就想做的一件关键事情，是对你能托管的东西的质量和复杂度绝不妥协。no-code 方案会走捷径，并设下很硬的上限，通常体现在性能和复杂度上。我们的使命，是为 agents 打造最容易使用、而且你永远不需要“毕业离开”的 cloud。附：欢迎来到 Vercel！
This is what education in the age of AI looks like. Start with the language: the linguistic surface is your roadmap. Much like an SDK has a set of function definitions, human language is the new API to the world. It also always has been… there's never been an expert in any field without mastery over its language. The distinction is that English alone could not produce tangible things. You had to translate it into machine instruction, through learning or delegation to others. You can now go direct.
这就是 AI 时代的教育应有的样子。先从语言开始：语言层面的表面结构就是你的路线图。就像一个 SDK 会有一组函数定义一样，人类语言正在成为通向世界的新 API（应用程序接口）。其实它一直都是……任何领域里从来都不存在那种不掌握该领域语言却能成为专家的人。区别在于，单靠 English 过去并不能产出有形的东西。你必须把它翻译成机器指令，要么自己学会，要么委托给别人。现在你可以直接到达那一步了。
Conductor has the distinct edge of being an IDE born for coding agents. An ADE if you will. Agents will take remote dev mainstream. At most large and sensitive organizations, “local” isn’t a thing. But this future wasn’t evenly distributed. Excited to power their Sandbox!
Conductor 的一个鲜明优势在于，它是一个专为 coding agents 而生的 IDE。或者如果你愿意，也可以叫它 ADE。Agents 将把 remote dev 带入主流。在大多数大型且敏感的组织里，“local” 根本不是一回事。但这个未来的分布并不均匀。很高兴为他们的 Sandbox 提供支持！
Unclear if a durable trend, but CEOs and CTOs are back to coding with a fury, thanks to coding agents. I have public company CEOs sliding into my DMs (and “InMail”) telling me about falling in love with shipping software again thanks to Claude Code and Vercel. “Dream accounts” that we always wanted to work with, where in the past the C-suite would hardly understand the infrastructure until much later in the game. Coding agents are the ultimate PLG-fication of the enterprise. Bad, legacy software can’t hide anymore. The stack that works is self-evident to the entire organization, from intern to CEO.
这是不是一种持久趋势还不清楚，但多亏了 coding agents，如今 CEOs 和 CTOs 又开始火力全开地写代码了。我有一些上市公司 CEO 直接滑进我的私信（以及“InMail”）里，告诉我多亏了 Claude Code 和 Vercel，他们又重新爱上了交付软件（shipping software）。还有那些我们一直梦寐以求想合作的“dream accounts”，过去在这类客户里，C-suite 往往要到项目进行到很后面，才勉强理解底层基础设施。Coding agents 是企业领域最彻底的 PLG-fication。糟糕的 legacy software 再也藏不住了。真正有效的技术栈（stack）对整个组织来说都变得一目了然，从 intern 到 CEO 都看得清清楚楚。
Ship the best product. Use lots of AI, some AI, maybe no AI. Just be the best.
做出最好的产品。可以大量使用 AI，也可以只用一些 AI，甚至完全不用 AI。关键是：只要做到最好。
Next.js Night · AMS 🇳🇱 · June 11 Learn what’s next + meet the team + share feedback
Next.js Night · AMS 🇳🇱 · 6 月 11 日 了解下一步的新进展 + 见见团队 + 分享反馈
Feedback is a gift. Critical feedback doubly so.
反馈是一份礼物，而批判性的反馈更是加倍珍贵。
This will bring AI to 42% of the web. Every model, every provider, every modality (text, image, video, audio).
这将把 AI 带到 42% 的 web（网络）上。每一个 model（模型）、每一个 provider（提供方）、每一种 modality（模态：text、image、video、audio）。
We’re shipping a CDN pricing model that “smooths over” traffic spikes and viral events. This allows us to remove stress and surprises, with uncompromising quality and network performance. Predictability without slow routes, degraded PoPs, or priority bandwidth tiers.
我们正在推出一种 CDN 定价模型，用来“平滑处理”流量高峰和爆火传播事件。这样一来，我们就能在丝毫不妥协质量与网络性能的前提下，消除压力和意外情况。实现可预测性，而不需要慢速路由、性能降级的 PoP，或分级优先带宽。
Claude Managed Agents 🤝 Vercel Sandbox
Claude Managed Agents 🤝 Vercel Sandbox
This is so cool: rerun.io Stack: @sveltejs / @threejs / @vercel
这太酷了：rerun.io 技术栈：@sveltejs / @threejs / @vercel
All Firewall mitigations are now fully free on @vercel. Not just DDoS and system-level mitigations, but also any rule you configure. Vercel now absorbs the computational and network costs of any size of attack or traffic mitigation for your peace of mind. (And more to come!)
现在，@vercel 上所有 Firewall（防火墙）缓解措施都已完全免费。不只是 DDoS 和系统级缓解，也包括你自行配置的任何规则。现在，无论攻击或流量缓解的规模有多大，Vercel 都会承担相应的计算和网络成本，让你更安心。（而且后面还会有更多！）
A 𝚏𝚒𝚛𝚎𝚠𝚊𝚕𝚕 your agents will love. One of the coolest things about Vercel’s Firewall is how hard the team worked on instant global propagation (~300ms). Imagine if 𝚒𝚙𝚝𝚊𝚋𝚕𝚎𝚜 took *minutes* to propagate? That’s the average industry CDN/WAF experience!
一个你的 agent（智能代理）一定会喜欢的 𝚏𝚒𝚛𝚎𝚠𝚊𝚕𝚕。Vercel 的 Firewall 最酷的一点之一，就是团队为“即时全球传播”下了很大功夫（约 300ms）。你可以想象一下，如果 𝚒𝚙𝚝𝚊𝚋𝚕𝚎𝚜 的传播要花上 *几分钟* 会怎样？这恰恰就是行业里一般 CDN/WAF 体验的平均水平！
Grok CLI has great support for Plugins and Skills. Installing the Vercel Plugin gives Grok cloud deployment superpowers. Watch this creative coding website be generated with Grok and hosted seamlessly on Vercel ↓
Grok CLI 对 Plugins 和 Skills 的支持非常出色。安装 Vercel Plugin 之后，Grok 就拥有了云端部署的超能力。看看这个创意编程网站如何由 Grok 生成，并无缝托管到 Vercel 上 ↓
Vercel protects your agents' deployments behind SSO like @Okta. Even Production ones, giving you a secure 'intranet' of apps generated with @v0, Codex, Claude, etc. It's all fun and games until your agent gets 𝟺𝟶𝟷 𝚄𝚗𝚊𝚞𝚝𝚑𝚘𝚛𝚒𝚣𝚎𝚍 from the deployment *it just made* though 😂 The solution: 𝚟𝚎𝚛𝚌𝚎𝚕 𝚌𝚞𝚛𝚕. By just prefixing 𝚟𝚎𝚛𝚌𝚎𝚕 or 𝚟𝚌 you unblock your agent and yourself to easily 𝚌𝚞𝚛𝚕 any URL you have access to within the Vercel ecosystem.
Vercel 会用像 @Okta 这样的 SSO（单点登录）把你的 agents 的部署保护起来。甚至包括 Production 环境的部署，从而为你提供一个由 @v0、Codex、Claude 等生成应用构成的安全“intranet（内联网）”。不过，乐子归乐子，要是你的 agent 被它自己刚刚创建的部署返回一个 𝟺𝟶𝟷 𝚄𝚗𝚊𝚞𝚝𝚑𝚘𝚛𝚒𝚣𝚎𝚍，那就尴尬了 😂 解决方案是：𝚟𝚎𝚛𝚌𝚎𝚕 𝚌𝚞𝚛𝚕。只要在命令前加上 𝚟𝚎𝚛𝚌𝚎𝚕 或 𝚟𝚌 作为前缀，你和你的 agent 就都能被“解锁”，从而轻松 𝚌𝚞𝚛𝚕 任何你在 Vercel ecosystem 中有权访问的 URL。
If you become exceptional at managing agents, but are also exceptional in your understanding of the fundamentals, you will be unstoppable. We all prefer to work with masters of their craft. What’s new: you can’t afford to miss out on the amplification agents have on your output
如果你在管理 agents 方面变得非常出色，同时对基础原理的理解也同样非常扎实，那么你将势不可挡。我们都更愿意与精通自己技艺的大师共事。新的现实是：你承受不起错过 agents 对你产出所带来的放大效应。
You can just render images on the terminal btw: ▲ ~/ npx ai-cli image 'a vercel ai sdk diagram' Run 𝚗𝚙𝚖 𝚒 -𝚐 𝚊𝚒-𝚌𝚕𝚒 and access every image, video & text model from @vercel AI Gateway instantly
顺便说一句，你完全可以直接在 terminal 里渲染图片：▲ ~/ npx ai-cli image 'a vercel ai sdk diagram' 运行 𝚗𝚙𝚖 𝚒 -𝚐 𝚊𝚒-𝚌𝚕𝚒，即可立刻通过 @vercel AI Gateway 访问所有 image、video 和 text model
Moms are the bedrock of civilization. I’m very thankful to my mom who put herself through engineering school in Argentina and set the bar so high for her children. Happy Mother’s Day!
妈妈们是文明的基石。我非常感谢我的妈妈——她在 Argentina 一边自力完成 engineering school（工程学院）学业，一边为她的孩子们把标准树立得如此之高。母亲节快乐！
𝚗𝚙𝚡 𝚍𝚎𝚎𝚙𝚜𝚎𝚌 We're introducing an open-source agent orchestrator for deep security reviews. We built it for internal use, and after running it against some major OSS projects, we gained conviction to share it with the world. Coding agents can now find critical vulnerabilities in minutes that would take teams of people months (if they can spot them at all). Since 𝚍𝚎𝚎𝚙𝚜𝚎𝚌 is optimized to work with Vercel Sandbox, you can effectively harness the power of thousands of agents scrutinizing your codebase in parallel. I encourage you to try this on your repositories. BTW: If you run an OSS project and want us to sponsor a run, my DMs are open.
𝚗𝚙𝚡 𝚍𝚎𝚎𝚙𝚜𝚎𝚌：我们推出了一个用于深度安全审查的开源 agent 编排器（orchestrator）。我们最初是为内部使用而构建它的；在将它运行到一些大型 OSS 项目之后，我们更加确信值得把它分享给全世界。如今，coding agents 可以在几分钟内发现关键漏洞，而这类问题原本可能需要团队花上数月才能找出来——前提还是他们最终能发现。由于 𝚍𝚎𝚎𝚙𝚜𝚎𝚌 针对 Vercel Sandbox 做了优化，你实际上可以并行调动成千上万个 agent 来审查你的代码库。我鼓励你在自己的仓库上试试它。顺便一提：如果你在运营一个 OSS 项目，并且希望我们赞助一次运行，欢迎私信我。
𝚗𝚙𝚡 𝚍𝚎𝚎𝚙𝚜𝚎𝚌 We're introducing an open-source agent orchestrator for deep security reviews. We built it for internal use, and after running it against some major OSS projects, we gained conviction to share it with the world. Coding agents can now find critical vulnerabilities in minutes that would take teams of people months (if they can spot them at all). Since 𝚍𝚎𝚎𝚙𝚜𝚎𝚌 is optimized to work with Vercel Sandbox, you can effectively harness the power of thousands of agents scrutinizing your codebase in parallel. I encourage you to try this on your repositories. BTW: If you run an OSS project and want us to sponsor a run, my DMs are open.
𝚗𝚙𝚡 𝚍𝚎𝚎𝚙𝚜𝚎𝚌：我们推出了一个用于深度安全审查的开源 agent 编排器（orchestrator）。我们最初是为内部使用而构建它的；在将它运行到一些大型 OSS 项目之后，我们更加确信值得把它分享给全世界。如今，coding agents 可以在几分钟内发现关键漏洞，而这类问题原本可能需要团队花上数月才能找出来——前提还是他们最终能发现。由于 𝚍𝚎𝚎𝚙𝚜𝚎𝚌 针对 Vercel Sandbox 做了优化，你实际上可以并行调动成千上万个 agent 来审查你的代码库。我鼓励你在自己的仓库上试试它。顺便一提：如果你在运营一个 OSS 项目，并且希望我们赞助一次运行，欢迎私信我。
Asked @v0 what it would look like if @vercel shipped @github 😁 (2 prompts)
我问了 @v0：如果 @vercel 推出了 @github，它看起来会是什么样子 😁（2 个 prompt）
We're growing the Vercel Labs team. Our mission is to build the devtools of the AI era. We used to build tools for humans, now we're building them for agents. @ctatedev & team have shipped the tools that agents love: 𝚊𝚐𝚎𝚗𝚝-𝚋𝚛𝚘𝚠𝚜𝚎𝚛, 𝚙𝚘𝚛𝚝𝚕𝚎𝚜𝚜, 𝚜𝚔𝚒𝚕𝚕𝚜, 𝚌𝚑𝚊𝚝, 𝚓𝚞𝚜𝚝-𝚋𝚊𝚜𝚑, 𝚓𝚜𝚘𝚗-𝚛𝚎𝚗𝚍𝚎𝚛… 22.8M+ downloads to date. Dream job. DMs open 😀
我们正在扩充 Vercel Labs 团队。我们的使命是打造 AI 时代的 devtools（开发工具）。我们过去为人类构建工具，现在则在为 agents（智能体）构建工具。@ctatedev 和团队已经推出了 agents 喜爱的工具：𝚊𝚐𝚎𝚗𝚝-𝚋𝚛𝚘𝚠𝚜𝚎𝚛、𝚙𝚘𝚛𝚝𝚕𝚎𝚜𝚜、𝚜𝚔𝚒𝚕𝚕𝚜、𝚌𝚑𝚊𝚝、𝚓𝚞𝚜𝚝-𝚋𝚊𝚜𝚑、𝚓𝚜𝚘𝚗-𝚛𝚎𝚗𝚍𝚎𝚛……迄今下载量已超过 22.8M。这是理想工作。欢迎私信 😀
I’m so encouraged by the way our team and industry peers have shown up to protect the internet. We’ve now shipped over 20 product improvements across Dashboard and CLI to help your security posture. Easier to set up MFA, audit your Environment Variables, Activity logs and more
我深受鼓舞，因为我们的团队和业内同行都挺身而出，共同保护互联网。我们现在已经在 Dashboard 和 CLI 上发布了 20 多项产品改进，以帮助提升你的安全态势。包括更容易设置 MFA（多因素认证）、审计你的 Environment Variables（环境变量）、查看 Activity logs（活动日志）等等
Getting lots of questions about how to learn more about the incident. We're actively maintaining the security bulletin. That's the source. The bulletin includes security best practices to take out of an abundance of caution. To reiterate, we directly contacted all Vercel customers that we believe to be impacted by the IOC shared in the bulletin. One misconception we've seen that I need to call out. Deletion (e.g.: of an env var, project, account…) does not imply Rotation. Rotating keys means *invalidating* the previous value with the vendor/service you're using, and getting a new one. Do that. i.e.: if you only delete the resource on the Vercel side, the associated key can "live on" with the other provider, and be mis-used
很多人都在问，怎样才能进一步了解这次 incident（安全事件）。我们正在持续维护 security bulletin（安全公告）。那就是信息来源。公告中还包含一些出于极度谨慎而建议采取的安全最佳实践。再次重申，我们已经直接联系了所有我们认为受到公告中共享的 IOC（妥协指标）影响的 Vercel 客户。我们看到的一个误解，我必须特别指出：Deletion（删除，例如删除某个 env var、project、account……）并不等于 Rotation（轮换）。Rotating keys 的意思是：在你所使用的 vendor/service 那里，将旧的值*作废*，并获取一个新的。请务必这样做。也就是说：如果你只是在 Vercel 一侧删除了相关资源，那么对应的 key 仍可能在另一个 provider 那里“继续有效”，并被滥用