Speaker 100:00 - 00:17
Fraudsters have figured out that in AI, you actually don't really need to steal money or credentials. You can just steal tokens. And the scale of this actually shocked me when I looked at the data. So more than one in six signups at AI companies are this kind of abuse. Whatever. Speaker 100:00 - 00:17
Fraudsters 已经发现，在 AI 领域，你其实根本不一定需要偷钱或 credentials（凭证）。你只需要偷 token（令牌）就行了。而当我去看数据时，这件事的规模其实让我很震惊。所以，AI 公司里超过六分之一的注册都是这种滥用行为。总之。
Speaker 100:17 - 00:41
The dine and dash, but it's for token. When I go and ask my friends and family whether they'd be comfortable letting an agent buy things on their behalf, they usually jump straight to like, well, is it gonna overspend, and is it gonna buy the wrong thing, and can I stop it? And those are actually all legitimate concerns. And it's not Emily permissioning an agent to buy on her behalf. It's Emily has an agent who's tasked with running a business, and that includes buying some things and selling some things and making some profits. Speaker 100:17 - 00:41
这就像吃霸王餐再跑路，只不过对象变成了 token。当我去问我的朋友和家人，他们是否愿意让一个 agent 代他们买东西时，他们通常会立刻想到：它会不会花超预算？会不会买错东西？我能不能叫停它？而这些其实都是完全合理的担忧。但这里并不是 Emily 给一个 agent 授权，让它代自己买东西。而是 Emily 拥有一个 agent，这个 agent 的任务是经营一家企业，而这其中就包括买入一些东西、卖出一些东西，以及赚取一些利润。
Speaker 100:41 - 00:45
And that'll be the world that I would like to be talking about twelve months from now. Speaker 100:41 - 00:45
而那会是我希望在十二个月后的今天能够讨论的世界。
Speaker 200:45 - 00:58
Hi. I'm Matt Turk. Welcome back to the Matt podcast. Today, I'm excited to be joined again by Emily Sense, head of data and AI at Stripe. The rise of agentic commerce has only accelerated since Emily and I talked about it last year. Speaker 200:45 - 00:58
嗨，我是 Matt Turk。欢迎回到 Matt podcast。今天，我很高兴再次请到 Stripe 的 data and AI 负责人 Emily Sense。自从我和 Emily 去年谈到 agentic commerce（代理式商业）以来，它的兴起只是在不断加速。
Speaker 200:58 - 01:40
And in this conversation, we go deep on the economic stack for AI agents, how agents buy and sell on your behalf, why per seat pricing is breaking, token monetization, token theft, which Emily calls the most under discussed topic in all of AI, and finally, whether agents may eventually run entire businesses on their own. Please enjoy my conversation with the always excellent Emily Sense. All right, so welcome back. You and I chatted about a year ago and the theme of the discussion we had was all about the rise of organic commerce. Keeping in mind that obviously this is a long term trend that's going to take a while to play out. Speaker 200:58 - 01:40
在这次对话中，我们会深入讨论 AI agents（AI 代理）的经济栈，agents 如何代表你买卖，为什么 per seat pricing（按席位定价）正在失效，token monetization（token 变现），token theft（token 盗窃）——Emily 认为这是整个 AI 领域里讨论最不充分的话题——以及最后，agents 是否最终可能完全自行运营整个企业。请欣赏我与一贯非常出色的 Emily Sense 的这场对话。好，欢迎回来。大约一年前我们聊过，而当时讨论的主题主要是 organic commerce 的兴起。要记住，显然这是一个长期趋势，还需要一段时间才会真正展开。
Speaker 201:40 - 01:49
I'm curious about what you've observed over the last twelve months. What has become reality and what is yet to be built? Speaker 201:40 - 01:49
我很好奇，在过去十二个月里你观察到了什么。哪些已经成为现实，哪些还尚待构建？
Speaker 101:49 - 02:14
Yeah. I mean, a year ago, we were talking about agents as buyers in a pretty hypothetical way. I think that the canonical experiences, the consumer experiences, for example, weren't defined. We were largely reasoning kind of from first principles about what this might look like. Fast forward a year, as you know, like, still early innings, still a lot to do, but we have actual infrastructure deployed. Speaker 101:49 - 02:14
是的。我是说，一年前我们是在一种相当假设性的层面上把 agents 当作买方来讨论的。我认为那些典型体验，譬如面向消费者的体验，当时还没有被定义出来。我们基本上是在从第一性原理出发，推演这件事可能会是什么样子。快进一年，到现在你也知道，仍然还处在很早期的阶段，仍然有很多事情要做，但我们已经有了实际部署的基础设施。
Speaker 102:14 - 02:37
We have real companies building on it. We have real patterns to learn from. And to be clear, like, the the shape of how this unfolds has become more clear and it's going to continue to evolve. So specifically, what we've come to believe is there's a full spectrum of how agentic commerce plays out. And that's actually really important for businesses in how they think about it. Speaker 102:14 - 02:37
我们已经有真实的公司在基于它构建业务。我们也已经有了真实的模式可以借鉴。并且要明确的是，这件事将如何展开，其整体形态已经变得更加清晰，而且还会继续演化。更具体地说，我们现在形成的判断是，agentic commerce 将会沿着一个完整的光谱展开。而这一点对于企业如何思考它，其实非常重要。
Speaker 102:37 - 03:08
So at one end, and this is where our machine payments protocol lives, but basically, have agents that are out autonomously discovering a service and deciding to buy it and handling the transactions, like, on their own. Right? Like, no human in the loop. And that's maybe what people think of when they say agent ecommerce, but that's just one end of the spectrum. There's also the whole other end of the spectrum where, like, people are looking for shoes for flat footed runners inside an AI surface, And the AI surface gives you an answer. Speaker 102:37 - 03:08
所以在光谱的一端，也就是我们的 machine payments protocol 所在的位置，基本上是让 agent 自主地去发现某项服务、决定购买它，并自行处理交易。对吧？也就是整个过程中没有 human in the loop（人类参与）。这可能是很多人提到 agent ecommerce 时想到的样子，但那只是这个光谱的一端。光谱的另一端则是，人们在某个 AI surface 里寻找适合扁平足跑者的鞋，而 AI surface 会给你一个答案。
Speaker 103:08 - 03:33
And increasingly, you know, that was true also in sort of traditional search, but now increasingly that answer comes with a buy button. And so, you know, this is already how a huge number of people are discovering products. If you're a business, you need to show up there. And we've been building the infrastructure to make it, a, easy for businesses to show up and, b, easy for agents, to execute those those transactions. So, you asked about sort of what's become more concrete. Speaker 103:08 - 03:33
而且越来越多时候，你知道，这在某种传统搜索里其实也成立，但现在这个答案会越来越多地直接附带一个 buy button。所以，实际上，已经有大量用户是通过这种方式发现商品的。如果你是一家企业，你就必须出现在那里。我们一直在建设相应的基础设施，让企业更容易出现在这些场景中，同时也让 agent 更容易执行这些交易。所以，你刚才问什么东西变得更具体了。
Speaker 103:33 - 03:59
We recently partnered with Google so merchants can sell, right inside AI mode and the Gemini app. So, you know, maybe you shop at JD Sports because I was on the topic of running shoes or Fanatics or Quints. Those were all early adopters. Microsoft and OpenAI, we're doing something similar with them, like helping businesses make their products discoverable inside Copilot and ChatGPT. Meta is another example, a little bit of a different flavor, but we're powering checkout right inside ads. Speaker 103:33 - 03:59
我们最近与 Google 达成合作，这样商家就可以直接在 AI mode 和 Gemini app 里销售商品。所以，比如你可能会在 JD Sports 购物，因为我刚才提到了跑鞋，或者也可能在 Fanatics 或 Quints 购物。这些都是早期采用者。对于 Microsoft 和 OpenAI，我们也在做类似的事情，帮助企业让自己的产品能够在 Copilot 和 ChatGPT 中被发现。Meta 是另一个例子，形式稍有不同，但我们正在为广告内的直接 checkout 提供支持。
Speaker 103:59 - 04:48
So there's the discovery and then the one click and the agent actually goes and executes the transaction on your behalf. But really, I'd say, like, what we've learned over the last year, the through line is like whether we're talking about, like, fully agent led transactions transactions with with MPP MPP or these very sort of human led purchases inside AI surfaces, there's just a new set of infrastructure that needs to work no matter where you are in the spectrum. And and that includes I mean, it's it's the premise behind our agentic commerce suite. But briefly, like, businesses need to be able to expose their products and their catalogs and their prices. And then consumers need to be able to authorize agents to pay on their behalf, and then agents need to be able to safely execute, that transaction. Speaker 103:59 - 04:48
所以这里既有发现商品，也有 one click，然后再由 agent 实际代表你执行交易。但我真正想说的是，过去一年里我们学到的一条主线是：无论我们谈的是通过 MPP 实现的完全由 agent 主导的交易，还是这些发生在 AI surface 内、明显更由人类主导的购买行为，不管你处在这个光谱的哪个位置，都需要一套新的基础设施来支撑。而这也正是我们 agentic commerce suite 背后的前提。简单说，企业需要能够暴露自己的产品、catalog 和价格；消费者则需要能够授权 agent 代表他们付款；然后 agent 还需要能够安全地执行这笔交易。
Speaker 104:48 - 05:05
So, that is the infrastructure we have built. Those are some of the partners that we've been working with. And I would say, like, you know, the companies building on it probably give you a good read of where commerce is headed. Right? So companies like Wix and Shopify and BigCommerce and Commerce Tools sort of on the platform side. Speaker 104:48 - 05:05
所以，这就是我们已经构建好的基础设施。这些也是我们一直在合作的一些伙伴。我会说，基于这套基础设施进行建设的公司，可能已经很好地反映了 commerce 的发展方向。对吧？比如在平台侧，有 Wix、Shopify、BigCommerce 和 Commerce Tools 这样的公司。
Speaker 105:06 - 05:47
And then, on the brand side, like Best Buy and Coach and URBN and Kate Spade. But again, it's still early and what the interaction patterns will be and how they'll evolve. And I'm particularly interested in, like, how quickly consumers will give up more of the decisioning process and really move from consumer happening over here where AI helps you find the product and the agent is the one sort of, helping you avoid going through cumbersome checkout flows, to a time when we say, I don't know, I have a $500 budget for back to school shopping, and you already know everything about my kids and their school and where I live. So, like, get it done. Or I don't even tell you when you just go do it for me. Speaker 105:06 - 05:47
而在品牌侧，则有 Best Buy、Coach、URBN 和 Kate Spade。不过，再次强调，现在仍然处于很早期的阶段，未来会形成什么样的交互模式、这些模式又会如何演化，都还有待观察。我尤其感兴趣的是，消费者会多快地放弃更多决策过程，真正从现在这种状态——AI 帮你找到产品，而 agent 只是帮你避免经历繁琐的 checkout 流程——过渡到未来某个时点，那时我们会说，我不知道，给你一个 $500 的 back to school shopping 预算，你已经了解我孩子的一切、他们的学校，以及我住在哪里。所以，直接帮我搞定。甚至我连这句话都不用说，你就直接替我完成了。
Speaker 105:47 - 05:50
But but I think we will need to learn that over the coming year or two. Speaker 105:47 - 05:50
不过我认为，这将是我们在未来一两年里需要逐步摸索清楚的事情。
Speaker 205:50 - 06:02
Is there a framework for Urgente Commerce, that you guys came up with or, like, somebody in the industry came up with that? So that would almost be, like, the levels of autonomy for self driving cars, like, you know, l one, you discover Speaker 205:50 - 06:02
关于 Urgente Commerce，有没有一个你们自己提出的 framework，或者说是行业里某些人提出的 framework？也就是说，它几乎可以类比为自动驾驶汽车的 autonomy levels（自主等级），比如说，l one，你去发现…
Speaker 106:03 - 06:08
We do. Literally. Yes. Oh my gosh. It's like we have, like, level one, level two, level three. Speaker 106:03 - 06:08
我们有。真的有。对。天哪。就像我们有，类似，level one、level two、level three。
Speaker 106:08 - 06:45
And it basically if you if you think about it, it's just like the the highest level is sort of the the MPP version that I talked about where the agent is, like, truly autonomous. And sort of the level one is, the human does basically all of the decisioning themselves, and it's the simple execution of the transaction. And I would say sort of on the consumer side, we're mostly hovering level two. People are delegating a little bit of the selection or leaning hard on the AI to help find the product, maybe a hint of level three, but, like, you know, we're not in a world where you're booking your summer vacation one shotting it with an LLM. Speaker 106:08 - 06:45
基本上，如果你这么想的话，最高级别有点像我刚才提到的那个 MPP 版本，也就是 agent（智能代理）真正实现自主运行的状态。而 level one 大致就是，人类自己完成几乎所有决策，系统只负责简单执行交易。我会说，在消费者这一侧，我们现在大多还徘徊在 level two。人们会稍微委托一点选择权，或者比较依赖 AI 帮忙找产品，也许带一点点 level three 的影子，但你知道，我们还没到那种你用一个 LLM 一次性就把暑假旅行全订好的世界。
Speaker 206:45 - 06:59
And what would you describe from a vendor standpoint? That's that would be sort of like level three, so that that's a reality as of today. Right? So you can already get a recommendation, and then you you press the the the button. That's sort of where are. Speaker 206:45 - 06:59
那从 vendor 的角度你会怎么描述？那大概就会是 level three，所以那其实今天已经是现实了。对吧？你现在已经可以先拿到推荐，然后再按下那个按钮。这大概就是我们现在所处的位置。
Speaker 206:59 - 07:04
And an example of this would be ChatGPT Instant Checkout, for example, where you get the recommendation. Speaker 206:59 - 07:04
举个例子，比如 ChatGPT Instant Checkout，你会先拿到推荐。
Speaker 107:04 - 07:08
Totally. Or you're in Gemini or, you know yeah. Exactly. Speaker 107:04 - 07:08
完全是。或者你在 Gemini 里，也是，你知道的——对，完全没错。
Speaker 207:08 - 07:25
You touch upon some some important developments that happened since we last chatted in terms of, like, overall maturation of the industry. You mentioned the organic commerce protocol, which I think came out last fall. What what is that? Do think that's something that you guys built in partnership with OpenAI? Speaker 207:08 - 07:25
你提到了自从我们上次聊天以来行业整体成熟过程中发生的一些重要进展。你提到了 organic commerce protocol，我记得那应该是去年秋天推出的。那是什么？你觉得那是你们和 OpenAI 合作开发的吗？
Speaker 107:25 - 07:51
Yes. We built it in partnership with OpenAI. The agentic commerce protocol is just a a standardized way for businesses to work with agents. And there's a couple of different components of it, and this is sort of wrapped in our broader agentic commerce suite. One is, how do businesses expose their product catalog, their inventory, their prices to agents? Speaker 107:25 - 07:51
是的。我们是和 OpenAI 一起做的。agentic commerce protocol 本质上就是一种让企业与 agent 协作的标准化方式。它包含几个不同的组成部分，而这也被纳入我们更大的 agentic commerce suite 之中。其中一个部分是，企业该如何向 agent 暴露自己的 product catalog、inventory 和 prices？
Speaker 107:52 - 08:27
And you could argue, oh, the agents could go out and you know, search or infer. But sort of inventory is a thing that you want, like, deterministically known. And we don't want businesses to need to kind of register their product catalog or register their inventory with every single new agent that comes online. Because in the same way you and I like to work with a lot of different model providers, or a lot of different models within those model providers, in many cases, both, we we similarly are seeing businesses not wanna place bets on just one agentic surface. They may be selling b to c and b to b. Speaker 107:52 - 08:27
你当然也可以说，哦，agent 也可以自己出去搜索，或者进行推断。但 inventory 这类信息，是你希望能够被确定性掌握的。我们也不希望企业每出现一个新的 agent，都得重新去注册自己的 product catalog，或者重新登记自己的 inventory。因为就像你我都喜欢和很多不同的 model provider 合作，或者在这些 provider 内部使用很多不同的 model，在很多情况下两者都会一样，我们同样也看到，企业并不想只押注在单一一种 agentic surface 上。他们可能同时面向 b to c 和 b to b 销售。
Speaker 108:27 - 09:11
They may be wanting to reach a wide swath of consumers across many different services. And so agentic commerce protocol lets them expose their product catalog once and then opt in to all of the agents who who work with with that protocol. It also includes the the shared payment token. And so this is about making sure that in the moment of transaction, the agent can securely pass the buyer's credentials over to the seller to execute the the transactions. And these are just tokenized credentials, so, you know, the agent doesn't have, access to the credentials in the way that you and I probably wouldn't want an agent, to have our credit card. Speaker 108:27 - 09:11
他们可能希望触达横跨许多不同服务的大量消费者。因此，agentic commerce protocol 让他们只需一次性开放自己的 product catalog（商品目录），然后选择接入所有使用该 protocol 的 agents。它还包含共享的 payment token（支付 token）。所以这件事的重点是，确保在交易发生的那一刻，agent 能够安全地把买家的 credentials（凭证）传递给卖家，以执行交易。而这些只是 tokenized credentials（token 化凭证），所以，agent 并不会以你我大概都不希望的那种方式，真正接触到这些凭证——比如拿到我们的信用卡信息。
Speaker 109:11 - 09:51
And one thing I love about, you know, both the, catalog component of this as well as the shared payment token is its, platform agnostic, payment processor agnostic. So all this works, you know, you mentioned that we co created with OpenAI. It works with OpenAI but also other providers. It works if Stripe processes your payments, but you can also pass on that that shared payment token, to any other, PSP. And for us, this is really about making it easy for businesses to reach their customers where they are, which is increasingly through AI tools, and to not have to reinvent their commerce infrastructure to do that. Speaker 109:11 - 09:51
还有一点我很喜欢，就是这里面的 catalog 组件以及共享 payment token 都是 platform agnostic（平台无关）的，也是 payment processor agnostic（支付处理方无关）的。所以这一整套都能运作，你提到这是我们与 OpenAI 共同创建的——它当然适用于 OpenAI，也适用于其他提供方。它可以在 Stripe 为你处理支付时使用，但你也可以把那个共享 payment token 传给任何其他 PSP。对我们来说，这真正关乎的是让企业能够轻松地在客户所在之处触达他们，而客户如今越来越多是通过 AI 工具出现的；同时，企业不必为了做到这一点而重新发明自己的 commerce infrastructure（商业基础设施）。
Speaker 109:51 - 09:58
Right? We wanna reinvent commerce infrastructure once, and then they, out of the box, can get these sort of new lines of demand. Speaker 109:51 - 09:58
对吧？我们希望把 commerce infrastructure 重做这件事只做一次，这样他们就能开箱即用地获得这类新的需求来源。
Speaker 209:58 - 10:15
So ACP is a little bit like MCP before commerce, right? Yes. Hence the name. What the status of that was launched at the end of September of last year, in terms of overall adoption, is that kind of like a work in progress to get commerce companies to to embrace it, or, where where are we? Speaker 209:58 - 10:15
所以 ACP 有点像商业领域出现之前的 MCP，对吧？对，所以名字也是这么来的。它是在去年 9 月底发布的；就整体采用情况而言，现在算是仍在推进、还需要让 commerce 公司去接受它的阶段吗，还是说，目前进展到哪一步了？
Speaker 110:15 - 10:22
We've actually seen a ton of demand from brands. So Best Buy's on it. Coach is on it. URBN's on it. Kate Spade's on it. Speaker 110:15 - 10:22
实际上，我们已经看到了品牌方非常强烈的需求。Best Buy 在上面，Coach 在上面，URBN 在上面，Kate Spade 也在上面。
Speaker 110:22 - 10:58
We've seen, Quince and Fanatics and JD Sports and a whole bunch more. We've seen a ton of demand from platforms, which sort of, you've probably long thought of platforms like Wix or Shopify or BigCommerce or whatever. You've probably long thought of them as you know, building technology for small businesses to do commerce. And now an important part of, technology for small businesses to do commerce is making sure those small businesses, are appearing in AI tools and can engage in this wave of agentic commerce. And so Wix and Shopify and BigCommerce and commerce tools on the platform side have all, adopted the protocol. Speaker 110:22 - 10:58
我们还看到了 Quince、Fanatics、JD Sports，以及更多很多公司。我们也看到了平台方的大量需求。你可能一直以来都会把 Wix、Shopify、BigCommerce 这类平台看作是为小企业构建 commerce 技术的平台。而现在，面向小企业的 commerce 技术里一个重要组成部分，就是确保这些小企业能够出现在 AI 工具中，并且能够参与到这一波 agentic commerce 浪潮中。因此，在平台这一侧，Wix、Shopify、BigCommerce 和 commerce tools 都已经采用了这个 protocol。
Speaker 110:58 - 11:36
And then on the AI side, on the agent side, we're working with all the all the big ones. So, with with Geminis and with Google, Microsoft, OpenAI, and, lots more coming online in in all three dimensions. But we really think of like the supply side is a combination of the large brands and the platforms who have the small businesses. And then on the sort of agent side, it's what you would think of as the traditional agent pay players. And then there's been some interesting nuance as well, for example, with Meta where, like, well, maybe ads are just becoming agentic buying too. Speaker 110:58 - 11:36
然后在 AI 这一侧、agent 这一侧，我们正在与所有那些大的参与者合作。所以，包括 Geminis、Google、Microsoft、OpenAI，而且在这三个维度上还会有更多参与者陆续接入。不过我们的看法是，供给侧其实是大型品牌和那些服务小企业的平台的组合；而在 agent 这一侧，则是你会认为属于传统 agent pay 领域的那些玩家。另外，也出现了一些有意思的细微变化，比如 Meta，那里的情况有点像——广告也许正在变成 agentic buying 的一部分。
Speaker 111:36 - 11:38
And so that's been an interesting extension. Extension. Speaker 111:36 - 11:38
所以这是一个很有意思的延伸。延伸。
Speaker 211:38 - 12:00
And I think we talked mostly so far about agents buying, at least like in the example we gave where the agent represents the the consumer. Presumably, there's a concept of agents selling as well. Like, what what's the what does this look like in the future? Like, if all technical problems are are solved and adoptions happen, is that basically two agents negotiating something? What is the ultimate vision? Speaker 211:38 - 12:00
我觉得到目前为止，我们主要谈的是 agent 在购买，至少像我们举的那个例子里，agent 代表的是消费者。按理说，也应该存在 agent 在销售这一概念。那么，未来这会是什么样子？如果所有技术问题都解决了、采用也发生了，本质上是不是就是两个 agent 在协商某件事？最终愿景是什么？
Speaker 112:00 - 12:09
So when I when when I when I, like, step back with my economist brain, I'm like, that would be really efficient. Right? Agents are really good at discovery. Right? We've already seen that. Speaker 112:00 - 12:09
所以，当我用经济学家的脑子退一步看时，我会觉得，那会非常高效。对吧？agent 非常擅长 discovery（发现、搜寻）。这一点我们已经看到了。
Speaker 112:09 - 12:30
Agents are really good at integration. Agents are pretty good at, like, finding optimal pricing, matching, negotiating. They're incredibly persistent. Their time is worth a lot less than human time, and they can get those get those back and forth done, much more much more quickly. And then they're also actually really good with, like, integrating and actually adopting the thing. Speaker 112:09 - 12:30
agent 也非常擅长 integration（集成）。它们在寻找最优定价、撮合、谈判这些方面也相当不错。它们极其有耐心。它们的时间价值远低于人类时间，因此能把那些来来回回的沟通更快地完成得多。而且，它们实际上也很擅长做 integration，真正把那个东西采用起来。
Speaker 112:30 - 13:25
So especially if you think of, like, b to b buying and maybe we can talk about Stripe projects a little bit later too. But just like, actually, like not just finding the service and negotiating it for the price and contracting on it and buying it, actually like getting all the way to integrating and using the product, I think agents are going to help with a lot. And so I'm I'm definitely, imagining an economy that is much more efficient because you have agents on the buy side and as you note, also on the sell side and they're kind of hyper efficient on all of those dimensions, which of course, know, in the in the in the, Ronald Coase, Nobel Prize winner Ronald Coase version of the world would basically, just like remove frictions for firms to work with each other, would make markets more efficient, would make competition higher, would serve consumers, would spur growth. I will say today, not a lot of agent to agent transactions happening. So like, you you asked at the top, like, what's become real and what's still in the future? Speaker 112:30 - 13:25
所以，尤其如果你想到 b to b 采购，而且也许我们稍后还能谈一点 Stripe 的项目。但不只是找到服务、为价格谈判、签合同并完成购买，而是真正一路走到 integration 并把产品用起来，我觉得 agent 会在很多方面提供帮助。所以我确实在设想一种效率高得多的经济体：买方有 agent，正如你提到的，卖方也有 agent，而且它们在所有这些维度上都处于一种超高效率状态。当然，如果用诺贝尔奖得主 Ronald Coase 的那套世界观来说，这基本上会消除企业彼此合作时的摩擦，会让市场更高效，会提高竞争，会服务消费者，会刺激增长。不过我得说，今天 agent 与 agent 之间的交易还并不多。所以就像你一开始问的，哪些已经成为现实，哪些还属于未来？
Speaker 113:25 - 13:32
I think that's still in the future. But I think you start with one side, you add the other, and and over time, we probably land with both. Speaker 113:25 - 13:32
我觉得这仍然属于未来。但我认为，一开始会先有一边，再加入另一边，随着时间推移，我们很可能会落到双方都存在的状态。
Speaker 213:32 - 13:54
Why does this matter so much? On the one hand, it's it's it's kind of cool. I can be, you know, doing something, and my agent does, like, look looks for a good product because he knows me, it's personalized, so it's convenient. But I think what you're saying is that it's it's much deeper than that. It's a it's a global kind of economy acceleration, productivity acceleration. Speaker 213:32 - 13:54
为什么这件事这么重要？一方面，它确实挺酷的。我可以在做别的事情时，我的 agent 因为了解我，会帮我寻找合适的产品，这是个性化的，所以很方便。但我觉得你的意思是，这件事的意义远不止于此。它其实是一种全球性的经济加速、生产率加速。
Speaker 213:54 - 13:57
Not to put words in your mouth, but, like, that that's what's at stake ultimately? Speaker 213:54 - 13:57
我不想替你说话，但最终利害攸关的，其实就是这个，对吗？
Speaker 113:57 - 14:19
Yes. And, actually, that, you know, that is true on kind of the consumption side. Right? If we make it easier for folks to discover and transact and integrate, then that will spur growth. And by the way, we're we're we're seeing, I think, some productivity, in the global numbers from AI. Speaker 113:57 - 14:19
是的。而且实际上，这一点在消费侧确实成立。对吧？如果我们让人们更容易去 discovery、交易和 integration，那么这就会刺激增长。顺便说一句，我觉得我们已经在全球数据中看到了一些由 AI 带来的生产率提升。
Speaker 114:19 - 14:45
But I think right now, it's not primarily about consumption because, like, the numbers are still very small. It's primarily about, like, oh, we're flooding the economy with a bunch of AI CapEx. Like, that's that's actually the consumption that's getting pumped into the economy. But I think that's that's definitely going to be, a driver over time. It is also true that, there there's a deeper change, which is agents are making it easier not just to buy things, but also to, like, start and run companies. Speaker 114:19 - 14:45
但我觉得现在，这主要还不是关于消费，因为目前的数据规模仍然非常小。它主要更像是，哦，我们正在把大量 AI CapEx（资本开支）灌进经济里。也就是说，真正被注入经济中的“消费”其实是这一部分。不过我确实认为，随着时间推移，这肯定会成为一个驱动因素。同样也确实存在一个更深层的变化，那就是 agent 正在让人们不仅更容易买东西，也更容易创办和经营公司。
Speaker 114:46 - 15:09
That's a whole other other angle, but we we see that very solidly in the macro data. You know, I don't know if you've seen, like, US business formations over time, but during the pandemic, they surged. That wasn't super surprising. But then if you look, they, like, plateaued over time, and then they're, like, accelerating again now over the last, couple quarters. And what's interesting to me isn't just that acceleration, but actually the composition. Speaker 114:46 - 15:09
这是另外一个完全不同的角度，但我们在宏观数据里非常清楚地看到了这一点。你知道，我不知道你有没有看过美国企业注册数量随时间变化的数据；在疫情期间，它们大幅上升。这并不算特别令人意外。但如果你继续看，会发现它们后来趋于平台期，而在最近几个季度又开始再次加速。而对我来说，有意思的不只是这种加速本身，实际上还有它的构成。
Speaker 115:09 - 15:40
Like, what are all of those incremental new businesses being created? And the incremental growth is coming entirely from, non employer firms is the literal language that the Census Bureau uses, but you and I would just call them solopreneurs. And so, you know, the number of people solopreneurs who are earning more than a $100,000 a year, has just gone like this since 2022. And now there's in America alone, 5,000,000 people making their living running solo companies. Not like, oh, I just said I was a solopreneur. Speaker 115:09 - 15:40
比如说，这些新增企业到底都是什么？而新增增长完全来自 non employer firms——这是 Census Bureau 的原话——但你我会直接把它们叫作 solopreneurs（单人创业者）。所以你知道，自 2022 年以来，年收入超过 100,000 美元的 solopreneurs 数量就一路这样往上走。现在仅在美国，就有 5,000,000 人靠经营 solo company 谋生。不是那种“哦，我说自己是个 solopreneur”而已。
Speaker 115:40 - 16:00
Like, literally that is my income supporting my family. And there are hundreds of thousands that are clearing a million a year. And so, you know, I think that's kind of interesting because it's like, with AI, can you build something? And then with AI, can you run the business around it? And I think vibe coding and and vibe deploying, by the way, are are really important for can you build something? Speaker 115:40 - 16:00
而是真的，这就是我的收入来源，靠它养家糊口。而且其中有数十万人年收入超过一百万。所以，我觉得这很有意思，因为问题变成了：有了 AI，你能不能把某个东西做出来？然后有了 AI，你能不能把围绕它的业务也跑起来？顺便说一句，我认为 vibe coding 和 vibe deploying 对于“你能不能做出东西”这件事真的非常重要。
Speaker 116:00 - 16:56
And then there's a bunch happening in AI, domain domain specific agents that are really solving for like, can you run that business on the accounting side and the customer support side and and and. And that's making these smaller companies very structurally viable. So anyway, I think like the economic enthusiasm I have around AI comes somewhat from the efficiency of markets and growing consumption and better matching and so on. But just as much, if not more, from the effect AI is having on business dynamism and the ability for individuals with an idea to get from an idea to a product that is in market and and meeting real user needs. So, anyway, I I think I think they'll both I think they'll both play into the macro numbers increasingly over the coming years. Speaker 116:00 - 16:56
另外，AI 里还有很多进展，尤其是 domain-specific agents（领域专用 agent），它们确实在解决这样的问题：你能不能把这家公司在会计、客户支持等等这些环节跑起来。而这正在让这些小公司在结构上变得非常可行。所以总之，我对 AI 的经济层面的热情，一部分来自市场效率的提升、消费的增长、更好的匹配等等。但同样多，甚至更多，来自 AI 对企业活力的影响，以及它如何让一个有想法的个人，能够从一个想法走到一个已经进入市场、并真正满足用户需求的产品。所以，总之，我认为未来几年这两者都会越来越多地体现在宏观数据里。
Speaker 216:56 - 17:29
So as we close this kind of, like, overview introduction section, just on the the reality of Agenda Commerce and the impact and including the future impact of Agenda Commerce. What are the biggest roadblocks right now? In in particular, do you think that the, issue ultimately is more just like technical capabilities? And talk in a second about some of the stuff that you guys have built. Or is that a human question of, like, trust and just accepting to have the machine do things for you, especially when your money is your personal money is at stake? Speaker 216:56 - 17:29
那么在结束这一类总览式的介绍部分时，回到 Agenda Commerce 的现实情况、它的影响，以及它未来可能带来的影响。现在最大的障碍是什么？特别是，你觉得问题归根结底更多只是技术能力的问题吗？我们待会儿也会谈谈你们已经构建的一些东西。还是说，那其实是一个人的问题，比如信任，以及是否愿意接受让机器替你做事，尤其是在涉及你的钱、你的个人资金的时候？
Speaker 117:29 - 17:48
Yeah. I think the the two primary blockers that, you know, we'll need to move through to really scale this up are, one, trust. And actually, we've we've done a lot on the trust side. We talked about the shared payment token. Agent doesn't have any access to the credentials. Speaker 117:29 - 17:48
对。我认为，要真正把这件事规模化，我们需要跨过去的两个主要障碍，第一就是信任。实际上，我们在信任这一侧已经做了很多工作。我们刚才谈到了 shared payment token（共享支付 token）。agent 完全无法访问凭证。
Speaker 117:48 - 18:05
Every shared payment token includes radar scores. Right? But is this a legitimate buyer and is this an agent acting in a legitimate way on behalf of the buyer? Maybe we'll talk a little bit about Link as the wallet for agents, but we've done a lot so that consumers can set guardrails around what the agent can spend. Right? Speaker 117:48 - 18:05
每个 shared payment token 都包含 radar scores（风控评分）。对吧？也就是：这是不是一个真实合法的买家？这个 agent 是不是在以合法方式代表买家行事？也许我们还可以聊一聊 Link 作为 agent 钱包这件事，但我们已经做了很多工作，让消费者能够对 agent 可以花多少钱设置 guardrails（约束边界）。对吧？
Speaker 118:05 - 18:49
So it's it's a little different than, like, a one time use virtual card, which are, like, pretty maniacally scoped credentials. But in the case of a LINK wallet, you you very much have have the guardrails to set. But even with the sort of trust layer from a from a technology or infrastructure perspective, like, I just think it takes time for any market to build trust, especially when you're talking about, making decisions for, you know, what I buy and spending my money. I think it's very natural for humans to kind of wanna build their way up to that. And so, I think that's a that's a big reason why on the consumer side, what we're mostly seeing is people are discovering things inside AI apps, but they're still choosing the exact thing. Speaker 118:05 - 18:49
所以，这和一次性使用的 virtual card（虚拟卡）有点不一样；后者基本上属于那种权限范围被限制得近乎偏执的 credentials（凭证）。但在 LINK wallet 的场景里，你其实是可以设置各种 guardrails（护栏/限制条件）的。不过即便从技术或基础设施角度看，已经有了某种 trust layer（信任层），我还是觉得，任何市场要建立信任都需要时间，尤其当你谈的是替我决定“我要买什么”和“怎么花我的钱”这种事时。我认为，人类会想循序渐进地走到那一步，这是很自然的。所以我觉得，这也是为什么在消费者端，我们现在主要看到的是：人们会在 AI apps 里发现商品，但最终还是自己来选定具体买哪一个。
Speaker 118:49 - 19:40
And they're still disproportionately buying low to mid priced stuff, and they'll need more trust. And honestly, also, to some extent, an evolution of the user experience in some of those apps if they're going to get to a place where they're handing off major decisions. And by the way, I wasn't super close to how people moved their spending from stores to online. But I bet in the first few years of sending on spending online, nobody was saying like, you know, I'm gonna go online and buy a couch or a mattress or a leather jacket, like a thing that I wanna feel or I'm gonna spend a lot of money on or where, like, quality is sort of hard to infer from things I can tell on the Internet. And over time, mechanisms built up for people to trust that that was the right product, that when they spent substantial money, it was gonna arrive at their door in good shape. Speaker 118:49 - 19:40
而且他们购买的，仍然明显更多是低价到中等价位的东西；如果要往上走，就还需要更多信任。坦率地说，在某种程度上，也还需要这些 apps 的 user experience（用户体验）继续演进，才能走到用户愿意把重大决策交出去的程度。顺便说一句，我并不是特别了解当年人们是怎么把消费从线下商店转到线上购物的。但我敢打赌，在最早几年开始在线消费的时候，应该没有人会说，“我要上网买个沙发、床垫，或者皮夹克”——这种我想亲手摸一摸、要花不少钱，或者仅凭网上信息很难判断质量的东西。后来，才慢慢建立起各种机制，让人们相信那确实是对的产品，也相信当他们花了一大笔钱后，东西会完好无损地送到家门口。
Speaker 119:40 - 20:02
And obviously, today, probably more mattresses bought online than than in person. So I I think we will get there, but I think trust is, an important enabler. And I would say, again, a lot of the technology foundations are in place, but it just takes reps. Like, humans just need reps for trust to be built. When it comes to financial stuff, basically, nobody enters with an assumption of trust. Speaker 119:40 - 20:02
很明显，今天买床垫的人里，在线购买的可能已经比线下更多了。所以我认为我们最终也会走到那一步，但我觉得 trust（信任）是一个重要的使能因素。再说一次，我会认为很多技术基础其实已经具备了，只是还需要反复实践。人类就是需要通过一遍又一遍的实际经历，信任才会建立起来。而一旦涉及金融相关的事情，基本上没有人会一开始就默认信任。
Speaker 120:02 - 20:03
You have to you have to earn it. Speaker 120:02 - 20:03
你必须靠自己把它赢得。
Speaker 220:03 - 20:19
I'm not gonna tell you that, but, there was a a time when, the idea of, like, entering a credit card, on the, internet was completely insane, which, of course, led to the unbelievable rise of Stripe and all the success. So okay. Fantastic. Speaker 220:03 - 20:19
我就不展开细说了，不过曾经有那么一个时期，像把 credit card（信用卡）信息输入到 Internet 上这种事，完全是不可想象的疯狂举动。而这当然也促成了 Stripe 的惊人崛起以及后续所有那些成功。所以，好吧，太棒了。
Speaker 120:19 - 20:58
But that's interesting. Right? It is there is still a I mean, it I still feel extremely uncomfortable entering my bank account details on the Internet. So I'm happy to enter my bank account details to my link wallet on Stripe so that it can make payments on my behalf, but I am not happy to enter still today on the Internet, my my bank account number. And, you know, and I think this is this is actually really interesting to reason about, will the Internet actually get safer to some extent as these foundations that we are sort of they're being pulled out of us because of agentic commerce. Speaker 120:19 - 20:58
但这很有意思，对吧？确实仍然存在这样一种情况——我的意思是，直到现在，我仍然会对把我的 bank account details（银行账户信息）输入到 Internet 上这件事感到极度不舒服。所以，我可以接受把我的 bank account details 输入到 Stripe 的 link wallet，让它代表我付款；但直到今天，我仍然不愿意把我的 bank account number 直接输入到网上。而且，你知道，我觉得这其实是一个很有意思的问题：随着这些基础设施因为 agentic commerce（agent 驱动的商业）而被某种程度上“逼”出来，Internet 会不会反而在某种程度上变得更安全？
Speaker 120:58 - 21:22
But as, you know, stored balances and wallets and linked bank accounts and whatever become more of the norm, I don't know. I need my agent to, like, be able to burn down my stablecoin balance with some guardrails. Like, does that actually reduce at least the types of fraud that we've seen in, traditional, traditional online commerce? Maybe. Maybe. Speaker 120:58 - 21:22
因为随着 stored balances（储值余额）、wallets（钱包）、linked bank accounts（已绑定银行账户）之类的东西越来越成为常态，我也说不好。我需要我的 agent 能够在有一些 guardrails 的前提下，去消耗掉我的 stablecoin（稳定币）余额。那这样一来，至少会不会减少我们在传统 online commerce（在线商业）里看到的那类欺诈？也许会。也许吧。
Speaker 221:22 - 21:34
You talked about the, user experience inside of the app not being great. Do you have is there anything specific that you have in mind in terms of, like, how it falls short? Speaker 221:22 - 21:34
你刚才提到，app 内部的 user experience 还不够好。就这点来说，你脑子里有没有什么更具体的例子？比如，它具体是怎么不够好的？
Speaker 121:35 - 22:36
I think there's a few things. One and, you know, part of what we're, working on with the AgenTek Commerce Suite is it should be really easy for, these, you know, AI tools, LLMs to accurately reflect and comprehensively reflect inventory. You know, absent, a standard protocol, it's actually, like, a little bit tricky to know exactly who's selling what, and is it real, and how's it priced, and how much is left, and what are the various parameters. And so, partly, the experience, I think, just needs to evolve to actually have the right inventory set to read over and the right deterministic metadata on that inventory set that the LLM can then do nondeterministic things on top of. But I think some of the experience is just we're all very familiar with the flow of going in and typing in search and getting some shopping results and choosing a thing. Speaker 121:35 - 22:36
我觉得有几件事。其一，也是我们正在用 AgenTek Commerce Suite 推进的一部分，就是应该让这些 AI 工具、LLMs 能够非常容易地准确且全面地反映 inventory（库存）。你知道，在缺少标准协议的情况下，实际上要弄清楚到底是谁在卖什么、是不是真的、价格如何、还剩多少，以及各种参数是什么，都会有点棘手。所以我认为，一方面，这种体验本身需要演进，先要有一套正确可读取的 inventory 集合，以及附着其上的、正确的 deterministic metadata（确定性元数据），这样 LLM 才能在其上做 nondeterministic（非确定性）的事情。但另一方面，我觉得也有一部分只是因为我们都太熟悉那种流程了：进去、输入搜索、看到一些购物结果，然后选一个东西。
Speaker 122:37 - 23:15
But I think that whole like, what part of consumption do we want to delegate? And what experience do we need? Which is a little bit of the discovery and a little bit of the guardrails and a little bit of the trust, but also just like the depth of understanding of us in order for that to be a great customer experience. There there's lots of great AI tools out there and there's lots of ways that they're accelerating, you know, the efficiency of consumption. But I haven't yet seen I guess what I was saying on experience is I haven't yet seen an experience where I'm like, that is magical and I know exactly the end things that I'm ready to offload. Speaker 122:37 - 23:15
但我认为，核心问题在于：我们到底想把消费过程中的哪一部分委托出去？又需要什么样的体验？这其中有一部分是 discovery（发现），一部分是 guardrails（护栏/约束），一部分是 trust（信任），但也包括系统需要对我们有足够深的理解，才能真正带来优秀的客户体验。现在确实已经有很多很棒的 AI 工具，也有很多方式在加速消费效率。但就像我刚才说到体验时提到的，我还没有见过一种体验让我觉得：这太神奇了，而且我已经非常明确知道，哪些最终环节是我愿意彻底交给它的。
Speaker 223:16 - 23:37
I wanna cover some of the stuff that, you mentioned in passing around what, you all have built and released in 2026, and especially around the concept of of giving agents money, safely. You mentioned, the Link Wallet. What is the Link Wallet for agents in simple terms? Speaker 223:16 - 23:37
我想聊聊你刚才顺带提到的一些内容，也就是你们在 2026 年构建并发布的东西，尤其是“安全地给 agents 钱”这个概念。你提到了 Link Wallet。用简单的话说，面向 agents 的 Link Wallet 是什么？
Speaker 123:37 - 23:57
Yes. So maybe even before we get to agents, Link is just Stripe's consumer wallet, and 300,000,000, users, use it today. We're making that the wallet for agents. So the idea is you can authorize an agent to make payments on your behalf, but you get these built in controls so you stay in the loop. Right? Speaker 123:37 - 23:57
好的。也许在讲 agents 之前，先说 Link 本身：它就是 Stripe 的 consumer wallet（消费者钱包），目前已经有 300,000,000 用户在使用。我们正在把它变成 agents 的钱包。这个想法是，你可以授权一个 agent 代表你付款，但同时你会获得内置的控制机制，所以你始终都在回路中。对吧？
Speaker 123:57 - 24:30
So you're not giving the agent, like, a blank check or unlimited access to your checking account. You're very distinctly, like, defining what it can do, and you can always pull it back. And, you know, we talked about trust a bit earlier. I actually think the trust dimension here is as underrated or under discussed as the product challenge. Like, when I go and ask my friends and family whether they'd be comfortable letting an agent buy things on their behalf, they usually jump straight to like, well, is it gonna overspend, and is it gonna buy the wrong thing, and can I stop it, and what happens if it buys the wrong thing? Speaker 123:57 - 24:30
所以你并不是在给 agent 一张“空白支票”，也不是给它无限制访问你的 checking account（支票账户）的权限。你会非常明确地定义它能做什么，而且你随时都可以把权限收回。然后，前面我们也谈到过一些 trust（信任）的问题。我实际上认为，这里的信任维度和产品挑战一样重要，只是被低估了，或者说讨论得不够多。比如我去问朋友和家人，是否愿意让 agent 代他们买东西时，他们通常会立刻想到：它会不会花超？会不会买错东西？我能不能让它停下来？如果它买错了会怎么样？
Speaker 124:30 - 24:55
And those are actually all legitimate concerns. And so, the controls within LinkWallet are really what make the whole thing, viable for for consumers. I would also say that, like, you and I and others have many different payment credentials. Right? Like, we have, bank accounts and we have credit cards that we may have debit cards and there are buy now pay leaders. Speaker 124:30 - 24:55
而这些其实全都是合理的担忧。所以，LinkWallet 里的这些控制机制，才是真正让整件事对消费者来说可行的关键。我还想说一点：你、我，还有其他人，其实都拥有很多不同的 payment credentials（支付凭证）。对吧？我们有 bank accounts（银行账户），有 credit cards（信用卡），可能也有 debit cards（借记卡），还有 buy now pay later。
Speaker 124:55 - 25:40
And one of and stablecoins and whatever else. One of the nice things about, a single wallet that can be backed by balances and many different payment credentials, fiat and crypto and whatever else, is, you know, there's just there's just you know, I don't like having a lot of things to reason over, and I especially don't like having a lot of things to reason over when I am passing it off to an agent and need to monitor it and make sure that what I think is happening is actually happening. And so I think part of the beauty of the link wallet is also just the the consolidation. You can back it by whatever the payment credentials are. But at the end of the day, your agent wallet is just a single wallet, which makes it much more, streamlined to reason around. Speaker 124:55 - 25:40
以及 stablecoins（稳定币）等等。一个很好的点在于：如果有一个单一钱包，可以由余额以及许多不同的 payment credentials 提供支持，不管是 fiat（法币）还是 crypto（加密货币）或其他形式，那么，你知道，我就是不喜欢有太多东西需要自己去梳理；尤其是在我要把这件事交给一个 agent、还得监控它，并确保我以为正在发生的事情确实正在发生的时候，我就更不喜欢了。所以我认为，link wallet 的一部分美感也在于这种整合能力。你可以用任何支付凭证为它提供支持；但归根结底，你的 agent wallet 就只是一个单一钱包，这会让整个事情在理解和管理上都更加 streamlined（简化、顺畅）。
Speaker 225:40 - 25:47
And technically, is that a one time usage credit card, or is that completely different? Speaker 225:40 - 25:47
从技术上说，那是一次性使用的信用卡吗，还是说它完全是另一回事？
Speaker 125:47 - 26:05
Good question. Okay. So the very like, when we were talking twelve months ago, the very, very first version of AgenTik Commerce on Stripe and I think in the world was these one time use cards. And, actually, the genesis of that was one time use cards that we used, for, like, platforms and marketplaces. Right? Speaker 125:47 - 26:05
问得好。好，差不多在我们十二个月前讨论这件事的时候，AgenTik Commerce on Stripe——我觉得可能也是整个行业里——最最早的版本，就是这种 one time use cards（一次性使用卡）。而实际上，它的起源是我们之前用于平台和 marketplaces（交易平台）的 one time use cards。对吧？
Speaker 126:05 - 26:50
Like, when I order, a salad from DoorDash, the driver has a one time use virtual card that can be used to pay for my salad, and that feels good to me because, neither the driver nor the restaurant, which I don't have any affiliation with, needs to see my payment credentials. And so in the first version of Agent eight Commerce, which for us, I believe the very first, meaningfully live volume was on Perplexity Shopping, we used, these these one time use virtual cards. And, basically, the the human consumer would say they wanted the thing. Their payment credentials were used to basically buy this, like, one time fund, this one time use virtual card. The agent was handed this one time use virtual card, and then they went off on the Internet and and and purchased with it. Speaker 126:05 - 26:50
比如说，我在 DoorDash 上点一份沙拉时，司机会拿到一张 one time use virtual card（一次性虚拟卡），用来为我的沙拉付款；这让我感觉很好，因为无论是司机还是餐厅——我和他们都没有直接的支付关联——都不需要看到我的 payment credentials（支付凭证）。所以在 Agent eight Commerce 的第一个版本里——我相信对我们来说，第一笔真正有意义的线上交易量出现在 Perplexity Shopping——我们用的就是这些 one time use virtual cards。基本上，人类消费者先表示他们想买某个东西，然后用他们的 payment credentials 去购买一笔一次性资金，也就是给这张 one time use virtual card 充值。接着，这张 one time use virtual card 会交给 agent，然后它就在互联网上用这张卡完成购买。
Speaker 126:50 - 27:23
And that was, very scoped by definition of being one time, and it was generally, you know, scoped to a single provider and scoped to a very fixed amount. I would think of, LinkWallet as much more flexible. Right? So I can set a budget, to be used across a set of providers or set of providers in a domain. I can permission at the individual transaction level, or I can permission sort of above some amount in aggregate individually. Speaker 126:50 - 27:23
而这种方式天然就很受限，因为它本身就是一次性的；通常也只限定给单一 provider（服务提供方），并且金额是非常固定的。我会认为，LinkWallet 要灵活得多。对吧？我可以设置一个 budget（预算），让它能在一组 provider 之间使用，或者在某个领域内的一组 provider 之间使用。我可以按单笔交易粒度去授权，也可以按累计金额超过某个数额的层级来分别授权。
Speaker 127:23 - 28:03
And so, one time use virtual cards were super valuable for sort of getting us off the ground, especially for these sort of single use consumer transactions, but the link wallet is much more flexible. Now I still think that we will be a few months hence before folks are actually, agreeing to, you know, agents using the wallet, beyond a particular scoped transaction. But making sure the infrastructure scales to that is really where we are with LinkWallet. And then as the consumers become ready for it, the technology is already online, and there's no holdup. Speaker 127:23 - 28:03
所以，one time use virtual cards 对我们起步来说非常有价值，尤其适合这类一次性、面向消费者的交易；但 LinkWallet 的灵活性要高得多。当然，我仍然认为，接下来几个月里，人们大概还不会愿意让 agents 在某一笔特定、受限的交易之外直接使用 wallet（钱包）。但我们现在的重点，就是确保这套基础设施能够扩展到那一步。这样一来，等消费者准备好了，技术本身已经在线，不会有任何卡点。
Speaker 228:03 - 28:06
You mentioned shared payment token. What what is that? Speaker 228:03 - 28:06
你提到了 shared payment token（共享支付 token）。那是什么？
Speaker 128:06 - 28:36
It's a new payment primitive. We built it, maybe six months ago, specifically for agent ecommerce. It is a way for a consumer to authorize an AI agent to pay on their behalf without handing over their actual card details. So the token encodes, exactly what the agent is allowed to do, right, which merchants can be charged up to what amount, in what currency, for how long. The agent presents the token to the merchant at checkout. Speaker 128:06 - 28:36
它是一种新的 payment primitive（支付原语）。我们大概在六个月前专门为 agent ecommerce 构建了它。它让消费者可以授权 AI agent 代表自己付款，而不需要交出自己真实的卡片细节。所以，这个 token 会精确编码 agent 被允许做什么——对吧——可以向哪些 merchant（商户）支付、最高支付到什么金额、使用什么货币、有效期多长。agent 在结账时把这个 token 出示给 merchant。
Speaker 128:36 - 29:02
The agent never sees the underlying credentials, and, shared payment tokens, you know, cover more payment methods, including buy now, pay later options like Affirm and Klarna. So it's not just cards. It can really represent, like, different payment methods, depending on what, the user has on file. And this is actually the payment primitive that makes Lynx Wallet for agents work. It also powers our machine, payments work. Speaker 128:36 - 29:02
agent 永远看不到底层 credentials（凭证）；而且 shared payment tokens 覆盖的支付方式更广，包括 buy now, pay later 这类选项，比如 Affirm 和 Klarna。所以它不只是卡。它实际上可以表示不同的 payment methods（支付方式），具体取决于用户档案里保存了什么。这其实就是让 Lynx Wallet for agents 运作起来的那个 payment primitive。它也支撑了我们的 machine payments 相关工作。
Speaker 129:02 - 29:29
It's really about, okay, how can agents transact, without taking on risk, and how can businesses remain in control, as the merchant of record. And so I would think of, like, one time use virtual cards were, like, a useful backstop as we went and built the shared payment token. Shared payment token is a new payments primitive that is wallet agnostic. And by the way, also payment processor agnostic, you can pass it over to Adyen or whoever else you have. It doesn't have to be processed on Stripe. Speaker 129:02 - 29:29
这件事的核心其实是：agents 怎样才能在不承担风险的情况下完成交易，以及企业作为 merchant of record（名义商户）如何继续保持控制权。所以我会这样看：one time use virtual cards 像是我们在构建 shared payment token 过程中的一个很有用的兜底方案。shared payment token 是一种新的 payments primitive（支付原语），它是 wallet agnostic（与钱包无关）的。顺便说一句，它也是 payment processor agnostic（与支付处理方无关）的；你可以把它交给 Adyen 或者任何别的服务商来处理，不一定非要在 Stripe 上处理。
Speaker 129:29 - 29:44
And then think of, like, Link Wallet as the consumer experience, the consumer wallet for agents, that leverages that same, shared payments token primitive, but is more fully featured for the consumer. Speaker 129:29 - 29:44
然后你可以把 Link Wallet 理解成面向消费者的体验，也就是给 agent 用的 consumer wallet；它利用的是同一种共享的 payments token 原语，但在面向消费者时会提供更完整的功能。
Speaker 229:44 - 29:55
And it's software, so it's fully programmable. So just to double click on some of what you just said, you can restrict certain categories of merchant, or you can only buy from Amazon, or you can only buy Speaker 229:44 - 29:55
而且它是软件，所以是完全可编程的。具体展开一下你刚才说的那些点：你可以限制某些 merchant 类别，或者只能从 Amazon 购买，或者只能购买
Speaker 129:55 - 29:59
in The US or in France or okay. Yes. Exactly. Speaker 129:55 - 29:59
在 The US 或者在 France 的商品之类的。对。完全正确。
Speaker 229:59 - 29:59
Okay. Fantastic. Speaker 229:59 - 29:59
好的。太棒了。
Speaker 129:59 - 30:36
Or or you can or you can literally set it to say, I need to approve every single transaction. Right? And, you know, when people are making reasonable sized transactions, that's fine. I think we're gonna move to a world very quickly sort of at the intersection of agent to commerce and what stablecoins enable in terms of sort of micro transactions and then reasoning about agents as the buyers. Agents are very you know, microtransactions never actually really made that much sense even in the context of content because nobody wanted to put in even if it was only 5¢, no one wanted to put in their credit card to buy a 5¢ article. Speaker 129:59 - 30:36
或者你也可以，真的就是把它设置成：我需要批准每一笔交易。对吧？而且，你知道，当人们进行的是金额还算合理的交易时，这没问题。我觉得我们会非常快地进入这样一个世界：在 agent-to-commerce 的交汇处，以及 stablecoins 所带来的 microtransactions 能力之下，再把 agent 视为买方。你知道，microtransactions 过去其实一直都没有真正讲得太通，哪怕是在内容消费场景里也是这样，因为即便只要 5¢，也没有人愿意掏出信用卡去买一篇 5¢ 的文章。
Speaker 130:36 - 31:09
Like, that was just too much friction, a. And b, nobody wanted to process a credit card for 5¢ because you would have negative margins. But I think when we enter a world where, like, the human is not doing any work to execute the transaction, you're not typing in any credentials, you're not navigating to any web pages, the agent's doing it for you, then microtransactions become viable. And you pair that with, okay, and now it's, you know, burning down some stablecoin balance, and we can talk about the work we're doing with Tempo there. Suddenly, microtransactions become very viable. Speaker 130:36 - 31:09
这主要是摩擦太大了，这是其一；其二，也没有人愿意为 5¢ 的交易处理一笔信用卡支付，因为那样利润率会变成负的。但我认为，当我们进入这样一个世界：人类根本不需要为执行交易做任何操作，你不用输入任何凭证，也不用打开任何网页导航，agent 会替你完成这一切，那么 microtransactions 就变得可行了。再结合上，好的，现在它只是从某个 stablecoin 余额里扣掉一点点，我们也可以聊聊我们和 Tempo 在这方面做的工作。这样一来，microtransactions 就突然变得非常可行了。
Speaker 131:09 - 31:40
And I think we're going to quickly move to a world where, especially in buying inference or tokens or sort of AI SASE products, there's gonna be you know, you are going to be using your link wallet to, make a ton of microtransactions and not approve everyone. But yes, if today you want to approve everyone, you can. That's not gonna scale because no one's gonna wanna approve the the 1¢ transaction for the little bit of data or the little bit of research or the token, that I needed for for this job. Speaker 131:09 - 31:40
我认为我们会很快进入这样一个世界：尤其是在购买 inference、tokens，或者某些 AI SASE 产品时，你会用你的 Link Wallet 发起大量 microtransactions，而且不会去逐笔批准。但没错，如果今天你想每一笔都批准，你是可以这么做的。只是这无法扩展，因为没人会愿意为了这项工作所需的一点点数据、一点点研究，或者一个 token，去专门批准那笔 1¢ 的交易。
Speaker 231:40 - 31:58
So what's our current thinking in terms of, like, everything that can go wrong? You know, you were saying, like, when you talk about this to your friend, they they say, well, what if the agent buys the wrong thing and so on and so forth? I mean, presumably, that's frontier everybody's trying to figure it out, but what's the current thinking? Speaker 231:40 - 31:58
所以，我们目前对于“所有可能出错的地方”大概是怎么想的？你刚才提到，比如你把这件事讲给朋友听，他们会说，那如果 agent 买错了东西怎么办，诸如此类。我是说，显然这还是前沿领域，大家都在摸索，但目前主流的思路是什么？
Speaker 131:59 - 32:26
Yeah. We're thinking about this on a few dimensions. One is, to your question on, you know, what if what if the wrong thing gets sent or the thing's not good or whatever, businesses have to remain the merchant of record. Like, that's a core design principle for us, which means, like, we see our job as having agentic transactions behave the way human transactions do. Now that doesn't mean that agency behave like humans or be constrained in the same way humans do. Speaker 131:59 - 32:26
对。我们会从几个维度来思考这件事。其中一个就是你刚才问的：如果发送了错误的东西，或者东西不好之类的情况怎么办。企业必须继续作为 merchant of record（登记商户/法定销售主体）。这对我们来说是一个核心设计原则。也就是说，我们的职责是让 agentic transactions（代理式交易）像人类交易那样运作。当然，这并不意味着 agent 的行为要像人类，或者要以和人类相同的方式受到约束。
Speaker 132:26 - 33:07
But from the perspective of the business, they remain the merchant of record. They are selling. You know, they happen to have that sale facilitated through an agent, but, like, the business is still, the business. Another design principle, which we've touched on a bit in different ways, is always provide appropriate granular programmable controls and guardrails to enable commerce to happen at scale. And that's basically saying, like, no matter how scoped in and in control you wanna be or how much you want it to be a free for all, like, you shouldn't have to adopt a new tool, move to a new wallet, change, you know, the underlying payment rails. Speaker 132:26 - 33:07
但从企业的角度看，它们仍然是 merchant of record。是它们在销售。只是这笔销售恰好是通过 agent 促成的，但企业本身仍然还是那个企业。另一个我们以不同方式稍微提到过的设计原则，是始终提供恰当、细粒度、可编程的控制和 guardrails（防护栏/约束机制），以支持商业大规模发生。基本上，这就是说：无论你希望范围收得多紧、控制得多严，还是希望它更开放自由，你都不应该因此不得不采用新工具、迁移到新 wallet，或者改变底层 payment rails（支付轨道/支付通道）。
Speaker 133:07 - 33:52
You shouldn't have to change your product catalog or where you expose it. Like, basically, out of the box, those those controls and guardrails should, should ebb and flow. And I think in the consumer case, that looks like consumers remaining in charge of how Link authorizes agents to buy. And in the business case, you know, I touched on this briefly in the context of shared payment token, but it looks like making sure that the information we have about the goodness of the underlying buyer and the agent operating on that buyer's behalf are passed fully to the business to action intelligently. And so, you know, radar is our fraud, protection product. Speaker 133:07 - 33:52
你也不应该不得不改变自己的 product catalog（商品目录）或改变它的展示位置。也就是说，默认情况下，这些控制和 guardrails 应该能够灵活收放。我认为，在消费者场景里，这意味着消费者仍然掌控 Link 如何授权 agent 进行购买。而在企业场景里——我刚才在 shared payment token（共享支付 token）的语境下稍微提到过——这意味着要确保：我们掌握的关于底层 buyer（购买方）质量，以及代表该 buyer 运作的 agent 质量的信息，能够完整传递给企业，让它们做出更智能的处理。所以，你知道，Radar 是我们的 fraud protection product（欺诈防护产品）。
Speaker 133:52 - 34:17
We've had it for, over a decade. It used to be really about transaction fraud. Now it looks at all kinds of, fraud and abuse and bot and multiple layers of goodness where there's not just the the the end customer, but also the agent operating on behalf of the customer. And so those radar scores are actually included out of the box in the shared payment token, for businesses to businesses to reason about it. But but, basically, I think business remains merchant of record. Speaker 133:52 - 34:17
我们已经做这个产品超过十年了。它以前主要关注 transaction fraud（交易欺诈）。现在它会看各种类型的 fraud、abuse（滥用）、bot，以及多层次的“可靠性”信号——因为现在不仅有终端客户，还有代表客户运作的 agent。所以这些 Radar 评分实际上已经默认包含在 shared payment token 里，供企业自行判断和处理。不过归根结底，我认为企业仍然是 merchant of record。
Speaker 134:17 - 34:55
Consumer has as fine grained guardrails and controls as they need. And then in the ecosystem, there is as much symmetric information as as we can create. I mean, we now look across just about 2% of of global GDP. In the context of AI, we look at a very, very large share, of that of that GDP because basically all AI buyers and all AI sellers are on link and on Stripe. And so, you know, that that allows us to actually understand when something is going sideways. Speaker 134:17 - 34:55
消费者拥有他们所需的细粒度 guardrails 和控制。然后，在整个生态系统中，我们会尽可能创造更多对称信息。我的意思是，我们现在看到的交易规模大约相当于全球 GDP 的 2%。而在 AI 的语境下，我们看到的是这部分 GDP 中非常非常大的一部分，因为基本上所有 AI 买家和所有 AI 卖家都在 Link 和 Stripe 上。所以，这实际上让我们能够理解什么时候事情开始出问题。
Speaker 134:56 - 35:03
And we see very much our job to protect the ecosystem by by providing that information up front. Speaker 134:56 - 35:03
而且我们非常明确地把自己的职责看作是：通过预先提供这些信息来保护整个生态系统。
Speaker 235:03 - 35:36
It's really fascinating as one unpacks this entire line of thinking around identity commerce because this what's the technology can do, that's what you can do as a responsible player in the ecosystem, but like ultimately a lot of this will have to go to court one way or another. Mean the legal system will need to adapt and evolve because if something goes terribly wrong in a large commerce transaction, you know, who's at fault? I mean, could it be the the model provider underneath because, like, the agent went haywire? I guess all of this is gonna take time to work its way through the ecosystem. Speaker 235:03 - 35:36
当你展开来看围绕 identity commerce（身份商业）的整条思路时，这真的很有意思。这里一部分是技术能做什么，一部分是作为生态系统中的负责任参与者你能做什么，但归根结底，其中很多问题最终都不得不以某种方式进入法庭。我的意思是，法律体系将需要适应并演进，因为如果一笔大型商业交易出了非常严重的问题，到底谁有责任？我的意思是，会不会底层的 model provider（模型提供方）也有责任，因为 agent 失控了？我想，这一切都需要时间，才能在生态系统中逐步理顺。
Speaker 135:36 - 35:50
Yeah. Not just in purchasing. Right? Like, you know, who is at fault for not just bad purchasing behavior, but bad behavior of all types when when there's an agent involved. So so I do think the the landscape is changing there. Speaker 135:36 - 35:50
对。而且不只是采购。对吧？比如说，当有 agent 参与时，谁该为不只是糟糕的购买行为，而是各种类型的不良行为负责？所以，我确实认为，这方面的格局正在发生变化。
Speaker 135:51 - 36:27
When it comes to sort of payments in particular, you you mentioned a bit ago that, back in the day, people didn't really trust just, like, putting their credit card on on the Internet. I think it's interesting to think, like, with things like shared payment tokens, agent never sees the underlying credentials. Each transaction is scored in real time by Stripe Radar. Merchants go and handle the transactions. Credentials, like, never travel through untrusted services the way they can when a human types a credit card number into a random website, not Stripe, but other random websites. Speaker 135:51 - 36:27
说到支付这类事情，尤其是 payments，你刚才提到一点：以前，人们其实并不太信任把自己的信用卡直接放到 Internet 上。我觉得一个很有意思的点是，像 shared payment tokens 这类东西，agent 根本看不到底层凭证。每一笔交易都会由 Stripe Radar 实时评分。Merchants 去处理交易。而这些 credentials 基本不会像人类把信用卡号手动输入某个随机网站——不是 Stripe，而是其他随机网站——那样，经过不受信任的服务流转。
Speaker 136:27 - 36:38
And so, anyway, I I think it's also interesting to ask, like, is there is there sort of meaningful payments trust and safety upside here? And and I could imagine, yes. Speaker 136:27 - 36:38
所以总之，我觉得还有一个有意思的问题是：这里是不是存在一种有实际意义的 payments trust and safety 提升空间？我觉得答案可能是有的。
Speaker 236:38 - 36:52
So to play back, agent payments, would eventually be safer than than human just typing cards in terms of fraud, in terms of mistake presumably as well. Right? Like, if you type in the wrong number or something like that. Speaker 236:38 - 36:52
所以换句话说，agent payments 最终在防欺诈方面会比人类手动输卡更安全，大概在减少失误方面也是，对吧？比如说，你输错了号码之类的。
Speaker 136:52 - 37:05
Yes. Done like, done right in the limit, I believe it should be much safer. Today, I think the first order effect is it's just more convenient. Like right? Like, usually, it's not a fraudster on my credit card. Speaker 136:52 - 37:05
是的。如果做得对，并且从长期极限来看，我相信它应该会安全得多。就今天而言，我觉得第一层面的效果其实只是更方便。对吧？通常并不是有什么 fraudster 在盗刷我的信用卡。
Speaker 137:05 - 37:21
It's like me fat fingering my own CVC. It's annoying, and I'm frustrated. But, you know, right now, I think most of what we're getting is is convenience. But in the limit, yeah, you could you could totally imagine, okay. Now nobody's passing these random payment credentials, over the Internet. Speaker 137:05 - 37:21
更多时候是我自己手滑，把自己的 CVC 输错了。这很烦人，也会让我很挫败。不过你知道，就现在来说，我觉得我们得到的大部分价值还是 convenience（便利性）。但从长期极限来看，没错，你完全可以想象这样一种情况：再也没有人把这些零散的 payment credentials 在 Internet 上传来传去。
Speaker 137:21 - 37:41
And by the way, people also aren't really necessarily reasoning about transactions coming in through 10 different payment credentials they have. Right? They really have a wallet. They see what happens in the wallet. And that wallet is used in a very tokenized, secure way, across sellers. Speaker 137:21 - 37:41
顺便说一句，人们其实也不一定真的会去逐一理解自己那 10 种不同 payment credentials 进来的交易，对吧？他们真正拥有的是一个 wallet。他们看的是 wallet 里发生了什么。而这个 wallet 会以一种高度 tokenized、很安全的方式，在不同 sellers 之间被使用。
Speaker 237:41 - 37:48
So another really interesting topic I wanted to cover and that you mentioned briefly is Vibe deployment. What does that mean? Speaker 237:41 - 37:48
所以我还想聊另一个非常有意思的话题，也是你刚刚简短提到过的：Vibe deployment。这是什么意思？
Speaker 137:48 - 38:05
It's what happens after you build the thing, and it's actually not talked about enough. But, like, the coding part, a lot of people talk about AI for coding. You know, when we were talking a year ago, that makes sense because it wasn't solved. Now that's basically solved. And we actually see that it's solved in our data. Speaker 137:48 - 38:05
它指的是你把东西 build 出来之后发生的事，而这件事实际上没有被讨论得足够多。比如说，coding 这一部分，很多人都在谈 AI for coding。你知道，一年前我们聊这个的时候，这很合理，因为那时这个问题还没被解决。现在这件事基本已经解决了。而且我们实际上也能从自己的数据里看到，它确实已经被解决了。
Speaker 138:05 - 38:19
These are semi random facts, but I I find them interesting. So, like, agent traffic to Stripe's documentation grew more than 10 x since we talked a year ago. Agent traffic is now about 40% of all our docs traffic. Speaker 138:05 - 38:19
这些算是一些半随机的事实，但我觉得它们很有意思。比如说，自从我们一年前聊过这件事以来，agent 流量访问 Stripe 文档的规模增长了 10 倍以上。现在，agent 流量已经占到我们全部文档流量的大约 40%。
Speaker 238:19 - 38:26
And meaning meaning that the agents are trying to figure it out, and they go to the Stripe technical documentation to understand Speaker 238:19 - 38:26
这意味着，agent 们在尝试自己搞明白问题，于是它们会去看 Stripe 的技术文档来理解。
Speaker 138:27 - 38:57
Meaning the coders now are are almost as much agents as they are developers. And in some segments, they are basically all agents and and, not at all developers. And another example actually is is our CLI, our command line interface, which historically was, like, a pretty niche tool used by, like, a pretty small group of developers. It now it just, like, exploded, and we were like, what is happening? And it's now, 70% of its API resource requests are from agents. Speaker 138:27 - 38:57
这意味着，现在写代码的人，几乎既是 agent，也是开发者，而且两者占比差不多。在某些细分领域里，他们基本上完全是 agent，而几乎不是开发者了。另一个例子其实是我们的 CLI，也就是 command line interface。历史上它一直算是个相当小众的工具，只被一小部分开发者使用。现在它的使用量完全爆发了，我们当时都在想，到底发生了什么？而现在，它 70% 的 API 资源请求都来自 agent。
Speaker 138:57 - 39:15
So, like, you could think of, like, the majorities the majority of entities that are using our CLI today, like, aren't people. So, anyway, I think though, you know, those are just two, like, fun Stripe anecdotes, but there's lots of them across the ecosystem you can look at that tells you, like, vibe coding is real. Right? Or just your own lived experience is the same. Right? Speaker 138:57 - 39:15
所以你可以认为，如今使用我们 CLI 的实体里，大多数都不是人。总之，我觉得这些只是两个挺有意思的 Stripe 小故事，但你也可以在整个生态系统里看到很多类似现象，它们都在告诉你，vibe coding 是真实存在的，对吧？或者说，你自己的亲身体验其实也是一样的，对吧？
Speaker 139:15 - 39:24
That part worked. But then what? So, like, an agent writes you, like, a complete working application in twenty minutes. Fabulous. Like, we love it. Speaker 139:15 - 39:24
这一部分是成功了。但接下来呢？比如，一个 agent 能在 20 分钟内给你写出一个完整可运行的应用。太棒了。我们当然喜欢这样。
Speaker 139:24 - 39:40
Okay. But you've still got this pretty big friction before that app is actually live. Right? So you gotta go, like I don't I mean, depends what you're doing. But you gotta probably create an account with your database provider, then your auth provider, then your hosting service, and you're, like, bouncing around. Speaker 139:24 - 39:40
好，但在这个 app 真正上线之前，你前面仍然有一段相当大的阻力，对吧？你得去——这个要看你具体在做什么——但你大概率得先去你的数据库提供商那里创建账号，然后是 auth 提供商，然后是 hosting 服务，然后你就在这些地方之间来回切换。
Speaker 139:40 - 40:06
You got all these dashboards. You're, like don't know how you do it, but copy pasting stuff by hand and managing credentials and API keys and whatever else. And, and every one of those services has its own onboarding flow. So like a little different flavor than the payments flow we talked about, but like, it's a pretty inefficient fill out a bunch of steps. And every one of them, just like the payments flows was designed for like you and me as humans sitting down and clicking through this like weird setup wizard thing. Speaker 139:40 - 40:06
你面前有一堆 dashboard，你会——我也不知道你具体怎么做——但总之就是手动复制粘贴各种东西、管理 credentials、API keys，以及其他乱七八糟的内容。而且，这些服务每一个都有自己的一套 onboarding 流程。所以它和我们刚才谈到的支付流程味道有点不一样，但本质上还是一种低效的、需要填写一堆步骤的流程。而且它们每一个，就像支付流程一样，都是为像你我这样坐在电脑前、一步步点击这个奇怪 setup wizard 的人类设计的。
Speaker 140:07 - 40:36
And I don't think it really bothered any of us that much because we weren't doing it that much because the hard part was the coding part. But now that, like, the app can be built in coded in 20, like, okay, the long pole is deploying the thing. And so, anyway, Vibe Coding was easy. Vibe deployment, has become, like, more of the binding constraint. And so we actually we we recently, like, in the last couple months, launched Stripe projects for this. Speaker 140:07 - 40:36
我觉得这件事以前并没有特别困扰我们，因为我们并不需要这么频繁地做这些，毕竟真正难的是写代码这部分。但现在，既然 app 能在 20 分钟内被构建、被编码出来，那么真正耗时最长的环节就变成部署它了。所以总之，Vibe Coding 已经很容易了，而 vibe deployment 反而变成了更核心的 binding constraint（约束瓶颈）。也正因为如此，我们其实在最近——大概过去几个月里——为此推出了 Stripe projects。
Speaker 140:37 - 41:03
But, basically, it's like agents should be able to, well, sign up for and configure and integrate all of the services they need to deploy an app, and they should be able to do that, like, right from the command line. And it's not just Stripe services. Right? It was like a whole ecosystem of, like, you get, like, Vercel and Supabase and Cloudflare and Twilio, like, whatever you need. Clerk, I think we announced, like, 16 more partners a week or two ago now. Speaker 140:37 - 41:03
但基本上，意思就是 agent 应该能够注册、配置并集成它们部署一个 app 所需的所有服务，而且它们应该能够直接从命令行完成这些事。而且不只是 Stripe 的服务，对吧？而是一个完整的生态系统，比如你会有 Vercel、Supabase、Cloudflare、Twilio，基本上你需要什么都可以。Clerk 也是，我记得我们大概一两周前又宣布了大概 16 个新合作伙伴。
Speaker 141:04 - 41:21
And, anyway, there's still a lot of enthusiasm from the ecosystem because it turns out that everyone kind of has the same problem. Developers all have the same problem, which is now, like, the long pole is, like, the deployment. And businesses have the same problem, which is, like, their onboarding was designed for humans, and now they need it to work for agents. Speaker 141:04 - 41:21
总之，生态系统对此仍然非常热情，因为事实证明，大家面临的某种程度上都是同一个问题。开发者面对的都是同一个问题，也就是现在真正拖后腿的长杆环节是 deployment（部署）。而企业面临的也是同一个问题，那就是它们的 onboarding（接入流程）原本是为人类设计的，但现在它们需要这套流程也能适用于 agent。
Speaker 241:22 - 41:26
And to bring it home, why does Stripe care? Speaker 241:22 - 41:26
那回到核心问题，Stripe 为什么会在意这件事？
Speaker 141:27 - 41:48
Okay. So I think there's a lot of things that Stripe cares about that on paper, you'd be like, why does Stripe care? I'm the the honest answer is we care because it was becoming the bottleneck. Like, the barrier to building is gone, but the barrier to deploying is like a real friction. And if you just zoom back, like, okay, separate from being a payments company, our mission is to increase the GDP of the Internet. Speaker 141:27 - 41:48
好。所以我觉得 Stripe 在意的很多事情，乍看之下你都会觉得，Stripe 为什么会在意这个？坦白说，原因是它正在变成瓶颈。构建的门槛已经消失了，但部署的门槛却成了一个真实存在的摩擦点。如果把视角拉远一点，先撇开我们是一家 payments company 这件事不谈，我们的使命是提升互联网的 GDP。
Speaker 141:48 - 42:15
A big part of that is how do we get more companies off the ground and how do we help them get kind of their first dollar faster and scale from there. And so if a person with an idea can create an app but can't deploy it, they can't sell. And so removing deployment friction directly obviously like expands, the the internet economy. And, you know, you can think of it primarily as, like, just just like orchestration. Like, all we're doing is orchestration. Speaker 141:48 - 42:15
这其中很大一部分在于：我们怎样让更多公司顺利起步，又怎样帮助它们更快赚到第一美元，并从那里继续扩张。所以，如果一个有想法的人能做出一个 app，却没法把它部署出去，那他就没法卖出去。因此，消除 deployment friction（部署摩擦）显然会直接扩大互联网经济。你也可以主要把这件事理解为 orchestration（编排）：我们所做的，本质上就是编排。
Speaker 142:15 - 42:32
So, if someone else wanted to and could and did orchestration, we'd be cool with that. But we were, like, looking at the developers, trying to get the thing live. And we were looking at the businesses, trying to get the thing, like, used by developers. And we're like, okay. I think we can we can make this market a little bit smoother. Speaker 142:15 - 42:32
所以，如果其他人想做、能做、也真的做了这种 orchestration，我们也完全可以接受。但我们当时看到的是开发者在努力把东西上线，也看到企业在努力让这些东西能被开发者用起来。于是我们就想，好吧，我觉得我们可以让这个市场顺滑一点。
Speaker 242:32 - 42:48
Yeah. And to playback orchestration because you have all these partners. So, ultimately, you do hosting observability, email, queue secrets, all the things, but that's provided by different vendors and you provide the glue to make sure that people can deploy their agents safely and efficiently. Speaker 242:32 - 42:48
对，回到 orchestration 这个点，因为你们有这么多合作伙伴。所以归根结底，hosting、observability、email、queue、secrets 这些能力都有，而且都是由不同的 vendor 提供，而你们提供的是把这些拼起来的 glue，确保人们能够安全、高效地部署他们的 agent。
Speaker 142:48 - 42:50
Yes, exactly. Exactly. Speaker 142:48 - 42:50
对，完全正确。完全正确。
Speaker 242:50 - 42:57
Very cool. Alright. So that's Vibe Deployment. Tokens as money is also a fascinating topic. It's like taking it from the the top. Speaker 242:50 - 42:57
很酷。好。那就是 Vibe Deployment。Tokens 作为货币也是一个很有意思的话题。这有点像是从最顶层开始看这个问题。
Speaker 242:58 - 43:01
What does it mean to monetize tokens from a Stripe perspective? Speaker 242:58 - 43:01
从 Stripe 的视角来看，把 tokens 变现意味着什么？
Speaker 143:01 - 43:22
It's a big question because what does it mean to monetize tokens from a world perspective is kinda like, okay. Actually, how's the whole next generation of b to b and some b to c gonna monetize? Look. Like, the the the last kind of decade plus of SaaS, had, like, pretty beautiful and simple to monetize economics. Right? Speaker 143:01 - 43:22
这是个很大的问题，因为如果从整个世界的视角来看，“把 tokens 变现”某种程度上其实是在问：好吧，下一代的 b to b 以及一部分 b to c 究竟要怎么变现？你看，过去十多年里的 SaaS，在变现经济模型上其实是相当漂亮而且简单的，对吧？
Speaker 143:22 - 43:59
And in particular with SaaS, like, you build a product once, and then you get one more customer, and it costs you basically nothing to serve them. Like, marginal costs are near zero. And that's why SaaS margins are really good, and that's why, you know, the fixed fee subscriptions or seat based licenses work really well, in SaaS. AI, and I say AI generally because, like, you could literally be selling LLMs, but you could also be selling, you know, some product that's a wrapper on top of LLMs or a product that's heavily powered by LLMs and and requires a lot of tokens. It breaks that model because, obviously, every prompt and every API call and every task has a a real marginal cost all of a sudden, which it didn't have in SAS. Speaker 143:22 - 43:59
特别是在 SaaS 里，你把一个产品做出来一次，然后多获得一个客户，服务这个客户基本上几乎不增加什么成本。也就是说，边际成本接近于零。这就是为什么 SaaS 的利润率非常好，也就是为什么固定费用订阅或者按 seat 计费的 license，在 SaaS 里效果很好。AI，我这里是广义地说 AI，因为你卖的可以直接是 LLM，也可以是某种封装在 LLM 之上的产品，或者是一个高度依赖 LLM、并且需要大量 tokens 的产品。它打破了那种模型，因为很明显，现在每一个 prompt、每一次 API 调用、每一项 task 都突然有了真实的边际成本，而这是 SaaS 里原本没有的。
Speaker 143:59 - 44:33
The inference isn't free. And so you now have all these businesses where how your customers use your product directly determines whether you make or lose money. And that's a very different game. And, you know, from our vantage point, like part of what that boils down to is the need for usage based billing as really critical for AI companies. Like, you need to be able to meter what customers are actually consuming in real time and then charge them, in a way, obviously, that that aligns with your with your underlying costs. Speaker 143:59 - 44:33
Inference（推理）不是免费的。所以现在你会看到很多企业，你的客户如何使用你的产品，会直接决定你是赚钱还是亏钱。这是一个非常不同的游戏。从我们的视角来看，这最终部分归结为：usage based billing（按使用量计费）对 AI 公司来说是非常关键的。你必须能够实时计量客户实际消耗了什么，然后以一种显然与你底层成本相匹配的方式向他们收费。
Speaker 244:34 - 44:51
And do you see from your perspective pretty much all the players in the AI economy from a vendor's start up standpoint, use usage based billing at this stage, or is that still a mix between per seat and per usage? Speaker 244:34 - 44:51
从你的观察来看，在 AI economy 里，站在 vendor 的 start up 这一侧，现在是不是几乎所有参与者都在使用 usage based billing，还是说目前仍然是按 seat 和按 usage 混合的状态？
Speaker 144:51 - 45:34
I see very few scaling or scaled AI companies that are still exclusively subscriptions or seat based. And I believe from my conversations with them that that is for the simple reason that the economics there don't make sense because, you know, you have some people that are, using a ton and some people that are not using very much. And these people cost you a ton and these people cost you not very much, and it is very hard to separate the sheep from the goats and price them appropriately without a usage based meter. That said, many of those businesses have a usage based offering, but it is complementary to what you can think of as like a fixed fee subscription. So like a great example is Lovable. Speaker 144:51 - 45:34
我看到的正在扩张或已经扩张起来的 AI 公司里，仍然完全只靠订阅制或按 seat 计费的非常少。根据我和他们的交流，我相信原因很简单：那样的经济模型说不通。因为有些人用得非常多，有些人则用得很少。前一类人让你付出很高成本，后一类人则几乎不怎么花你的钱；如果没有一个 usage based 的计量机制，就很难把不同类型的客户区分开来，并给他们制定合适的价格。话虽如此，这些企业中很多都有 usage based 的产品方案，但它通常是对某种固定费用订阅的补充。一个很好的例子就是 Lovable。
Speaker 145:34 - 45:41
When they launched, they had like a simple subscription through Stripe Billing. Makes total sense. They were early. They're moving fast. They wanted to monetize quickly. Speaker 145:34 - 45:41
他们刚发布时，是通过 Stripe Billing 提供一种简单的订阅方案。这完全说得通。那时候他们还处于早期阶段，推进速度很快，也希望尽快开始变现。
Speaker 145:41 - 45:57
Subscriptions are also very familiar and easy for consumers. You know, if you think about a lovable, like, some of their, target users are, like, not very technical. Like, how are they gonna feel about, like, oh, how should I reason about a token or a credit? Right? Okay. Speaker 145:41 - 45:57
订阅模式对消费者来说也非常熟悉、很容易理解。你想想看，Lovable 很讨喜，而且他们的一些目标用户并不是特别懂技术。那他们会怎么看待“我该怎么理解 token 或者 credit？”这种问题呢？对吧？好。
Speaker 145:57 - 46:23
So so they they started with subscriptions. But then as the as the company grew and and some of their costs grew, their billing needs evolved. And, they needed to charge at least somewhat based on actual, token consumption. And so what they did is a hybrid billing model, which which actually we're seeing a large number of businesses, especially businesses that have a b to c component do, where they have usage based billing on top of their subscriptions. So customers hit some threshold. Speaker 145:57 - 46:23
所以他们一开始采用的是订阅模式。但随着公司增长、一些成本也上升，他们的计费需求发生了变化。他们至少需要在一定程度上根据实际的 token 消耗来收费。于是他们采用了一种混合计费模型，而这其实正是我们看到很多企业——尤其是带有 to c 业务成分的企业——在做的事：在订阅之上再叠加 usage based billing（按使用量计费）。也就是客户达到某个阈值之后开始进入另一层计费。
Speaker 146:23 - 46:54
You know? And by the way, many people have freemium thresholds, but then above that, they often have, like, a fixed fee, you know, $25 a month or $100 a month, up to some number of credits. And then above that, you you have usage based billing. So once customers hit the threshold, Lovable, in this case, charges, like, precisely based on the number of tokens consumed above that. And, you know, I think that that that alignment is good because you get a bunch of people in the door comfortable with subscriptions, but then sort of your at at any volumes that matter, your revenue is scaling very directly with, like, how customers are using the product and correspondingly with, like, how much it is costing you to provide that product. Speaker 146:23 - 46:54
你知道吧？顺便说一句，很多人会设置 freemium（免费增值）阈值；但超过这个阈值之后，他们通常还是会先收一个固定费用，比如每月 25 美元或每月 100 美元，对应某个 credit 数量。再往上，才进入 usage based billing。也就是说，一旦客户达到这个阈值，像 Lovable 这样的公司就会非常精确地按照超出部分实际消耗的 token 数量来收费。我认为这种对齐关系是好的，因为它能先让一大批用户以自己熟悉的订阅方式轻松进入；但当使用量真正变得有意义时，你的收入就会非常直接地随着客户使用产品的方式而扩张，同时也会与您提供该产品所承担的成本相对应。
Speaker 146:55 - 47:15
So customers pay only for what they get and you make sure you monetize, for, for the underlying cost that you're gonna have to bear. And Eleven Labs is another example. Like, went through the exact same thing, literally, like started with subscriptions. They recently moved to this sort of pay go plan. And and this is just a pattern that we see playing out pretty much everywhere in the AI space right now. Speaker 146:55 - 47:15
这样一来，客户只为自己实际获得的东西付费，而你也能确保把自己必须承担的底层成本变现。Eleven Labs 也是另一个例子。他们经历了完全相同的过程：字面意义上就是从订阅起步，最近转向了这种 pay as you go plan（随用随付方案）。而这正是我们现在几乎在整个 AI 领域都看到的一种普遍模式。
Speaker 247:15 - 47:19
Does billing change much, once you start charging agents? Speaker 247:15 - 47:19
一旦你开始向 agent 收费，计费方式会有很大变化吗？
Speaker 147:20 - 48:24
So, I think so for a couple reasons. One is agents can consume at machine speed. And so even if you're, like, even if you're doing a usage based thing, but you're like charging at the end of the month, by the time you've gotten to the end of the month, an agent can have spent bet a human can too to some extent with agents working for them, but like especially an agent can have spent like an egregious amount. And so actually, I think in the world of agents, what we're going to see more and more of is real time metering, like what have you consumed and real time billing, which is actually what we have co built between Metronome, so real time metering, usage based billing for very complex models, and Tempo, blockchain, where, you know, agents are consuming tokens in real time and, paying down, for the cost of those tokens. And that's important because agents can buy machine speed. Speaker 147:20 - 48:24
我觉得会，原因有几个。第一，agent 可以以 machine speed（机器速度）消耗资源。所以即使你采用的是 usage based 的模式，但如果你是在月底统一收费，那么等到月底时，一个 agent 已经可能花掉——人类在 agent 替自己工作时某种程度上也会这样——但尤其是 agent，已经可能花掉一笔非常夸张的金额。所以我认为，在 agent 的世界里，我们会越来越多地看到 real time metering（实时计量），也就是“你已经消耗了什么”，以及 real time billing（实时计费）。而这实际上正是我们在 Metronome 和 Tempo 之间共同构建的东西：Metronome 负责实时计量，以及针对非常复杂模型的 usage based billing；Tempo 则是 blockchain。在这个体系里，agent 会实时消耗 token，并实时支付这些 token 的成本。这很重要，因为 agent 的购买与消耗能力都可以达到机器速度。
Speaker 148:24 - 48:41
It's viable because, of a bunch of the the infrastructure between Tempo and Metronome. And, agents are happy to just, like, pay as they go in a very literal way, plus businesses need them to be paying as they go so that they don't, rack up a bunch of spend and then and then go dark. Speaker 148:24 - 48:41
之所以这件事可行，是因为 Tempo 和 Metronome 之间有一整套基础设施支撑。而且 agent 也很乐于真正意义上地 pay as they go（边用边付）；同时，企业也需要它们这样做，免得它们先累积出一大笔支出，然后突然消失不见。
Speaker 248:42 - 48:51
Yeah. It's it's fascinating also in terms of downstream consequences for what it means for, like, finance and accounting and, like, systems that need also to move at that speed. Speaker 248:42 - 48:51
是的。而且如果再往下游看，这对 finance、accounting，以及那些同样需要以这种速度运转的系统意味着什么，也非常令人着迷。
Speaker 148:51 - 49:32
Totally. I mean, we have a a revenue recognition accounting product, and it's and it's mostly needed by I mean, we we provide it to a bunch of people who a bunch of businesses who have traditional subscriptions. But where there's the most acute pain is actually, like, traditional accounting in spreadsheets, like, does not work when you have this just like, like, proliferation of rows because these are like, microtransactions are are truly happening. I also think it's changing what it means to be an accountant. I I I will, avoid naming the company, but I was talking to the sort of number two in accounting at a pretty successful AI company. Speaker 148:51 - 49:32
完全同意。我的意思是，我们有一款做 revenue recognition（收入确认）会计的产品，它主要是给那些采用传统订阅模式的企业使用的——我们确实服务了不少这样的公司。但痛点最强烈的地方其实在于，传统那种用 spreadsheets（电子表格）做会计的方式，在你面临这种行数爆炸式增长时根本行不通，因为这些 microtransactions（微交易）是真正在持续发生的。我也觉得，这正在改变“会计”这份工作的含义。我就不点公司名了，但我之前和一家相当成功的 AI 公司里负责会计的二号人物聊过。
Speaker 149:32 - 49:51
And a couple of things were interesting. One is they were like a hybrid accountant engineer, which they needed to be because of the scale of the data they were dealing with. And two, they weren't doing, like, accounting in the traditional you and me sense of, like, just close the books. Like, they have to close the books for sure. But their job was also to find, like, weird anomalous stuff happening. Speaker 149:32 - 49:51
其中有几件事很有意思。第一，他们其实是会计和工程师的混合体，而他们必须这样，因为他们要处理的数据规模太大了。第二，他们做的也不是你我传统意义上理解的那种会计工作——比如只是 close the books（结账、关账）。当然，他们肯定也要关账。但他们的工作还包括发现各种异常、反常的情况。
Speaker 149:52 - 50:04
And, I was talking to them because they were, like, identifying some fraud patterns and wanted help with them. But anyway, just just very interesting. Like, okay. What is actually what does it mean to do accounting at, like, one of these AI companies? Definitely needs new tools. Speaker 149:52 - 50:04
我当时和他们交流，是因为他们正在识别一些 fraud（欺诈）模式，想让我们帮忙。但不管怎样，这真的很有意思。就是说，在这些 AI 公司里，做会计这件事到底意味着什么？显然，它需要新的工具。
Speaker 150:04 - 50:20
Definitely needs new systems. Probably a different different skill set. And then, like, what you're accountable for really isn't just closing the books. It's like looking across the the whole rev rec stack and saying, what does this tell me about the health of our business? What does this tell me about fraud and abuse in our business? Speaker 150:04 - 50:20
也显然需要新的系统，可能还需要一套不同的技能组合。而且，你真正需要负责的，已经不只是把账关上而已。更像是要纵览整个 rev rec（收入确认）技术栈，然后问：这些信息告诉了我什么——关于我们业务的健康状况？关于我们业务中的 fraud（欺诈）和 abuse（滥用）？
Speaker 150:20 - 50:22
And what does this tell me about breakages in our product? Speaker 150:20 - 50:22
以及，这些信息又告诉了我什么——关于我们产品里哪些地方出了问题？
Speaker 250:22 - 50:36
Yeah. All of which becomes a massive data problem that needs to be treated in in in in real time. Right? Hence, the rise of the click houses of the world. Like, whoever can process massive amounts of data in real time becomes the core part of the required infrastructure. Speaker 250:22 - 50:36
对，而这一切最终都会变成一个巨大的数据问题，而且必须以实时方式来处理。对吧？这也就是为什么会出现 clickhouse 这一类系统。谁能实时处理海量数据，谁就会成为这类必需基础设施的核心组成部分。
Speaker 250:36 - 50:45
Okay. You wrote somewhere about token theft, and I wanna make sure that we cover that. So what is that? What is token theft as a new form of fraud, I guess? Speaker 250:36 - 50:45
好，你之前在某处写到过 token theft（token 窃取），我想确认我们把这个话题也讲到。所以那到底是什么？我想，它算是一种新的 fraud（欺诈）形式，对吗？
Speaker 150:45 - 51:02
Yeah. It's it's a new form of fraud. It's, I think, one of the most under discussed topics in AI right now, maybe by a lot. Fraudsters have figured out that in AI, you actually don't really need to steal money or credentials. You can just steal tokens. Speaker 150:45 - 51:02
对，这是一种新的 fraud（欺诈）形式。我觉得它可能是当下 AI 领域里讨论最不充分的话题之一，而且可能远远不止“一点点”被忽视。诈骗者已经发现，在 AI 里，你其实并不一定需要偷钱或者凭证。你只需要偷 tokens。
Speaker 151:02 - 51:17
And tokens have real value. Right? You can use them to build things. You can resell them on marketplaces. You can wrap a new product on top, without paying a cent and go and sell that real that new product. Speaker 151:02 - 51:17
而且 token 确实有真实价值。对吧？你可以用它们来构建东西。你可以在 marketplace 上转卖它们。你还可以在其之上包装一个新产品，不花一分钱，就把那个真正的新产品拿去卖。
Speaker 151:17 - 51:38
So it was kinda like resell, but it doesn't doesn't look like you're selling the subscription to the thing. It looks like you're selling this other product that you've come up with on your own, but really in the back end, it's completely powered by someone else's thing. And for AI companies, maybe this is implied by our earlier conversation on SaaS, but like the risk is existential. Right? If someone stole a little bit of your SaaS, like, didn't matter because it didn't really cost you anything on the margin. Speaker 151:17 - 51:38
所以这有点像转售，但看起来又不像是在卖那个东西的订阅。它看起来像是在卖这个你自己想出来的另一种产品，但实际上在后端，它完全由别人的东西驱动。对于 AI 公司来说，也许这在我们前面关于 SaaS 的讨论里已经有所暗示，但这里的风险是生存级的。对吧？如果有人偷走了你一点 SaaS，其实没太大关系，因为边际上并不会真的让你多花什么成本。
Speaker 151:38 - 52:06
When someone steals your tokens, if your fraud rate is high enough, the economics of your product actually, break really fast. So anyway, because we, work with basically all the all the AI companies, we have this interesting front row seat into, what's actually happening and the fraud patterns, playing out. And right now, there's three. And I don't think these will be the three forever because we've already we've already largely burned down these three over the last three months. But new ones will new ones will pop up. Speaker 151:38 - 52:06
但当别人偷的是你的 token 时，如果你的 fraud rate（欺诈率）高到一定程度，你产品的经济模型实际上会非常快地崩掉。总之，因为我们基本上和所有 AI 公司都有合作，所以我们得到了一个很有意思的一线视角，能看到实际上正在发生什么，以及这些 fraud pattern（欺诈模式）是如何展开的。眼下有三种。我不认为它们会永远只是这三种，因为在过去三个月里，我们已经在很大程度上把这三种压下去了。但新的还会不断冒出来。
Speaker 152:06 - 52:29
I think these three are just, instructive for reasoning about sort of the breadth of the space. One is multi account abuse. So this is like bad actors just, like, hammering your sign up again and again and again and again so that they can get those, like, new user credits. And the scale of this actually shocked me when I looked at the data. So, like, more than one in six signups at AI companies are this kind of abuse. Speaker 152:06 - 52:29
我认为这三种情况很有启发性，有助于我们理解这个领域的广度。其中一种是 multi account abuse（多账号滥用）。也就是说，不法分子会一遍又一遍地猛刷你的注册流程，这样他们就能拿到那些新用户 credits（额度）。而且当我看到数据时，这种行为的规模真的让我很震惊。比如说，在 AI 公司里，超过六分之一的注册都属于这种滥用。
Speaker 152:29 - 52:45
And by the way, that's also very confusing for the company. Like, are these good customers? Who are these people? But it's very expensive for the company because, like, I've spent a bunch of tokens on these free products for, you know, a relatively small number of people who are spinning up a massive number of accounts and actually, consuming quite a bit of tokens. Okay. Speaker 152:29 - 52:45
顺便说一句，这对公司来说也会非常令人困惑。比如，这些是优质客户吗？这些人到底是谁？但这对公司来说又非常昂贵，因为我已经在这些免费产品上花掉了大量 token，而实际上只是相对少数的人开出了海量账号，并且确实消耗了相当多的 token。好。
Speaker 152:45 - 52:54
So that's like very, very top of funnel. Another example that I think is interesting is is free trial abuse. So these are like fraudsters come in. They create a free trial. They put down a payment method. Speaker 152:45 - 52:54
所以这算是非常、非常漏斗顶端的问题。另一个我觉得有意思的例子是 free trial abuse（免费试用滥用）。这类情况通常是 fraudster（欺诈者）进来，创建一个免费试用，然后填入一种 payment method（支付方式）。
Speaker 152:55 - 53:23
They they drain through, like, all of the credits, but they never have any intent to convert. And, you know, free trial abuse has always existed in various forms across the Internet, AI or not, but it's more than doubled on Stripe in the last six months. And most of that doubling is coming from AI and, AI businesses. And just to give you a sense, like, how lucrative this is, like there's whole cottage industries built around it. So I don't know if you've ever been marketed a free trial card. Speaker 152:55 - 53:23
他们会把所有 credits 都耗光，但从来没有任何转化付费的意图。你知道，free trial abuse 一直以来都以各种形式存在于互联网中，不管是不是 AI 场景，但它在过去六个月里在 Stripe 上已经翻了一倍多。而这部分增长中的大多数，都来自 AI 和 AI businesses（AI 企业）。为了让你感受一下这件事有多赚钱，围绕它甚至已经形成了完整的小型产业。所以我不知道你有没有被推销过 free trial card（免费试用卡）。
Speaker 153:23 - 53:45
I was recently marketed one. It's basically like literally, it's like, oh, you can like spin up free trial cards. They expire in twenty four hours so you'll never have to pay. And if you or I had one, we'd probably use it for some legitimate purpose to try out a service and just not have to go through the pain of canceling. But literally, fraudsters can just explode these things and spin up a bunch of free trials and spend a bunch of tokens with no intent to convert. Speaker 153:23 - 53:45
我最近就被推销过一次。它基本上就是——真的就是那种——“哦，你可以批量创建 free trial card，它们会在 24 小时后过期，所以你永远不用付钱。” 如果你或者我拿到这种卡，我们大概会把它用于某种正当用途，比如试用一下某个服务，顺便省去取消订阅的麻烦。但欺诈者真的可以把这类东西大规模滥用，批量开出一堆免费试用，消耗大量 token，而且根本没有任何转化意图。
Speaker 153:45 - 54:16
And then the third is is a little further downstream of this, which is basically like we talked about the usage based, billing. Folks are racking up, like, thousands of dollars of costs and then being billed at the end of the month, and never paying. And so, like, whatever. The dine in dash that happens in restaurants is like a dine in dash, but it's but it's for for tokens. And, unfortunately, at that point, again, this is AI, not SaaS, So, like, the company has borne the cost, of of those of those tokens. Speaker 153:45 - 54:16
然后第三种会更发生在这个链条的下游一点，本质上就是我们说的 usage-based billing（按使用量计费）。有些人先累计了几千美元的成本，到月底才出账单，然后就直接不付钱了。所以，怎么说呢，这就像餐厅里那种 dine and dash（吃霸王餐后逃单），只不过这里逃的是 token 的账。可惜的是，到那个时候，这又是 AI，不是 SaaS，所以这些 token 的成本已经由公司自己承担了。
Speaker 254:16 - 54:41
And sorry if that's obvious, and I guess I would be a terrible fraudster because it's it's not even super obvious to me. So, if I get tokens for free, what do I actually do with them? So I guess if I get tokens from a general purpose LLM like Claude or ChatGPT, I could see what I do with it. If I if I go to Cursor or Lovable or eleven Labs. So what is it that I actually do with this token? Speaker 254:16 - 54:41
抱歉，如果这问题很显然的话，不过我想我大概会是个很差的 fraudster（诈骗者），因为这对我来说甚至也不是特别显然。所以，如果我免费拿到了 token，我实际能拿它做什么？我想，如果我从像 Claude 或 ChatGPT 这样的通用 LLM 那里拿到 token，我大概能理解怎么用。如果我去用 Cursor、Lovable 或 eleven Labs，那我实际到底会拿这个 token 去做什么？
Speaker 154:41 - 54:57
Okay. So what you do with it very much, changes based on, like, what's the service? Like, what exactly were these were these tokens, meant for? You nail on the head for like, okay. You just, like, get the tokens from the the underlying LLM. Speaker 154:41 - 54:57
好，具体怎么用，其实会随着服务本身而有很大变化。比如说，这些 token 到底是给什么服务用的？你刚才其实说到了关键点：如果你拿到的只是底层 LLM 的 token。
Speaker 154:57 - 55:13
Fine. For those sort of, businesses that are a layer above that, a bunch of resale abuse. So, you literally sell it sometimes in other for a slightly discounted price a discounted price on the base price that you should have paid but you didn't pay. Speaker 154:57 - 55:13
那也行。对于那些处在更上层一层的 business（业务）来说，就会出现大量 resale abuse（转售滥用）。也就是说，有时候你会直接把它卖掉，而且还是以略低于你本来应该支付、但实际上并没支付的基础价格去打折转卖。
Speaker 255:13 - 55:24
So there's a dark like, a dark web of marketplaces where you say, like, use a cursor or lovable for $2, but my my cost is zero. Therefore, I make money? Speaker 255:13 - 55:24
所以会存在一个有点像 dark web（暗网）的 marketplace（交易市场），有人会说，比如“用 Cursor 或 Lovable 只要 2 美元”，但我的成本是零，因此我还能赚钱？
Speaker 155:24 - 55:47
Yeah. Exactly. So so I just I give you my login credentials, whatever. There's also but but then there's just, like, there's just this, like, long tail makes it sound small, but a very domain specific fraud pattern. So for example, people steal tokens to create content that they then use to extract money in all sorts of scammy ways. Speaker 155:24 - 55:47
对，完全是这样。所以我就是把我的登录凭证给你之类的。不过除此之外，还有一类——说它是 long tail（长尾）听起来像规模不大，但其实是非常 domain-specific（特定领域）的 fraud pattern（欺诈模式）。比如说，有人偷 token 来生成内容，然后再用这些内容通过各种带骗术性质的方式变现。
Speaker 155:47 - 56:21
So a simple example, we see people going in and, like, mass generating music tracks and then uploading them to Spotify and Apple Music and then getting fake streamers and then collecting royalties. Or then for, like, basically, all of the rapper businesses, there's this whole other layer of wrapper on a wrapper where, like, rather than just resell the subscription on places like Taobao, they'll, like, literally clone the AI company where you can just, like, vibe code a website. The back end is just, like, spitting out exactly what you got from the service that you're stealing from, And then you, like, sell the product as yours, but cheaper. Speaker 155:47 - 56:21
一个简单的例子是，我们看到有人进去大批量生成音乐曲目，然后上传到 Spotify 和 Apple Music，再找 fake streamers（虚假播放流量）去刷播放量，接着领取版税。再比如，基本上所有这些 wrapper businesses（套壳业务）上面，还有另一层“套壳套壳”的玩法：他们不只是像在 Taobao 这样的地方转卖订阅，而是会直接克隆那家 AI 公司。比如你可以用 vibe code（氛围编程）快速做一个网站，后端其实只是把你从被盗用的服务里拿到的结果原样吐出来，然后你把这个产品当成自己的来卖，只是价格更便宜。
Speaker 256:21 - 56:27
Are you gonna give it to people that they're creative? I mean, that sounds almost harder than starting an actual company. Speaker 256:21 - 56:27
你是不是都得夸他们还挺有创造力的？我的意思是，这听起来几乎比真的去创办一家公司还难。
Speaker 156:27 - 56:33
Yeah. I I don't know. I mean, a bunch yeah. They they definitely they definitely put a lot of time into it. Tokens are also very valuable. Speaker 156:27 - 56:33
对啊。我也不太知道。我的意思是，很多，是的。他们肯定、绝对在这上面投入了很多时间。token 也非常有价值。
Speaker 156:34 - 57:02
And, you know, what's been what's been interesting as we sort of started seeing these trends, maybe at this point six, nine months ago in various flavors, but then they they escalated a bunch. You know, talking to AI companies, AI company I mean, the the large AI companies are all over this. Like, it's, like, one of the most existential things for their margins. They, have been, like, in the trenches with us identifying the issues. They have been, like, you know, we we literally, like, we see a problem. Speaker 156:34 - 57:02
而且，你知道，比较有意思的是，随着我们开始看到这些趋势——大概是在六到九个月前，以各种不同形式出现——后来它们又进一步加剧了。你知道，和 AI 公司交流时，我说的 AI 公司，是那些大型 AI 公司，它们全都非常关注这件事。因为这几乎是对它们利润率最具生存性影响的问题之一。它们一直和我们一起在一线排查这些问题。它们会说，你知道，我们是真的亲眼看到问题了。
Speaker 157:02 - 57:08
We build a model. We like you know, multi account abuse. Okay. Like, at the time of login, there's an API. You send us what you know. Speaker 157:02 - 57:08
我们构建一个模型。比如说，你知道，多账号滥用（multi account abuse）。好，在登录时会有一个 API（应用程序接口）。你把你已知的信息发给我们。
Speaker 157:08 - 57:20
We send you back a score. You block if they're bad. At the time of free trial start, same thing. As people are accumulating usage, you send us all the metadata. We send you back whether they're fraudy and you can require a top up or a cutoff service or whatever. Speaker 157:08 - 57:20
我们会回传给你一个评分。如果他们风险高，你就拦截。到了开始免费试用的时候，也是同样的流程。随着用户不断累积 usage（使用量），你把所有 metadata（元数据）发给我们。我们再回传给你他们是否有欺诈倾向（fraudy），然后你就可以要求他们先 top up（充值），或者直接 cutoff service（停止服务），或者采取别的处理方式。
Speaker 157:20 - 57:33
Like, literally, each of those, as we've seen them, we've, like, order of weeks gotten you know, generally, this isn't even like a full fledged product. It's like an API. Like, you send us some stuff. We send you some stuff. And the adoption has just been like, okay. Speaker 157:20 - 57:33
实际上，这里面的每一项需求，只要我们看到之后，通常都是几周量级就做出来了。一般来说，这甚至都不算一个完整成型的产品，更像是一个 API。就是，你发一些东西给我们，我们回一些东西给你。而 adoption（采用）速度就是那种：好，直接上。
Speaker 157:33 - 57:59
So the the AI companies are are are all over this. I think what's interesting is, there's like, every company is going to become an AI company. And I don't think the industry at large is thinking about this yet or has really even reasoned that, like, actually, most of the fraud that's happening is not traditional, like, credential or payment fraud. It's like what we would historically have called, like, first party abuse. Right? Speaker 157:33 - 57:59
所以这些 AI 公司都非常重视这个。我觉得有意思的是，每一家公司都会变成 AI 公司。但我不认为整个行业现在已经在认真思考这一点，甚至还没有真正推导到这样一个结论：实际上，现在发生的大多数 fraud（欺诈），并不是传统意义上的 credential（凭证）欺诈或 payment fraud（支付欺诈）。它更像是我们过去会称为 first party abuse（第一方滥用）的东西。对吧？
Speaker 157:59 - 58:37
Like, resale abuse or account sharing or multi accounting or free trials, like first party abuse. And I think it wasn't that first party abuse didn't happen before. It's just at least in sort of SASE stuff, first party abuse didn't cost you anything. And so, a, it wasn't that useful to get, like, a little bit of Salesforce for free where, you know, the examples we talked about wouldn't be relevant in in most SASE. And and b, even if it was useful to figure out a way to, like, skim off the top, and you could get a little bit of of profit for it as the fraudster, it didn't actually cost the business that much because their marginal cost was zero. Speaker 157:59 - 58:37
比如转售滥用、账号共享、多账号、免费试用滥用，这些都属于 first party abuse。我认为，不是说以前没有 first party abuse，只是至少在 SASE 这类东西里，first party abuse 以前不会给你带来任何成本。所以，第一，像白嫖一点 Salesforce 这样的事，本来也没那么有用；而且我们前面谈到的那些例子，在大多数 SASE 里其实也不太适用。第二，就算它确实有用，欺诈者也确实能通过某些方式从中抽一点油水、赚到一点利润，它实际上也不会给企业造成多大成本，因为它们的 marginal cost（边际成本）是零。
Speaker 158:38 - 59:12
And so, you know, as every business becomes an AI business, I think, we've we've been, in the context of our work on on radar, really reasoning about our, fraud prevention product as moving from transaction to full customer life cycle and moving from traditional fraud to end to end abuse. But I don't think, like, the whole industry is is there yet. And, you know, you and I talked about much of the economic upside of AI, but I think that'll really only be realized, if it can happen safely. So for example, six, nine months ago, I was talking to some of these AI companies. They they'd be like, oh, I know. Speaker 158:38 - 59:12
所以，随着每一家企业都变成 AI 企业，我认为，结合我们在 radar 上的工作背景，我们一直在思考：我们的 fraud prevention product（欺诈防护产品）正在从只关注 transaction（交易）转向覆盖完整 customer life cycle（客户全生命周期），并且从传统 fraud（欺诈）转向 end to end abuse（端到端滥用）防控。但我不觉得整个行业已经走到这一步了。而且，你知道，你和我谈过 AI 的很多经济上行空间，但我认为，只有在这件事能够安全发生的前提下，这些价值才真正能被释放。举个例子，六到九个月前，我和其中一些 AI 公司聊天时，他们会说，哦，我知道。
Speaker 159:12 - 59:40
I'm gonna solve my free trial abuse problem by cutting off free trials. Like, I'm gonna solve my free trial abuse problem by only having a sales sold motion and only going after enterprises and not having PLG and right? And, like, you know, I hear less of that today. Why? Not because the fraud's totally solved, but because, everyone knows they need agents to also be their buyers. Speaker 159:12 - 59:40
我打算通过砍掉 free trial（免费试用）来解决 free trial abuse（免费试用滥用）问题。比如，我要通过只保留 sales-led motion（销售驱动模式）、只做 enterprise（企业客户）、不做 PLG（产品驱动增长）来解决这个问题，对吧？不过，你知道，我今天已经很少听到这种说法了。为什么？不是因为 fraud（欺诈）已经被彻底解决了，而是因为大家都知道，他们也需要让 agents（智能体）成为自己的买家。
Speaker 1 | 59:40 - 1:00:03 And if agents are gonna be their buyers, they better have a self serve motion. They better have a PLG motion. Like, there's no way they wanna siphon off that source of growth, and sort of only double down on, like, a highly secure sales sold motion. But, it's been it's been interesting to see what's happened with token theft. I I totally agree the fraudsters are are creative, but I think that's, like, a manifestation of how valuable the tokens are.
Speaker 1 | 59:40 - 1:00:03 如果 agents 要成为他们的买家，那他们最好得有 self-serve motion（自助式增长模式）。他们最好得有 PLG motion。就像，他们不可能愿意把那部分增长来源直接切掉，只去加码那种高安全性的 sales-led motion。不过，看到 token theft（token 盗窃）这件事的发展还挺有意思的。我完全同意 fraudsters（欺诈者）很有创造力，但我觉得，这恰恰体现了 tokens 有多值钱。
Speaker 1 | 1:00:03 - 1:00:06 And so I actually don't think that creativity is is going anywhere.
Speaker 1 | 1:00:03 - 1:00:06 所以我其实不觉得这种创造力会消失。
Speaker 2 | 1:00:06 - 1:00:14 And radar is, one, real time and, two, presumably entirely AI driven as well? Mean, it's like AI fighting AI kind of a kind of thing?
Speaker 2 | 1:00:06 - 1:00:14 那 Radar 一是 real time（实时）的，二是按理说也完全由 AI 驱动，对吗？我的意思是，这有点像 AI 对抗 AI 的那种情况？
Speaker 1 | 1:00:14 - 1:00:31 Yeah. It's real time. It's AI driven. And then I actually, like, most importantly here, for its, like, differentiation is just looks across the Stripe network. So, like, there's basically no good AI buyer we haven't seen before, and there are very few bad AI buyers we haven't seen before.
Speaker 1 | 1:00:14 - 1:00:31 对。它是 real time 的，是 AI 驱动的。然后其实我觉得这里最重要、也是它的差异化所在，是它能看到整个 Stripe network。也就是说，基本上没有什么好的 AI buyer（AI 买家）是我们以前没见过的，而坏的 AI buyer 里，我们以前没见过的也非常少。
Speaker 1 | 1:00:31 - 1:00:57 And so, that combination, you know, it's it's, yes, the size of the network and there's 2% of global GDP, flowing through Stripe. But really when it comes to AI, it's the density of the network. And we talked about LINK briefly, but to give you a sense, like, Lovable is a good example. As an AI company, 58% of Lovable's volume flows through LINK. Like, LINK is, like, you know, LINK is an extremely dense network when it comes to AI.
Speaker 1 | 1:00:31 - 1:00:57 所以，这种组合的价值，没错，一方面来自网络规模——有全球 GDP 的 2% 在 Stripe 上流转。但说到 AI，真正关键的是网络密度。我们刚才简单提到过 LINK，举个例子，Lovable 就很典型。作为一家 AI 公司，Lovable 58% 的交易量是通过 LINK 流动的。可以说，LINK 在 AI 这件事上是一个密度极高的网络。
Speaker 1 | 1:00:57 - 1:01:15 And so, you know, you can sort of extrapolate with them. But if if we know who all the good buyers are, and we've seen the bad guys be bad somewhere else, then combine that with YesAI and sort of these sort of real time these real time APIs, and you get, pretty good fraud defenses.
Speaker 1 | 1:00:57 - 1:01:15 所以，你可以顺着这个逻辑推下去。如果我们知道所有好的 buyers 是谁，也见过坏人曾经在别处作恶，那再把这些和 YesAI 以及这类 real time APIs（实时 API）结合起来，你就能得到相当不错的 fraud defenses（欺诈防御）。
Speaker 2 | 1:01:15 - 1:01:22 You mentioned Tempo at some point. Should we cover Tempo? Like, is is how relevant is Tempo to the Genetic Commerce conversation?
Speaker 2 | 1:01:15 - 1:01:22 你刚才某个时候提到了 Tempo。我们要不要讲讲 Tempo？就是说，Tempo 跟 Genetic Commerce 这个话题到底有多相关？
Speaker 1 | 1:01:22 - 1:01:51 So I think there's there's a couple components of our work with Tempo that I think are, really interesting. So, when you think about Agentic Commerce on the business side, we launched the machine payments protocol or MPP, and we built it with Tempo, And it's an open standard, and the way it works is, quite elegant. Right? Like, an agent requests access to a service. It can be whatever an API or an MCP server or whatever.
我认为我们与 Tempo 的合作里有几个部分非常有意思。比如从业务侧来看 Agentic Commerce，我们推出了 machine payments protocol，也就是 MPP，而且这是和 Tempo 一起构建的。它是一个开放标准，运行方式也相当优雅。也就是说，一个 agent 会请求访问某项服务，这项服务可以是 API、MCP server，或者别的什么。
Speaker 1 | 1:01:51 - 1:02:39 And then the service responds with a payment request, and then the agent pays. And there's no kind of account creation or checkout UI or human in the loop or sort of the way you and I would traditionally, engage on the Internet. It's just this very kind of machine readable, standardized way for agents to to buy from businesses. And that is really, MPP is really the primary, mechanism that we're seeing businesses use, for kind of agents agents as buyer. The other collaboration with, with Tempo that I am super bullish on is is more related to to fraud, because, you know, agents are increasingly becoming the users of AI products, and agents can burn through tokens very, very quickly.
然后该服务会返回一个 payment request（支付请求），接着 agent 付款。整个过程中没有什么账户创建、checkout UI、human in the loop（人工介入），也不是你我传统上在 Internet 上那种交互方式。它就是一种机器可读、标准化的方式，让 agents 可以向企业购买服务。我们现在看到，MPP 确实是企业在“由 agents 作为买方”这种场景下使用的主要机制。另一个我对与 Tempo 的合作非常看好的方向，则更多与 fraud（欺诈）有关。因为你知道，agents 正在越来越多地成为 AI 产品的用户，而 agents 消耗 token 的速度会非常、非常快。
Speaker 1 | 1:02:39 - 1:03:16 And so we were talking a little bit about this, like, dichotomy that a business faces where either you can, like, siphon off self serve and be, really safe but grow slowly or open it up, including to agents, but then, like, you know, be putting yourself at risk of for quite a bit of abuse and monetary losses. Neither of those is great. What you actually wanna do, especially when the agents are the buyers, is you wanna, track the tokens as they're consumed. And you mentioned the infrastructure you wanna track them in real time at substantial scale. And then as importantly, you don't just wanna track them as they're consumed.
所以我们之前也谈到过，企业会面临这样一种两难：要么你把 self serve（自助服务）这部分收紧一些，虽然更安全，但增长会比较慢；要么你把入口开放，包括向 agents 开放，但这样又会让自己面临相当多的滥用风险和资金损失。这两种都不理想。真正应该做的，尤其是在 agents 作为买方时，是去跟踪 token 的消耗情况。你刚才提到了基础设施——你需要以大规模、实时的方式来跟踪它们。而且同样重要的是，你不只是想在 token 被消耗时进行跟踪。
Speaker 1 | 1:03:16 - 1:04:01 You actually wanna, like, collect payments as they're consumed. And so, we call this streaming payments, and it's what Metronome and Tempo, which is the blockchain optimized for payments that Stripe helped co build, are are making possible together. And so just Metronome's job is, like, track the usage in real time, and Tempo's job is just enable this, like, fast, low cost, high volume micro payments, that settle instantly. Obviously, they settle in in stables. And so put together, it's like AI companies, can charge generally agent buyers as tokens are consumed, instead of having to choose between kind of, closing off business, you know, or or getting stiffed on on the invoice.
你实际上还希望在它们被消耗的同时就完成收款。所以我们把这叫做 streaming payments（流式支付），而这正是 Metronome 和 Tempo——一个由 Stripe 帮助共同构建、专为支付优化的 blockchain——正在一起实现的能力。Metronome 的工作就是实时跟踪 usage（使用量），而 Tempo 的工作则是支持这种快速、低成本、高吞吐的 micro payments（小额支付），并且能够即时结算。显然，它们是用 stables，也就是 stablecoin 来结算的。所以把两者结合起来，效果就是 AI 公司可以在 token 被消耗的过程中，持续向买方——通常是 agent 买方——收费，而不用在“把业务入口关掉”和“最后 invoice 收不回来”之间二选一。
Speaker 1 | 1:04:01 - 1:04:12 So, we we've, yeah. We're we're very bullish about tempo and stables in general in in the agent economy at large and for the purposes of, of agent ecommerce.
所以，是的，我们非常看好 Tempo，也非常看好 stables 在更广义的 agent economy 以及 agent ecommerce 中的作用。
Speaker 2 | 1:04:12 - 1:04:32 Super interesting. Alright. So, perhaps as the last topic, you guys have all sorts of interesting stats about the AI economy in general, but, like, in particular, AI startups. I think we covered some of that, last time. What have you seen in the last year in terms of AI startups, trends, and facts, and growth rates?
非常有意思。好，那也许作为最后一个话题，你们手上有很多关于整个 AI economy 的有趣统计数据，尤其是 AI startups。我们上次应该也聊过其中一些。那么在过去一年里，关于 AI startups 的趋势、事实和增长率，你们看到了什么？
Speaker 2 | 1:04:32 - 1:04:33 What have you seen?
你看到了什么？
Speaker 1 | 1:04:33 - 1:05:07 So it's interesting. When we when we talked last year, we talked about, the growth of AI startups and, how they looked different than traditional startups. And what I would say today is, like, know, there's definitely a delta between AI startups and non AI startups. But what's more striking to me is how AI is just changing the startup ecosystem generally. And so, you know, in the vein of Vibe coding and Vibe deployment and all of that, like, new business registrations are, well so I think, in general, the the business formation story is underappreciated.
这很有意思。我们去年聊天时，谈到了 AI startups 的增长，以及它们与传统 startups 有什么不同。而我今天想说的是，AI startups 和非 AI startups 之间当然仍然存在明显差异。但对我来说，更显著的是 AI 正在如何改变整个 startup ecosystem。所以，顺着 Vibe coding、Vibe deployment 以及这一类趋势来看，new business registrations（新企业注册）正在增长。更广泛地说，我觉得 business formation（企业创办）这个故事，其实是被低估了的。
Speaker 1 | 1:05:07 - 1:05:43 So, new business registrations are up basically around the world. At least for advanced economies, they're up, like, 40% in The Netherlands and 70% in Finland and 80% in France. And so there's this sort of, like, surge in dynamism that, like, yes, we see and feel in here in The US, but it's happening, across advanced economies. When we look at this with the Stripe lens, the pace of new businesses launching has doubled since we talked last year. And not all those businesses are AI businesses, but, many, many of them were, made possible because of because of AI.
所以，新企业注册量基本上在全球都在上升。至少在发达经济体里是这样：The Netherlands 大约上涨了 40%，Finland 上涨了 70%，France 上涨了 80%。所以现在出现了一波这种活力激增的趋势——对，这种趋势我们在 The US 内部确实能看到、也能感受到，但它其实正在整个发达经济体范围内发生。用 Stripe 的视角来看，自从我们去年聊这件事以来，新企业成立的速度已经翻倍了。这些企业并不全是 AI 企业，但其中非常非常多的企业，确实是因为 AI 才得以成立。
Speaker 1 | 1:05:43 - 1:06:03 And they're not just getting started. They're also scaling. So, like, Atlas is our, product for founders to incorporate. And Atlas startups from the 2026 cohort you know, it's only June, so it's early in their life cycle. But they're tracking to, like, five times the revenue of last year's class at at the same, number of months.
而且它们不只是刚刚起步，它们也在扩张。所以，Atlas 是我们面向创始人提供公司注册的产品。Atlas 里 2026 cohort 的创业公司——当然，现在才 6 月，它们还处在很早的生命周期阶段——但按目前趋势看，在相同的成立月数下，它们的收入大约会是去年那一届的五倍。
Speaker 2 | 1:06:03 - 1:06:03 Five times.
五倍。
Speaker 1 | 1:06:04 - 1:06:25 Five times. And some of that is there's more of them. But a bunch of it is they are getting to their first dollar faster, and then they are scaling more up more quickly. And getting to their first dollar faster, a lot of that is for sure AI. And then scaling up more quickly, the you know, the a big part of this is actually how they're going global.
五倍。其中一部分原因是它们的数量更多了。但很大一部分原因是，它们更快就赚到了第一笔收入，之后扩张速度也更快。更快赚到第一笔钱，这里面很多无疑都要归因于 AI。而扩张更快，其中一个很大的因素，其实是它们走向全球的方式。
Speaker 1 | 1:06:25 - 1:06:52 And I don't know how much of that is AI or not, but, like, sort of the old model people had was you get big, and then once you're big, you deserve to go global. And what we're seeing with, like, increasingly over the last year is you literally go global from it doesn't litter necessarily literally mean every country, but it's like you're in dozens of countries on day one, like your launch day. And that is how you get big. You get big by being global. And so, I guess we're talking about AI.
这里面有多少要归因于 AI，我也不完全确定，但以前人们的旧模式是：你先做大，等你足够大了，你才“有资格”全球化。而我们看到的是，尤其是过去一年越来越明显的趋势是，你几乎从一开始就在全球化——当然这不一定字面上意味着每个国家，但你的上线第一天、发布当天，就已经进入了几十个国家。而你也正是通过全球化来做大的。你是靠全球化变大的。所以，嗯，我想这也算是在谈 AI。
Speaker 1 | 1:06:52 - 1:07:13 So I could use, like, Emergent Labs as an example. Right? AI platform for, you know, you build and deploy these kind of full stack apps. So they were founded in 2024 in The US. 70% of their revenue, comes from international sales, and they do material business in 16 countries.
所以我可以拿 Emergent Labs 举例。对吧？它是一个 AI 平台，用来构建和部署这类 full stack app（全栈应用）。它于 2024 年在 The US 成立，70% 的收入来自国际销售，并且在 16 个国家都有相当实质性的业务。
Speaker 1 | 1:07:13 - 1:07:56 Like, a substantial share of their revenue comes from many, many countries. So, anyway, I I think there's this sort of, like, yes, there's AI companies and, you know, it's moved from, you know, just being sort of the underlying providers to, like, a lot of rapper businesses, proliferation of rapper businesses across every single vertical. But I think what's kind of more interesting from the tops down macro perspective is just you used to have to be a developer to be a builder and therefore build a business. Now you kind of have to have an idea plus vibe coding, plus vibe deployment, plus reasonable economic infrastructure. And then, you know, we see this proliferation of, new businesses, and they, aren't, their their development is not easily arrested.
比如说，它有相当大一部分收入来自很多很多国家。总之，我觉得现在的情况有点像这样：对，确实有 AI 公司，而且它已经从最底层的基础提供商，发展到在每一个垂直领域里都涌现出大量 wrapper business（封装型业务）。但如果从更高层的宏观视角来看，我觉得更有意思的是：以前你得先是个 developer（开发者），才能成为 builder（构建者），进而创办一家企业。现在你大概只需要一个想法，加上 vibe coding，加上 vibe deployment，再加上还算合理的经济基础设施。然后，你就会看到新企业大量涌现，而且它们的发展并不容易被遏制。
Speaker 1 | 1:07:56 - 1:08:16 Right? Like, even with one employee, they, you know, start monetizing early. They, grow quickly. They, expand across a bunch of markets. And, I think that's almost certainly helped by all of the operational work that can also be done with AI, is less than sort of the day to day core wheelhouse of Stripe.
对吧？哪怕只有一名员工，它们也会很早开始 monetizing（变现）。它们增长很快，扩张到很多市场。而我认为，这几乎肯定也得益于大量运营工作现在同样可以用 AI 来完成，尽管这部分不算是 Stripe 日常核心能力范围内最直接的领域。
Speaker 1 | 1:08:16 - 1:08:46 We help with some of that on the accounting or rev rec or whatever. But, you know, much of the customer support and other operations are are obviously done by other businesses. But I think it's an interesting time. And there's a lot of discussion of like, is AI going to lead to a small number of firms, with heavy market share and not a lot of competition? And I think one of the reasons I'm bullish on the AI economy is at least so far, for sure, there are big guys who have things that are highly complementary to AI.
我们会在其中一部分上提供帮助，比如 accounting、rev rec（收入确认）之类的工作。但你知道，很多 customer support（客户支持）和其他运营工作显然还是由别的企业来完成。不过我觉得现在是个很有意思的时点。现在有很多讨论在说，AI 会不会导致只剩下少数几家企业占据很高的 market share（市场份额），而竞争却不多？我之所以看好 AI economy（AI 经济），至少到目前为止，一个重要原因是，当然，确实有一些大公司拥有与 AI 高度互补的东西。
Speaker 1 | 1:08:46 - 1:09:01 We we don't need to go through that list. Everyone knows them, that are exploding. But there's also an explosion of little guys coming in the market and, not just being created, but, like, reaching customers and growing quickly. And so I think that bodes well for for competition and for and for economic growth.
我们没必要把那份名单再过一遍了，大家都知道那些正在爆发增长的公司。但与此同时，也有大量小公司涌入市场，而且不只是被创建出来而已，它们还真的在触达客户、快速增长。所以我觉得这对竞争、以及对经济增长，都是个好兆头。
Speaker 2 | 1:09:01 - 1:09:27 All of this is obviously, extraordinarily exciting. But since we talked about, token costs and, you know, usage base and all the thing, do you worry that, like, a part of this is a little bit people, on the on the on the buying side and on the usage side got a little carried away, quite didn't quite realize the dollar amount that, using AI represented and that there might be some kind of backlash against that, hypergrowth.
这一切显然都让人无比兴奋。但既然我们刚刚谈到了 token 成本、usage base（使用基础）以及诸如此类的事情，你会不会担心，这里面有一部分是买方和使用方有点过于上头了，没有完全意识到使用 AI 实际代表着多大的美元支出，因此可能会对这种 hypergrowth（超高速增长）出现某种反弹？
Speaker 1 | 1:09:27 - 1:10:09 So, I mean, we've all read the stories of companies who have accidentally gone bananas on token spend because they had, like, no control over what their employees were doing. I think the companies who have truly gone bananas are by and large the companies with pretty deep pockets, which isn't to say that it's not, like, gonna be a problem in the economy if that spending continues. But there are smart companies that are well run, well managed. They can handle a month or two of poor decisions and runaway costs. And, you know, by and large, I I don't exactly know what we're talking about at these companies, but we're talking oh, 2%, 3%, 4% of their headcount costs are going into tokens and, 30% of that or 40% of that is inefficient.
我的意思是，我们都看过那种故事：有些公司因为根本没有控制员工在做什么，结果 token 支出意外地飙得离谱。我认为，真正把支出搞到离谱的公司，整体上大多都是那些资金相当雄厚的公司。这并不是说，如果这种支出持续下去，就不会对经济造成问题。但那些聪明、运营良好、管理到位的公司，可以承受一两个月的错误决策和失控成本。而且，总体来说，我并不完全清楚这些公司的具体数字，但我们讨论的大概是这样一个量级：它们 2%、3%、4% 的 headcount costs（人力成本）流向了 tokens，其中 30% 或 40% 是低效的。
Speaker 1 | 1:10:09 - 1:10:33 I think they can, like, pretty quickly get back to an an efficient frontier, and and it's not really anything existential. For the little guys, actually, don't don't think that's happening. You know, many of them are sort of making fixed fee fixed fee purchases, and they're still on small plans and and whatever. So, I mean, the stories are out there. And by the way, just to be clear, like, I think we have a lot to learn about the efficient frontier of AI use.
我觉得它们可以相当快地回到一个 efficient frontier（效率前沿），这其实不是什么生死攸关的问题。至于那些小公司，事实上，我并不认为这种情况正在发生。你知道，它们很多都是在做 fixed fee（固定费用）的采购，仍然还在用小套餐之类的。所以，这类故事确实存在。顺便说清楚一点，我认为我们对于 AI 使用的 efficient frontier 还有很多东西要学。
Speaker 1 | 1:10:33 - 1:11:08 And there's a component there around, like, model routing and what models use like for the job. And there's a component around just, like, observability, which I think many companies have find found themselves to be behind on. And then there's just like like norms and controls and guardrails. And, you know, I think we all want high ROI, usage of LLMs. But we also want employees to know when they're, you know, rack it up really substantive costs, and we want them to be able to tell us, whether they anticipate the ROI, of those costs is gonna be there.
这里面有一部分和 model routing（模型路由）有关，也就是针对不同任务该用什么模型。还有一部分只是 observability（可观测性）问题，我觉得很多公司都发现自己在这方面落后了。然后还有 norms、controls 和 guardrails（规范、控制与护栏）的问题。你知道，我想我们都希望 LLMs 的使用能有高 ROI（投资回报率）。但我们也希望员工知道，什么时候他们实际上正在累计相当可观的成本；我们也希望他们能够告诉我们，他们是否预期这些成本最终会带来相应的 ROI。
Speaker 1 | 1:11:08 - 1:11:19 So, yeah, I think there'll be like a little bit of a recalibration, but I don't think anything existential has happened that's gonna annul the sort of long run upside here.
所以，是的，我觉得会有一点 recalibration（重新校准），但我不认为发生了什么根本性的问题，会抹掉这里长期上行空间的那种前景。
Speaker 2 | 1:11:19 - 1:11:38 All right. So as a last question, when we talk again in a year from now, where do you think we are in terms of agentic commerce, maybe using L1 to L5? And I I won't hold you to the prediction, but, like, directionally, what do you think realistically is going to happen in the next twelve months?
好的，那最后一个问题：如果我们一年后再聊，当我们谈到 agentic commerce（agent 驱动的商业）时，你觉得我们会处在什么位置？也许可以用 L1 到 L5 来衡量。我不会拿这个预测来要求你兑现，但从方向上说，你觉得未来十二个月里，现实地看，最有可能发生什么？
Speaker 1 | 1:11:38 - 1:12:10 I think the most interesting thing in the next twelve months is I mean, I think we'll I think we'll move up, and I don't know if it's gonna be to four or to five. But I I think the more interesting thing is actually that, like we talked about agents as economic actors mostly in the context of buying. But I think we'll start to see and I'm not saying this will be proliferated everywhere, but I think we'll start to see agents that are, like, multifaceted economic actors. They're buying and they're selling, and they're provisioning infrastructure, and they're running businesses. And they're, like, doing the thing end to end.
我觉得未来十二个月里最有意思的事情是——我的意思是，我觉得我们会继续往上走，至于会到四还是到五，我也不知道。但我认为更有意思的其实是，之前我们谈到 agent 作为 economic actors（经济行为体）时，主要是在“购买”这个语境下讨论的。不过我觉得我们会开始看到——我不是说这会到处大规模出现——一些 agent 变成多面向的 economic actors。它们既在买，也在卖，还在配置基础设施，也在运营业务。也就是说，它们会把整件事从头到尾做完。
Speaker 1 | 1:12:10 - 1:13:04 And, again, I don't know how many of these there will be or how they'll operate or whether it'll be, like, with each other or, like, in these, like, weird niche silos. But I think all of that's just gonna continue to demand more purpose built infrastructure, including financial infrastructure versus just the sort of old, human centric commerce stack. And so that that's that's kinda where like, what is what does it look like when, okay, I'm vibe deploy, I'm I'm vibe coding and I'm vibe deploying, and then, my content is my offerings are default exposed. And, actually, we just we just quietly went to public preview on, on Stripe directory, which is just a really easy way for agents to, discover providers. And then through Stripe projects, they can integrate them directly.
再说一次，我不知道这样的 agent 会有多少，会如何运作，或者它们是彼此协作，还是存在于一些奇怪的小众 silo（孤岛）里。但我觉得，这一切都会继续催生更多 purpose-built（专门构建的）基础设施需求，包括金融基础设施，而不只是那种旧式的、以人为中心的 commerce stack（商业技术栈）。所以我现在在想的是：当“好吧，我在 vibe deploy，我在 vibe coding，我在 vibe deploying，然后我的内容、我的 offerings（服务供给）默认就是公开暴露出来的”时，这到底会是什么样子。实际上，我们刚刚也低调地把 Stripe directory 推到了 public preview，这是一种让 agent 更容易发现 provider（服务提供方）的方式。然后通过 Stripe projects，它们可以直接完成集成。
Speaker 1 | 1:13:04 - 1:13:39 But, like, okay. And then an agent's, like, also discovering everything and integrating it and buying it and then, you know, creating a service out of the combination of things it has provisioned and integrated and bought and is starting to sell a thing. Anyway, like, this this idea of, an agent as a micro firm, I think would probably be the most interesting, thing to see twelve months from now. And, again, I don't think the median, firm is going to be, or forget a solopreneur, a what would it be? A solo agent.
但再进一步想，假如一个 agent 还能主动发现所有东西，把它们集成起来、买下来，然后你知道，把它所 provision（配置）、integrate（集成）、buy（购买）来的这些能力组合成一个服务，并开始出售某种东西。总之，我觉得，把 agent 看作一家 micro firm（微型公司）这个想法，可能会是十二个月后最值得观察的事情。再说一次，我并不认为典型的 firm（公司）会变成这样，或者甚至别说 solopreneur（个体创业者）了，那该叫什么？solo agent。
Speaker 1 | 1:13:41 - 1:14:06 I don't think that's I don't think that's a world to live in, but I think twelve months from now, we could totally see some examples of that that sort of pave the path for, like, the the whole thing end to end. And, you know, this always happens with the new technology. Right? You take, like, your current processes or market or whatever you have, and you figure out how does the new technology make that 5% more efficient or 10% more efficient. Keeps me from having to type in my credit card number.
我不认为那会是一个适合长期生活于其中的世界，但我觉得在十二个月后，我们完全可能看到一些这样的例子，为整套端到端模式铺路。你知道，新技术总是这样。对吧？你会拿当前的流程、市场，或者手头已有的东西，然后思考：新技术怎样能让它提升 5% 的效率，或者 10% 的效率；怎样让我不用再手动输入信用卡号。
Speaker 1 | 1:14:06 - 1:14:31 Right? But, like, where it actually gets interesting is where, we we start to reimagine, like, how the system works. And it's not Emily permissioning an agent to buy on her behalf. It's Emily has an agent who's tasked with running a business, and that includes buying some things and selling some things and making some profits. And, that'll that'll be the world that I that I would like to be talking about 12 from now.
对吧？但真正开始变得有意思的地方在于，我们开始重新想象整个系统是如何运作的。那不再是 Emily 给一个 agent 授权，让它代表她去购买；而是 Emily 拥有一个 agent，这个 agent 的任务是经营一门生意，而这其中包括买一些东西、卖一些东西，并赚取一些利润。而那会是我希望在 12 个月后我们正在讨论的世界。
Speaker 2 | 1:14:31 - 1:14:35 Well, Emily, it's been another amazing conversation. Thank you so much. Really enjoyed it.
嗯，Emily，这又是一次非常精彩的对话。非常感谢你。我真的很享受这次交流。
Speaker 1 | 1:14:35 - 1:14:36 Thank you.
谢谢。
Speaker 2 | 1:14:37 - 1:14:55 Hi. It's Matt Turk again. Thanks for listening to this episode of the mad podcast. If you enjoyed it, we'd be very grateful if you would consider subscribing if you haven't already or leaving a positive review or comment on whichever platform you're watching this or listening to this episode from. This really helps us build a podcast and get great guests.
大家好，我是 Matt Turk。感谢你收听这一期 mad podcast。如果你喜欢这一期内容，我们会非常感激你考虑订阅——如果你还没订阅的话——或者在你观看或收听本期节目的平台上留下积极的评价或评论。这对我们建设这个 podcast、邀请到优秀嘉宾真的很有帮助。
Speaker 2 | 1:14:55 - 1:14:57 Thanks and see you at the next episode.
谢谢，我们下期节目见。