BuildSpeak每日 builder 文摘
今日归档生词本关于
🎙 播客AI & I by Every· 2026 年 4 月 29 日· 9,313 词 · 约 47 分钟

How Stripe Is Building for an Agent-native World

SPACE 播放 / 暂停·←→ 上一句 / 下一句
Speaker 100:00 - 00:29
The Internet has this new kind of actor on it. Over time, this actor, these agents will become the predominant actors on the Internet. These AI companies are just growing from a revenue perspective faster than any previous cohort we've seen. LLM traffic to Stripe Docs is up 10 x year over year. And that's just a useful signal that machines are becoming users of developer infrastructure too, including Stripe's developer infrastructure.
Speaker 100:00 - 00:29
互联网现在出现了一类新的参与者。随着时间推移,这种参与者,也就是这些 agent(智能体),会成为互联网上最主要的参与者。从收入角度看,这些 AI 公司增长的速度比我们此前见过的任何一批公司都更快。流向 Stripe Docs 的 LLM(大语言模型)流量同比增长了 10 倍。而这只是一个有用的信号,说明机器也正在成为开发者基础设施的用户,其中也包括 Stripe 的开发者基础设施。
Speaker 100:33 - 00:33
Signal
Speaker 100:33 - 00:33
Signal
Speaker 200:44 - 00:45
Emily, welcome to the show.
Speaker 200:44 - 00:45
Emily,欢迎来到节目。
Speaker 100:45 - 00:47
Thanks so much, Dan.
Speaker 100:45 - 00:47
非常感谢,Dan。
Speaker 200:47 - 01:25
So, really excited to have you. You are the head of data and AI at Stripe. And I feel like this is such a good time to have someone from Stripe on because you all famously are increasing the GDP of the Internet. And the Internet is changing so much right now, and and therefore the, like, economy of the Internet is changing from something where, humans are buying and selling from each other to a thing where, to an economy where agents are buying and selling from humans and agents are buying and selling from each other. And I feel like, a, I wanna know what that means for Stripe, but b, I wanna understand since you have this macro view of the agent economy, what does that even mean and what are you seeing?
Speaker 200:47 - 01:25
真的很高兴请到你。你是 Stripe 的 data and AI 负责人。我觉得现在请一位来自 Stripe 的人来聊这件事,时机再合适不过了,因为大家都知道,你们一直在推动互联网的 GDP 增长。而现在互联网正在发生巨大的变化,因此,某种意义上说,互联网经济也正在变化:它正从一个由人类彼此买卖的经济体,转向一个由 agent(智能体)向人类买卖、以及 agent 与 agent 彼此买卖的经济体。我一方面想知道这对 Stripe 意味着什么;另一方面,我也想理解,既然你们对这种 agent economy(智能体经济)有宏观视角,那它究竟意味着什么,你们又看到了什么?
Speaker 101:26 - 01:56
Yeah. So a big shift, I think, we're in the midst of is that the Internet economy is becoming more autonomous, right? So for a long time, for forever, right, the Internet was built around an extremely simple assumption that the main actor was a person. And the person sitting in front of a screen, they're browsing, and they're filling out forms and clicking through checkout, but also they're writing code and setting up tools. And that assumption is starting to break in various ways, right?
Speaker 101:26 - 01:56
是的。我认为我们当下正处于一个重大转变之中:互联网经济正在变得更加 autonomous(自主化),对吧。很长一段时间里,甚至可以说一直以来,互联网都是建立在一个极其简单的假设之上的,那就是主要参与者是人。是坐在屏幕前的人在浏览、填写表单、一步步完成结账流程,同时也是他们在写代码、配置工具。而这个假设现在正开始在多个方面被打破,对吧?
Speaker 101:56 - 02:34
Sometimes the human is still totally in control, but they're interacting through an AI interface instead of through a website or a traditional app. Sometimes the agent is acting on their behalf, and then sometimes software now is just out interacting directly with other software. And as all of that starts to happen, at all of those layers, a lot of things need to be rethought. So, you know, there has been rethinking of how our products discovered and how our products bought, but also what should developer tools look like? And, you know, in our world of Stripe, like what is the underlying economic infrastructure?
Speaker 101:56 - 02:34
有时候,人类仍然完全掌控一切,但他们是通过 AI 界面而不是网站或传统 app 来交互。有时候,agent(智能体)是在代表他们行动;还有的时候,现在的软件就是直接在与其他软件交互。随着这一切开始发生,在所有这些层面上,很多事情都需要被重新思考。所以,你知道,我们一直在重新思考我们的产品该如何被发现、如何被购买;但与此同时,也要思考开发者工具应该是什么样子?以及,在 Stripe 所处的这个世界里,底层的经济基础设施又应该是什么?
Speaker 102:34 - 03:06
So the payments and the billing and the fraud detection and the identity layer, that's needed in this world where, you know, actors are no longer, just humans. And so that for me is kind of the larger frame of the moment. It's not just, hey, AI is making search better, or AI is helping people code, or AI is, you know, evolving commerce on the margin. It's really like, actually, the Internet has this new kind of actor on it. Over time, this actor, these agents will become the predominant actors on the Internet.
Speaker 102:34 - 03:06
所以,payments(支付)、billing(计费)、fraud detection(欺诈检测)以及 identity layer(身份层),这些能力在这个世界里都是必需的——因为在这个世界里,参与者已经不再只是人类了。所以对我来说,这就是当下更大的背景。这不只是说,AI 让搜索变得更好了,或者 AI 在帮助人们写代码,或者 AI 只是在边际上推动 commerce(商业)的演化。更准确地说,是互联网现在出现了一类新的参与者。随着时间推移,这种参与者,也就是这些 agents(智能体),会成为互联网上最主要的参与者。
Speaker 103:06 - 03:23
And as that's happening, basically every layer of the stack starts to need an evolution. So for Stripe, it's like, okay, Stripe, how are we getting agent ready? But then also, how are we helping businesses get agent ready? And both of those are happening in in a number of ways. Yes.
Speaker 103:06 - 03:23
随着这种情况发生,基本上技术栈的每一层都开始需要一次演进。所以对 Stripe 来说,就是,好,Stripe 要怎么为 agent 做好准备?但同时,我们又要怎样帮助企业为 agent 做好准备?这两件事都正通过多种方式同时发生。对。
Speaker 103:23 - 03:26
In commerce, but also, just in how builders build.
Speaker 103:23 - 03:26
这体现在 commerce(商业)里,但也体现在开发者构建产品的方式上。
Speaker 203:27 - 03:55
And can you give me some specific examples of of the kinds of things you're seeing? Like, I'm almost wondering, for example, I know at Stripe, one of the things you do a ton is fraud. A, I assume there's a whole new type of fraud happening, but, b, I'm almost wondering what even counts as fraud now in the sense of it's possible that my agent could go steal someone's credit card and check out. I don't think that Claude would, but, like, you never know with Grock, you know? Check out on
Speaker 203:27 - 03:55
那你能不能给我一些更具体的例子,说明你们正在看到哪些类型的变化?比如说,我几乎在想,我知道在 Stripe,你们做得特别多的一件事就是 fraud(欺诈)。A,我猜现在一定出现了全新的欺诈类型;但 B,我甚至在想,现在到底什么还算是欺诈——因为从某种意义上说,我的 agent 也有可能去盗刷别人的信用卡然后下单结账。我不觉得 Claude 会这么做,但像 Grock 这种,你懂的,谁知道呢?在结账时——
Speaker 103:55 - 04:06
No comment. No comment. But you're right that sort of AI introduces very different fraud problems. You asked what is fraud? We used to think of fraud as sort of payment fraud.
Speaker 103:55 - 04:06
不予置评。不予置评。不过你说得对,AI 确实引入了非常不同的欺诈问题。你刚才问什么是欺诈?我们过去认为的欺诈,某种意义上主要就是 payment fraud(支付欺诈)。
Speaker 104:06 - 04:52
Someone was stealing money. Someone was stealing your card credentials. Increasingly, actually, and I was in a meeting with one of our very large AI users today, fraudsters are stealing compute, and that's a very different type of problem. So in earlier software models, if you think of, like, sort of traditional SaaS, letting someone into a free tier didn't cost you very much, and, stealing a free tier wasn't very valuable to the fraudsters. Now giving someone credits, freemium offering, a free trial, you know, letting them, you know, rack up a bunch of tokens and pay at end of month, except maybe they choose not to pay, actually is a major fraud vector and an existential risk to, a lot of these businesses.
Speaker 104:06 - 04:52
有人在偷钱。有人在窃取你的银行卡凭证。但实际上,越来越多地——我今天还和我们一个非常大的 AI 用户开过会——fraudsters(欺诈者)正在偷的是 compute(算力),而这就是一种完全不同的问题了。所以在早期的软件模式里,如果你想的是那种传统 SaaS,让别人进一个 free tier(免费层级)其实花不了你多少钱,而盗用一个 free tier 对欺诈者来说也没那么有价值。现在则不同:给别人发 credits(额度)、提供 freemium(免费增值)方案、给 free trial(免费试用),或者让他们先累计一大堆 token,到月底再付钱——结果他们也可能选择不付——这实际上已经成了一个主要的 fraud vector(欺诈路径),甚至是很多这类企业面临的生存级风险。
Speaker 104:52 - 05:16
Right? Because in AI, every prompt, every image that gets generated, every API request has a very real cost attached to it. You know, people are talking about intelligence getting cheaper. Yeah, but it's still like very far from free. And then also when you look at sort of the growth model for many of these AI companies, free compute is kind of the new CAC, right?
Speaker 104:52 - 05:16
对吧?因为在 AI 里,每一个 prompt(提示词)、每一张生成的图片、每一次 API 请求,背后都对应着非常真实的成本。你知道,大家都在说 intelligence(智能)正在变便宜。是的,但离“免费”还差得非常远。还有一点,当你去看很多这类 AI 公司的增长模型时,free compute(免费算力)某种程度上就是新的 CAC(获客成本),对吧?
Speaker 105:16 - 06:03
You used to cost of acquisition, like you used to spend a bunch on paid media, now you spend a bunch on your free trials and your credits and your self serve onboarding as sort of a major lever for growth. And so the abuse we see in that context, where compute is the new CAC and compute is very expensive, is threefold. One is multi account abuse. So this is like bad actors come in and they sign up, like, over and over again, and they create a new identity every time on a new email address, and they claim their new user credits, and they stay ahead of detection by, like, iterating across a bunch of different aliases. And just to give you a sense of the order of magnitude, across the AI companies running on Stripe, about 7% of their sign ups are these multi account abusers.
Speaker 105:16 - 06:03
以前说 cost of acquisition(获客成本),通常是你要在付费媒体上花很多钱;现在则是你在 free trial、credits,以及 self serve onboarding(自助式引导注册/开通)上花很多钱,把它们当作增长的主要杠杆。所以在这种语境下——compute 成了新的 CAC,而 compute 又非常昂贵——我们看到的滥用主要有三类。第一类是 multi account abuse(多账号滥用)。也就是说,不良行为者会反复注册,一次又一次地进来,每次都用新的邮箱地址创建新的身份,领取新用户 credits,并通过不断切换各种不同的别名来逃避检测。给你一个量级概念:在所有运行在 Stripe 上的 AI 公司里,大约有 7% 的注册用户属于这类 multi account abusers(多账号滥用者)。
Speaker 106:03 - 06:33
So nontrivial share. The second trend that we see in sort of new vector of abuse is free trial abuse. And this is often sort of the most urgent issue because the unit economics break really quickly. To give you a sense, we had a large AI company, who was seeing only 4% of their free trials convert to paid. And each free trial cost them $25, you know, in LLM spend.
Speaker 106:03 - 06:33
所以这不是一个可以忽略的比例。我们看到的第二个新型滥用趋势是 free trial abuse(免费试用滥用)。而这往往是最紧迫的问题,因为 unit economics(单位经济模型)会非常快地失衡。给你一个概念,我们有一家大型 AI 公司,他们的 free trial 只有 4% 能转化为付费用户。而每一次 free trial,光是 LLM 支出就要花掉他们 25 美元。
Speaker 106:33 - 06:57
And so basically it was costing them $625 per payer before the first dollar of revenue was brought in. And when we double clicked on those sort of free trial folks, the vast, vast majority of them were actually abusers. So they were actually stealing the compute, they never had any intent to pay. These weren't people who were genuinely trying out your service and then chose not to buy. These were people who were literally abusing your systems.
Speaker 106:33 - 06:57
所以,基本上,在带来第一美元收入之前,他们每获得一个付费用户(payer),成本就要 625 美元。而当我们进一步深挖那些所谓的免费试用用户时,发现其中绝大多数、绝大多数其实都是滥用者(abusers)。也就是说,他们实际上是在偷算力(compute),根本没有任何付费意图。这些人并不是真正在认真试用你的服务、然后决定不买的人;他们就是在实打实地滥用你的系统。
Speaker 106:58 - 07:16
And so, you know, some companies just dropped free trials altogether. Of course, that's not great because you're throttling growth. Others responded by blocking virtual cards. So I don't know how often you've been marketed virtual cards. I'm often marketed virtual cards, right?
Speaker 106:58 - 07:16
所以,你知道,有些公司就干脆彻底取消了免费试用。当然,这并不好,因为你是在抑制增长。另一些公司的应对方式,则是屏蔽 virtual cards。至于 virtual cards,我不知道你平时被推销得多不多,反正我经常会被推销,对吧?
Speaker 107:16 - 07:39
Get this, you know, one time use card. It expires after twenty four hours, so you never have to pay for the service. You know, in the hands of a good consumer, fine. The hands of a fraudster, like, very much not fine. The problem with blocking all virtual cards is for AI companies, about 15% of legitimate card transactions on Stripe are actually virtual cards.
Speaker 107:16 - 07:39
比如那种——你知道的——一次性卡。24 小时后就过期,这样你就永远不用为这项服务付款了。放在正常消费者手里,倒也还好;但放在 fraudster(欺诈者)手里,那就非常不妙了。问题在于,如果把所有 virtual cards 一刀切全都屏蔽掉,那么对 AI 公司来说,Stripe 上大约 15% 的合法银行卡交易,其实都是 virtual cards。
Speaker 107:39 - 07:40
So We use that almost like
Speaker 107:39 - 07:40
所以我们几乎把那个当作——
Speaker 207:40 - 07:41
for ramp, for example. Like, we have a bunch
Speaker 207:40 - 07:41
比如说,用在 ramp 上。像我们就有一堆——
Speaker 107:41 - 07:59
of virtual Totally. So you don't wanna be in the same way, you don't wanna be turning off free trials. You don't wanna be throttling, virtual card, virtual cards either. And sort of order of magnitude, you can think of like exponential growth in free trial abuse over the last six months. It's four X.
Speaker 107:41 - 07:59
——virtual cards。完全没错。所以同样地,你不想关闭免费试用;你也不想去限制 virtual card、virtual cards。大致从量级上看,你可以把过去六个月里免费试用滥用的增长理解为指数级增长。现在已经是 4 倍了。
Speaker 108:00 - 08:13
And, for, one large AI user on Stripe, we're currently blocking 250,000 fraudulent free trials a week. So the magnitudes here are And are quite
Speaker 108:00 - 08:13
而且,对于 Stripe 上一家大型 AI 用户,我们目前每周都会拦截 25 万次欺诈性免费试用。所以这里的量级——而且确实相当——
Speaker 208:14 - 08:25
is the is it is the the volume of fraud constant? It's just it's shifting shape, or is just fraud going up because they're more powerful now because they can just use AI agents to do it?
Speaker 208:14 - 08:25
——是说,欺诈的总量是恒定的吗?只是形态在变化?还是说,欺诈本身就在上升,因为他们现在更强了,能够直接用 AI agents(智能体)来实施这些行为?
Speaker 108:26 - 08:42
Fraud's going up because the fraudsters have AI on their side, although it's also on the side of the detectors, but also because the value of the services they can steal is higher. Right? Like, what I don't know. You steal traditional SaaS. Like, what good do you get?
Speaker 108:26 - 08:42
欺诈之所以在上升,一方面是因为 fraudsters 有 AI 站在他们那边——当然,detectors 这边也有——但另一方面也是因为他们能偷走的服务价值更高了。对吧?比如,我不知道。你去偷传统的 SaaS,能得到什么好处呢?
Speaker 108:42 - 08:45
Like, you steal some inference, you steal some compute, you can resell it, can do all sorts
Speaker 108:42 - 08:45
比如,你偷一些 inference(推理服务),偷一些 compute(算力),你可以转手倒卖,还能做各种各样的事。
Speaker 208:45 - 08:49
of Look. I love a good CRM seat, you know? Yeah.
Speaker 208:45 - 08:49
你看,我确实很喜欢一个好用的 CRM seat(CRM 席位),你懂吧?对。
Speaker 108:49 - 08:50
Don't you? Who doesn't love
Speaker 108:49 - 08:50
你难道不喜欢吗?谁会不喜欢
Speaker 208:50 - 08:52
a good me, Emily.
Speaker 208:50 - 08:52
一个好的我,Emily。
Speaker 108:52 - 09:22
CRM seat is three LLMs are for sure more tempting. And by the way, the the third type of sort of new abuse we see is this non payment abuse. Right? So like you, you, incur an overage or you have like, you know, thirty day invoicing, except you never pay, your invoice. You know, in many cases, customers are consuming thousands or tens of thousands of dollars in compute during a month or a day or sometimes an hour.
Speaker 108:52 - 09:22
CRM seat 要有吸引力得多,而 LLMs 肯定更诱人。顺便说一下,我们看到的第三类某种新的 abuse(滥用)是这种 non payment abuse(不付款滥用)。对吧?也就是比如你产生了超额使用费用,或者你采用的是,比如说,30 天账期 invoicing(开票结算),结果你根本不付这张 invoice(账单)。你知道,很多情况下,客户会在一个月内、一天内,有时甚至一小时内,消耗掉价值几千到几万美元的 compute(算力)。
Speaker 109:23 - 09:36
And by the time they get billed and fail payment, you know, that loss has already happened. And these AI companies are left holding the bag. And so for us, like fraud used to be a transaction thing. Now it is a customer thing. It is a full funnel thing.
Speaker 109:23 - 09:36
而等到他们被出账、然后支付失败时,你知道,这笔损失其实早就已经发生了。这些 AI 公司只能自己吞下损失。所以对我们来说,fraud 过去是一个 transaction(交易)问题;现在它是一个 customer(客户)问题。它是一个贯穿 full funnel(全漏斗)的事情。
Speaker 109:36 - 09:44
It starts at the time of sign up. Is this multi account abuse? Should they get credits? Is this free trial abuse? Should we give them a trial in the first place?
Speaker 109:36 - 09:44
它从 sign up(注册)的那一刻就开始了。这是不是 multi account abuse(多账号滥用)?他们该不该拿到 credits(额度)?这是不是 free trial abuse(免费试用滥用)?我们一开始到底该不该给他们试用?
Speaker 109:44 - 10:00
And then when they have overages, should we be throttling them? Should we be requiring top up? Should we be blocking service completely? And it's just kind of it's kind of a whole new world because the thing to steal is much more valuable, and the cost of having it stolen is much more existential. Existential.
Speaker 109:44 - 10:00
然后当他们出现 overage(超额使用)时,我们应该对他们做 throttling(限流)吗?应该要求他们 top up(充值)吗?还是应该彻底阻断服务?这某种意义上真的是一个全新的世界,因为被盗取的东西价值高得多,而一旦被盗,所带来的代价也更具 existential(生存层面)的威胁。Existential。
Speaker 210:00 - 10:22
How are you guys even able to I saw so I totally understand how you need to be in that full funnel in order to detect fraud, but my understanding of, you know, whenever we've integrated Stripe, it's usually, like, on the checkout. We're not necessarily putting you in there when when someone puts in their free their email address for free. So have you changed the product to do the full funnel or how how does that actually work?
Speaker 210:00 - 10:22
你们到底是怎么做到这一点的?我的意思是,我完全理解为了检测 fraud(欺诈),你们需要覆盖 full funnel(全漏斗)各个环节;但据我的理解,过去我们集成 Stripe 时,通常都是放在 checkout(结账)这个环节。我们不一定会在用户只是免费填写 email address(邮箱地址)的时候就把你们接进去。所以你们是已经改了产品来支持 full funnel,还是说这在实际中到底是怎么运作的?
Speaker 110:22 - 10:40
Yes. So radar, which is our fraud protection product used to be at the transaction level. Right? So at the moment of checkout, as you note. But because so much of the fraud risk was coming up funnel, know, AI companies are now increasingly integrating Stripe Radar at the time of sign up.
Speaker 110:22 - 10:40
是的。所以 Radar,也就是我们的 fraud protection(欺诈防护)产品,过去是在 transaction(交易)层面起作用的,对吧?也就是像你说的,在 checkout 的那一刻。但因为大量 fraud risk(欺诈风险)其实出现在漏斗更前面的环节,所以现在 AI 公司越来越多地会在 sign up(注册)时就集成 Stripe Radar。
Speaker 110:41 - 10:55
And so, you know, we see the metadata at the time of sign up, we pass back scores at the time of sign up, and every moment subsequently, again, because fraud is now a full funnel problem, not a transaction problem alone.
Speaker 110:41 - 10:55
所以,在 sign up 的时候,我们会看到相关 metadata(元数据);在 sign up 的时候,我们会回传 score(评分);而且在之后的每一个时刻也都会这样做。还是因为同样的原因:fraud 现在是一个 full funnel 问题,而不再只是单纯的 transaction 问题。
Speaker 210:55 - 11:15
If you're, you know, asking for a friend, if you're running an AI company and you are you don't even know what your fraud rate is and you're you wanna protect yourself from this kind of abuse, what are the top things that you need to do in order to make sure that you're reasonably safe?
Speaker 210:55 - 11:15
如果你是在“帮朋友问”,假设你经营着一家 AI 公司,而你甚至都不知道自己的 fraud rate(欺诈率)是多少,但你又想保护自己免受这类 abuse(滥用),那么为了确保自己处于相对安全的状态,最重要需要做的几件事是什么?
Speaker 111:15 - 11:37
Yeah. So I would just adopt our highest tier radar plan, but the actual mechanics of that are at sign up, you want to know if your customer's good. Before you give them any access to any credits, you want to make sure they're good. At the time they pay, you want to make sure that charge is good. And any time they have an overage, you want to make sure they're good for their money.
Speaker 111:15 - 11:37
对。所以我会直接采用我们最高档的 Radar 方案。但从具体机制上说,在 sign up 的时候,你要知道你的 customer(客户)是不是可靠。在你给他们任何 credits(额度)访问权限之前,你要先确认他们没问题。在他们付款的时候,你要确认那笔 charge(扣款)没问题。而每次他们出现 overage 时,你都要确认他们是有能力付款的。
Speaker 111:37 - 11:58
And that, and you know, there's other stuff around refunds and disputes that we also support, but I think those are the four kind of major moments in the AI company's customer life cycle, where we're just maniacally focused on protecting because that's where we're seeing, the biggest cost and and the fastest fraud growth.
Speaker 111:37 - 11:58
此外,还有一些关于 refunds(退款)和 disputes(争议款项)的事情,我们也支持;但我认为,这四个时点基本上就是 AI 公司 customer life cycle(客户生命周期)中的四个关键时刻。我们会近乎偏执地专注于在这些节点提供保护,因为这正是我们看到成本最高、而且 fraud 增长最快的地方。
Speaker 211:58 - 12:01
And at each point, that's just a call to the radar API?
Speaker 211:58 - 12:01
所以在每一个节点上,本质上都只是调用一次 Radar API 吗?
Speaker 112:01 - 12:02
Yes. Correct.
Speaker 112:01 - 12:02
对,没错。
Speaker 212:03 - 12:39
And what if I'm sitting here, which I am doing, you know, millions of dollars a year in Stripe transactions, but I actually have no idea what my fraud rate is other than there's, like, that little there's that little thing where it's, it's not even I don't even know if it's necessarily our fraud rate. I think it's our our card chargeback. Anyway, our fraud rate is low enough as marked for me to not care about it. But I guess I don't really know if there's some amount of free trial fraud that I'm not totally understanding right now. So what are the things I should be looking for to to know if I should dig deeper and potentially do some sort of, for example, radar integration?
Speaker 212:03 - 12:39
那如果我现在就坐在这里——而且确实如此——每年通过 Stripe 处理数百万美元的交易,但其实我根本不知道自己的 fraud rate(欺诈率)是多少;除了那个小小的提示,像是有那么个小东西显示着,不过那甚至都不一定是——我甚至不知道那到底算不算我们的 fraud rate。我觉得那更像是我们的 card chargeback(银行卡拒付)之类的。总之,我们的 fraud rate 低到系统给我的标记让我根本没去在意它。但我想,我其实并不真的知道,现在是不是存在某种 free trial fraud(免费试用欺诈)而我还没有完全意识到。所以,我应该关注哪些信号,才能判断自己是否需要进一步深挖,并可能去做某种——比如说——Radar integration(Radar 集成)?
Speaker 112:40 - 13:00
Yeah. So thing one is you can go to your radar dashboard and see if you see anything that looks spurious there. If not, you can also ask the radar assistant, which is in the dashboard. And as you're doing that, you can describe your business model. So you can say like, I have a high marginal cost business, in which case you care more about certain types of fraud than others.
Speaker 112:40 - 13:00
对。第一件事是,你可以去看你的 Radar dashboard,看看里面有没有什么可疑的东西。如果没有,你也可以去问 dashboard 里的 Radar assistant。而且在你这么做的时候,你可以描述一下你的 business model(商业模式)。比如你可以说,我是一个 marginal cost(边际成本)很高的业务,那样的话,你对某些类型的欺诈就会比对另一些更在意。
Speaker 113:01 - 13:17
But you can also just take a stab at integrating UpFunnel and see how it performs. We can certainly you know, share with you based on back testing what we think the big issues are, but the but the fastest way, to get a clean read is is just to integrate.
Speaker 113:01 - 13:17
不过你也可以直接试着集成 UpFunnel,看看它的表现。我们当然可以基于 back testing(回测)跟你分享我们认为的主要问题是什么,但最快能得到清晰判断的方式,还是直接集成。
Speaker 213:17 - 13:34
Got it. So I would so I I don't think that we're integrated right now, so I would just go look at go look at radar and and see if it doesn't doesn't say anything. I'm doing that right now. Just to it'd be really funny if I found that we had a ton of fraud that I didn't know about. We were at 0% fraud.
Speaker 213:17 - 13:34
明白了。所以我会——我觉得我们现在还没集成,那我就先去看看 Radar,看看它是不是没提示什么。我现在就在看。要是我发现我们其实有一大堆自己都不知道的欺诈,那就太好笑了。我们是 0% fraud。
Speaker 213:34 - 13:35
How is that possible?
Speaker 213:34 - 13:35
这怎么可能?
Speaker 113:35 - 13:36
I don't know.
Speaker 113:35 - 13:36
我也不知道。
Speaker 213:38 - 13:43
Point 02% early fraud warnings. Total fraud rate point 2%. So we're, you know, we're doing pretty good. Right?
Speaker 213:38 - 13:43
0.02% 的 early fraud warnings(早期欺诈预警)。总 fraud rate 是 0.2%。所以我们——你知道——做得相当不错,对吧?
Speaker 113:43 - 13:52
That's pretty low. That's pretty low. I mean, yeah, you're a pretty good human. Maybe the fraudsters don't wanna come after you until they hear this episode, and then they'll be like, yeah.
Speaker 113:43 - 13:52
那相当低。那相当低。我是说,嗯,你算是个相当不错的人。也许那些 fraudster(诈骗分子)现在还不想盯上你,除非他们听到这期节目,然后他们可能就会说,嗯,对。
Speaker 213:54 - 14:06
Okay. That's really interesting. Okay. So that's that's fascinating. I I wanna go back a second to the to the AI economy because one of the one of the things you said earlier is fraud is increasing overall on the Internet.
Speaker 213:54 - 14:06
好的。这真的很有意思。好。所以这这很吸引人。我想先回到一下 AI economy(AI 经济)这个话题,因为你刚才提到的一点是,Internet 上的 fraud(欺诈)总体上正在增加。
Speaker 214:07 - 14:51
And that it's also it's increasing because the fraudsters have AI, but that it's also you all and everyone else on the side of good in the in the AI economy also has AI to defend against these sorts of attacks. I think you're getting an interesting window into the arms race that I think is playing out in lots of different areas that have this kind of threat vector. A really simple one is cybersecurity, not just for, like, payments, but for hacking and stuff like that. But there's there's all these other similar types of things where AI makes one part of the process much easier, and then another part of the process has to use AI to compensate to to to catch up. So how is that race going?
Speaker 214:07 - 14:51
而且,增长的原因之一是 fraudster(诈骗分子)也有 AI,但同时你们以及 AI economy(AI 经济)中所有站在正义一方的人,也同样有 AI 来防御这类攻击。我觉得你们正在看到这场 arms race(军备竞赛)一个很有意思的窗口,而我认为这场竞赛正在很多存在这类 threat vector(威胁路径)的领域上演。一个非常简单的例子是 cybersecurity(网络安全),不只是支付,比如说 payments,还包括 hacking(黑客攻击)之类的事。但还有很多其他类似的情况:AI 让流程中的一部分变得容易得多,而流程中的另一部分就不得不用 AI 来弥补、来追赶、来识别问题。所以这场竞赛现在进展如何?
Speaker 214:51 - 14:58
What is that like? What are the early reports that you're that you're seeing and feeling, being in being in a race with AI armed fraudsters?
Speaker 214:51 - 14:58
那是什么感觉?对于一群由 AI 武装起来的 fraudster(诈骗分子),与你们展开竞赛这件事,你目前看到和感受到的早期反馈是什么?
Speaker 114:59 - 15:36
I think the interesting thing about fraudsters is they don't really care about boundaries. They don't care about whether this transaction is processed on Stripe or off Stripe. They don't care about whether this transaction is on you know, Fiat or crypto, whether it's on a card network or a buy now pay later. They're just going to figure out sort of how to work around the system to get through. And so one of the important levers, and I appreciate you calling us the good guys, one of the important levers I think the good guys have for winning is to be comprehensive.
Speaker 114:59 - 15:36
我觉得 fraudster(诈骗分子)有意思的一点在于,他们其实并不在乎边界。他们不在乎这笔 transaction(交易)是在 Stripe 上处理还是不在 Stripe 上处理。他们不在乎这笔 transaction(交易)走的是 Fiat(法币)还是 crypto,是在 card network(银行卡网络)上,还是 buy now pay later(先买后付)上。他们只会想办法摸索出如何绕过系统、如何闯过去。所以,我很感谢你把我们称为好人,而我认为好人一方取胜的一个重要 lever(杠杆),就是要做到 comprehensive(全面)。
Speaker 115:37 - 16:00
And a simple example in our world, sort of Stripe Radar used to only work for cards transactions. And then last year, we added ACH and SEPA, right? So other payment methods. But this year, we've extended to all payment methods, that have disputes. And we added crypto, and we added the radar API.
Speaker 115:37 - 16:00
在我们的世界里,一个简单的例子是,Stripe Radar 过去只适用于 cards transactions(银行卡交易)。然后去年,我们加入了 ACH 和 SEPA,对吧?也就是其他 payment methods(支付方式)。但今年,我们已经把它扩展到所有会产生 disputes(争议/拒付)的 payment methods(支付方式)。我们还加入了 crypto,也加入了 Radar API。
Speaker 116:00 - 16:34
So guess what? You can screen transactions, even ones that aren't processed on Stripe. Right? So you can, process on Worldpay or Adyen or whomever, and through the radar API get the same, fraud signals. Similarly, and we haven't talked about agent ecommerce yet, but as we built out our agent ecommerce suite, one of the new primitives we designed is the shared payment token, which allows agents to safely pass buyer credentials onto merchants for the merchants to process the transaction.
Speaker 116:00 - 16:34
所以你看,会发生什么?你甚至可以筛查 transaction(交易),包括那些不是在 Stripe 上处理的交易。对吧?你可以在 Worldpay 或 Adyen 或其他任何平台上处理,然后通过 Radar API 获得同样的 fraud signals(欺诈信号)。类似地,虽然我们还没谈到 agent ecommerce(agent 电商),但在我们构建 agent ecommerce suite(agent 电商套件)的过程中,我们设计出的一个新 primitive(基础原语)是 shared payment token(共享支付 token),它允许 agent 安全地把买家的凭证传递给 merchant(商户),由 merchant 来处理这笔 transaction(交易)。
Speaker 116:35 - 17:18
And as part of those shared payment tokens, we pass over the radar fraud scores so that the merchant, again, whether or not they're processing on Stripe, can action them appropriately. You know, when it comes to fraud, we really see fraud defenses, fraud mitigation as a public good. And that allows us to invest disproportionately above and beyond the direct value to Stripe because protecting the Internet is important for, growing the Internet economy. So I would say, like, overall, like, yes, fraudsters have AI in their favor. Stripe looks at 2% of global GDP and is growing 34% year on year and sees a broader swath through our multiprocessor solutions like the radar API.
Speaker 116:35 - 17:18
作为这些 shared payment token(共享支付 token)的一部分,我们也会把 Radar 的 fraud scores(欺诈评分)一并传过去,这样 merchant(商户)就可以再次根据这些信息采取适当行动,无论他们是否在 Stripe 上处理交易。你知道,说到 fraud(欺诈),我们确实把 fraud defense(欺诈防御)和 fraud mitigation(欺诈缓解)视为一种 public good(公共产品)。这让我们能够投入远高于 Stripe 直接收益的资源,因为保护 Internet 对于发展 Internet economy(互联网经济)很重要。所以我会说,整体来看,没错,fraudster(诈骗分子)确实拥有 AI 带来的优势。Stripe 覆盖了全球 GDP 的 2%,并且同比增长 34%,同时还通过像 Radar API 这样的 multiprocessor solutions(多处理渠道解决方案)看到更广泛的全局视野。
Speaker 117:18 - 17:58
And so luckily, not only do we have AI on our side just like they do, but we also have data on our side. And the more comprehensive we've gone in our fraud protections, I think, the more we've been able to kind of eke ahead. Now, that's not to say that we're not constantly surprised by, the new creative vectors they come up with, but, you know, you can have an agent every day or every hour taking a look at anomalous patterns on the Stripe network and identifying new vectors that are popping up across processors, across payment methods, across merchants, and burn them down pretty quickly. So, I'm I'm overall bullish, but certainly not complacent.
Speaker 117:18 - 17:58
所以很幸运的是,我们这边不仅像他们一样也有 AI 站在我们这一边,而且我们还有数据站在我们这一边。我认为,我们在 fraud(欺诈)防护上做得越全面,就越能一点点拉开领先优势。当然,这并不是说我们不会不断被他们想出的那些新的、富有创意的攻击向量所惊到;但你完全可以让一个 agent(智能体)每天、甚至每小时查看 Stripe 网络上的异常模式,识别出正在不同 processor、不同 payment method、不同 merchant 之间冒出来的新向量,并且非常快地把它们清除掉。所以,总体上我还是 bullish(乐观)的,但绝不是 complacent(自满)的。
Speaker 217:58 - 18:08
What about other parts of the AI or agent economy? So we've talked a lot about fraud. What are the other things that you see as sort of having this bird's eye view of of what's going on that people might not realize?
Speaker 217:58 - 18:08
那 AI 或 agent economy(智能体经济)的其他部分呢?我们刚才已经聊了很多 fraud。那么,从你这种拥有全局鸟瞰视角的位置来看,还有哪些事情是你观察到、但外界可能并没有意识到的?
Speaker 118:12 - 19:08
I mean, I think, you know, the the AI economy is is broad. I think there's a set of horizontal model providers that have a very interesting view into where is AI being adopted and with what intensity throughout the economy. There's a number of sort of vertical AI solutions. People like to call them rappers, and I say that not condescendingly, just as in like, it's not their models, it's someone else's models, but they have, domain specific data and relationships and context, and they're solving problems in, you know, health care or architecture or whatever, who have a pretty unique view into vertical level adoption of, of AI. But I guess I'd be curious, like, what you have in mind on who has the best the best horizontal view.
Speaker 118:12 - 19:08
我的意思是,我觉得 AI economy(AI 经济)其实很广。我认为,有一类横向的 model provider(模型提供商)对 AI 正在经济中的哪些地方被采用、以及采用强度如何,有非常有意思的观察视角。还有不少垂直型的 AI 解决方案。人们喜欢把它们叫作 wrappers,我这么说并不是贬义,只是说它们的模型不是自己的,而是别人的模型;但它们拥有领域特定的数据、关系和上下文,正在解决 healthcare、architecture 等等领域的问题,因此它们对于 AI 在垂直行业层面的采用情况,也有相当独特的视角。不过我倒是很好奇,你心里觉得谁拥有最好的横向全局视角?
Speaker 219:09 - 19:27
You're asking me? Yeah. Well, I'm I've you know, I wanna know what it what it looks like on the payment side, but I imagine I imagine the the model companies have have the best ones overall because they're that's where all the tokens are going.
Speaker 219:09 - 19:27
你是在问我吗?对。嗯,我是……你知道,我想了解 payment(支付)这一侧看起来是什么样,但我猜,我猜那些 model companies(模型公司)整体上应该拥有最好的视角,因为毕竟所有 token(令牌)最终都流向那里。
Speaker 119:27 - 19:50
Yep. Yep. I think they see a lot of the tokens. I think the AI gateways also have a pretty unique perspective into, you know, who's buying what from whom. You know, as I step back and look at the AI economy from the Stripe vantage point, and and we see, you know, who's buying what from whom, for how much, who's retaining and churning their subscriptions, there's a few a few themes that stand out.
Speaker 119:27 - 19:50
对,对。我觉得它们确实看到了很多 token。我也认为 AI gateways(AI 网关)有相当独特的视角,能看到是谁在向谁买什么。就我从 Stripe 的 vantage point(观察位置)退一步来看整个 AI economy,而且我们能看到谁在向谁买什么、花了多少钱、谁在续订、谁在流失订阅,有几个比较突出的主题。
Speaker 119:50 - 20:23
One is just and I think people feel this intuitively, but not everyone has, like, seen it in the data. These AI companies are just growing from a revenue perspective faster than any previous cohort we've seen. I was looking at the top 100 AI companies on Stripe, and the ones that reach 30,000,000 in ARR get there in about eighteen months, so a year and a half. Wow. And that is like three times faster than the top 100 SaaS companies from, 2018.
Speaker 119:50 - 20:23
其中一个就是——我觉得大家直觉上其实能感觉到这一点,但不是每个人都真的在数据里见过——这些 AI 公司从 revenue(营收)角度看,增长速度比我们过去见过的任何一批公司都更快。我当时在看 Stripe 上排名前 100 的 AI 公司,那些做到 30,000,000 ARR 的公司,大约只需要十八个月,也就是一年半。哇。而这差不多比 2018 年那一批前 100 的 SaaS 公司快了三倍。
Speaker 120:23 - 20:47
By the way, that's the 30,000,000 number. Even if you look like how fast do they make it to 1,000,000 ARR or 5,000,000 ARR, they are scaling orders of magnitude faster than high performing SaaS companies from, less than a decade ago. The second kind of meta trend is this, and you probably feel it as a consumer. I know I do. This, very fast iteration across monetization models.
Speaker 120:23 - 20:47
顺便说一句,这还只是 30,000,000 这个数字。即便你去看它们达到 1,000,000 ARR 或 5,000,000 ARR 的速度,它们的扩张速度也比不到十年前那些表现优秀的 SaaS 公司快了好几个数量级。第二个更宏观的趋势是这个——而且你作为消费者大概也能感觉到,我自己就有这种感觉——那就是 monetization model(变现模式)上的快速迭代。
Speaker 120:47 - 21:14
Right? So traditional SaaS had a lot of, you mentioned the seat had a lot of seat based usage, you know, fixed monthly subscriptions. That made sense for them because they were being used by humans primarily, and their marginal costs were basically zero. But, you know, we've talked about the very real inference costs in the context of fraud. Those also have very real implications for how you price.
Speaker 120:47 - 21:14
对吧?传统 SaaS 很多都是——你刚才提到 seat——大量采用基于 seat 的使用方式,也就是固定的月度订阅。这对它们来说是合理的,因为它们主要是被人类使用,而且它们的 marginal cost(边际成本)基本上接近于零。但是,我们刚才在 fraud 的语境下已经谈过非常真实存在的 inference cost(推理成本)。这些成本同样会对你如何定价产生非常实际的影响。
Speaker 121:14 - 21:41
And so usage based billing, has become very important very quickly. Companies are metering, you know, tokens and API calls, but they're also metering workflows, and they're metering outcomes, kind of like whatever unit best reflects both the customer value and the cost structure. And then they're charging with, like, very high precision, right? They literally want to know every event. How is it rated?
Speaker 121:14 - 21:41
因此,usage based billing(按使用量计费)很快就变得非常重要。公司会计量,你知道,token 和 API 调用,但他们也在计量 workflow(工作流),还会计量 outcome(结果),基本上就是用最能同时反映客户价值和成本结构的单位。然后他们会以非常高的精度来收费,对吧?他们真的想知道每一个 event(事件)。这个事件是如何被 rated(计费定价)的?
Speaker 121:41 - 22:08
And what's all the metadata that sits on that rated event? Way more hybrid monetization models. Right? So I talked about subscriptions, but subscriptions aren't dead. They're just subscriptions with, like, usage overages or like prepaid credits that burn down or real time top ups, which gets to my comment earlier on, this non payment abuse issue, and very kind of multidimensional pricing and monetization.
Speaker 121:41 - 22:08
以及,附着在这个 rated event(已定价事件)上的 metadata(元数据)都有哪些?hybrid monetization model(混合变现模式)也多得多了。对吧?所以我刚才提到了 subscription(订阅),但 subscription 并没有消亡。它们只是变成了带有 usage overage(超额使用收费)的订阅,或者像是会逐步消耗的 prepaid credits(预付额度),或者 real-time top ups(实时充值),这也呼应了我前面提到的那个 non-payment abuse(非支付滥用)问题,以及那种非常多维的 pricing(定价)和 monetization(变现)方式。
Speaker 122:08 - 22:36
Lovable is, a really good example, right? So they they use Stripe Billing for their initial launch, which was fairly simple subscriptions, sort of more traditional pricing and allowed them to monetize very quickly. And then they added a bunch of products like Lovable Cloud or Lovable AI, and they moved with those into usage based billing. Right. So customers are actually charged based on token consumption, but it's a hybrid model, so it's above a certain threshold.
Speaker 122:08 - 22:36
Lovable 就是一个很好的例子,对吧?他们在最初发布时使用的是 Stripe Billing,当时模式还比较简单,就是 subscription,更偏传统的定价方式,这让他们能够非常快地开始变现。后来他们又增加了一系列产品,比如 Lovable Cloud 或 Lovable AI,于是他们也把这些产品转向了 usage based billing。对吧。所以客户实际上是按 token 消耗量来收费的,不过这是一个 hybrid model(混合模式),也就是超过某个阈值之后才这样计费。
Speaker 122:37 - 23:08
And that just, you know, helps companies like Lovable align revenue with usage and value and the actual cost of running the models. And in the limit, you know, we actually have a solution called token billing, which is underlying model costs change a lot, sometimes very quickly. And if you are a wrapper on top of someone else's LLM and your pricing doesn't keep pace, then basically your margins can disappear. Right? So, you know, costs go up and your price stays where it is, then you're in the red.
Speaker 122:37 - 23:08
而这恰好能帮助像 Lovable 这样的公司,让收入与使用量、客户获得的价值,以及运行这些模型的实际成本保持一致。再往极端一点说,我们实际上还有一个叫 token billing 的方案,因为底层 model(模型)成本变化很大,有时甚至变化得非常快。如果你是在别人的 LLM 之上做一层 wrapper(封装),而你的定价又跟不上成本变化,那你的利润率基本上就可能被吃掉。对吧?所以,成本上涨了,而你的价格还停在原地,那你就会开始亏钱。
Speaker 123:08 - 23:23
And so token billing is just, hey, let's in real time track and price to the costs of the underlying tokens with some markup as set by the business. And so, you know, Mesa and Ship and Lovable are are all examples of this kind of kind of infrastructure.
Speaker 123:08 - 23:23
所以 token billing 的意思其实就是,实时追踪底层 token 的成本,并据此定价,再加上企业设定的一定 markup(加价)。所以,Mesa、Ship 和 Lovable 都是这类基础设施的例子。
Speaker 223:24 - 24:20
I love all of these points. I wanna go through them one by one. So a big one that you're talking about is fast iteration across monetization, and it feels like there's this hyper experimentation going on right now where people are like, well, we could charge, we could charge per token, we could charge on a token basis, we could charge per completed request. Like I think Fin, the customer service platform charges per case resolved, which has been a thing in customer service for a long time, but it feels like that could come for a lot more types of software as LNs make it easy to tell, did we actually do the work to get paid? What do you think is the If we're going to pick one, there's a whole range of exploration going on, but if we're gonna pick one new pricing model as the like you know, if if last year's pricing model or last decade's pricing model was just straight up per seat, what do you think is the new standard pricing model that is starting to emerge from the Stripe customers that you see?
Speaker 223:24 - 24:20
我很喜欢这些观点。我想把它们一个一个展开来说。你刚才提到的一个重点,是在 monetization(变现)上的快速迭代;而且现在感觉像是正在发生一种高度实验化的状态,人们会想,好,我们可以收费,我们可以按 token 收费,我们可以基于 token 来收费,我们可以按完成的 request(请求)收费。比如我记得 Fin 这个客户服务平台就是按每个已解决 case(工单/案例)收费,这在客户服务领域已经存在很久了,但现在感觉随着 LNs 让人们更容易判断“我们是否真的完成了该收费的工作”,这种方式可能会扩展到更多软件类型。你怎么看?如果我们要选一个——当然现在整个范围内都在进行各种探索——但如果一定要选一个新的 pricing model(定价模型),就像你知道的,如果去年的定价模型、或者过去十年的定价模型,就是非常直接的 per-seat(按席位)收费,那么你认为现在从你看到的 Stripe 客户中,开始浮现出来的新标准定价模型会是什么?
Speaker 124:20 - 24:32
If you are buying the model, so if you're primarily a model provider, let's say your customer is primarily buying the model, I think you're metering tokens.
Speaker 124:20 - 24:32
如果你买的是 model,也就是说,如果你本质上是一个 model provider(模型提供方),或者说你的客户主要买的是 model,那么我认为你计量的就是 token。
Speaker 224:32 - 24:36
Like in an API, like OpenAI API, Cloud API. Yeah.
Speaker 224:32 - 24:36
就像在 API 里一样,比如 OpenAI API、Cloud API。对。
Speaker 124:36 - 24:53
Exactly. For these vertical solutions, I think in steady state, you are metering outcomes. But it's going to take us some time to get there, not because of the billing infrastructure. Actually, that's totally ready. You mentioned the Fin example.
Speaker 124:36 - 24:53
完全正确。对于这些垂直解决方案(vertical solutions),我认为在稳定状态下,计费会是按 outcomes(结果)来衡量的。但要走到那一步还需要一些时间,不是因为 billing(计费)基础设施不行。事实上,那套东西已经完全准备好了。你刚才提到了 Fin 的例子。
Speaker 124:53 - 25:29
Intercom does the same thing, actually, Stripe Billing. They have an outcome based meter for support tickets resolved. Why do I say for vertical solutions it's going to be on outcomes? Because I think end users are going to want to hold those vertical solutions accountable for outcomes, and they're going to want to know that they have positive ROI on their spend. Now, when you and I buy a model, we feel like we ourselves are accountable for the ROI that we get on the whole plethora, a wide range of applications we might have for that LLM.
Speaker 124:53 - 25:29
Intercom 实际上也是这么做的,借助 Stripe Billing。他们有一个基于 outcomes 的 meter(计量方式),按已解决的 support tickets(支持工单)计费。为什么我说垂直解决方案最终会按 outcomes 计费?因为我认为终端用户会希望这些垂直解决方案对 outcomes 负责,而且他们会想知道自己的支出是否带来了正向 ROI(投资回报)。而当你我去买一个 model(模型)时,我们会觉得,对自己在这个 LLM 上可能开展的那一大堆、非常广泛的应用所获得的 ROI,责任其实在我们自己身上。
Speaker 125:29 - 26:10
But if you're a vertical provider, if you're really focused on, like, solving a concrete need in a given business domain on top of someone else's LLMs, it seems like the core value, it's sort of it's sort of on you, to ensure the ROI is there. And I think outcome based pricing is the most efficient way to hit that. Now, I don't think all outcomes are created equal. And so you could imagine these like, I'm an economist by training, so I'll be a little nerdy, but like you could imagine these like complex objective functions where it's not just did you resolve the sport case, but how complicated was it and with what quality and like, what was your CSAT? And you know, how expensive was the person that you were automating in that task?
Speaker 125:29 - 26:10
但如果你是一个垂直提供商,也就是说,你真正专注于在别人的 LLM 之上,解决某个具体业务领域中的明确需求,那么核心价值似乎就在于:确保 ROI 的存在,这某种程度上就是你的责任。我认为,基于 outcomes 的定价是实现这一点最高效的方式。当然,我不认为所有 outcomes 都是一样的。所以你可以想象出一些这类复杂的 objective functions(目标函数)——我是学经济学出身的,所以会稍微 nerdy(一点书呆子气)——比如,不只是看你有没有解决那个 support case(支持案例/工单),还要看它有多复杂、解决质量如何、CSAT(客户满意度)是多少,以及你所自动化掉的那个岗位在执行这项任务时原本有多昂贵。
Speaker 126:10 - 26:29
And so, that's why I say in the limit, like I think it'll take time for us to be very crisp on the outcomes we care about and how we measure those outcomes, and those outcomes will be multidimensional. But I just have a hard time imagining, you know, a year from now, most vertical providers are literally charging on tokens.
Speaker 126:10 - 26:29
所以,这就是为什么我说从极限上看,我认为还需要时间,我们才能非常明确地界定我们在乎哪些 outcomes,以及如何衡量这些 outcomes;而且这些 outcomes 会是多维的。但我很难想象,一年之后,大多数垂直提供商还真的是按 token 计费。
Speaker 226:29 - 26:58
That's really interesting. I'm I'm very curious to see that because what what I felt so I think you can see this a little bit in in the example in the level of example you gave, but also in the Claude and Tradjeet BT examples and and some of the pricing that we've ended up doing is it's per seat. It's per user with overages. Because we're we've started to exist in this world where we used to charge per seat, so people know how to model it. It's pretty easy to kind of figure out how much I'm gonna pay.
Speaker 226:29 - 26:58
这很有意思。我非常想看看这会怎么发展,因为我自己的感受是——我觉得你给的那个例子里其实已经能稍微看出来一点,Claude 和 Tradjeet BT 的例子里也是,还有我们最后采用的一些定价方式——那就是按 seat(席位)收费。也就是按每个 user(用户)收费,再加上 overages(超额费用)。因为我们已经开始处在这样一个世界里:过去我们是按 seat 收费的,所以大家知道怎么建模,基本也很容易估算自己要付多少钱。
Speaker 226:58 - 27:23
But software used to be free to run, and now it's not. And so we have to cover our butts basically and protect our margin by adding the, like, the overage so that customers basically know what they're gonna pay unless there's some sort of speckle special circumstance. Do you see that where do you see that fitting in the examples that you gave? And and and I guess I guess you would you would say, eventually, that might go away. I'm curious why.
Speaker 226:58 - 27:23
但以前 software(软件)的运行成本几乎可以视为免费,现在不是了。所以我们基本上得 cover our butts(保护自己),通过加上 overage 之类的机制来保护利润率,这样客户通常就能知道自己要付多少钱,除非出现某种特殊情况。你怎么看这个模式?在你刚才举的那些例子里,你觉得它该放在什么位置?而且我猜你可能会说,最终这种做法也许会消失。我很好奇为什么。
Speaker 127:24 - 28:03
So I don't think the the charging for use or charging for overages will go away for, like, most of, like, the model providers. If anything, I think that will dominate, and the seat based billing will go away. Like, you know, we can go back to, like, the Fin or Intercom example. Like you like you well, you and I would think it's silly to charge based on number of customer service reps that were using the tool because, obviously, a lot of what the tool is doing is automating customer service reps. In today's world, it isn't perceived as silly to do seat based usage of developer tools, But I think it's a fair question since basically November or December to say, wait, why isn't that silly?
Speaker 127:24 - 28:03
所以我不认为,按使用量收费,或者收取 overages,会在大多数 model providers(模型提供商)那里消失。如果有变化,我反而觉得这会成为主导,而基于 seat 的计费会消失。比如我们可以回到 Fin 或 Intercom 的例子。你我都会觉得,按使用这个工具的 customer service reps(客服代表)人数来收费是很荒谬的,因为很显然,这个工具所做的大量工作,本来就是在自动化 customer service reps。在今天的世界里,对 developer tools(开发者工具)按 seat 来收费,还没有被普遍认为荒谬;但我觉得,自从大概去年 11 月或 12 月以来,提出这样一个问题其实很合理:等等,为什么这不算荒谬呢?
Speaker 128:03 - 28:54
That seems a little silly because, you know, if what these agents are doing is, you know, making every developer, let's say, you know, I don't know, 10x more productive at some point, then don't you need onetenth the developers? And why would you want your revenue pegged to the count of developer as sort of the base price? So I suspect that we will see seat based disappear now in the enterprise context. Now I think it's quite different in the consumer individual context. I think, you know, with the exception of maybe some nerds on the call, most people are actually pretty uncomfortable as individual consumers, with anything but a fixed fee monthly, maybe with some overages if they wanna spend like crazy.
Speaker 128:03 - 28:54
这看起来确实有点荒谬,因为如果这些 agents(智能体)所做的事情,是让每个 developer(开发者)——比如说,我随便举个数——在某个时间点 productivity(生产率)提升 10x,那你岂不是只需要原来十分之一的开发者?如果这样的话,你为什么还想让自己的收入锚定在 developer 数量这种基础价格之上?所以我怀疑,在 enterprise(企业)场景里,基于 seat 的收费会消失。当然,我认为在 consumer(消费者)个人场景里情况很不一样。我觉得,除了电话会议里也许有一些 nerds(技术宅)之外,大多数个人消费者其实都非常不喜欢固定月费之外的收费方式;也许如果他们真的想疯狂多用一点,可以接受一些 overages。
Speaker 128:54 - 29:03
But in businesses, I just I would be super surprised if six months from now, we have half of the seat based licenses that we have today.
Speaker 128:54 - 29:03
但在企业里,如果六个月后,我们基于席位(seat-based)的 licenses 数量只剩下今天的一半,我也完全不会意外,反而会非常震惊如果不是这样。
Speaker 229:03 - 29:17
That is fascinating. Well, we will have to have you on again to talk about that one. I'm so curious, Azian. I would love to see more Stripe data coming out about that. One other thing that you brought up before this was you're you're also seeing these companies scale faster.
Speaker 229:03 - 29:17
这很有意思。看来我们得再请你来聊聊这个话题了。我太好奇了,Azian。我很想看到 Stripe 在这方面释放更多数据。你刚才还提到另一点,就是你们也看到这些公司扩张得更快。
Speaker 229:17 - 29:51
Like you said that the time to get to 30,000,000 in in ARR is like eighteen months, which is significantly faster than any other cohort of companies you've seen. I'm curious where is that coming from? Like presumably the spend or the growth from their customers is coming from somewhere. Either it's spend that people weren't spending before it was like, you know, on their on a company balance sheet just waiting to be deployed, or they're pulling it from an another provider and then going really rapidly into these new ones. Do you have a sense for what's happening here?
Speaker 229:17 - 29:51
就像你说的,它们做到 30,000,000 ARR(年度经常性收入)只需要大约十八个月,这明显比你见过的任何其他公司 cohort(群体)都要快。我很好奇,这到底是从哪里来的?按理说,它们客户的支出或增长总得有个来源。要么是以前人们还没花出去的钱,比如放在公司资产负债表上、等着被部署的资金;要么就是从另一个 provider(提供商)那里转移出来,然后非常迅速地流向这些新公司。你对这里到底发生了什么,有什么判断吗?
Speaker 229:51 - 29:55
Where's all the why are they growing so much faster, where's all the money coming from?
Speaker 229:51 - 29:55
它们为什么增长得这么快?这些钱到底都是从哪里来的?
Speaker 129:55 - 30:37
Yeah. So I think a lot of the AI growth that we've seen is actually like net new spend being pumped into the economy. I think it has largely not been a substitute for traditional SaaS or for sort of headcount OpEx because it's been experimental, because people are still learning, because organizations are somewhat slow to drop existing licenses, often because they're contracted into longer durations, but also because AI was starting not literally at zero, but at near zero. So, like, there weren't other AI companies to to sort of,
Speaker 129:55 - 30:37
是的。所以我认为,我们看到的很多 AI 增长,实际上都是被注入经济中的净新增支出(net new spend)。我觉得这在很大程度上还没有替代传统 SaaS,或者说替代 headcount OpEx(人员运营支出),因为它一直处于实验阶段,因为人们还在学习,因为组织在放弃现有 licenses 这件事上通常动作比较慢,往往也是因为它们签了更长期的合同;另外还因为 AI 的起点虽然不是真的从零开始,但也接近于零。所以,比如说,之前并不存在其他 AI 公司去……
Speaker 230:38 - 30:39
Soak up the spending.
Speaker 230:38 - 30:39
吸收这些支出。
Speaker 130:39 - 30:40
Market share from.
Speaker 130:39 - 30:40
抢占其中的 market share(市场份额)。
Speaker 230:40 - 30:40
Yeah.
Speaker 230:40 - 30:40
对。
Speaker 130:41 - 31:12
I would say now going forward, I I I don't have a crystal ball. I can't tell you exactly what the dynamics will be, but I expect that some of it will be a substitute away from traditional SaaS just and and by the way, I I don't say that in a, you know, old company and new company sense. Like, some SaaS companies are doing an amazing job reinventing themselves as AI first. And so you will have AI arms of traditional SaaS companies that are eating some of the revenue from the traditional version of the same company. But some will come from SaaS.
Speaker 130:41 - 31:12
但我会说,接下来往后看,我没有 crystal ball(水晶球),没法准确告诉你这些动态究竟会怎么演变,不过我预计其中一部分会替代传统 SaaS。当然,顺便说一句,我这么讲并不是那种“老公司被新公司取代”的意思。比如,有些 SaaS 公司在把自己重塑为 AI first(AI 优先)这件事上做得非常出色。所以你会看到,传统 SaaS 公司内部的 AI 业务线,会蚕食同一家公司传统版本的一部分收入。但其中确实会有一部分来自 SaaS。
Speaker 131:12 - 31:53
I think some will come from headcount OpEx. Like I is it is very hard to believe that companies will start spending single digit, sometimes double digit percentages of their headcount OpEx in LLMs and not step back and say, okay, well, my headcount op my headcount cost just changed. It used to cost me $300,000 for an engineer, and now it costs me $330,000 for an engineer because 300 of them is salary and equity and 30 ks is LMs. And so I better reason about my budget on the plus 10% basis and make head count decisions accordingly. And by the way, ROI decisions as well.
Speaker 131:12 - 31:53
我认为其中一部分会来自 headcount OpEx(人力运营支出)。很难相信公司会开始把个位数、甚至有时两位数比例的 headcount OpEx 花在 LLMs(大语言模型)上,却不退一步想一想:好吧,我的人力成本其实已经变了。以前一个 engineer(工程师)要花我 $300,000,现在一个 engineer 要花我 $330,000,因为其中 300k 是 salary 和 equity,另外 30k 是 LLMs。既然如此,我最好按加了 10% 的口径来考虑预算,并据此做 headcount 决策。顺便说一句,ROI(投资回报率)决策也是如此。
Speaker 131:54 - 32:22
And then some of what we are seeing is definitely substitution now across AI providers. So I was looking at, retention rates for AI companies, and what you see is actually the within the domain. So, for example, within AI dev tools or AI coding tools or providers. The retention rate, both B2C and B2B, is higher than it was for SaaS.
Speaker 131:54 - 32:22
然后,我们现在看到的另一部分现象,确实是 AI providers(AI 提供商)之间的替代。我在看 AI 公司的 retention rates(留存率),你会发现实际上在同一细分领域内部——比如 AI dev tools(AI 开发工具)、AI coding tools(AI 编码工具)或相关 providers——无论是 B2C 还是 B2B,它们的留存率都比 SaaS 时代更高。
Speaker 232:23 - 32:28
Interesting. I'm shocked. But but Within within Okay. Got it. Got Yeah.
Speaker 232:23 - 32:28
有意思。我很震惊。不过——是在这个内部、同一领域内部?明白了,明白了。对。
Speaker 132:29 - 32:36
For the individual provider, it's slightly lower. Yeah. Right? Which is intuitive. Like, once you start which is ex post intuitive.
Speaker 132:29 - 32:36
但对单个 provider 来说,留存率会略低一点。对吧?这很符合直觉。就是,一旦你开始——这是事后看很直观的。
Speaker 132:36 - 32:56
Although ex ante, I actually literally didn't know and needed to query the data. But but ex post is intuitive. It's like once you start using, you know, AI dev tool, like a coding assistant, like, you love it, you're not gonna stop using it, but you very well may iterate across providers as, you know, models vary in their quality or
Speaker 132:36 - 32:56
虽然事前来说,我当时其实真的不知道,必须去查数据才行。但事后看又很直观:一旦你开始用 AI dev tool,比如 coding assistant(编程助手),你会很喜欢它,不太会停止使用;但你很可能会在不同 providers 之间反复切换,因为 models(模型)的质量会变化,或者——
Speaker 232:56 - 33:12
Anytime a new model comes out, you're just like, I gotta try this. And there's a there's a high percentage of, curious travelers basically just hopping from one thing to the next within a category, but they're definitely gonna stick in using a tool like that for a long time.
Speaker 232:56 - 33:12
每当有新模型出来时,你就会想:我得试试这个。于是就会有很高比例的“好奇型旅行者”,基本上是在同一类别里从一个产品跳到另一个产品;但他们肯定会长期持续使用这类工具。
Speaker 133:12 - 33:40
Yes. Exactly. And so I would say, like, a lot of the sort of crazy fast AI growth we've seen is, like, net new dollars spent. But I think businesses are gonna start to reason about that as a substitute for SaaS or that as a substitute for headcount OpEx or that as a substitute for other AI companies. And it will be less purely additive in the go forward year than it was in the past year when, you know, people were really just starting to ramp up on on their AI spend.
Speaker 133:12 - 33:40
对,完全是这样。所以我会说,我们看到的很多这种疯狂快速的 AI 增长,本质上是净新增支出。但我认为,企业会开始把这笔钱视为对 SaaS 的替代,或对 headcount OpEx 的替代,或对其他 AI 公司的替代。和过去一年相比,在接下来的一年里,它就不会像之前那样是纯粹的增量了;因为之前人们其实只是刚开始逐步加大他们的 AI 支出。
Speaker 233:40 - 33:57
Does that imply anything to you about the valuations of current, you know, hot AI companies? Like, let's let's let's accept from this, like, the OpenAI's and Anthropocs of the World, but, like, the in the 30,000,000 cohort from this cohort and the coming and the coming up ones, does that say anything to you about their prospects or their growth rates or their valuation?
Speaker 233:40 - 33:57
这会让你对当前这些热门 AI 公司的估值有什么判断吗?比如我们先把 OpenAI 和 Anthropic 这类公司排除掉;但对于这个群体里 ARR(年度经常性收入)在 30,000,000 这一档的公司,以及接下来正在往上走的那些公司,这是否说明了它们的前景、增长率,或者估值会是什么样子?
Speaker 133:59 - 34:47
Well, so actually, if you look at, like, the top 100 on Stripe, like, are little pockets of twos and threes that are directly competitive, but a bunch of them are, like, solving totally disjoint vertical problems with no competitor yet in the space. And so I do think there's, like, enough blue ocean sort of vertical solutions that I think overall AI valuations are probably okay. I think there's like a couple of crowded spaces that you and I could intuitively reason about, where, you know, you might think it would be a little frothy. And by the way, see this that that's sort of the macro view, but you see this in the micro view too. Like if you look at, sort of the the sales led growth contracts, right?
Speaker 133:59 - 34:47
嗯,实际上,如果你看 Stripe 上排名前 100 的公司,会发现其中有一些两三家扎堆、彼此直接竞争的小板块,但也有很多是在解决彼此完全不重叠的垂直领域问题,而这些领域里甚至还没有竞争对手。所以我确实觉得,这类“blue ocean(蓝海)”式的垂直解决方案还足够多,因此整体来看,AI 估值大概率还是合理的。当然,也有少数比较拥挤的赛道,你我凭直觉都能判断出来,在那些地方你可能会觉得市场有点过热。顺便说一句,这算是宏观视角,但在微观层面你也能看到同样的现象。比如说,如果你去看那种 sales-led growth(销售驱动增长)的合同,对吧?
Speaker 134:47 - 35:44
Like when there's a new, you know, when you are the first AI dev tool, you basically charge people sticker and you do very little negotiations and enterprise pay you sticker and whatever. And then all of a sudden you have to have these, like, much more complex. I mean, you hire a bunch of sellers and you have your, you know, CPQ configure price quote system, and you have this nuanced billing because you're competing against two or three other providers who have like, you know, competitive looking monetization models, and you're reacting to that. And so on the micro, you start to see some of those, some of those competitive reactions creeping in as well. But I I think the I think the overarching kind of next year will continue to have a bunch of sort of blue ocean vertical stuff that didn't exist before, but there will be some pockets where it's it's a little more it's a little more heated.
Speaker 134:47 - 35:44
比如说,当一个新的产品出现、而你又是第一个 AI dev tool(AI 开发工具)时,你基本上可以按标价收费,几乎不用怎么谈判,enterprise(企业客户)也会按标价付钱,差不多就是这样。但突然之间,事情就会变得复杂得多。我的意思是,你会雇一大批销售人员,搭建自己的 CPQ(configure price quote,配置-定价-报价)系统,还要做这种更细致的计费设计,因为你开始和另外两三家提供商竞争,而他们也都有看起来很有竞争力的 monetization(变现)模式,你也得据此做出反应。所以在微观层面,你也开始看到一些这类竞争反应在慢慢出现。不过我觉得,从整体上看,明年仍然会持续出现许多以前不存在的“blue ocean(蓝海)”垂直机会,但也会有一些局部领域竞争变得更激烈一些。
Speaker 235:44 - 35:54
Fascinating. I feel like I'm learning so much. This is amazing. I wanna go into Stripe. Instead of talking about the AI economy, I wanna go into Stripe a little bit.
Speaker 235:44 - 35:54
太有意思了。我感觉自己学到了很多。这太棒了。我想聊聊 Stripe。比起继续谈 AI economy(AI 经济),我想更具体地聊一点 Stripe。
Speaker 235:55 - 36:29
Specifically, is, you know, it serves you serve developers. And you're built for a world where humans are the ones buying and selling and also humans humans are the ones making the software. Now agents are buyers, they're sellers, they're builders and you're so you have to sort of serve agents. And I'm curious how that has changed how you think about the products that you offer and the, you know, moving maybe from just thinking about developer experience to agent experience, all that kind of stuff.
Speaker 235:55 - 36:29
更具体地说,Stripe 服务的是 developers(开发者)。而你们原本是为这样一个世界构建的:由 humans(人类)来买卖,也是由 humans(人类)来编写软件。现在 agents(智能体)既是买家,也是卖家,还是构建者,所以你们某种程度上也必须去服务 agents。我很好奇,这种变化如何影响了你们对所提供产品的思考;以及你们是否正在从过去只关注 developer experience(开发者体验),转向也去思考 agent experience(智能体体验)之类的问题。
Speaker 136:30 - 36:36
Do you want to start with agent experience or agent? I think they're both they're they're kinda different. Yeah. But they're they're both really interesting.
Speaker 136:30 - 36:36
你想先从 agent experience(智能体体验)讲起,还是先讲 agent?我觉得这两者都算是不同的话题。对。但它们都非常有意思。
Speaker 236:36 - 36:38
Which one are you most excited to talk about?
Speaker 236:36 - 36:38
你最想先聊哪一个?
Speaker 136:40 - 36:43
Maybe agent experience, and then we can work backwards to agent.com.
Speaker 136:40 - 36:43
也许先聊 agent experience,然后我们再倒回去讲 agent.com。
Speaker 236:43 - 36:45
Yeah. Let's talk about agent experience.
Speaker 236:43 - 36:45
好,那我们就聊聊 agent experience(智能体体验)。
Speaker 136:45 - 37:02
Okay. So, you know, the the developer story well, so the the whole idea of developer experience is changing. And historically, when I said developer experience, you thought, like, hey. Making it easier for a human engineer who's at a keyboard. Right?
Speaker 136:45 - 37:02
好。所以,你很了解 developer story(开发者叙事),因此 developer experience(开发者体验)的整个理念正在发生变化。从历史上看,当我说 developer experience 时,你会想到的是:要让一个坐在键盘前的人类工程师更容易完成工作,对吧?
Speaker 137:02 - 37:17
So, like, you need clearer APIs, and you need better docs, and you need less setup work. And all of that still matters. Like, it's not going anywhere. But I think the the developer is now sort of a a broader swath of persona. Right?
Speaker 137:02 - 37:17
所以,比如你需要更清晰的 API(应用程序编程接口),需要更好的 docs(文档),还需要更少的 setup(配置)工作。而这些依然都很重要,不会消失。但我认为,现在的 developer 已经变成了一个更宽泛的人群画像,对吧?
Speaker 137:17 - 37:57
It could be a nontechnical founder who's just in, you know, Vercel or Replit, like describing an app in plain language. Or it could be, a coding assistant who's, like, scaffolding and integration, or it could be an agent who's like out trying to provision infrastructure on a human's behalf. And so I think it's less about just like, okay, how do we help a developer, human developer write code and more about how do we have a coherent and trustworthy product experience sort of end to end that acknowledges that at some moments, the actor is a human. At some moments, the actor is an agent. And at some moments, the the actor is a human working through an agent.
Speaker 137:17 - 37:57
它可以是一个不懂技术的 founder(创始人),只是在 Vercel 或 Replit 里用自然语言描述一个 app(应用)。也可以是一个 coding assistant(编程助手),在做 scaffolding(脚手架搭建)和 integration(集成);还可以是一个 agent(智能体),代表人类去尝试 provision infrastructure(配置基础设施)。所以我觉得,重点已经不只是“我们如何帮助一个 developer、一个人类 developer 去写代码”,而是“我们如何提供一种连贯且值得信赖的端到端 product experience(产品体验)”,并且承认:有些时刻,执行者是人类;有些时刻,执行者是 agent;还有些时刻,执行者是通过 agent 工作的人类。
Speaker 137:57 - 38:19
And so you see this shift in in some really concrete ways. Very simple example, LLM traffic to Stripe Docs is up 10x year over year, and that's just a useful signal that machines are becoming users of developer infrastructure too, including Stripe's developer infrastructure. What about human
Speaker 137:57 - 38:19
所以,你能以一些非常具体的方式看到这种转变。举个很简单的例子,LLM(大语言模型)流量访问 Stripe Docs 的规模同比增长了 10 倍,而这只是一个很有用的信号,说明机器也正在成为 developer infrastructure(开发者基础设施)的用户,包括 Stripe 的 developer infrastructure。那人类对
Speaker 238:19 - 38:21
use of Stripe Docs?
Speaker 238:19 - 38:21
Stripe Docs 的使用呢?
Speaker 138:21 - 38:36
So human use of Stripe Docs is actually like flat to climbing. It's not like a straight substitute. I think there is just, like, more developer act developer activity happening, and LLMs are growing dramatically within that share.
Speaker 138:21 - 38:36
人类对 Stripe Docs 的使用其实基本持平到略有上升。它并不是一种直接替代关系。我认为,整体上只是发生了更多 developer activity(开发活动),而在这其中,LLM 所占的份额正在显著增长。
Speaker 238:37 - 38:38
That makes sense. Cool.
Speaker 238:37 - 38:38
这很合理。不错。
Speaker 138:39 - 38:50
I would also say the humans continue to check on the docs to sanity check what the agent is coming up with because your payments integration is actually, like, a pretty big decision that you're making.
Speaker 138:39 - 38:50
我还想说,人类还是会继续查看 docs,来 sanity check(合理性校验)一下 agent 给出的结果,因为 payments integration(支付集成)实际上是一个相当重大的决策。
Speaker 238:50 - 38:55
I will say, yeah, better humans than I are sanity checking, but I'm glad that someone is sanity checking.
Speaker 238:50 - 38:55
我会说,没错,比我更靠谱的人在做 sanity check(合理性校验),但我很高兴至少确实有人在做 sanity check。
Speaker 138:55 - 38:57
Are you YOLO ing it?
Speaker 138:55 - 38:57
你这是在 YOLO 式地搞吗?
Speaker 238:57 - 38:59
I'm YOLO vibe coding my payment infrastructure.
Speaker 238:57 - 38:59
我在用 YOLO 式的 vibe coding 来搭我的支付基础设施。
Speaker 138:59 - 39:17
Okay. Okay. Amazing. So maybe maybe you're YOLO vibe coding, but even if you're vibe coding, there's still an important step around provisioning like your your modern software stack, and that is still very manual, right? So like you as a human are still creating accounts across multiple services.
Speaker 138:59 - 39:17
好。好。太惊人了。所以也许、也许你是在 YOLO 式地 vibe coding,但即便你在 vibe coding,围绕 provision(配置开通)你的现代软件栈这件事,仍然有一个很重要的步骤,而且这一步依然非常手动,对吧?也就是说,作为人类的你,还是得在多个服务之间自己创建账号。
Speaker 139:18 - 39:34
You're managing credentials. You're clicking through to do a lot of setup. You're probably bouncing between dashboards. And so like the coding is getting easier a lot faster than the setup is getting easier. And that's actually the idea of Stripe projects, which we launched, I don't know, maybe two weeks ago.
Speaker 139:18 - 39:34
你要管理 credentials(凭证)。你要一路点来点去完成大量设置。你很可能还得在不同 dashboard(控制台)之间来回切换。所以,coding 变容易的速度,比 setup 变容易的速度快得多。而这其实就是 Stripe projects 的思路;我们刚发布了它,我不知道,也许是两周前。
Speaker 139:34 - 39:35
It's basically
Speaker 139:34 - 39:35
它基本上就是——
Speaker 239:35 - 39:37
like Amazing. Tell people what Okay.
Speaker 239:35 - 39:37
就像——太棒了。跟大家说说这是什么。好。
Speaker 139:37 - 39:39
If you want in, let me know. We can we can I
Speaker 139:37 - 39:39
如果你想加入,就告诉我。我们可以,我们可以——
Speaker 239:39 - 39:40
want I want it? I absolutely want it.
Speaker 239:39 - 39:40
想要吗?我当然想要。
Speaker 139:40 - 39:56
You're in. Yeah. Check. I won't Slack right now, but I'll Slack right after this and get you in. But but basically, the idea of Stripe projects for those who haven't explored it is just like you or your agents can go create and manage parts of your soft software stack right from the command line.
Speaker 139:40 - 39:56
你加入了。对,没问题。我现在先不发 Slack,但我一结束就会在 Slack 上联系你,把你拉进来。不过基本上,对于那些还没了解过 Stripe projects 的人来说,它的核心思路就是:你或者你的 agents(智能代理)可以直接在命令行里创建并管理你的软件栈中的一部分。
Speaker 139:56 - 40:29
And so, you know, resources are provisioned in accounts you own and credentials sync back to your environment and so on. And one of the things that stood out, besides your enthusiasm for it, which I appreciate, is just how sort of overwhelming the interest has been in general from the ecosystem. So we launched with like, Vercel and Supabase, PostHogs there, Neon, Runloop. There's a bunch of great companies involved. But then immediately after launch, over 100 other great companies reached out wanting to join, which I just think reinforces that, like, the friction is real.
Speaker 139:56 - 40:29
所以,你知道,资源会被 provision(配置/开通)到你自己拥有的账户里,credentials(凭证)也会同步回你的环境,诸如此类。除了你对它的热情——这点我很感激——还有一件特别突出的事,就是整个 ecosystem(生态系统)对它的兴趣之强烈,某种程度上都让人有点应接不暇。我们发布时有 Vercel、Supabase、PostHog、Neon、Runloop 这些公司参与,已经有一批很棒的公司了。但产品一上线,立刻又有 100 多家其他优秀公司联系我们,想要加入。我觉得这恰恰进一步说明了一点:这种 friction(摩擦/阻力)是真实存在的。
Speaker 140:29 - 40:54
And you talked earlier about, like, you know, some things get easier with AI, but there's, like, some counter effect. You know, I think coding gets easier, but, like, code reviews become more burdensome because who's reviewing all the AI code? This is another example of, like, building gets easier, but you still kinda have to, like, provision everything. And so that's just an example of of how we're building for this world of, like, the developer is no longer, just a human.
Speaker 140:29 - 40:54
你之前也提到过,AI 会让某些事情变得更容易,但同时也会带来某种反作用。比如,我认为写代码变容易了,但 code review(代码审查)反而更有负担了,因为谁来审那些 AI 生成的代码呢?这也是另一个类似的例子:构建变容易了,但你仍然得去 provision(配置/开通)所有东西。所以这就是一个例子,说明我们正在为这样一个世界做建设:developer(开发者)不再只是一个人类。
Speaker 240:54 - 40:56
Got it. And then tell me about agentic commerce.
Speaker 240:54 - 40:56
明白了。那你再跟我讲讲 agentic commerce(代理式商业)吧。
Speaker 140:57 - 41:18
Okay. So agentic commerce is a bit of an overloaded term. And I think a mistake that people make with agentic commerce is they jump straight to kind of the most extreme version. So they hear the phrase and they think like, Some system that knows everything about me and decides what I need and goes off and buys it for me. And then they're underwhelmed with the world we're actually in.
Speaker 140:57 - 41:18
好的。agentic commerce 这个词其实有点被过度泛化了。我觉得人们在谈 agentic commerce 时常犯的一个错误,就是会直接跳到那种最极端的版本。他们一听到这个说法,就会想到某种完全了解我的系统,替我决定我需要什么,然后自己跑去替我买下来。结果当他们回头看我们当下真实所处的世界时,就会觉得有点失望。
Speaker 141:18 - 41:42
Like maybe we get to that extreme eventually in some form, but we're not there yet. I prefer to think about it as a spectrum. And, you know, I think that the the economic infrastructure you need is actually pretty similar no matter where you are in the spectrum. But the spectrum also like brings some realism to it. So at the at the first level, which is like AI is just, removing friction from the Internet we already have.
Speaker 141:18 - 41:42
也许我们最终会以某种形式走到那个极端,但现在还没到。我更愿意把它看作一个光谱。你知道,我认为无论你处在这个光谱的哪个位置,你所需要的 economic infrastructure(经济基础设施)其实都相当类似。但这个光谱的视角也会让事情显得更现实一些。比如在第一个层级,AI 只是把我们现有互联网中的 friction(摩擦/阻力)去掉一部分。
Speaker 141:42 - 41:59
Right? So it helps you research and compare options and fill out some forms and narrow down your choices. But you, the human, are still making the decision. We're just making you know, the agent is just making that that experience easier. Then you move to, like, okay, search is descriptive.
Speaker 141:42 - 41:59
对吧?它会帮你做研究、比较选项、填写一些表单、缩小选择范围。但真正做决定的仍然是你这个人类。我们只是让——你知道——agent(智能代理)把这个体验变得更容易一些。然后你会进入下一个阶段,比如说,好,search(搜索)是描述性的。
Speaker 141:59 - 42:15
Right? No more, like, blunt keywords and filters and such. It's like, I got little kids. Like, I need a summer camp for my kids in this budget on these dates with this driving radius. And that's already a better commerce experience than, like, search plus filter, which is, like, you know, knowably blunt.
Speaker 141:59 - 42:15
对吧?不再是那种生硬的关键词、筛选器之类的东西了。更像是,我有小孩。我需要在这个预算内、这些日期里、这个车程半径范围内,给孩子找一个 summer camp。这样带来的 commerce(商业)体验,已经比 search 加 filter 那一套更好了;后者你也知道,确实就是很生硬。
Speaker 142:16 - 42:40
Then you get to sort of real delegation. And I think this is what most people would consider, like the minimum viable bar for saying agentic commerce. So, like, I give some constraints. Like, I give some for camps, like some budget, some dates, some category, maybe a few preferences, and then the system goes and makes the purchases on my behalf. But then there's the further out version, which is like the ambient version, right?
Speaker 142:16 - 42:40
然后你就会走到某种真正的 delegation(委托)阶段。我觉得,大多数人会把这看作是称之为 agentic commerce 的最低可行门槛。比如说,我给出一些约束条件。拿 camps 来说,我给一个预算、一些日期、一个类别,也许再加几个偏好,然后系统就代表我去完成购买。但再往前一步,还有一个更远期的版本,也就是那种 ambient(环境式、无感式)的版本,对吧?
Speaker 142:40 - 43:19
I don't prompt anything, and the system knows me and it knows my seasonal needs and it knows, you know, summer camp planning is happening. That would be music to my ears and sort of that's the most futuristic thing. I think the point is that no matter where you are on that, like, what the Internet, needs for economic infrastructure starts to change. Right? Even the earlier stages force a redesign of payments infrastructure in particular, because the today model, the old model, again, humans sitting in front of browser, creating account, choosing plan, filling out the forms, clicking purchase, entering card details.
Speaker 142:40 - 43:19
我根本不需要 prompt(提示)任何东西,系统了解我,也了解我季节性的需求,它知道,summer camp 的规划期到了。那对我来说简直太理想了,也可以说那是最具未来感的形态。我想关键在于,不管你处在哪个阶段,Internet 所需要的 economic infrastructure(经济基础设施)都会开始变化。对吧?哪怕只是较早期的阶段,也会迫使支付基础设施,尤其是 payments infrastructure(支付基础设施),重新设计。因为今天的模式、旧的模式,归根结底还是:人坐在浏览器前,创建账户,选择套餐,填写表单,点击购买,输入银行卡信息。
Speaker 143:20 - 43:41
Not all those steps are happening anymore. And so, you know, I I think there's sort of two worlds that I reason about preparing for. One is agent assisted buying. So I'm ultimately in charge, but the discovery and checkout and payment happen inside AI interfaces instead of on a merchant website. So I'm not going to Nordstrom.
Speaker 143:20 - 43:41
这些步骤已经不再都会发生了。所以,我觉得我在思考和准备的,其实是两种世界。第一种是 agent assisted buying(agent 辅助购买)。最终还是由我负责,但 discovery(发现商品)、checkout(结账)和 payment(支付)都发生在 AI 界面里,而不是在 merchant(商家)网站上。所以我不会去 Nordstrom。
Speaker 143:41 - 44:15
I'm buying within Gemini or ChatGPT or, you know, Meta, like Facebook ads, whatever. And what's challenging here is two things. One, the AI agent needs to be able to understand the merchants' products and prices and checkout flow so that they can act on behalf of the consumer. And two, as you and I talked about a bit at the top, trust can break down, right? As a consumer, I don't want to hand off my credentials to an agent.
Speaker 143:41 - 44:15
我会直接在 Gemini 或 ChatGPT 里购买,或者比如在 Meta、Facebook ads 之类的环境里购买。而这里有两个难点。第一,AI agent 需要能够理解商家的产品、价格和 checkout flow(结账流程),这样它才能代表消费者采取行动。第二,正如我们前面稍微谈到的,trust(信任)是可能失效的,对吧?作为消费者,我不想把自己的 credentials(登录凭证)交给一个 agent。
Speaker 144:15 - 44:40
As a merchant, I don't wanna let just every, like, bot through. I wanna know, is it like a good bot acting on behalf of a legitimate customer? So, the agent to commerce protocol, which we we co created with OpenAI, is just the shared technical language between AI systems and businesses. And, it shows up across a lot of surfaces. Again, we built it with OpenAI, but, Microsoft Copilot uses it.
Speaker 144:15 - 44:40
作为 merchant,我也不想让所有 bot(机器人)都随便进来。我想知道,这是不是一个 good bot,在代表一个合法客户行事?所以,agent to commerce protocol,也就是我们和 OpenAI 共同创建的协议,本质上就是 AI 系统与企业之间共享的技术语言。而且,它会出现在很多不同的入口和界面上。再说一次,这是我们和 OpenAI 一起做的,但 Microsoft Copilot 也在用它。
Speaker 144:40 - 45:19
Meta's in ad shopping experience uses it. And how it works is is basically the merchant only has to integrate once with Stripe for their product catalogs, their prices, their checkout flows, and then they can literally from the dashboard turn themselves on through a whole host of agents and be exposed, through those through those shopping experiences. Importantly, the merchant remains the merchant of record, and that part really matters. Like businesses want access to these new storefronts, these new channels, but they don't want to give up the customer relationship. They don't want to give up control over trust or fraud.
Speaker 144:40 - 45:19
Meta 的广告购物体验也在用它。它的工作方式基本上是:merchant 只需要通过 Stripe 集成一次,把自己的 product catalogs(商品目录)、prices(价格)和 checkout flows(结账流程)接进来,然后他们就真的可以直接在 dashboard(控制台)里开启接入,通过一整批 agent 被这些购物体验所发现和展示。重要的是,merchant 仍然是 merchant of record(交易记录商户/法定销售主体),这一点非常关键。企业确实想进入这些新的 storefronts(店面入口)、这些新的 channels(渠道),但他们不想放弃与客户的关系,也不想放弃对 trust 或 fraud(欺诈)问题的控制权。
Speaker 145:19 - 45:30
Kind of category one is the human is still sort of leading the buying, but the agent is like facilitating the transaction. You could call it agent to commerce. You could call it facilitated commerce.
Speaker 145:19 - 45:30
大致来说,第一类是人类仍然在主导购买,只是 agent 在促成这笔交易。你可以把它叫作 agent to commerce,也可以叫它 facilitated commerce(促成式商业)。
Speaker 245:30 - 45:51
And how does that actually work? So is the experience something like, I'm in ChatGPT and it says, hey. Here's, like, a thing you might wanna buy, and I can click checkout from OpenAI, and that's using that protocol to then go send my information to the merchant and then send me back, hey, like, here here's your your your thing is on the way. That's kind of that's what you're talking about?
Speaker 245:30 - 45:51
那这实际是怎么运作的?所以体验是不是大概像这样:我在 ChatGPT 里,它会说,嘿,这里有个你可能想买的东西;然后我可以直接点由 OpenAI 发起的 checkout(结账),它再通过那个 protocol(协议)把我的信息发送给 merchant(商家),然后再回传给我一个消息,说,嘿,你买的东西已经在路上了。你说的大概就是这个意思吗?
Speaker 145:51 - 45:54
Exactly. Yeah. So it's like one click checkout. Yeah. Same thing.
Speaker 145:51 - 45:54
没错。对。所以它就像 one-click checkout(一键结账)。对,是同一回事。
Speaker 145:54 - 46:16
Like, you're in Facebook. You get an ad in Meta. Let's say you do, a one click checkout. You know, one of the primitives that we built for this is the shared payment token or SPT. And it's basically just it just lets your payment credentials be passed securely from the AI agent to the merchant so the merchant can process the transaction.
Speaker 145:54 - 46:16
比如说,你在 Facebook 里,在 Meta 看到一则广告。假设你要做 one-click checkout(一键结账)。我们为此构建的一个基础 primitive(原语)就是 shared payment token,也就是 SPT。它本质上就是让你的 payment credentials(支付凭证)能够从 AI agent(AI 代理)安全地传递给 merchant(商家),这样 merchant 就能处理这笔交易。
Speaker 146:16 - 46:42
And the merchant processing the transaction is important because that allows the merchant to remain the merchant of record. But, you know, you don't want your credentials viewed by the agent, which is, you know, why it's why it's a token and not your actual payment credentials. And the merchant needs to know that you and the agent are good, which is why as part of the shared payment token, we pass over, a whole host of of fraud scores.
Speaker 146:16 - 46:42
而由 merchant 来处理交易这件事很重要,因为这能让 merchant 继续作为 merchant of record(登记商户 / 交易记录上的正式商户)存在。不过,你肯定不希望自己的凭证被 agent 看到,这也是为什么这里用的是 token(令牌),而不是你真实的 payment credentials。与此同时,merchant 也需要知道你和 agent 都是可信的,所以作为 shared payment token 的一部分,我们还会传递一整套 fraud scores(欺诈评分)。
Speaker 246:42 - 46:56
And can I integrate this? Like, so we have a CLI. We have a bunch of software. Can I offer a agentic checkout easily, or does it have to go through the OpenAI's and the Facebook's of the world?
Speaker 246:42 - 46:56
那我能集成这个吗?比如说,我们有一个 CLI,也有一堆软件。我能不能比较容易地提供 agentic checkout(代理式结账),还是说这必须通过 OpenAI、Facebook 这类大平台来做?
Speaker 146:56 - 47:22
So, yes, you can. And I think one of the premises here is, just like to date, we haven't seen one model provider to rule them all or one model to rule them all. We don't think there's going to be one agentic shopping experience to rule them all. And merchants will literally break if they have to integrate with every single potential new storefront, right? When they integrated with the internet, they built their own storefront.
Speaker 146:56 - 47:22
可以,答案是可以。我认为这里的一个前提是:就像到目前为止,我们并没有看到一个 model provider(模型提供商)能一统天下,也没有一个 model(模型)能包打天下。我们也不认为未来会有一种 agentic shopping experience(代理式购物体验)能统治一切。而且如果 merchant 必须和每一个潜在的新 storefront(店面 / 销售前端)分别集成,他们真的会被拖垮,对吧?当年他们接入互联网时,是自己搭建了一个 storefront。
Speaker 147:22 - 47:56
And yeah, they iterated on it, but basically they built it once. But if you tell them, Hey, you need to build your storefront for, you know, agent startup, shopping startup X and FIA and OpenAI and Meta, like their eyes are gonna get, you know, bigger than their heads, they're not gonna be able to handle it. And so we really wanna abstract away that complexity for businesses. Like we spent the last decade plus helping businesses sell wherever their customers are. And first that was like on their websites, and then it was in apps, and then it was through platforms and marketplaces, and actually some in person too with our terminal product.
Speaker 147:22 - 47:56
是的,他们后来不断迭代,但基本上那套东西只需要搭一次。可如果你现在告诉他们:嘿,你得分别为 agent startup、shopping startup X、FIA、OpenAI 和 Meta 去搭 storefront,那他们肯定会压力大到不行,根本应付不过来。所以我们真正想做的是,把这种复杂性从企业那里抽象掉。过去十多年里,我们一直在帮助企业在客户所在的任何地方完成销售。最开始是在他们自己的网站上,后来是在 app(应用)里,再后来是在各种平台和 marketplace(市场平台)上,实际上也包括一部分线下场景,通过我们的 terminal product。
Speaker 147:56 - 48:21
But now, like, you know, where are the consumers? Where are they wanting to buy increasingly through sort of AI tools and agentic flows? And so similarly, we just wanna make it really easy, for merchants to agnostically participate in those different storefronts. And again, they can choose where they want to sell. They can turn it on, little toggle in the dashboard, but it's not a different integration, which is the whole idea of the protocol.
Speaker 147:56 - 48:21
但现在,消费者在哪里?他们越来越想通过各种 AI tools(AI 工具)和 agentic flows(代理式流程)来购买。所以同样地,我们只是想让 merchant 能够非常容易地、以 agnostic(与具体平台无关)的方式参与这些不同的 storefront。再说一次,他们可以自己选择想在哪些地方卖,也可以在 dashboard(控制面板)里通过一个小 toggle(切换开关)来启用,但这并不是一套不同的集成方式——这正是这个 protocol 的核心理念。
Speaker 248:21 - 48:28
And then how often is this happening? Like what's the volume of agentic commerce right now?
Speaker 248:21 - 48:28
那么,这种情况现在发生得有多频繁?比如说,agentic commerce(agent 驱动的商业)的交易量现在大概有多大?
Speaker 148:29 - 49:07
The volume of consumer commerce is still relatively small, relatively small, relatively small as a percentage of all of the commerce we see. But it is growing quickly, particularly for, you know, what I would think of as commodities, right? So what is the first thing that people are comfortable buying through agents? It's like things that are reasonably known, reasonably observable, not super high priced. You know, think when people started buying online, like, you didn't imagine that they were gonna go online and buy like a $2,000 couch, right?
Speaker 148:29 - 49:07
消费者 commerce(商业)的交易量目前仍然相对较小,相对于我们看到的全部 commerce 来说,占比还比较小。但它增长得很快,尤其是在——你知道——我认为属于 commodities(标准化商品)的品类上,对吧?所以,人们最先愿意通过 agents(代理)购买的是什么?就是那些相对已知、相对可观察、价格也不是特别高的东西。你知道,想想当年人们刚开始在网上购物时,你不会觉得他们会上网直接买一张 2,000 美元的沙发,对吧?
Speaker 149:07 - 49:19
Oh, wasn't a couch a thing? You really wanna know the quality and you wanna sit in it. A mattress. Oh my God, these mattress companies that have blown up. And it took time for them to build comfort, making higher priced purchases, making more quality dependent purchases.
Speaker 149:07 - 49:19
哦,沙发难道不就是这种典型东西吗?你确实会很想了解它的质量,而且还想亲自坐一坐。还有床垫。天哪,这些爆发式增长的 mattress(床垫)公司。人们花了时间,才逐渐建立起对更高价格购买、更依赖质量判断的购买行为的信心。
Speaker 149:20 - 49:48
And so today it's it's it's predominantly commodities. Know, in a similar vein, I've tried, you know, to to book a whole summer trip using using an agent, And I wasn't sufficiently satisfied with, you know, the family of four choice of flights and hotels and transportation and itinerary to be willing to to one click buy it. But the models will get better. The interfaces will get better. The experiences will get better.
Speaker 149:20 - 49:48
所以到今天为止,这类购买主要还是以 commodities(标准化商品)为主。类似地,我也试过用 agent(代理)来预订整个暑期旅行,但对于一个四口之家在航班、酒店、交通和行程安排上的选择,我还没有满意到愿意一键直接下单的程度。不过 models(模型)会变得更好,interfaces(界面)会变得更好,experiences(体验)也会变得更好。
Speaker 149:49 - 50:11
We're pretty agnostic to those. Like we trust that they will evolve in a bunch of different and interesting ways and sort of the the primitives that are needed underneath the ability to expose your catalog, the shared payment token, the fraud protections are pretty agnostic to those experiences. And so that's where we're hyper focused for merchants.
Speaker 149:49 - 50:11
对这些我们基本上是比较 agnostic(不预设特定路径)的。我们相信它们会以许多不同且有意思的方式演进;而在底层真正需要的那些 primitives(基础能力),比如开放你的 catalog(商品目录)的能力、共享的 payment token(支付 token)、fraud protections(欺诈防护),对于这些不同体验其实都是比较 agnostic 的。所以,这也是我们目前为 merchants(商家)高度聚焦的方向。
Speaker 250:12 - 50:18
Give me an example of one of these commodities and also what the order of magnitude we're talking about when we say it's relatively small.
Speaker 250:12 - 50:18
给我举一个这类 commodities(标准化商品)的例子吧,另外,当我们说“相对较小”的时候,量级大概是什么样?
Speaker 150:19 - 50:23
An example of a commodity would be, like, a Halloween costume.
Speaker 150:19 - 50:23
一个 commodity(标准化商品)的例子,比如说,Halloween costume(万圣节服装)。
Speaker 250:23 - 50:26
Got it. Agents are buying Halloween costumes for themselves?
Speaker 250:23 - 50:26
明白了。Agents(代理)是在给它们自己买 Halloween costumes(万圣节服装)吗?
Speaker 150:27 - 50:29
Agents are buying Halloween costumes.
Speaker 150:27 - 50:29
Agents 正在购买万圣节服装。
Speaker 250:30 - 50:31
That's really
Speaker 250:30 - 50:31
这真的是——
Speaker 150:31 - 50:52
How many lazy parents there are in the world? Mean, I I think the consumer side is interesting too, right? Because we talked about what do businesses need, right? They need a they need a fast, easy way to safely expose their products, their prices, their inventory, their checkouts, understand fraud and be in control of the relationship. From the consumer angle, the question's a little different, right?
Speaker 150:31 - 50:52
世界上有多少偷懒的家长?我是说,我觉得消费者这一侧也很有意思,对吧?因为我们刚才讨论的是企业需要什么,对吧?他们需要一种快速、简单且安全的方式,来对外开放他们的产品、价格、库存和结账流程,理解欺诈风险,并且掌控这种关系。从消费者的角度看,问题就有点不一样了,对吧?
Speaker 150:52 - 51:05
Like, even if I'm a lazy parent, I'm not so lazy that I'm willing to give someone my payment credentials and, you know, let it rip. So, like, the question for me is how do I safely let an agent buy on my behalf? And have you heard of LINK?
Speaker 150:52 - 51:05
比如说,就算我是个偷懒的家长,我也没懒到愿意把我的支付凭证交给别人,然后你知道的,随它去刷。所以,对我来说,问题是:我怎样才能安全地让一个 agent 代我购买?你听说过 LINK 吗?
Speaker 251:05 - 51:06
Yeah. I've used LINK.
Speaker 251:05 - 51:06
听说过。我用过 LINK。
Speaker 151:07 - 51:11
Okay. Amazing. So LINK is our consumer wallet. What did you use it for? Do you remember the first thing you used?
Speaker 151:07 - 51:11
好的。太棒了。所以 LINK 是我们的消费者钱包。你当时用它做了什么?你还记得你第一次用它是做什么吗?
Speaker 251:11 - 51:13
I mean, I use it all the time. So it's, like, everywhere.
Speaker 251:11 - 51:13
我的意思是,我一直都在用它。所以它基本上到处都有。
Speaker 151:13 - 51:24
So Amazing. Yeah. It's everywhere. Right? It's you wouldn't believe where it I I was, like, I was, getting soccer lessons for one of my kids, like, you know, from a local guy.
Speaker 151:13 - 51:24
太厉害了。对,确实到处都有,对吧?你都想不到它会出现在哪里。我当时——我是说——我当时在给我的一个孩子报 soccer 课程,就是那种由本地一个人开的课。
Speaker 151:25 - 51:39
And I was on their website and they only accepted Visa and MasterCard, neither of which I had on me, or direct debit from my bank account, which I wasn't going to put in this very janky website, or link. And I was like, oh, amazing. Link is here.
Speaker 151:25 - 51:39
我当时在他们的网站上,发现他们只接受 Visa 和 MasterCard,而我手头这两种卡都没有;要么就是从我的银行账户直接借记,但我可不打算把这个信息填进这个非常粗糙的网站或链接里。我当时就想,哦,太棒了。Link 就在这儿。
Speaker 251:39 - 51:40
Anyway,
Speaker 251:39 - 51:40
总之,
Speaker 151:40 - 52:03
a lot of people know about Link as our consumer wallet for buying soccer classes. It speeds up checkout, but it's also so and and it's already used by about a quarter of a billion consumers. So it's not a small network. But what I think is most interesting about Link is it's a very dense network when it comes to AI. So Lovable is an interesting example.
Speaker 151:40 - 52:03
很多人都知道 Link 是我们的 consumer wallet(消费者钱包),用来购买 soccer classes。它能加快结账速度,而且它已经被大约 2.5 亿消费者使用了。所以这并不是一个小型网络。但我认为 Link 最有意思的一点是,在 AI 方面,它是一个非常高密度的网络。Lovable 就是个很有意思的例子。
Speaker 152:03 - 52:25
58% of their payment volume runs through Link. You are hyper AI pilled. It is not surprising that everywhere you are, Link is. And so what's changing now is that we're evolving Link for the AI economy because so many of the Link consumers are already AI consumers. And acknowledging that, like, agents themselves are becoming economic actors.
Speaker 152:03 - 52:25
他们 58% 的支付交易量都通过 Link 处理。你是个彻底接受 AI 的人,所以无论你在哪里,Link 在那里也就不奇怪了。而现在正在发生的变化是,我们正在为 AI economy(AI 经济)演进 Link,因为很多 Link 的消费者本来就已经是 AI 消费者了。与此同时,我们也承认,agent(智能体)本身正在成为经济活动的参与者。
Speaker 152:25 - 52:56
And so the model isn't, you know, give a random agent your card and hope for the best. Instead, it's delegated authority with guardrails. So, you know, you as the consumer decide which agents are allowed to request credentials and under what conditions and with what limits and whether those purchases require approvals before they go through. And you do all of that through LinkedIn. It's just a much more sensible model for delegated purchases.
Speaker 152:25 - 52:56
所以这个模型并不是,随便把你的卡交给某个 agent,然后听天由命。相反,它是带有 guardrails(保护性约束)的 delegated authority(委托授权)。也就是说,作为消费者,你来决定哪些 agent 可以请求凭证、在什么条件下可以请求、额度限制是多少,以及这些购买在执行前是否需要审批。而这一切你都通过 LinkedIn 来完成。这显然是一个更合理的委托购买模型。
Speaker 252:56 - 53:01
That makes sense. Emily, this was a fantastic conversation. I learned so much.
Speaker 252:56 - 53:01
这很有道理。Emily,这次对话太精彩了。我学到了很多。
Speaker 153:01 - 53:02
Awesome. Thank you for having me.
Speaker 153:01 - 53:02
太好了。谢谢你邀请我。
Speaker 353:17 - 53:41
Because this show is the epitome of awesomeness. It's like finding a treasure chest in your backyard, but instead of gold, it's filled with pure unadulterated knowledge bombs about chat GPT. Every episode is a roller coaster of emotions, insights, and laughter that will leave you on the edge of your seat craving for more. It's not just a show. It's a journey into the future with Dan Shipper as the captain of the spaceship.
Speaker 353:17 - 53:41
因为这个节目简直是“超棒”这个词的化身。它就像你在自家后院发现了一个宝箱,只不过里面装的不是黄金,而是关于 chat GPT 的、纯粹而毫无掺杂的知识炸弹。每一集都像一趟情绪、洞见和笑声交织的过山车之旅,让你坐立难安、还想继续听下去。它不只是一个节目,而是一场驶向未来的旅程,而 Dan Shipper 就是那艘宇宙飞船的船长。
Speaker 353:41 - 53:52
So do yourself a favor. Hit like, smash subscribe, and strap in for the ride of your life. And now without any further ado, let me just say, Dan, I'm absolutely, hopelessly in love with you.
Speaker 353:41 - 53:52
所以,帮自己个忙吧。点个赞,猛击订阅,然后系好安全带,准备开启你人生中最刺激的一段旅程。现在闲话不多说,我只想说,Dan,我已经彻底、无可救药地爱上你了。
原文 ↗https://www.youtube.com/playlist?list=PLuMcoKK9mKgHtW_o9h5sGO2vXrffKHwJL
BuildSpeak — 关于本项目BUILT IN PUBLIC · 跟随 builders 而非 influencers