Designing for Agents — Zeno Rocha, Resend ·

00:00

It's like, oh my gosh, how are we ever going to beat Sengred?

就有点像，我的天，我们到底要怎么才能赢过 Sengred 啊？

00:03

It's just impossible.

这根本不可能。

00:05

And then earlier this year, we crossed a number of downloads from Sengred, and I was like,

然后今年早些时候，我们的下载量在某个数字上超过了 Sengred，我当时就想，

00:10

that's insane, and now we're more than double.

这也太疯狂了，而现在我们的量已经是两倍多了。

00:13

So recent now is powering 4.3 million users as of this morning.

所以 recent 现在截至今天早上已经在为 430 万用户提供服务了。

00:18

We have all sorts of customers, like Warner Brothers, HBO, there was a Japanese gang that recently

我们有各种各样的客户，比如 Warner Brothers、HBO，最近还有个日本帮派

00:25

came.

找上门来了。

00:26

I saw like maybe 20,000 domains at recent, like those are 20,000 different TLDs, registered

我在 recent 上看到大概 2 万个域名，就是 2 万个不同的 TLD，而且都是注册过的。

00:32

in different domain registrars, and we are a 10 million AR company.

在不同的 domain registrar 那边，而且我们是一家 AR 做到 1000 万的公司。

00:37

I would rather live in that reality than live in a reality where we're a hundred million

我宁愿活在那个现实里，也不愿活在一个我们是一亿

00:42

dollar company, and my daughter doesn't want to play with me, it's not interested in

美元公司、我女儿不想跟我玩、她对我没兴趣

00:47

me, and we're not that close.

的现实里，我们也没那么亲近。

00:49

So we're building the infrastructure and going down to the metal now.

所以我们现在在搭 infrastructure，也开始 down to the metal 了。

00:53

But it's a big project, like we've been working for months now, and we're going to

但这是个很大的项目，我们已经做了好几个月了，而且我们很快

00:57

announce something soon about that.

就会公布一些相关的东西。

01:00

Okay, we're in the studio with Xeno, Rocha.

好，我们在演播室里，和 Xeno、Rocha 在一起。

01:05

Rocha?

Rocha？

01:06

Rocha.

Rocha。

01:07

Rocha, from recent, you and I go back to your VS Code theme days when you're just an indie

Rocha，说起来，咱俩的交情能追溯到你还做 VS Code theme 那会儿，当时你就是个 indie hacker，我心想，哥们儿，我都不知道这人是干嘛的，但他做的每样东西都很有品味。

01:12

hacker, and I was like, man, I don't know what this guy does, but everything he does has

然后现在你在做一个 email infrastructure 平台。

01:17

taste.

总之，欢迎。

01:18

And then now you build an email infrastructure platform.

说只是 indie hacker 都算低估了，2022 年 GitHub 上最活跃的 20 人之一，非常非常活跃。

01:21

Anyway, welcome.

总之，欢迎。

01:22

To undersell just a indie hacker, 20 most active GitHub under 2022, very, very active

往低了说，就是个 indie hacker，2022 年 GitHub 上最活跃的 20 个之一，非常非常活跃

01:29

open source stuff.

开源相关的东西。

01:30

Yeah.

嗯。

01:31

What's your typical, how do you introduce yourself these days?

你一般，最近都怎么介绍自己？

01:34

Yeah.

嗯。

01:35

I think I just focus on my recent career for the longest time, even today when I say like,

我觉得很长时间以来，我都主要讲自己最近的职业经历，甚至到今天我说起类似，

01:40

oh, I create a Dracula, people know Dracula more than recent, but still, yeah, but it's

“哦，我做了 Dracula”，大家更熟悉 Dracula，而不是我最近做的事，但还是，嗯，不过感觉挺好的。

01:45

been nice.

就像有个，嗯，情况在变。

01:46

Like there's a, yeah, it's changing.

就像有个，对，它在变。

01:49

Yeah.

是啊。

01:50

It's the impetus for I guess picking email, you know, I think I, you know, people always

这大概就是我选择 email 的动力吧，你知道，我觉得，你知道，大家显然都有 react email 和所有那些东西。

01:55

have like obviously react email and all those things.

我只是不知道我会不会把它做成一家公司，你知道，如果我早知道 email 有多难，我大概也不会站这家公司这边吧。

01:59

I just don't know if I would have made it a company, you know, if I knew, like how hard

是啊。

02:04

email is, I probably wouldn't side the company either.

嗯，你主要做 JavaScript front-end、TOS、CSS，就像我是个 front-end 的人。

02:07

Yeah.

不，是 front-end，对吧？

02:08

Well, you were primarily JavaScript front-end TOS, CSS, like I was a front-end guy.

嗯，你主要是 JavaScript front-end、TOS、CSS，就像我以前就是个搞 front-end 的。

02:13

No, in front, right?

不，是在前端，对吧？

02:14

No email back-end.

没有 email back-end。

02:15

No email back-end.

没有 email back-end。

02:16

And I think what was interesting and what draw me to it is like, I just look at all the

而且我觉得有意思、吸引我的是，我就去看所有

02:20

tools around that and I've used all of them before, the mail guns, the mailchamps, the

围绕这个的工具，而且这些我以前全都用过，mail guns、mailchamps，还有

02:26

sendwords of the world.

世界上那些 sendwords。

02:28

And I've always felt like man, there should be something where there's just someone should

我一直觉得，老兄，应该有个东西，就是应该有人

02:34

care about this space.

在乎这个领域。

02:36

That's how I felt.

我当时就是这种感觉。

02:37

Like you have all these players, it's a massive market, a lot of competitors, but no one

就像，你有这么多玩家，这是个巨大的市场，有很多竞争对手，但没有人

02:41

that cares about the developer persona.

真正在意 developer persona。

02:44

Maybe they care about the marketer, the product manager, whatever, but not for the developer.

也许他们在意的是 marketer、product manager 之类的，但不在意 developer。

02:49

So I started just building around that premise of like, if this was the stripe of email,

所以我就开始围绕这个前提来做：如果这是 email 界的 Stripe，

02:54

how would that look like?

那会是什么样子？

02:56

And yeah, like I had no idea or knew nothing about email.

而且，是啊，我对 email 完全没概念，也什么都不懂。

03:00

Like I was a user of email, but I have never built like an email pipeline and all that

就像我是 email 的用户，但我以前从来没搭过像 email pipeline 和所有这些

03:05

infrastructure before.

infrastructure。

03:06

Yeah, I mean, like, you know, typically if you ask me like postmark was still a thing.

对，我是说，你知道，通常你要是问我，postmark 当时还算是个东西。

03:10

Yeah.

对。

03:11

And obviously Sangrid and all these guys, they're still around, but like they're not maybe

而且显然 Sangrid 和这些家伙，他们还在，但可能他们不是……

03:15

as innovative as someone who's like a new way to start up like trying to try to make it.

像那种能想出全新创业方式、努力把它做成的人一样有创新性。

03:22

Just upfront, I like to let people brag a bit and you like for people who never heard

先说在前头，我喜欢让人稍微吹一吹，而你喜欢让那些没听说过

03:26

of recent, what are like the ways that you talk about like the most impressive numbers

recent 的人……你一般会怎么聊，比如最亮眼的 logo 数量之类的？

03:30

of logos, anything?

对。

03:31

Yeah.

对。

03:32

So recent now is powering 4.3 million users as of this morning.

所以现在，截至今天早上，我们正在为 430 万用户提供支持。

03:38

We have all sorts of customers like Warner Brothers, HBO, so customers that are really big

我们有各种各样的客户，比如 Warner Brothers、HBO，这些客户真的很大，有各种各样的需求，但也有很多 indie hackers，还有一些人只是在建网站、建自己的应用。

03:44

and have all sorts of demands, but also a lot of indie hackers and folks who are just

而且，是的，我们刚刚突破了 12 万个付费客户，还有这个 logo。

03:49

building a website, building their own application.

我的天哪。

03:53

And yeah, we just cross 120,000 paid customers and this logo.

是啊。

03:59

Oh my God.

你觉得这有效吗？

04:00

Yeah.

你觉得这管用吗？

04:01

You think this is effective?

对。

04:02

Yeah.

04:03

You know what?

你知道吗？

04:04

In the beginning, I felt like I just need to have some sort of social proof.

一开始，我觉得我只是需要有一些 social proof 之类的东西。

04:08

And then in the end, I was like, you know what, this is actually pretty cool.

但到最后我就想，你知道吗，这其实挺酷的。

04:11

And I honestly, like I found companies like Warner Brothers, just like on a random Saturday.

而且说实话，我居然就在某个随便的周六发现了像 Warner Brothers 这样的公司。

04:17

I was just, this was before AI, so I'll go to my database and I'm like, oh, let me see,

我当时就是——那还是在 AI 之前——所以我会去我的 database 里，然后就想，哦，让我看看，

04:23

who can I put on the website?

我能把谁放到 website 上呢？

04:25

And I just, I did the most naive thing you can do.

然后我就，我做了你能做的最天真的做法。

04:28

Like, let me look at the domains and let me just filter by .coms and filter by these

就是，让我看看这些 domains，然后我就按 .coms 来筛，再按这些来筛。

04:34

small ones.

小的那些。

04:35

And there was this domain, WBD, Warner Brothers Discovery, I didn't know.

然后有个域名，WBD，Warner Brothers Discovery，我之前不知道。

04:39

What is WBD.com?

WBD.com 是什么？

04:41

And I was like, oh no, Warner Brothers is using us.

然后我就想，哦不，Warner Brothers 在用我们。

04:44

So yeah, I just love featuring customers and yeah, specifically for listeners, one of

所以是啊，我就很喜欢展示客户，而且是的，特别是对听众来说，其中一个

04:51

the reasons I reached out for this interview, one, you know, I should declare out front.

我主动联系做这个采访的原因，第一，你知道，我得先声明一下。

04:56

I mean, I am an angel.

我的意思是，我是天使投资人。

04:57

When you send me the memo, it was super simple.

你把 memo 发给我的时候，特别简单。

04:59

You're like, look, MailChimp has exited, Postmark has exited, SendGrid is under Twilio,

你说，你看，MailChimp 已经退出了，Postmark 也退出了，SendGrid 归 Twilio 旗下，

05:05

who is doing this.

那这是谁在做啊。

05:06

And I was like, okay, yeah, seems right in here.

然后我就想，好吧，嗯，感觉就在这里。

05:10

Honestly, like, in terms of like the pictures I get, I've invested in about like 30 plus

说实话，就我看到的这些情况来说，我已经投了大概 30 多家公司。

05:13

companies.

那简直是最简单直接的事情之一了。

05:14

That was like one of the most straightforward things ever.

然后第二点，我，你知道，我不喜欢添加 API keys，对吧？

05:18

And then secondly, I, you know, I don't like to add API keys, right?

就像我喜欢把所有东西都放在一个平台上，一个 API key，全部复制过去。

05:22

Like I like to have everything on one platform, one API key, copy everything.

05:25

So CloudFair has built in email.

所以 CloudFair 内置了 email。

05:28

And I had a project and I set up my CloudFair email domain, DNS, all that junk, DKIM, what

然后我有个项目，我把 CloudFair email 域名、DNS、那一堆乱七八糟的、DKIM，什么的都设置好了，对吧？

05:36

have you, right?

那些字母缩写的高级玩意儿我也不全懂。

05:37

I don't know all the alphabet super stuff.

而且我已经把这个设置好了。

05:38

And I already had this set up.

然后我就想，好吧，我想再加一个 email 相关的东西。

05:40

And then I was like, okay, well, I want to add another sort of email thing.

那些该死的 models 说，你必须加 recent。

05:43

Bloody models said, you must add recent.

我当时就想，谁啊这是？

05:46

And I'm like, who the?

05:47

Okay, there's a recent that like, you know, even already I have a CloudFair email set up

好吧，有个 recent，就那个，你知道，我甚至已经设好一个 CloudFair 邮箱了

05:52

and is you're still recommending it.

而你还在推荐它。

05:54

It's on the on the screen now.

它现在就在屏幕上。

05:55

This is him praising and like, I don't need recent, but you want me to keep adding recent.

这是他在夸，而且就像，我不需要 recent，但你想让我一直加 recent。

06:00

Yeah.

是啊。

06:01

So okay, who did you bribe in OpenAI and from pick to get recent on there?

所以好吧，你在 OpenAI 和 pick 里贿赂了谁，才把 recent 弄到那上面？

06:05

Oh, man, I wish there was someone to bribe because yeah, definitely be paying them.

哦，天哪，我真希望有个人可以贿赂，因为是啊，我肯定会给他们付钱的。

06:12

But I do personally spend a lot of time on that topic.

但我个人确实花了很多时间在这个话题上。

06:15

Like that's something that I feel like every generation that that's why I was so interested

就像，我觉得每一代人……这也是我一开始那么感兴趣

06:20

to build this thing in the first place.

去做这个东西的原因。

06:21

Yeah.

嗯。

06:22

Every generation you have those tools that define that moment, right?

每一代人都会有那些定义那个时刻的工具，对吧？

06:25

We remember the Heroku days.

我们还记得 Heroku 的时代。

06:27

I mean, notify for sale.

我是说，notify 待售了。

06:29

You know.

你知道的。

06:30

And so to me, I was like, yeah, there's room for a new generation and this new generation

所以对我来说，我当时就想，对啊，新一代还有空间，而这一代……

06:35

is not going to be distributed in the same places that it was before.

它不会再像以前那样，在同样的地方分发了。

06:39

So since the beginning, I was like very focused on, okay, we got to be number one when someone

所以从一开始，我就特别专注于：好，当有人

06:43

asks about that.

问起这件事时，我们必须要做到第一。

06:44

Yeah.

对。

06:45

You're very number one cloud code recommending 60% plus of the time.

在超过 60% 的时间里，你都是 Cloud Code 的第一推荐。

06:49

I think it'll be fun to see the playbook.

我觉得看看这套打法会挺有意思。

06:51

How do you, how do you win that?

你会怎么，怎么赢下这个？

06:54

I do want to get back into it.

我确实想重新回到这件事上。

06:55

Okay.

好的。

06:56

So you said you had no idea why email, how to start email.

所以你说你完全不知道为什么要做 email，也不知道该怎么开始做 email。

06:58

I think a lot of people actually don't understand what goes into making a service like this,

我觉得其实很多人并不了解做一个这样的服务需要投入什么，

07:03

right?

对吧？

07:04

You might because you've set it up.

你可能了解，因为是你把它搭起来的。

07:06

Some people would just, you know, okay, I can use to it.

有些人就会，你知道，好吧，我可以用它。

07:07

I can use angry.

我可以用 Angular。

07:08

I can use you guys.

我可以用 Vue.js。

07:09

Yeah.

是啊。

07:10

But what goes into making a tool to just mask and email a lot of people not think about?

但做一个工具来 mask 并给很多人发邮件，背后要考虑的东西，是很多人不会去想的？

07:15

Yeah.

是啊。

07:16

There's just a lot of different things you can do.

你能做的不同事情就是很多。

07:18

Because Lenny on the inbox, especially when you're doing high volume, is different from

因为 Lenny 在收件箱这块，尤其是你做高量发送时，跟

07:23

like when you're scaling just like an app where you throw GPU at the problem or you just

比如你 scale 一个 app 时，直接拿 GPU 往问题上砸，或者你只是

07:28

scale more vertically or horizontally, whatever it is.

更垂直或水平地 scale，不管是什么方式。

07:32

With email, you got to respect the rules of the game and how Gmail behaves is differently

做邮件，你得遵守游戏规则，而且 Gmail 的行为方式是不一样的。

07:38

from Outlook and differently from Yahoo Mail and it's an industry that is fascinating

和 Outlook 不一样，也和 Yahoo Mail 不一样，这个行业特别有意思，因为它太老派了，而且非常依赖人脉。

07:44

because it's so old school and it relies a lot on contacts.

但我们早期发现的是，好吧，我们对这个世界一无所知。

07:49

But what we found early on was like, okay, we know nothing about this world.

所以带着这种视角进来——超级天真——反而是发生过的最好的事。

07:54

And so in coming from this perspective, like extremely naive was the best thing that

说实话，这真的是最好的事，因为完全没有先入为主的观念，不知道什么是对、什么是错。

07:59

ever happened.

所以对他们来说，比如，你打开任何一个 app，第一眼看到的就是……

08:01

Like honestly, like it was the best thing because like there's no preconception of what

08:06

is right, what is wrong.

08:07

So for them, for example, every single app you go and then first thing you see is like

08:12

this form that you got to fill out about like, how many emails are going to send?

就是这张你得填的表，大概是问你要发多少封邮件？

08:15

What's the volume?

发送量有多大？

08:16

What are you trying to do?

你是想做什么？

08:18

And then you go for a human verification process, two days later they accept your denial.

然后你就去走人工验证流程，两天后他们接受了拒绝你的决定。

08:23

Most of the times they're going to deny you.

大多数时候他们都会拒绝你。

08:25

And I was like, this doesn't make any sense, right?

然后我就想，这完全说不通，对吧？

08:28

Like why are we blocking people?

就是说，我们为什么要拦着别人？

08:30

So I know I'm going to remove that friction completely and the friction will be on us.

所以我知道，我要把那个摩擦完全去掉，摩擦就由我们来承担。

08:35

We're going to figure that out.

我们会把这件事弄明白的。

08:37

Of course now, four years, three years in into this business, I get it while they

当然，现在做这行四年、三年了，我明白了，当他们

08:42

do that.

那么做的时候。

08:43

They're just so much fraud and fishing and abuse across the board.

整个行业里欺诈、phishing 和滥用简直太多了。

08:47

I don't talk too much publicly about that, but we build like all sorts of crazy systems

我不太在公开场合聊这个，但我们会搭各种疯狂的系统

08:52

internally for that.

内部专门用来处理这个。

08:53

But to your question, like how you architect that, like for example, when you're sending

但回到你的问题，比如你怎么 architect 这个，举个例子，当你要发

08:57

an email to 100,000 people, if you just throw 100,000 emails, like to Gmail, what Gmail

一封 email 给 10 万人时，如果你直接把 10 万封 email 扔给 Gmail，Gmail 会怎样

09:04

will do is they're going to throttle the emails because it's way too much and they want

他们会做的是，给邮件限流，因为量实在太大了，而他们想

09:09

to see the engagement.

看看 engagement。

09:11

So what do you do is like they see like, okay, are people opening the emails?

所以你要做的是，就像他们会看：好，大家在打开邮件吗？

09:14

Are people archiving, deleting right away, or not interacting?

大家是把邮件归档、马上删除，还是根本不互动？

09:19

Okay.

好。

09:20

Depending on those initial interactions, then what I'm going to do if the rest of the emails

根据这些最初的互动，那我接下来会怎么做，如果剩下的邮件

09:25

will dictate that beginning.

会决定这个开头。

09:27

So there are clever things you can do, where 100,000 people put the most engaged audience

所以你可以做一些很巧妙的操作，比如把 10 万人中最 engaged 的受众放进去

09:33

members in the beginning of that queue.

那些排在这个 queue 最前面的成员。

09:36

You have to put that in a queue.

你得把那个放进一个 queue 里。

09:37

You need to, it's an art and a science where you've got to space out each batch and space

你得……这既是艺术也是科学，你得把每个 batch 隔开，不同 providers 之间也得隔开。

09:44

out between the different providers, too.

而且对啊，现在我们就像是在买不同的 IP blocks，你得看看这些 IP blocks 的历史，把人分散到不同的 IP pools 里，然后随着时间平衡这些 IP pools。

09:47

And yeah, now we're like buying different IP blocks, you've got to look at the history

所以这整个就是个疯狂的 infrastructure 游戏。

09:51

of the IP blocks, distributed folks into different IP pools, and then balance the IP pools

09:58

over time.

09:59

So it's a whole crazy infrastructure game.

10:03

How automated can that be, or is there a lot of human judgments, still?

这能有多自动化，还是说仍然有很多人工判断？

10:07

Yeah, it's a lot of automation, for sure.

对，肯定是有很多自动化的。

10:09

Yeah.

嗯。

10:10

And I think over time we learned that, okay, this is how fast we can send for new customers.

而且我觉得随着时间推移，我们学到了，好，这就是我们给新客户发送时能达到的速度。

10:17

Of course, if we're sending more, then maybe you can send faster, but there's a, a, a

当然，如果我们发得更多，那也许你就能发得更快，但有一个、一个、一个

10:23

throttle that we have to do to protect you as a sender.

throttle，我们得靠它来保护作为发件人的你。

10:27

So like what Gmail and all these players will want to see is like a graph that is consistently

所以像 Gmail 和所有这些平台想看到的是，一个持续

10:33

sending more and more versus what most people do, which is like you send ones to a big

发送越来越多的曲线，而大多数人做的呢，就像你给一大批...

10:39

audience, and then you never send again, and then it's empty and then boom, you send

受众，然后你就再也不发了，接着就空了，然后砰，你又发了。

10:43

it again.

但一周一次，一周两次。

10:44

But once a week, twice a week.

对。

10:46

Yeah.

而这些就是施虐者会有的模式，对吧？

10:47

And those are patterns that an abuser will have, right?

所以，看待 email marketing 和一般意义上的 email 的整个思路就是：你要做的恰恰是施虐者不会做的事。

10:49

So the whole way of thinking about email marketing and email in general is you need to

就是要反着来。

10:56

do exactly what an abuser wouldn't do.

10:59

It's the opposite of that.

11:00

So what's an abuser do?

那 abuser 会怎么做？

11:02

They buy a domain.

他们会买一个 domain。

11:03

We have now, like there was a Japanese gang that recently came.

现在的情况是，比如最近来了一个日本帮派。

11:07

They brought like maybe 20,000 domains at recent.

他们最近大概弄来了 20,000 个 domain。

11:10

Like those are 20,000 different TLGs registered in different domain registrars.

就是那种 20,000 个不同的 TLGs，注册在不同的 domain registrars 里。

11:16

And we look for every single factor there.

我们会在那里检查每一个因素。

11:19

So they come in, it's a domain that was about two days ago.

所以它们进来时，这是个大概两天前才有的 domain。

11:22

So very new domains, they get bad reputation.

所以非常新的 domain，reputation 会很差。

11:27

And you look at the website, is it returning to 100 or basic things like that?

然后你看一下网站，它是不是恢复到 100，或者类似这种基础的东西？

11:32

Some people don't even set that up, or they do like a look alike domain.

有些人甚至根本没设置那个，或者他们搞一个看起来很像的 domain。

11:37

So they're like, oh, maybe I'll do another domain for code outreach.

所以他们会想，哦，也许我另做一个 domain 来做 code outreach。

11:42

That's terrible because it looks like you're fishing, right?

那很糟糕，因为看起来就像你在搞 fishing，对吧？

11:45

Because you're using a domain that looks like the same.

因为你用的 domain 看起来是一样的。

11:48

So all these little things, and they just do, they send a lot of emails all at once,

所以所有这些小细节，他们就是会一次性发一大堆 email，

11:53

and then they never send again.

然后就再也不发了。

11:54

So you kind of got to reverse engineer how fishing works and build a platform that does

所以你得 reverse engineer 一下 fishing 是怎么运作的，然后建一个能做到这个的平台。

12:01

exactly the opposite of that.

完全相反。

12:03

Do you have a list somewhere, or do you recommend people go to some well-known lists?

你们哪里有清单吗，还是建议人们去看一些知名清单？

12:07

Yeah, what I've learned is, again, another thing I didn't know, email is the type of

是啊，我学到的是，又一件我原先不知道的事，email 是那种

12:13

world where it requires a lot of education, but people are not interested in learning,

需要大量教育，但人们没兴趣去学习的领域，

12:18

right?

对吧？

12:19

Like the platform should do all those things for them.

就像平台应该替他们把那些事都做了。

12:22

So we added like insights inside the product, so people can just apply the insights.

所以我们在产品里加了类似 insights 的东西，这样人们直接应用这些 insights 就行。

12:27

So we look at sending patterns, and now with the eyes pretty cool too, like we have link

所以我们看发送模式，现在有了 eyes 也挺酷的，比如我们有 link

12:31

checkers and spell checkers, like all these checkers before you send.

checkers 和 spell checkers，就像你发送前用的所有这些 checkers。

12:36

So yeah, it's a mix of your teaching, your audience, but at the same time, the product needs

所以是的，这是你的讲解和你的受众的一种混合，但与此同时，产品需要

12:41

to do the teaching, and the more you can bake in the best practices, the better.

承担起讲解的部分，而且你能把最佳实践内置得越多，就越好。

12:46

So one example, too, when you send an email, you can send an HTML version and a plain

再举个例子，当你发送邮件时，你可以发送一个 HTML 版本和一个 plain

12:50

text version.

text 版本。

12:52

Very early on, we were like, no watch, like this, the whole world is changing, there

很早的时候，我们当时就说，不，看，像这样，整个世界都在变，而且

12:57

are way more agents than before.

现在的 agents 比以前多多了。

13:00

Let's just generate, auto-generate the plain text version of each email.

我们就直接生成，auto-generate 每封邮件的 plain text 版本吧。

13:04

We can extract from the HTML the markdown and then just send that.

我们可以从 HTML 里把 markdown 提取出来，然后直接发那个就行。

13:09

So we start sending that automatically, and that increased deliverability a ton.

于是我们开始自动发送它，这让 deliverability 提升了一大截。

13:13

You know, it's so much easier for spam filters to just look at the content directly without

你知道，spam filters 直接看内容要容易得多，不用

13:18

having to parse that.

去 parse 它。

13:20

So those things that we're like, oh no, let's bake into that product in.

所以那些我们会说“哦不，把这个直接做进产品里”的东西。

13:24

Most people don't do that, but we're like, no, that doesn't make sense.

大多数人不这么做，但我们会说，不，那说不通。

13:27

But with plain text, there's no tracking, right?

但用 plain text 的话，就没有 tracking，对吧？

13:28

There's no tracking.

没有 tracking。

13:29

Because you try to embed the image.

因为你会尝试 embed 图片。

13:31

That's right.

没错。

13:32

So you don't have open rates.

所以你们没有 open rates。

13:33

Yeah.

对。

13:34

How do you know you have deliverability if you don't have open rates?

如果你没有 open rates，你怎么知道你的 deliverability 如何？

13:36

Yeah, this is very interesting because the industry, there are two ways of tracking.

对，这很有意思，因为在这个行业里，追踪方式有两种。

13:41

You have open tracking and click tracking.

有 open tracking 和 click tracking。

13:44

Open tracking, what it does, it includes a one by one gift image, a transparent image,

Open tracking 的作用是，它会包含一张 1x1 的 GIF 图片，一张透明图片，

13:50

and then if that image is loaded, then you detect that as open.

然后如果那张图片加载了，你就把它判定为 open。

13:55

The problem with that is that all the changes that Apple did recently in terms of transparency

问题在于，Apple 最近在透明度方面做的所有改动

14:01

and data privacy, they read the emails before it lands.

以及 data privacy 方面，他们会在邮件到达之前先读取。

14:07

If they're doing the summaries and all those kinds of things, so the open rate as a metric,

如果他们做摘要和所有那些类似的事情，那么 open rate 作为一个指标，

14:13

as a whole, isn't really as reliable as it used to be because all these spam filters

整体上已经不像以前那么可靠了，因为所有这些 spam filters

14:18

are checking that before.

都会提前检查。

14:21

So then you rely on click tracking.

所以你就得依赖 click tracking。

14:23

Now click tracking can also be difficult because you want the URLs to match the sending

现在 click tracking 也可能很难，因为你会希望 URLs 和发送时匹配。

14:30

domain.

domain。

14:31

You also want the images URLs to match the sending domain.

你还希望图片 URL 和发送 domain 匹配。

14:35

That's another pattern that an abuser will do.

这是恶意用户会用的另一种模式。

14:38

They would host images somewhere else, point links somewhere else, and then they are sending

他们会把图片托管在别处，把链接指向别处，然后却从

14:42

from this domain.

这个 domain 发送。

14:43

Yeah.

对。

14:44

I have used email services that would rewrite, like for the link tracking, right?

我用过一些邮件服务，它们会做 rewrite，比如为了 link tracking，对吧？

14:48

They would rewrite to their own domain.

它们会 rewrite 到自己的 domain。

14:50

Yeah.

嗯。

14:51

You're saying that's bad?

你是说那样很糟？

14:52

That's bad because what happened, we did that too at first, and then that's normal.

那很糟，因为当时的情况是，我们一开始也那么做过，后来那就变得很正常了。

14:56

Yeah.

嗯。

14:57

But what's bad is when you have a noisy neighbor problem too, where if one of the customers

但糟糕的是，你还会遇到 noisy neighbor problem，就是如果某个使用 click tracking 的客户是个滥用者，发了一大堆 phishing，那个 domain 可能会被列进 blacklist，然后可能要几天甚至几周才能从 blacklist 里出来。

15:03

using that click tracking is an abuser and sends a bunch of phishing, that domain might

15:09

get listed on a blacklist and then it takes maybe days or weeks to get out of the

15:14

blacklist.

15:15

Then you're screwed as a business.

那作为企业你就完蛋了。

15:16

Is that not pretty common?

这不是挺常见的吗？

15:18

You make it very easy to just sign up, free tier, send a bunch of you.

你让人很容易就能注册，free tier，然后发送一堆你。

15:22

I would assume you have a lot of spam, a lot of phishing, but you're able to avoid, and

我会以为你会有很多 spam，很多 phishing，但你能避免，而且

15:27

this is why.

这就是为什么。

15:28

Yeah, I think what was different to me, and then this is more like a philosophy, right?

是啊，我觉得对我来说不同之处在于，然后这更像是一种哲学，对吧？

15:33

So I saw those other products and the way they behaved, and I felt like, okay, it's

所以我看到了那些其他产品以及它们的行为方式，我感觉，好吧，这

15:39

almost like if you were born in a society where you first go to prison and then you have

几乎就像你出生在一个社会里，你先去监狱，然后你有

15:44

to spam like 18 years in prison, you've got to prove that you're a good citizen, and

发spam就像要坐18年牢，你得证明自己是个好公民，然后你出狱了，现在你可以做你想做的事了。

15:48

then you get out of prison, and now you can do the things you want to do.

我一直觉得，不，我想生活在一个我生来自由的社会里，如果我犯了错，那我可能只会受到一点点惩罚，或者如果我做了非常糟糕的事，就会受到严厉的惩罚，对吧？

15:53

I always felt like, no, I want to live in a society where I'm born free, and if I make

所以我对那件事也是这么感觉的。

15:58

a mistake, then I might be punished just a little bit, or if I do something like really

你进来，我们默认信任你。

16:03

bad, a bad punishment, right?

当然，这又是一种天真的想法，但你知道吗，我们会信任你的。

16:05

So that's how I feel about that, too.

16:07

You come in, we trust you by default.

16:10

Of course, again, a naive way of thinking, but you know what, we're going to trust you

16:15

by default versus make the whole platform worse for everyone because of those bad actors,

默认情况下，还是因为那些恶意行为者，让整个平台对所有人都更糟，

16:22

right?

对吧？

16:23

So we trust you by default, but then what we do differently from others is like at

所以我们默认信任你，但我们跟别人不一样的地方在于，就在

16:26

the very first moment, like when we notice something is off, like by looking at the

最开始的那一刻，比如当我们发现哪里不对劲，比如看

16:31

bounce rates, complain rates, all the feedback we get from mailbox providers, then we block

bounce rates、complain rates，以及我们从 mailbox providers 收到的所有反馈，然后我们就封禁

16:36

you like very quickly, and then with Ciao, if you would try to understand, and most of

你，非常快，然后换成 Ciao 的话，如果你试着去理解，大多数

16:40

the time is just misuse, it's not abuse.

时候只是 misuse，不是 abuse。

16:43

So we learned that we need to differentiate, yeah, I think we have skill issues, we wouldn't

所以我们学到的是，我们需要做出区分，嗯，我觉得是我们能力不行，我们不会

16:51

have expected.

本来就预料到了。

16:52

Yeah, you know, it's so important, and it's so dirty, and it's so secretive, and it's

是啊，你知道，这事太重要了，又那么脏，还那么神神秘秘，简直就是那种……我几乎就想让你干脆把一切都告诉我们。

16:57

just such a, I almost just want you to like just tell us everything.

我得说，你有个挺有意思的页面，就是你说的那套哲学，我光是在这上面到处读就读晕了，然后还有一个像是呼吁的东西，你承诺要让 email 更安全，如果你对这些东西感兴趣的话，你知道，你这里内容还挺多的。

17:01

I will say, you have a fun page, literally your philosophy as you call it, I got lost

是啊。

17:07

just reading around this, and then there's like a call, you have a pledge that you will

17:11

make email safer, like if you're interested in this stuff, you know, you have quite a bit

17:16

up here.

17:17

Yeah.

17:18

Okay, so actually, I do want to, there's a how we scale support there.

好，其实我确实想，那个，就是我们在那边怎么 scale support 的问题。

17:22

I think that's, this is a, you remind me of another investment that I had Jake from

我觉得那是，这是个，你让我想起我之前另一个投资，Jake 来自

17:26

real way.

17:27

Yeah.

是啊。

17:28

Because also another like very, very permissive, very developed oriented, but also you have

因为另外那个也像是，非常非常宽松，非常开发者导向，但你也得

17:32

to skill support.

去做 skill support。

17:33

Yeah, right?

是啊，对吧？

17:34

Because the moment someone gets blocked, they come to you and say, hey, like, you know, don't

因为一旦有人被卡住了，他们就会来找你说，嘿，就那个，你知道的，别

17:38

keep me off your platform.

别让我上你们的平台。

17:39

Yeah.

是啊。

17:40

How do you skill support?

你们怎么搞 skill support？

17:41

Man, I think we, if support was a software, we're probably on major version four or something.

哥们，我觉得，如果 support 是一款软件，我们大概已经到 major version 4 之类的了吧。

17:48

Okay.

好。

17:49

Like the very first version was like, you meet me, and I was so stressed out, I'm like,

就像最早那版，你跟我一见面，我压力特别大，我就想，

17:55

dude, I just need someone to be sitting on the chair, like, it's that simple.

哥们，我就需要有人坐在那把椅子上，就这么简单。

18:00

I cannot be doing this anymore.

我真不能再干这个了。

18:02

And it was like just anyone, anyone, I don't care.

而且就像随便谁都行，谁都可以，我根本不在乎。

18:06

Then you do that.

然后你就这么干了。

18:07

Of course, it doesn't go well because you realize very quickly, you need someone that will

当然，这事不会顺利，因为你很快就会意识到，你需要一个

18:12

actually try to prevent the tickets from happening in the first place.

真的会从一开始就设法防止这些 tickets 发生的人。

18:16

So then we went to V2, which was like a new version, like a new person that we, that

所以后来我们换到了 V2，就像一个新版本，也像是一个我们、我们

18:21

we ended up hiring that was like, okay, we're answering the tickets.

最终招进来的人，他的态度就像是，行，我们来处理这些 tickets。

18:25

They have domain expertise, but they were still like not doing the proactive work of like,

他们有领域专业知识，但还是没做那种主动的工作，比如，

18:30

no, let's make sure this is documented before we talk about.

不行，我们先确保这件事被记录下来，再讨论。

18:33

So one policy that we, we set, it was pretty cool actually is like, you can only respond

我们定过一个 policy，其实挺酷的，就是：你只能用一个 link 来回复 ticket。

18:39

to a ticket with a link.

所以你不能直接把答案给出来。

18:41

So you cannot just give the answer right away.

你得先停下来，写一篇 knowledge based article，写一份 docs，然后，所以你回复一个 ticket 可能要花一个半小时左右。

18:45

You had to stop, write a knowledge based article, write a docs, and then, so it takes like

但你猜怎么着？

18:52

a hour and a half maybe for you to answer a ticket.

现在你又把那个 link 到处发。

18:55

But guess what?

所以那真的超酷，而且它帮我们建起了那个 knowledge base。

18:56

Now you're sending that link all over again.

18:58

So that was super cool and it helped us build that knowledge base.

19:01

But even so like after that, when we crossed maybe one million users, so we had one million

但即便如此，就是在那之后，当我们跨过大概一百万用户的时候——我们去年十二月有一百万用户，现在有四百万。

19:07

users back in December, and now we have four.

所以我不能靠堆人来解决这个问题，但那样根本解决不了，对吧？

19:11

So I can't throw people at the problem, but that's just not going to solve it, right?

就像我只是不停加人手，情况就是这样，不行，所以我们真的得重新想，好吧，一切从头再来。

19:15

Like I'm just going to keep staffing and that's how it is like, no, so we really have to

而且我们想保留，就是，那个 free tier 不是……我不想把那个 free tier 去掉。

19:21

rethink like, okay, everything from scratch again.

我一开始当工程师，就是因为我能用那些免费产品，对吧？

19:25

And we want to keep like that free tier is not, I don't want to get rid of that free

19:29

tier.

19:30

I started as an engineer because I was able to use those free products, right?

19:33

Like we all, like if it wasn't for Heroku's free tier and all these other products,

就像我们所有人一样，如果不是因为 Heroku 的 free tier 和所有这些其他产品，

19:38

like we wouldn't be here.

我们也不会走到今天。

19:39

So I want to keep recent that way, but it took as many iterations to, yeah, to get to

所以我想让 recent 保持那种方式，但确实花了很多次 iterations，对，才走到

19:45

it.

这一步。

19:46

I imagine this is like, okay, it's a domain people don't know a lot about.

我想这就像是，好吧，这是一个人们了解不多的 domain。

19:49

There are actually unique questions that need human support versus, you know, Command

实际上有些独特的问题需要人工支持，而不是，你知道，Command

19:54

K with just, you have docs, you have the model, customer service plots exist, right?

K 那种，只要你有 docs、你有 model，客服 plots 是存在的，对吧？

19:59

Yeah.

对。

20:00

I think right now is much better.

我觉得现在好多了。

20:02

There are a few things where man, there's so much friction, DNS, oh my gosh, that's

有几件事，天哪，摩擦太大了，DNS，我的天，那

20:08

always DNS.

永远是 DNS。

20:09

It's always DNS, right?

永远是 DNS，对吧？

20:11

Like the majority of the questions we get today is just like people not able to set

比如我们今天收到的大多数问题，就是人们没法把

20:16

that up and it's fascinating because we're serving a new audience, right?

那个设置好，而且这挺有意思的，因为我们在服务一群新受众，对吧？

20:21

It's an audience that doesn't even know how to troubleshoot.

这是一群甚至不知道怎么排查问题的受众。

20:25

Like I remember last year when lovable came and bold and VZero, I could tell these are

比如我记得去年 lovable 出现的时候，还有 bold 和 VZero，我能看出来这些是

20:32

different users.

不同的用户。

20:34

Like they don't know when they have a problem, they don't know how to go past that.

就像他们不知道自己什么时候遇到问题，也不知道该怎么解决。

20:39

And with DNS, that was always the case of like, yeah, they just simply don't know what

而在 DNS 这块，情况一直都是这样：是啊，他们就是压根不知道

20:44

is a CNAME.

CNAME 是什么。

20:45

What is an A record?

A record 是什么？

20:46

TXT?

TXT 呢？

20:47

Where do I put those things?

这些东西我该填到哪里？

20:48

So they would paste the wrong value or silly things.

所以他们就会粘贴错误的值，或者一些很傻的东西。

20:51

By the way, right now I use computer use like codex or cloud code and it just completely

顺便说一句，现在我用 computer use，比如 codex 或 cloud code，它就完全

20:57

manages my DNS records.

管理我的 DNS records。

20:58

It's beautiful, right?

这太漂亮了，对吧？

21:00

It debugs it.

它会去 debug 它。

21:01

It like digs, you know, and tries all the DNS stuff.

它会去 dig，你懂的，把各种 DNS 的东西都试一遍。

21:04

Like I don't have to worry about it.

就像我完全不用操心它。

21:05

Yeah.

对。

21:06

One thing that we did, like I found this super obscure protocol that like very few people

我们做过的一件事是，比如我发现了一个特别冷门的 protocol，几乎没多少人

21:12

in the industry were like investing on, it's called domain connect.

业内当时在投的，叫 domain connect。

21:16

And it's super interesting, by the way.

顺便说一句，这特别有意思。

21:18

And it's something that started like from some domain providers.

而且它最早是从一些 domain 提供商那边起来的。

21:23

And oh, yeah, yeah, yeah, this is like the OAuth for domains, right?

哦，对对对，这就像是给 domain 用的 OAuth，对吧？

21:27

Exactly.

没错。

21:28

Yeah.

对。

21:29

This is like Cloudflare oriented.

这有点像是以 Cloudflare 为导向的。

21:30

Yeah.

对。

21:31

So some people have like Cloudflare is there, then we convinced we're still to be there

所以有些人那边已经有 Cloudflare 了，然后我们又说服自己，我们也还得在那儿。

21:35

as well.

然后我们就开始拉更多人来。

21:36

And we started like getting more people.

所以这有两边，对吧？

21:38

So you have the two sides, right?

所以 DNS providers、service providers，比如我们自己就是 service providers，然后 DNS providers 在做实现。

21:40

So DNS providers, the service providers, we are the service providers, for example, and

但这行业真是疯狂，疯狂得很。

21:45

then the DNS providers are implementing.

就像 go daddy 为这个收一大笔钱。

21:48

But it's a crazy, crazy industry.

21:52

Like go daddy charges like just a lot of money for that.

21:56

It's so weird.

这太奇怪了。

21:58

Like Jonas, it took us like the first one, it took us like months to get after that person

就像 Jonas，第一个我们花了好像好几个月才找到那个人

22:04

too.

也是。

22:05

Like no one really maintains the protocol and things like that.

就像根本没人真的在维护这个 protocol 之类的。

22:08

But it's beautiful like having a protocol that you can just do an OAuth for domains.

但很棒的是，有一个 protocol，你可以直接对 domains 做 OAuth。

22:13

Yeah.

是啊。

22:14

So for listeners, if you're not familiar, if you haven't tried it, something like that,

所以对听众来说，如果你不熟悉，或者还没试过类似的东西，

22:18

it's literally like you do use plaid or use like OAuth, you just only click buttons, you

就真的跟你用 plaid 或者用 OAuth 一样，你只要点按钮，你

22:22

don't type in anything, there's nothing to get wrong, you basically just like sign in.

什么都别输入，不会出错的，你基本上就是登录一下。

22:26

And as long as you trust this thing called domain connect, which you have confirmed that

而且只要你信任这个叫 domain connect 的东西，你已经确认过

22:29

it is legit, then like it's trust.

它是靠谱的，那基本上就是信任了。

22:31

I literally, I used it for the first time, like two weeks ago, I was like, what is this?

我真的，我大概两周前第一次用它，我当时就想，这是什么？

22:34

I've never seen it before.

我以前从来没见过。

22:36

Yeah.

嗯。

22:37

So it's really cool.

所以真的很酷。

22:38

One of the things that I'm trying to figure out as well is how you like start this company.

我还在琢磨的一件事也是，你到底是怎么把这家公司办起来的。

22:43

Like you said, you tried to get in touch with the Ionus guy.

就像你说的，你试着联系过 Ionus 那个人。

22:45

I'm like, I wouldn't dare to start recent without knowing, like having like some guy on

我当时就想，我要是不知道情况，可不敢贸然开始最近这个事，就像得有个

22:51

the inside of Gmail.

Gmail 内部的人似的。

22:52

Even now it's like, okay, can we vibe code our own email thing?

就算是现在，也会想，好吧，我们能不能 vibe code 一个自己的邮件东西？

22:56

Yeah.

是啊。

22:57

Clearly no, but that's, that's even with tools today.

显然不行，但这事，这事就算放到今天有这些工具也一样。

23:00

You did this years ago, like I, you know, you want people at domain networks like you need

你几年前就干过这事了，就像我，你知道的，你想在 domain networks 里有人，就像你需要

23:04

the guy, right?

那个人，对吧？

23:05

Yeah.

对。

23:06

Yeah, pick up the phone.

对，拿起电话打过去。

23:07

Yeah.

对。

23:08

There were many times that we had to pick up the phone.

有很多次，我们都不得不拿起电话打过去。

23:11

It's fascinating.

这挺有意思的。

23:12

Like it's an industry that relies on relationship a lot.

你知道，这个行业非常依赖关系。

23:16

And we didn't have those relationships.

而我们没有那些关系。

23:18

So we hired for a few key roles, like deliverability roles, folks that have been at postmark since

所以我们招了几个关键岗位，比如 deliverability 岗位，还有那些自从……就一直在 postmark 的人

23:24

the beginning or 15 years ago at MayoChamp, and then we reliverged that network.

最开始，或者说15年前，在MayoChamp，然后我们又重新利用了那个network。

23:30

But yeah, it's pretty crazy.

不过是啊，这挺疯狂的。

23:32

But what's interesting now, of course, I didn't know that when I started.

但现在有意思的是，当然，我刚开始的时候并不知道这一点。

23:36

Like, I could never predict what's going to happen there.

就像，我根本没法预测那里会发生什么。

23:39

You didn't do research?

你没做研究吗？

23:40

Yeah.

是啊。

23:41

No.

没有。

23:42

Like I could never predict, like, okay, what AI was going to do and all that, right?

就像，我根本没法预测，好吧，AI会做什么之类的，对吧？

23:45

But I think we are in a space where me and I'm glad we chose infrastructure.

但我觉得我们现在处在这样一个领域，我很高兴我们选了 infrastructure。

23:50

I'm glad we're infrastructure as a service and not a just, yeah, software as a service.

我很高兴我们做的是 infrastructure as a service，而不只是一个，对，software as a service。

23:54

You had a little bit, right?

你之前也稍微接触过一点，对吧？

23:55

You had, you made React email the year before, so you did dip your toes in open work.

你之前，你前一年做了 React email，所以你确实在 open work 里试过水。

24:01

Yeah.

对。

24:02

And then came to come.

然后 came to come 就来了。

24:03

Yeah.

对。

24:04

React email is our next yes, right?

React email 是我们的下一个 yes，对吧？

24:06

So we wanted like an open source channel for top of the funnel.

所以我们想要一个 open source 渠道，来做 top of funnel。

24:11

And I remember doing some things very intentionally.

我记得我当时有些事是特意去做的。

24:13

If you go to the docs, for example, you'll find there are docs about how do you set this

比如说你去看文档，会发现里面有文档讲怎么用 send grid、MayoChamp 和 MayoGun 把它搭起来。

24:19

up with send grid and MayoChamp and MayoGun.

那些文档里的每一篇都是我写的。

24:23

I wrote every single one of those docs over there.

我就是想说，好吧，这是给所有人的，我们是在试着重新定义这件事该怎么做。

24:27

And I just want to like, okay, this is for everyone and we're trying to redefine how

那是一种策略，就是，我们要给、给、给。

24:32

this is.

24:33

It was a strategy of like, we're going to give, give, give.

24:36

And then we ask.

然后我们就去问。

24:38

And if people are Gary Vee, yeah.

如果这些人是 Gary Vee，对。

24:41

And if people are like, building the, like the email template using React, like now do

如果大家是在用 React 搭那个 email template，那现在你需要一个平台来发它。

24:47

you need a platform to send it.

他们在哪儿做这个？

24:49

Where do they do that?

这是最近的情况。

24:50

Here's recent.

所以那是一种，有点像讲故事的方式去切入市场。

24:51

So it was a, like a storytelling kind of way of approaching the market.

就像我们要解决 journey 的第一段，也就是搭 template。

24:56

Like we're going to fix the first part of the journey, which is building the template.

25:01

And then once you have the template, now that that problem is fixed, you need a platform

然后一旦你有了模板，既然那个问题已经解决了，你就需要一个平台来发送它。

25:05

to send it.

平台就是这个。

25:06

Here's the platform.

但你开始做 React email 的时候就已经有 recent 的想法了吗，还是它本来是一个网站项目，后来才变成一家公司？

25:07

But did you have the idea for recent when you started React email or was it a site

我们创办了 recent。

25:10

project that then became a company?

而且在打造 recent 的过程中，我做了很多用户研究，大家会

25:13

We started recent.

说，哎呀，写模板太痛苦了，我拿到那个，那个 figma 文件是从

25:15

And as we were building recent, I was like doing a lot of user research and people were

25:19

saying, like, oh, the pain of writing the template, I get the, the figma file from

25:23

my designer.

我的设计师。

25:24

It sucks.

太烂了。

25:25

Like it's HTML tables and everything.

就像全是 HTML 表格之类的。

25:28

They would cut the images, right?

他们会切图，对吧？

25:30

Yeah.

是啊。

25:31

Oh, my gosh.

我的天哪。

25:32

Because there's nothing.

因为什么都没有。

25:33

And then there's like a special syntax of HTML that only works on email.

然后还有一种只在邮件里生效的特殊 HTML 语法。

25:37

Yeah.

嗯。

25:38

Right?

对吧？

25:39

Or CSS?

还是 CSS？

25:40

Or CSS?

还是 CSS？

25:41

Yeah.

嗯。

25:42

Or CSS?

还是 CSS？

25:43

Yeah.

嗯。

25:44

And then like for outlook, you have like some special comments and then you got a, it's

然后比如说对于 outlook，你会有一些特殊的注释，然后你还有一个，它是

25:48

such a mess.

真是一团糟。

25:49

It's like a, it's a mess.

就像是，呃，一团糟。

25:51

And so to me, the reason that React works is that it is basically a compiler.

所以对我来说，React 能起作用的原因在于，它基本上就是一个编译器。

25:55

It's like, well, somebody writes this once and then as long as you fit the API, it's going

就像，嗯，有人把它写一次，然后只要你对上 API，它就能到处都跑得通。

25:59

to work everywhere.

而且你可以从 Tailwind 里 import 你的 tokens。

26:00

And you can import your tokens from Tailwind.

现在你用 TypeScript 就有了一种 type definition。

26:03

Now you have a type of definition with TypeScript.

所以不用再有 hay、空逗号，你懂的，就像你邮件里那样，你其实可以真正拥有

26:06

So no more hay, empty comma, you know, like on your emails, like you can actually have

26:12

like a string validation or whatever you want.

比如 string validation 之类的，或者你想怎么叫都行。

26:14

Yeah.

对。

26:15

I think one of the initial people that push back on React email is like, well, you're not

我觉得最早质疑 React email 的人里，有一个就说，嗯，你其实并没有

26:18

actually using React as no state.

真正以没有 state 的方式在用 React。

26:20

Yeah.

对。

26:21

And like, yeah, React is supposed to be a, views as a function of state.

然后就像，对啊，React 本来应该是，views 是 state 的函数。

26:24

So it's mostly JSX is templating.

所以主要就是 JSX 在做 templating。

26:26

Yes.

是的。

26:27

Like animations, is there interactivity?

像动画那样，有交互性吗？

26:30

No.

没有。

26:31

Expanding like audience, no, nothing like that.

扩大受众之类的？不，没有那种东西。

26:33

So you cannot ship JavaScript on an email.

所以你不能在邮件里塞 JavaScript。

26:36

So then what you do is, yeah, we basically use the template.

所以你要做的就是，对，我们基本上就是用 template。

26:40

It's turning complete.

它是 Turing complete。

26:41

Yeah.

对。

26:42

Yeah.

对。

26:43

Okay.

好的。

26:44

Okay.

好的。

26:45

So cool.

太酷了。

26:46

I actually do want to like continue on this email path, but I see so many names I used

其实我真的很想继续顺着这封邮件聊下去，但我看到好多我以前认识的名字。

26:50

to know.

我就是有点想问，你能不能稍微迁就我们一下，带我们过一遍其中一些名字。

26:51

I was just kind of wondering if you can indulge us a bit and just go through like some

嗯。

26:55

of these names.

26:56

Yeah.

26:57

And from them, if you have any like cool stories, if just just anything, you know, me and so

然后从他们那儿，如果你有什么很酷的故事，哪怕只是随便什么都行，你知道，我这边还有好多。

27:02

much.

是啊。

27:03

Yeah.

太多了。

27:04

So much.

我记得 recent 的第一年，我当时想，我见到了 Mailgun 的创始人，因为我那时还是 Mailgun 的用户。

27:05

I remember the very first year of recent, I was like, I got to meet the founder of Mailgun

然后我和他聊了聊，呃，他，他创建了，我的天哪。

27:10

because I was still a user of Mailgun.

Blogger、medium。

27:13

And I talked to have, um, he's, he built, oh my gosh.

27:18

Blogger, medium.

27:19

Um, no, it's a, it's a different product.

呃，不是，那是一个，那是一个不同的产品。

27:23

Um, but anyways, like him, it's a security product.

呃，但不管怎样，跟他一样，那是个安全产品。

27:27

Uh, that, that he's building now.

呃，那个，那个他现在正在做的。

27:29

And I remember talking to him.

我记得跟他聊过。

27:30

I was like, man, why'd you sell so quickly your company or what was the thing?

我就说，老兄，你怎么这么快就把公司卖了，或者到底是怎么回事？

27:35

And he told me like, oh yeah, 90% of our R&D was justified abusers.

他跟我说，哦对，我们 90% 的 R&D 都是有正当理由的滥用者。

27:41

So it was a, it's a terrible business.

所以那是个，那是个糟糕的生意。

27:43

Like you shouldn't go there.

就是你不该碰这个。

27:45

And I was like, uh, like, I don't want to live in that world at all, uh, so I was very worried

我当时就想，呃，像，我完全不想生活在那样一个世界里，呃，所以我特别担心这件事。

27:50

about that.

呃，这对我来说是个很大的教训，就是，好吧，我们得把这东西自动化。

27:51

Uh, and it was a big lesson for me of like, okay, we got to automate this thing.

就，我们不可能只是去建一个庞大的组织，呃，因为，我们现在确实有很多人对 safety 感兴趣，但这不像是会成为我们的核心业务。

27:56

Like there's no way we're just going to build like a massive org, uh, for, we do have

不，我们要，我们必须推动整个行业往前走。

28:02

a lot of people interested in safety now, but it's not like that's going to be our

但那是其中一个，而 send grit 在那些里面一直是我最喜欢的，而且一直

28:05

core business.

28:06

No, we're going to, we have to be pushing the industry forward.

28:10

But that was one and send grit has always been my favorite out of those ones and always

28:16

been the behemoth, like the, the biggest, the, it's like, oh my gosh, how are we ever

一直是那个巨无霸，就像，那个最大的，感觉就像，天哪，我们到底要怎么才能赢过它？

28:22

going to beat?

这根本不可能。

28:23

It's just impossible.

然后今年早些时候，我们从 San Grid 的 downloads 突破了一个数字，我当时就想，这也太疯狂了。

28:25

And then earlier this year, we crossed a number of, uh, downloads from San

而现在我们已经翻了一倍还多。

28:29

Grid and I was like, that's insane.

所以，就 unique users 来说，那个，那个数字是多少？

28:31

And now we're more than double.

然后，嗯，downloads，你知道，就是，呃，SDK，呃，那是最好的 metric 吗？

28:33

So in terms of unique users, so what's the, what's the number?

28:35

And yeah, I'm downloads, you know, it's, uh, SDK, uh, is that the best metric of it?

28:40

Yeah.

嗯。

28:41

Like I think there, there's something about that metric in terms of adoption, but then

我觉得吧，就采用率这个指标来说，确实有点意思，但接下来

28:46

if you look at how many emails they sent is obviously much more than us because they

如果你看他们发了多少邮件，那显然比我们多得多，因为他们

28:50

have like the big enterprise customers yet in, yeah, that kind of thing.

有那种大企业客户，嗯，就那种情况。

28:54

I think one thing that's surprising is the whole point of the Twilio San Grid acquisition

我觉得有一点挺让人意外的，就是 Twilio San Grid 这笔收购的重点

29:00

is that I now know you are no longer an email service.

就是我现在知道了，你们不再是一个邮件服务了。

29:03

You're a communications platform, right?

你们是一个通信平台，对吧？

29:05

You, you integrate across everything and it's not, it's not just about cross sell.

你们，你们是把所有东西都集成起来了，而且这并不，并不只是 cross sell 的事。

29:09

It is about, I can reach you across multiple channels, something like that.

大概是说，我能通过 multiple channels 联系到你，之类的。

29:12

Yeah.

是啊。

29:13

Or that you have the size, you know, Twilio's like, I don't know what the size is now.

或者说你有那个规模，你知道，Twilio 大概，我也不知道现在规模多大了。

29:17

Let's call it 13, 15 billion dollar company.

就说它是家 130 亿、150 亿美元的公司吧。

29:20

You should have all the relationships, all the talents, all these, all these things, right?

你应该有所有的人脉、所有的人才、所有这些，所有这些东西，对吧？

29:23

So how do you compete with it?

那你怎么跟它竞争？

29:26

Yeah.

是啊。

29:27

Man, I, I admire Twilio so much.

老兄，我，我太佩服 Twilio 了。

29:30

They kind of invented dev rel or, yeah, ask your developer, you know, ask your developer.

他们算是发明了 dev rel，或者说，对，ask your developer，你知道的，ask your developer。

29:35

Like, and now when I, I, I, I go past that one on one billboards and I see a customer

就比如，现在当我，我，我，我经过那个一对一广告牌，然后看到一个客户 AI

29:41

AI or whatever, I'm pissed.

还是什么的，我就很火大。

29:44

I'm like, why, like, you neglected the thing that got you there.

我就想，为什么啊，你把让你走到今天的东西给忽略了。

29:50

And I think now, like, Jeff is not there anymore.

而且我觉得现在，Jeff 已经不在那儿了。

29:54

The CFO is...

CFO 是……

29:55

He bought the onion.

他买下了 The Onion。

29:56

Yeah.

是啊。

29:57

That's right.

没错。

29:58

That's right.

没错。

29:59

So I feel like they're just sleeping on it and first they did that with San Grid and

所以我觉得他们就是没把这事当回事，而且他们最早对 San Grid 就是这么干的，然后

30:04

they're doing that with SMS, too.

他们对 SMS 也是这么干的。

30:06

And yeah.

嗯，对。

30:07

I think my, my lesson here is I'm trying to figure out the economics of scale of this industry,

我觉得我、我在这里的体会是，我想搞清楚这个行业的规模经济，

30:12

right?

对吧？

30:13

Like, clearly, you know, let's, let's call it there 100 times larger than you when you

就像，很明显，你知道，咱们就、就说它比你大 100 倍吧，当你

30:17

used to, when you just got started, right?

以前，当你刚起步的时候，对吧？

30:18

Yeah.

是啊。

30:19

How come they cannot use their size to just win, right?

他们怎么就不能利用自己的体量直接赢呢，对吧？

30:22

Yeah.

是啊。

30:23

So in, in other words, maybe there is no benefits to scale.

所以，换句话说，也许 scale 并没有什么好处。

30:27

Maybe you can just, you know, you just make a better product than you can slowly.

也许你只要，你知道，你只要做出更好的产品，然后你就可以慢慢来。

30:30

For sure.

当然。

30:31

And I've always thought about you like, why did Twilio acquire San Grid and not the other

而且我一直想问你，为什么是 Twilio 收购了 San Grid，而不是反过来？

30:37

way around?

有办法绕过去吗？

30:38

Uh, Twilio was bigger.

呃，Twilio 规模更大。

30:40

I mean, SMS.

我是说，SMS。

30:41

Yeah.

对。

30:42

It's big.

它很大。

30:43

But like, if email and communication is a commodity, like SMS, your margins are even worse.

但你想，如果 email 和通信都成了商品，像 SMS 一样，那你的利润率就更糟了。

30:49

It's worse.

更糟。

30:50

Yeah, even worse.

对，甚至更糟。

30:51

So...

所以……

30:52

Have you looked into SMS?

你研究过 SMS 吗？

30:53

Yeah.

嗯。

30:54

For SMS?

SMS 方面的？

30:55

I think one day.

我觉得总有一天吧。

30:56

Yeah.

嗯。

30:57

Yeah.

嗯。

30:58

What are the channels that matter?

哪些渠道重要？

30:59

Email.

Email。

31:00

SMS.

SMS。

31:01

What's app?

What's app？

31:02

What's app?

What's app？

31:03

What's app?

What's app？

31:04

What's app is free, right?

What's app 是免费的，对吧？

31:05

You don't pay anything.

你什么都不用付。

31:06

And now business messages are going to be charged.

而现在商务消息要开始收费了。

31:11

So there is some friction there, for sure.

所以那儿确实有一些摩擦。

31:15

Phone calls?

电话通话？

31:16

Phone calls is a good one.

电话通话是个不错的例子。

31:17

I don't know what that's interested in.

我不知道那个对什么感兴趣。

31:19

I never take any verbal calls, right?

我从不接任何语音通话，对吧？

31:20

Yeah.

是啊。

31:21

That's true.

没错。

31:22

I'm less interested about that too.

我对那个也没那么感兴趣。

31:23

What I care is like, what are the things that makes me able to connect with my audience?

我关心的是，哪些东西能让我和我的受众建立连接？

31:29

Like that's what I ultimately care.

这才是我最终在乎的。

31:31

Like I don't care about email.

我不在乎电子邮件。

31:33

DMs.

私信。

31:34

Yeah.

对。

31:35

DMs.

私信。

31:37

It's hard because you cannot DM a big crowd.

这很难，因为你没法给一大群人发私信。

31:40

But Slack maybe is one, right?

但 Slack 也许算一个，对吧？

31:42

As you're Slack.

就像你们用 Slack 一样。

31:43

I have so many Slack connects and I need to share the news about this one new feature

我在 Slack 上有好多联系人，我得分享一下我做的这个新功能的消息。

31:49

that I built.

嗯。

31:50

Yeah.

Thailand。

31:51

Thailand.

我不知道你们用不用 Thailand。

31:52

I don't know if you guys use Thailand.

对。

31:53

Yeah.

我觉得他们在这种 Slack 支持功能上算是有点创新。

31:54

I think they kind of innovated this Slack support thing.

31:55

Yeah.

对。

31:56

But you can obviously trivially build your own.

但你显然可以很轻松地自己做一个。

31:58

Okay.

好。

31:59

And then I think the other one I want to touch on is AWS SES.

然后我想提到的另一个是 AWS SES。

32:01

Yeah.

对。

32:02

So my connection to this is I learned to code with Free Code Camp.

我和这个的关联是，我是通过 Free Code Camp 学会编程的。

32:06

Thanks.

谢谢。

32:07

You can see Larson and everything.

你能看到 Larson 之类的。

32:08

Yeah.

是啊。

32:09

They have like five million subscribers on their newsletter.

他们的 newsletter 大概有五百万订阅者。

32:11

Yeah.

是啊。

32:12

Sent through Amazon SES for like $5,000 a month.

通过 Amazon SES 发送，每月大概 5,000 美元。

32:14

Mm-hmm.

嗯。

32:15

Which is cheap, cheap.

这很便宜，很便宜。

32:16

I don't know how.

我不知道怎么做到的。

32:17

I don't know what you will cost.

我不知道你那边会花多少钱。

32:18

Yeah.

是啊。

32:19

Mm-hmm.

嗯嗯。

32:20

More than that.

不止是这样。

32:21

So basically like, to me it's just like the low cost player, right?

所以基本上，对我来说，它就像是那个低成本玩家，对吧？

32:23

Yeah.

是啊。

32:24

That is, I mean, it's Amazon.

也就是说，我是说，就是 Amazon。

32:25

It's like there.

就像，它就在那儿。

32:26

They're super established.

他们非常成熟。

32:28

What do you take from them and then what do you, what should people know to stay away from

你从它们那儿学到什么，然后你……大家应该知道什么才能避开它们？

32:32

them?

我们也是从 SES 起步的。

32:33

We started with SES too.

我们知道，好吧，如果我们要在这里创新，我想在 DX layer 上创新。

32:35

We knew that, okay, if we're going to be innovating here, I wanted to innovate on the DX

因为我觉得那是个服务不足的领域。

32:40

layer.

所以我就给每一种语言都配了 SDKs。

32:41

And because I felt like that was the underserved space.

我记得在 serverless 的时候，我们强调了 servers，因为好像没人有关于……的指南

32:44

So let me have SDKs for every single language.

32:48

I remember with serverless, we highlighted servers because like no one had like guides on

32:53

how to send an email with lambda or something.

怎么用 lambda 之类的发邮件。

32:55

So there's just this underserved space in DX.

所以 DX 里就有这么一个没被充分满足的空间。

32:59

So we started with like the most high level layer.

所以我们一开始是从最高层的那层开始做的。

33:03

And that served us really well, was able to like scale because we were using SES behind

而且这对我们特别有效，能 scale，因为背后用的是 SES。

33:08

the scenes.

结果发现，在你怎么组织 mains 这块儿限制特别多，而且他们还有 configuration sets 这个概念。

33:09

Turns out there's just a lot of limitations when it comes to the way you structure the

而且如果你用 SES，隐性成本在于，好吧，你可能只付 5K，

33:15

mains and they have this concept of configuration sets.

33:20

And the hidden cost if you're going with SES is that, okay, you might be paying only 5K,

33:25

for example, but now you need SNS for the events.

比如说，但现在你需要用 SNS 来处理这些 events。

33:30

So you need like all the feedback from bounces, complaints, open clicks, all that.

所以你需要像 bounces、complaints、open clicks 这些所有的反馈，所有那些。

33:35

And now so you get the events great.

现在你拿到了这些 events，很好。

33:38

Now you need cloud watch for the logs.

现在你需要用 cloud watch 来处理 logs。

33:40

Now you need a queue because it's going to be a lot of events.

现在你需要一个 queue，因为会有很多 events。

33:43

So now you add SQS.

所以现在你加上 SQS。

33:45

And now that you have SQS, so you start like adding all these different layers, oh, you

现在你有了 SQS，于是你开始加所有这些不同的 layers，哦，你想要 web hooks，好吧，现在你又得再建一个东西。

33:49

want web hooks, okay, now you've got to build another thing.

33:52

So if you go with SES, it's a great platform, it's fine, but you're going to notice

所以如果你用 SES，它是个很好的平台，没问题，但你会注意到，你的 footprint 很快会比你想的大得多，因为你需要它周围所有那些配套的东西。

33:59

like your footprint will be much bigger than you would imagine very quickly because you

对，我想说，你知道，邮件领域里又冒出来一个 YC startup，叫 Agent Mail，然后他们——我也跟那个人聊过，他说，对，我们不打算——我们打算一直构建到 VCS，而不是真的跟他们竞争。

34:03

need all the apparatus around it.

对。

34:07

Yeah, I would say, you know, another YC startup that came up in the mail areas, Agent

但对我们来说，现在情况正在改变。

34:12

Mail, and then they, I talked to the guy as well and he was like, yeah, we're not going

34:16

to, we're going to build until VCS and not really compete with them.

34:19

Yeah.

34:20

But for us, it's changing that now.

34:22

So we are, we're building the infrastructure and going down to the metal now.

所以我们现在是在搭 infrastructure，而且已经深入到最底层了。

34:27

And for us, it feels like we got to the point where now this makes sense to innovate

对我们来说，感觉我们已经到了这个阶段：现在来做创新是说得通的，

34:32

on the, the most basic layer.

就是在最基础的那一层。

34:35

And now we can get better footprints, better deliverability if we just focus on that.

现在只要我们把重点放在这上面，就能拿到更好的 footprints、更好的 deliverability。

34:40

But it's a big project.

但这是个很大的项目。

34:41

Like we've been working for months now and we're going to announce something soon about

我们现在已经做了好几个月了，而且很快会宣布一些东西，关于

34:45

the problem.

这个问题。

34:46

What is the hardest thing to, is it like trusted IPs?

最难的是什么，是 trusted IPs 之类的吗？

34:50

Is it compute in data centers or what, what is the blocker?

是数据中心的 compute 吗，还是什么，卡点到底在哪？

34:54

Yeah, definitely IPs you're buying like a whole IP block and then now you're managing

对，肯定是你买的那些 IP，比如直接买一整个 IP block，然后现在你得自己管理

35:00

the IP pulls yourself.

这些 IP pulls 本身。

35:02

That's one.

这是其中一个方面。

35:03

And then you're almost doing like traffic control and you're controlling the fruit put into

然后你几乎是在做类似 traffic control 的事，还要控制投放到

35:09

each mailbox provider.

每个 mailbox provider 的吞吐量。

35:11

And then you need the relationships with the providers.

然后你还需要和这些 provider 搞好关系。

35:14

But then like once you have that established, that infrastructure now have way more control

但一旦你把那套建立起来，那套基础设施现在就能带来多得多的控制力。

35:19

into what you can give to users.

转化成你能提供给用户的东西。

35:22

So one of the biggest things were like observability, you know, like that as an engineer, I want

所以最大的点之一就是 observability，你知道，作为工程师，我想

35:28

to see exactly what's happening.

确切地看到到底发生了什么。

35:30

And I've always felt like in this space, there wasn't enough information.

我一直觉得在这个领域，信息不够。

35:34

Like once when an email goes out, is it really delivered?

比如一旦 email 发出去，它真的送达了吗？

35:38

Is it really on this platform?

它真的在这个平台上吗？

35:39

Is it in the promotional staff?

它在 promotional staff 里吗？

35:41

Like I never know.

我从来都不知道。

35:42

Like you just trust that the email goes out.

就像你默认 email 会发出去一样。

35:44

And then you always hear from some users like, oh, it didn't get the email.

然后你总会听到一些用户说，哦，我没收到 email。

35:47

I'm like, what?

我就想，啊？

35:48

Like what happened?

就是，到底发生了什么？

35:49

There's that empty space between those things.

这两件事之间就有一块空白。

35:52

And the more we're close to the metal, the more we have in terms of data that we can surface

而且我们越 close to the metal，就越有更多 data 可以 surface 给用户。

35:57

for users.

所以我真的很兴奋。

35:58

So I'm really excited.

36:00

It's like the flagship change that we're making out to.

这就像我们要做出的旗舰级改变。

36:05

I mean, you're becoming, it's really interesting for all like DX type businesses.

我是说，你们正在变成……这对所有类似 DX 类型的业务来说都特别有意思。

36:09

You started like a thin layer, like a react email.

你们一开始就像是一个薄层，比如 react email。

36:12

And then it's like, okay, well, we'll build on other people's info.

然后就会想，好吧，我们会在别人的 info 基础上构建。

36:14

And then once you get sized, then you can actually build your own info, right?

然后一旦你规模做大了，你就能真正构建自己的 info，对吧？

36:17

Yeah.

对。

36:18

I think differently because now you own everything.

我的想法不一样，因为现在你拥有这一切。

36:21

Yeah.

对。

36:22

I think what we're doing is, um, and by the way, obviously your dashboards are beautiful.

我觉得我们正在做的，嗯，顺便说一句，很明显你们的 dashboard 很漂亮。

36:27

Yeah.

是啊。

36:28

It's funny, man.

挺有意思的，老兄。

36:29

Everyone I talk to now, like all the customers are like, oh, I never opened the dashboard.

现在我跟每个人聊，所有客户都是这种反应：哦，我从来没打开过 dashboard。

36:34

But we can talk about that later.

不过这个我们可以之后再说。

36:36

But it's interesting.

但这挺有意思的。

36:37

Like I spend so much time trying to make the dashboard beautiful and now people don't

就是我花那么多时间想把 dashboard 做得漂亮，结果现在大家都不去那儿了。

36:41

go there.

36:42

It's kind of frustrating, but I get it.

这有点让人沮丧，但我理解。

36:43

But yeah, we're just trying to approach that with the same naive mentality.

但没错，我们就是想用同样天真的心态去应对。

36:48

That we approach all the other problems.

就像我们处理其他所有问题那样。

36:50

And we're like, okay, this shouldn't be this way.

然后我们会说，好吧，这不该是这样。

36:53

So you're challenging the most basic things of like, the shouldn't be this way.

所以你是在挑战最基础的东西，比如“这不该是这样”。

36:57

The P99 shouldn't take this long.

P99 不该花这么长时间。

36:59

How we can make it faster?

我们怎么才能让它更快？

37:01

Okay.

好。

37:01

And so that's one of the things we're doing now, like focus very, very deep into performance.

所以这就是我们现在在做的事情之一，就是非常非常深入地聚焦性能。

37:07

And we have always been like a no JS shop and we have like, just type script everywhere.

而且我们一直算是那种不用 JS 的团队，基本上到处都是 TypeScript。

37:13

And most recently, like in the sending path, we no longer use a JavaScript.

最近呢，比如在 sending path 里，我们已经不再用 JavaScript 了。

37:18

We have go there and we're changing other parts to be more, more performative in other

我们在那儿用了 Go，而且我们也在把其他部分改得更、更高性能，在别的

37:23

ways.

方面。

37:24

Because no languages don't matter.

因为不，语言并不重要。

37:25

Exactly, right?

完全正确，对吧？

37:26

It's just so cheap to, yeah, like pick a problem and yeah, you can rewrite the whole

因为成本太低了，是啊，就是挑一个问题，是啊，你就能把整个重写掉

37:31

thing in whatever the best languages.

用最好的语言来做东西。

37:33

I'm like running Russ projects and I've never, I don't really know Russ.

我好像在跑 Rust 项目，但我从来没有，我真的不太懂 Rust。

37:37

Like I know like some theoretical things about Russ, but I've never really bothered to learn

我大概知道一些 Rust 理论上的东西，但我从来没真的费心去学

37:41

it.

它。

37:42

Like we are not really covering the recent product surface area.

就像我们其实没有真正覆盖最近的 product surface area。

37:44

Yeah.

嗯。

37:45

Maybe you want to go.

也许你想先来。

37:46

No, no, I mean, you know, there's everything we've talked about, which is why it's hard,

不不，我是说，你知道，我们聊过的所有东西都在那儿，所以这才难，

37:50

what all you're solving.

你都在解决哪些问题。

37:51

And then there's also just send emails from CLI, super agent forward.

然后还有直接用 CLI 发邮件，super agent forward。

37:56

You have some takes on this.

你对此有些看法。

37:58

I think the other day you tweeted like MCP is so back, right?

我记得前几天你发推说 MCP 又回来了，对吧？

38:01

Like over a million MCP calls like, what's the, what's the, let's say, I cite a female?

比如超过一百万次 MCP 调用，那什么，那什么，比如说，我引用一个女性？

38:07

Yeah.

对。

38:08

I was saying MCP was dead, right?

我之前不是说 MCP 已经死了吗，对吧？

38:10

A couple of months ago, like, yeah, the only thing that matters is the CLI.

几个月前，就像，对，唯一重要的就是 CLI。

38:14

What I've noticed is that I stop paying attention to what, what's in, what's out.

我注意到的是，我不再关注什么在流行、什么已经过时。

38:21

And I really think about these problems is like, okay, how do I want my product to be?

而我真正思考这些问题的时候，就是想，好吧，我希望我的产品是什么样？

38:26

Well, I want my product to be formless.

嗯，我希望我的产品是没有固定形态的。

38:28

Whatever they want to do with the product is should be possible.

不管他们想用这个产品做什么，都应该能做到。

38:31

So that, that's why we've always been API first.

所以，这就是为什么我们一直坚持 API first。

38:36

I had no idea how much this would help us down the line.

我当时完全不知道这会给我们之后带来多大帮助。

38:39

But I knew, yeah, like, if I will never be able to cover all the use cases.

但我知道，是啊，就像，我不可能覆盖所有 use cases。

38:43

So let me just give the API, you do whatever you want.

所以干脆我就把 API 给出去，你想怎么用就怎么用。

38:46

And with the MCP is just fascinating, like every month, like, at least 2x growth in

而且 MCP 这块真的很有意思，感觉每个月都至少有 2 倍增长，到现在八月也是这样。

38:52

now in August.

这些数字还更高。

38:53

These numbers are even more.

这还是在八月底之前。

38:54

This was before the end of August.

我觉得到结束的时候是 120 万。

38:56

I think it was 1.2 million by the time it ended.

不过是啊，大家开始从 cloud desktop 或 codecs 来工作了。

38:59

But yeah, like people are starting to work from cloud desktop or codecs.

而如果工作是从那里开始的，那你就需要那些 integrations，对吧？

39:08

And if work starts from there, then you need those integrations, right?

这真的超级重要。

39:12

It's just super important.

39:14

I just want to clarify as well.

我也想澄清一下。

39:16

So, you know, you show less sort of like the month by month numbers on the MCP thing.

你知道的，你在MCP那个东西上展示的逐月数字比较少。

39:19

It picked up from like, let's say 50% month a month ago, which is already super high.

它从一个月前的，比如说50%的月环比增长开始，这已经非常高了。

39:23

Then it starts to double basically every month.

然后它基本上每个月都开始翻倍。

39:26

Did you do investment into MCP?

你们对MCP做投资了吗？

39:28

Or you just had an MCP, and it just, after some time, it gets picked up.

还是你们只是有个MCP，然后过了一段时间，它自己就起来了。

39:32

Well, the fun fact about that, our very first MCP was built by Yoko from A16Z, our investor.

嗯，有趣的是，我们第一个MCP是我们的投资人A16Z的Yoko做的。

39:39

Wow, very bad.

哇，太糟了。

39:40

Yeah.

是啊。

39:41

Yeah.

是啊。

39:42

So she built the open source version, it was a local MCP.

所以她做了个 open source 版本，是个本地的 MCP。

39:47

And then it started to grow like, like, crazy, and I was like, man, why are people using

然后它就开始像疯了一样增长，我当时就想，天哪，为什么大家会用

39:52

it?

它？

39:53

It's like so hard to set up MCP.

MCP 明明那么难配置。

39:54

This was last year.

这是去年的事。

39:56

And then Yoko.

然后就是 Yoko。

39:58

Yeah, exactly.

对，没错。

39:59

So she built that, and then I just asked her, hey, can you transfer to our org?

所以她把这个做出来了，然后我就直接问她，嘿，你能不能转到我们 org 来？

40:04

It was like, it's just like an engineer building something.

感觉就像，就是一个工程师在搭个东西。

40:07

And it was nice to see after the traction it was getting, but to do the MCP right, to

看到它后来获得的 traction 还挺不错，但要把 MCP 做对、把远程版本做出来，我们就得回头重新审视整个权限模型，因为我们需要把它全部关掉。

40:14

do the remote one, we had to go back and revisit our entire permissions model, because we

然后我们就想，好吧，现在我们有一个 block 了。

40:21

needed it all off.

或者我们就基于它来搭。

40:22

And then we were like, okay, now we have a block.

40:24

Or let's build it off.

40:26

So then we build it off, and it's like, okay, our back, let's look at that, like, let's

然后我们就把它构建出来，然后就像，好，我们的 backend，来看看那个，就像，我们

40:30

look at how users have access to each one of the resources.

看看用户是如何访问每一个资源的。

40:34

So it was a whole journey to build it off, and then after we had it off, okay, now we

所以把它做出来是一整段旅程，然后在我们把它做出来之后，好，现在我们

40:39

have the remote MCP, and then it's distribution.

有了 remote MCP，然后就是 distribution。

40:42

So then, okay, now we have the remote MCP, okay, great, great, great, observability is

然后，好，现在我们有了 remote MCP，好，太好了，太好了，太好了，observability

40:46

in place, great, great, great, okay, all the data dog dashboards, okay, now all in,

到位了，太好了，太好了，太好了，好，所有 data dog dashboards，好，现在全都投入了，

40:51

let's talk to everyone, and that was hard too, like, it's not as easy, I wish I had all

我们跟所有人聊聊吧，而且那也很难，就像，没那么容易，我希望我有你所有的

40:56

your connections, it's not as easy to talk to me, but like talking to the labs and trying

connections，跟我聊没那么容易，但就像和 labs 沟通、试着

41:03

to get into like the official stores and plug-in marketing, so you think that actually matters?

比如说进官方商店、做 plug-in 营销，你觉得这真的重要吗？

41:10

Oh, for sure.

哦，那当然。

41:11

Yeah, so I think sometimes like, I've been on the other side where I try so hard to get

是啊，所以我觉得有时候吧，我也在另一边待过，就是我会拼命去争取

41:16

these partnerships, we launch it, and then nothing, all right, so which one matters?

这些合作，我们上线了，然后啥也没有，好吧，那到底哪个重要？

41:22

So as a consumer, I make decisions now based on, do I have the access to that data or not?

所以作为消费者，我现在做决定，看的就是：我到底能不能访问到那个数据？

41:29

So as a buyer of B2B software, if I go to the cloud connectors, and I don't see the

所以作为 B2B software 的买家，如果我去看 cloud connectors，却没看到

41:36

company listed there, with an official MCP, like, we had this, like, we were using plausible

那家公司列在那里，有官方 MCP，就像，我们之前有这种情况，我们当时在用 plausible

41:43

and post hog, plausible had an MCP built by century, I don't care.

还有 post hog，plausible 有一个由 century 构建的 MCP，我不在乎。

41:49

If the plausible doesn't maintain that, and that's not official, I cannot trust that

如果 plausible 不维护这一点，而且那也不是官方的，我就没法那么信任它

41:54

as much, even though I love the century team and I trust that team, but I need the owner

尽管我很喜欢 century team，也信任那个团队，但我需要这个软件的所有者

41:58

of the software to own this piece of integration too, and I made the call like, no, we are getting

也来负责这块 integration，然后我就做了决定，说不行，我们要

42:04

rid of plausible, and we're just going to use post hog here because I have the MCP there,

把 plausible 去掉，这里就只用 post hog，因为我那边有 MCP，

42:11

and I did that as a B2C consumer too, I'm like, into running, and then I'm like, oh,

而且我作为 B2C 消费者也这么做了，我挺喜欢跑步的，然后我就想，哦，

42:16

let me, like, Strava Strava has an MCP, I can like access my rungs, so I use Strava because

让我，那个，Strava Strava 有 MCP，我能访问我的 rungs，所以我用 Strava，就是因为

42:23

of that.

这个。

42:24

So, and if there's like, oh no, we have an MCP, but you have to connect like this, like

所以，如果有人说，哦不，我们有 MCP，但你得这样连接，像

42:29

the X MCP, for example, I have to do like the local one, it's just so frustrating, it works

比如那个 X MCP，我得弄成本地的那种，真的太烦了，它在桌面端能用，在我手机上却不行，对我来说，唉，要是它不在那儿，我就不用了。

42:34

on the desktop line, not on my phone, like, to me, just, yeah, if it's not there, I don't

另外，好吧，它是由供应商官方维护的还是怎样，那到底是什么让一个好的 MCP 区别于一个 social MCP 呢？

42:40

use it.

甚至好几个小时，我跟它交互的时候，我总是，就是，它总是崩掉，对吧？

42:41

Apart from, okay, it is officially maintained by the vendor or whatever, what makes a good

那些 tool calls，我总是打开来试着理解，就是，它到底做了什么？

42:46

MCP versus a social MCP?

就在几天前，有一次 call，它没法把所有东西都放进去。

42:49

Even for hours, as I interact with that, I always, like, it's always collapse, right?

42:56

The tool calls, I'm always opening and trying to understand, like, what did it do?

43:00

And just a couple of days ago, like, there was one call where it couldn't put everything

43:05

on the contacts for some reason, and then, like, no, this shouldn't happen.

不知为什么是在 contacts 上，然后就，呃，不对，这不应该发生。

43:10

Like, how much data are we sending?

就像，我们到底在发多少 data？

43:12

Maybe we're sending way too much data to the client, so that's something that I noticed.

可能我们往 client 发了太多太多 data，所以这是我注意到的一点。

43:16

Like, the amount of tools, this was a bigger concern earlier, like, oh, if you have way too

就像 tools 的数量，这个之前是更大的顾虑，就像，哦，如果你有过多

43:21

many tools that get confused, I don't see that as much anymore, and in fact, we're just

过多的 tools 会搞混，我现在不太看到这种情况了，而且实际上，我们只是

43:26

adding more to the MCP, like, we had, like, maybe, I don't know, 20 tools when we launched,

在往 MCP 里加更多，就像，我们上线的时候，大概，呃，我也不知道，有 20 个 tools，

43:34

or even less than that, now we have like 80-something tools.

或者还更少，现在我们有 80 多个 tools。

43:37

What the fuck?

搞什么鬼？

43:38

Yeah.

对。

43:39

And, like, I'm adding more and more, like, the long tail of the footprint, that's what

而且我越来越多地在加东西，像是 footprint 的长尾，这就是

43:46

I've noticed, like, you can do, like, only 90% of what your dashboard does, and that's

我注意到的，像是，你可以只做 dashboard 功能的 90%，这是一种

43:52

one way of doing it, or you can do 100.

做法，或者你可以做到 100%。

43:56

Getting to 100% coverage of what your dashboard can do is actually very, very difficult, because

要做到 100% 覆盖 dashboard 能做的事，其实非常非常难，因为

44:01

now you're doing usage, billing, team invites, account provisioning, deletion, should that

现在你要做 usage、billing、team invites、account provisioning、deletion，这些

44:09

be exposed as an API or not, like, it's really hard, like, it's a difficult problem.

要不要暴露成 API，像是，真的很难，像是，这是个很难的问题。

44:14

Yeah.

对。

44:15

In fact, you know, you are one of the few liens based guests that I can ever talk about

其实，你知道，你是我能真正聊这个的少数几个 liens based 的嘉宾之一。

44:19

this with.

我觉得这是 Next.js 的误区之一。

44:20

I think this is one of the mistakes of next year's.

哦，真的吗？

44:22

Oh, really?

因为 Next.js 的 server components 和 pages API 或 apps API，它们让你基本上可以很容易地把数据绑定到页面上，所以你永远不会暴露 public API。

44:23

Because next year's and server components and pages API or apps API, they make it easy

是的。

44:30

for you to basically bind data to your page, so you never expose the public API.

默认情况下，出于性能考虑，他们鼓励你做的那个东西，你不会……

44:35

Yeah.

44:36

By default, the thing that is performance and they encourage you to do, you will not

44:41

have an API.

有一个 API。

44:42

That's right.

没错。

44:43

Then you have to just do, like, separate call paths or whatever.

然后你就得做，像是，单独的 call paths 或者什么的。

44:46

Yeah.

对。

44:47

In fact, you should just always have, like, completely headless back end, and then your

其实，你就应该一直有一个，嗯，完全 headless 的 back end，然后你的

44:51

front end just happens to be one of many clients that connects to your back end.

front end 恰好只是连接到你的 back end 的众多 clients 之一。

44:55

And we started with our API on next.

而且我们是从 next 上的 API 开始的。

44:58

I remember that.

我记得。

44:59

First of all, the view was just so high that we were like, there's no way we can pay for

首先，那个 view 实在太高了，我们当时就觉得，这我们绝对付不起。

45:03

this.

后来我们换到了传统的 node server，server 随便你怎么叫吧。

45:04

Then we moved to a traditional node server, server for whatever you call it.

然后我们的 dashboard 一直就像个 private API，消费的也是同一套东西。

45:12

And then our dashboard was always like a private API, consuming the same thing.

我们得复制一大堆代码。

45:16

We had to duplicate a lot of code.

简直一团糟，对吧？

45:18

It was a mess, right?

现在好多了，所有东西都在一个 monorepo 里，东西可以复用。

45:20

Now it's so much better, everything on a monorepo, you can reuse stuff.

但 yeah，默认情况下，理想的话，那些 framework 直接给我们这些 components 就行了，对吧？

45:24

But yeah, by default, ideally, those frameworks just gives us those components, right?

45:31

We had to build a public API separately from that.

我们不得不在那之外单独构建一个 public API。

45:35

And it sucks to maintain both.

而且同时维护两套真的很烦。

45:36

Yeah.

是啊。

45:37

Okay.

好。

45:38

And then one more thing about just a new SAP front end, so you're 80 tools in there.

然后还有一点，就是单说一个新的 SAP front end，你在里面已经有 80 个 tools 了。

45:43

You're growing more tools.

你们的 tools 还在不断增加。

45:44

You are very counter-consensus compared to cloud storage, only as two tools, search and

跟 cloud storage 相比，你非常反共识，只有两个 tools，search 和

45:48

execute or something like that.

execute 之类的。

45:50

So far, you haven't felt that issue, but what would make you change your mind, right?

到目前为止，你还没感觉到那个问题，但什么情况会让你改变想法，对吧？

45:55

It seems like a number is going up, but do you actually know, do you talk to your customers

看起来数字是在涨，但你真的知道吗？你会去跟你的客户聊，

46:00

and say they're like, oh my god, it's like the best MCP on Earth?

然后他们说，“天哪，这简直就是地球上最好的 MCP”？

46:04

I do try to get a lot of feedback on that.

我确实会在这方面尽量多收集一些反馈。

46:06

And then the way I like to build products is I've got to use the product on my own, right?

然后我喜欢做产品的方式是，我得自己先用这个产品，对吧？

46:13

So if I'm like the investor update that I sent you yesterday, that was built using the

所以，就像我昨天发给你的那份投资人更新，就是用

46:19

MCP.

MCP

46:20

It's like fetching data from one place, from another place, and then-

做的。它就像从一个地方取数据、从另一个地方取数据，然后——

46:24

You're making me feel bad now.

你现在让我有点难受了。

46:25

I don't use MCP now, basically, it's what I'm hearing.

我现在基本不用 MCP，反正我听到的是这样。

46:29

I mean, I guess the other thing is also like, you do have CLI, you do have regular SDK.

我是说，我觉得另一个东西也类似，你确实有 CLI，也有普通 SDK。

46:34

Are those also just going crazy?

那些也都在疯涨吗？

46:36

Is everything just from a lot of growth or-

是全都因为增长很猛，还是——

46:38

Yeah.

对。

46:39

The CLI is definitely growing, but not as much as the MCP.

CLI 肯定在增长，但没 MCP 涨得那么猛。

46:44

But I have a lot of fun with the CLI, too.

不过我用 CLI 也玩得挺开心的。

46:46

If you go to our career spage, I'm actually proud of this, because I built it on a weekend

如果你去我们的招聘页面，其实我挺为这个自豪的，因为我好像是一个周末就把它做出来了。

46:51

or something.

如果你去那里看任意一个职位，然后试着申请，往下翻的话，现在你有三种申请方式——

46:53

If you go to any job there and you try to apply for a job, and you have those to go down

你可以用 CLI 申请，或者如果你是……

47:00

now, you have three ways of applying-

所以你可以直接登录，它有一个 interactive mode，或者我们在 CLI 里内置了 skills，你直接跟 Claude 说，比如申请一个产品设计师岗位，带上 recent。

47:02

You can apply if you have CLI, or if you are.

好。

47:06

So you can just log in, it has an interactive mode, or we have skills baked in into the CLI,

47:13

and you can just ask Claude, like apply for a product designer role with recent.

47:18

Okay.

47:19

So you're filling yourself with Slop, especially if you're-

所以你就一直用 Slop 把自己塞满，尤其是如果你——

47:22

Yeah, yeah.

对，对。

47:23

No, I mean, I think maybe that's also the people you want, right?

不，我是说，我觉得那可能也正是你想要的那类人，对吧？

47:26

Totally.

完全同意。

47:27

I actually have a filter if the source is the CLI, I looked differently, because I know-

其实如果 source 是 CLI，我会有个 filter，我会换个角度看，因为我知道——

47:33

Someone went through it.

有人确实走过那套流程。

47:34

People are going the extra mile, like setting that up, like, oh, no, I got to download

大家真是多走了一步，比如把那个设置好，就像，哦不，我得下载

47:38

this thing, and I have to authenticate, and then I have to go through that flow.

这个东西，还得 authenticate，然后还得走完那个 flow。

47:42

But guess what?

但你猜怎么着？

47:43

People who go through that, they always tell me like, man, I've never applied for a job

经历过这个的人，总是跟我说，天哪，我以前从来没

47:47

using a CLI before.

用 CLI 申请过工作。

47:49

So even if they are not accepted, there's just something, it's just a different experience

所以就算他们没被录用，就是有种特别的东西，这跟其他求职申请

47:53

than other job applications.

是完全不一样的体验。

47:55

Even API, apply for one job through it.

甚至 API，都能通过它申请一个职位。

47:58

Exactly.

没错。

47:59

I mean, it's the same thing as you see, like-

我是说，就跟你看的那样，是一回事，就像——

48:01

Yeah.

是啊。

48:02

So I think one of the things that I'm trying to figure out as well is product philosophy,

所以我觉得，我也在努力搞清楚的其中一件事，就是产品哲学，

48:07

right?

对吧？

48:08

You're clearly just maxed everywhere, every SDK, every language, every surface area, obviously

很明显你在所有地方都已经拉满了，每个 SDK、每种语言、每个 surface area，显然

48:13

easier with agents.

用 agents 会更容易。

48:16

But I think some other founders that I would know would say, hey, this doesn't matter.

但我觉得，我认识的一些其他创始人会说，嘿，这根本不重要。

48:21

Give that to greenhouse or ashbee.

把这个交给 greenhouse 或 ashbee 就行了。

48:23

Yep.

对。

48:24

It's your job application.

这是你的工作申请。

48:25

You're not going to differentiate like this, right?

你不会靠这种方式做出差异化吧？

48:27

Yeah.

对。

48:28

You clearly have a very different philosophy of you care about everything.

很明显，你的理念非常不同，你什么都在乎。

48:31

Even though you know, probably some of these stuff is just wasted.

即使你知道，这里面有些东西可能只是白费功夫。

48:34

The way I see it is, I'm trying to optimize not only for the privileged users, like

在我看来，我努力优化的不只是那些特权用户，比如

48:43

the users who have the greatest Wi-Fi, the MacBooks that are here in SF or New York, right?

那些 Wi-Fi 最好、人在 SF 或 New York 用着 MacBooks 的用户，对吧？

48:51

All of these people in the margins that are in other countries with a terrible Wi-Fi

还有所有这些边缘群体，他们在其他国家，Wi-Fi 差得要命

48:56

Brazil.

巴西。

48:57

Brazil growing up, that's how it was for me, a terrible Windows computer and terrible Internet

在巴西长大，对我来说就是这样：一台糟透了的 Windows 电脑，还有糟透了的 Internet access。

49:04

access.

所以这就是为什么我们有 free tier。

49:05

So that's why we have the free tier.

我们做 free tier 不是因为，你知道的，这背后是有原因的。

49:07

We don't have the free tier because of the, you know, there's a reason behind it.

而且我们所有 SDK 都有，所以我在 SDK 上已经拉满了，哥们，弄 dotnet SDK，你觉得那好玩吗？

49:13

And we have all the SDKs, so I'm maxing on the SDKs, man, buting the dotnet SDK, do you

就比如，我们还研究过 stainless、stainless、fern，第三个是啥来着？

49:20

think that was fun?

49:22

Like, and we looked into stainless, stainless, fern, what's the third one?

49:28

But like, we look at all these companies, right?

但我们看看所有这些公司，对吧？

49:30

We didn't like what was generated.

我们不喜欢生成出来的东西。

49:32

So every one of those SDKs is humanly generated, the Java one, the Rust, all those ones, right?

所以那些 SDK 里的每一个都是人工生成的，Java 的那个、Rust 的，所有那些，对吧？

49:41

And I do that because I feel like, yeah, if you're using a product, you should have the

我这么做是因为我觉得，对，如果你在用一款产品，你就应该在你使用的语言里有最好的体验。

49:46

best experience in that language that you're using.

哦，但从公式来看，只有 500 个人。

49:48

Oh, but there's only 500 people out of the formulae.

我在乎那 400 个人。

49:52

I care about those 400 people.

哇。

49:55

Wow.

49:56

You know?

你知道吧？

49:57

Well, why not?

嗯，为什么不呢？

49:58

You know?

你知道吧？

49:59

Why not is you have 4 million users.

不行的原因是你有400万用户。

50:01

That's why not.

这就是为什么不行。

50:02

There's a lot of stuff to do, but, you know, it's good to have that opinionated side towards

有很多事情要做，但你知道吧，在打造这个时有那种有主见的一面

50:08

building this, right?

是好事，对吧？

50:09

Any other product surface area stuff that we should cover?

还有什么其他 product surface area 相关的东西需要我们覆盖吗？

50:11

You covered CLI, you covered templates, covered MCP.

你讲了 CLI，讲了 templates，也讲了 MCP。

50:15

I think it's good to get a high level perspective from you of how much of this is very traditional.

我觉得挺好的是，你能从 high level 的角度讲讲，这里面有多少其实非常传统。

50:21

You even have dashboards, templates, the manually write emails.

你甚至还有 dashboards、templates，还有手动写邮件。

50:24

You have programmatic.

你也有 programmatic 的方式。

50:25

You also have the full agent pill.

你还有完整的 agent pill。

50:27

Like, where do you see it going?

比如，你觉得它会往哪儿走？

50:29

How do you think about building this for a pretty fast changing world, right?

在这样一个变化特别快的世界里，你会怎么考虑构建这套东西，对吧？

50:33

Yeah.

对。

50:34

I think it's still a component of email moving out of this single player mode and going

我觉得它仍然是 email 的一部分，让 email 从这种单人模式走出来、走向

50:39

more into multiplayer.

更多多人模式。

50:42

We even built our editor like this.

我们甚至把编辑器都做成了这样。

50:44

It's multiplayer.

它是多人协作的。

50:45

It fix all the conflicts.

它能解决所有冲突。

50:47

If you have two or three people, we actually run if I have a private recent account with

如果你有两三个人，我们其实会运行——如果我有一个跟

50:53

all the teams and we run all of our all hands there.

所有团队一起的 private recent account，而且我们所有的 all hands 都在那儿开。

50:58

So the all hands is pretty much, you go there, it's a huge email.

所以 all hands 基本上就是：你进去，那就是一个巨大的 email。

51:01

You make an email.

你写一封邮件。

51:02

It's an email.

就是一封邮件。

51:03

We send the email together at the end of the all hands.

我们在 all hands 结束时一起把这封邮件发出去。

51:06

It's super fun.

特别好玩。

51:07

And you have like 20 people in the same document.

而且同一份文档里大概有 20 个人。

51:11

This is super interesting.

这特别有意思。

51:12

I feel like 99% of people have not thought about multiplayer email, right?

我感觉 99% 的人都没想过 multiplayer email，对吧？

51:16

But yes.

但确实是这样。

51:17

You have a lot.

你有很多。

51:18

Yeah.

是啊。

51:19

And it was really fun, man.

而且真的挺好玩的，兄弟。

51:20

That editor alone, it's so freaking hard to maintain that thing.

光是那个 editor 本身就特别难维护。

51:24

But I love it.

但我很喜欢它。

51:25

I love that we have the UI for people who need the UI.

我很喜欢我们给需要 UI 的人提供了 UI。

51:29

Like some teams hate to have the email template pushed into their source code, then they just

比如有些团队很讨厌把 email template 塞进他们的 source code，然后他们就

51:35

want to build a visual editor, a template, give to the marketing team and just call the

想做一个 visual editor、一个 template，交给市场团队，然后直接调用那个

51:40

ID.

ID。

51:41

I don't care about the template.

我不在乎template。

51:42

I just want to call the ID with variables and I'm done.

我只想用变量调用ID，然后就完事了。

51:46

Other teams, they do like, no, I want to commit.

其他团队，他们会说，不行，我要commit。

51:49

I want to have source control over this and own this piece of things.

我要对这个有source control，并且拥有这部分东西。

51:52

So I tried to tailor to those different audiences and build a product that feels that we care,

所以我试着针对这些不同的受众，打造一个让人感觉我们很在意的产品，

52:02

which is not the feeling I get elsewhere.

这是我在别处没有得到的感觉。

52:04

Yeah.

是的。

52:05

But what does the recent team use to read email?

不过最近的团队用什么读邮件？

52:08

Superhuman?

Superhuman？

52:09

Yeah.

对。

52:10

Only in that.

只在那上面。

52:11

Have you seen the last superhuman update this morning?

你今天早上看到 Superhuman 最新的更新了吗？

52:14

I have it.

看到了。

52:15

I have it.

看到了。

52:16

I have it.

看到了。

52:17

I have it.

我拿到了。

52:18

And for listeners, we are recording this on the day of the Astro launch, which I was pulled

听众朋友们，我们是在 Astro 发布当天录的，我被拉进去了，而且，我都没睡觉。

52:21

into and like, I am not sleeping.

但这有什么问题呢？

52:24

But what's wrong with it?

他们只是稍微更新了一下 UI。

52:26

They just updated the UI a little bit.

是啊。

52:28

Yeah.

就有一种感觉，像是，哦，这会走上其他所有收购都会走的那条路了，对吧？

52:29

There's just a sentiment of like, oh, this is going to the direction of every other acquisition

52:36

goes, right?

52:37

Like, now you're feeding a brand that has other things.

比如说，你现在是在给一个还有别的业务的品牌投喂。

52:40

Grammarly and other things.

Grammarly 以及其他东西。

52:41

Yeah.

是啊。

52:42

Okay.

好。

52:43

Let's just email.

我们就用 email 吧。

52:44

And I do want to keep up to date on email tech in general.

而且我确实想总体上跟上 email tech 的发展。

52:48

We are on the sending side, other people on the reading and receiving side.

我们是在发送端，其他人是在阅读和接收端。

52:51

Yes.

是的。

52:52

And then I think, you know, increasingly, I would say agents reading my email, right?

然后我觉得，你知道，越来越多时候我会说，是 agents 在读我的 email，对吧？

52:57

Yeah.

对。

52:58

I actually don't read most of my email.

其实我大部分 email 都不读。

52:59

I have poke and two humans helping me with my own eyes.

我有 poke 和两个真人帮我用我自己的眼睛看。

53:02

I love that.

我很喜欢这一点。

53:04

Poke is pretty good.

Poke 挺不错的。

53:05

I honestly like instinct now super hot, $2.5 billion company with no revenue.

说实话我现在挺喜欢 instinct 的，超级火，一家没有收入的 25 亿美元公司。

53:13

And like, it's just like people trying to read email of agents.

而且就像，人们只是在试着读 agents 的 email。

53:16

Yeah.

对。

53:17

We are definitely looking to that space.

我们肯定在关注那个领域。

53:19

That's another thing that I'm spending a lot of my time, like how you can programmatically

这也是我花很多时间在做的另一件事，比如怎么 programmatically

53:24

curate dozens of inboxes.

管理几十个 inbox。

53:27

And then how are those inboxes accessed by agents?

然后那些 inbox 又是怎么被 agents 访问的？

53:30

Because you create, you enabled receiving email.

因为你创建了、你启用了接收 email。

53:33

What is it called?

那叫什么来着？

53:34

Inbound.

Inbound。

53:35

Yeah.

嗯。

53:36

Email.

邮件。

53:37

So now we're going another abstraction.

所以现在我们要再往上抽象一层。

53:38

Inbound is like the most basic abstraction.

Inbound 就像是最基础的抽象。

53:40

Now we're going one step further and be like, oh, you need 100 inboxes for any reason.

现在我们要再进一步，变成：哦，你需要 100 个 inbox，不管什么原因。

53:46

And here are the APIs that you need on top of that, right?

然后在它之上，你需要这些 APIs，对吧？

53:50

So you need a read API, markers and reds, you need star API, labeling.

所以你需要一个 read API、markers 和 reds，你需要 star API、labeling。

53:54

I see.

我明白了。

53:55

So we're going to be launching a product around that soon.

所以我们很快会围绕那个推出一款产品。

53:58

Yeah.

对。

53:59

But again, it's all API driven because that's the fully, yeah.

但再说一次，它完全是 API 驱动的，因为那才是完整的，对。

54:02

Because I would say the comparable stuff that I've been looking at is misive.

因为我会说，我一直在看的可比产品是 misive。

54:07

Yep.

对。

54:08

And there's other like a help scout or like a sharing box step of product.

还有其他的，比如 help scout，或者像 sharing box 这类产品。

54:13

Front.

Front。

54:14

Front.

Front。

54:15

Hopefully this can be a version of that.

希望这能成为那个的一个版本。

54:18

Okay.

好。

54:19

I can share with you.

我可以分享给你。

54:20

There's a feature flag for you.

有个 feature flag 是给你用的。

54:21

If you want to try that out.

如果你想试试看的话。

54:22

Yeah.

嗯。

54:23

Yeah.

嗯。

54:24

I'm very interested in email tech.

我对 email tech 非常感兴趣。

54:25

Okay.

好的。

54:26

And then to me, again, I'm just like asking all these questions here.

然后对我来说，我又是在这儿问这一堆问题。

54:30

What percent of your email goes to a G Suite or, you know, like it was the market share,

你的邮件有多大比例是发到 G Suite 的，或者说，你知道，就像市场份额那样，

54:36

right?

对吧？

54:37

Like Google versus Outlook versus Apple versus Hotmail.

比如 Google 对 Outlook 对 Apple 对 Hotmail。

54:41

Yeah.

嗯。

54:42

I don't have the number of stuff on my head, but the vast majority, like at least 60%,

我脑子里没记着这些数字，但绝大多数，比如至少 60%，

54:48

60% Gmail.

60% 是 Gmail。

54:49

And then you have like maybe the big tree after like the Outlook, the Yahoo mails of the

然后你大概就有，比如在 Outlook、Yahoo 邮箱这些全球性的东西之后紧接着出现的那个 big tree。

54:54

world right after.

然后你还有一个长尾，Apple 没那么大。

54:56

And then you have a long tail of Apple is not that big.

Apple 很大。

55:00

Apple is big.

是啊。

55:01

Yeah.

尤其是他们现在搞的隐私那套，让注册 Apple 账号变得更难了。

55:02

And especially if the privacy thing that they did now, it's harder to sign up to an Apple.

什么东西？

55:08

What things are?

是啊。

55:09

Yeah.

55:10

How do you generate an email address for you?

你是怎么给自己生成 email 地址的？

55:12

Oh, yeah.

哦，对。

55:13

Yeah.

是啊。

55:14

Yeah.

是啊。

55:15

I used that.

我用过那个。

55:16

Yeah.

是啊。

55:17

Is there a problem for you guys?

你们那边有问题吗？

55:18

They're more strict in terms of what can get into those inboxes.

它们对哪些内容能进那些收件箱更严格。

55:22

So sometimes as a domain owner, we actually have like an in-product wizard to help you like

所以有时候，作为 domain owner，我们其实会有一个类似 in-product wizard 的东西来帮你，比如

55:29

register with Apple so that you can get your emails delivered there.

向 Apple 注册，这样你的邮件就能投递到那边。

55:33

Yeah.

嗯。

55:34

Yeah.

嗯。

55:35

Okay.

好。

55:36

I understand.

我明白。

55:37

Obviously this is your job, right?

显然这是你的工作，对吧？

55:38

If you don't have the ability, you don't have a business.

如果你没这个能力，你就做不成生意。

55:40

Exactly.

没错。

55:41

Yeah.

嗯。

55:42

That's kind of the.

差不多就是那个……

55:43

Yeah.

对。

55:44

We can have the most beautiful Rubik's cube in the website.

我们可以在网站上放最漂亮的 Rubik's cube。

55:47

The most cool like coolest handbook ever, if it's down, it's all for enough.

最酷的，简直是有史以来最酷的手册，如果它挂了，那就全白费了。

55:52

And how correlated are they, are they, right?

那它们之间的相关性有多大，对吧？

55:54

Like if you get banned or down ranked or put in promotions tab in Google, does that transfer

比如你在 Google 被封了、被降权了，或者被放进 promotions tab，这种影响会传导过去吗？

56:00

over to the other guys?

那另一拨人呢？

56:01

Are they very decentralized?

他们非常 decentralized 吗？

56:03

Hopefully now, like in the beginning of the email industry, it was very IP based.

希望现在，就像电子邮件行业早期一样，它非常基于 IP。

56:08

And if one IP pool was damaged, like now you are getting the consequences of that.

而且如果一个 IP pool 坏了，就像现在你正在承受它带来的后果。

56:15

Now there's a lot of reputation associated with the domain.

现在很多 reputation 是和 domain 绑定的。

56:19

So then what people do is you create different subdomains depending on what type of emails

所以人们会做的是，你会根据 emails 的类型创建不同的 subdomains，

56:25

you're sending.

也就是你要发送的。

56:26

Like one subdomain for transactional emails that you know that maybe it doesn't have

比如一个 subdomain 专门用于 transactional emails，你知道它可能没有

56:30

click and open tracking.

点击和打开 tracking。

56:32

And then you have another domain for your marketing emails.

然后你的营销邮件还有另一个 domain。

56:35

And those subdomains help segregate the reputation of whatever it would be like transactions.resend.com.

那些 subdomains 有助于隔离信誉，比如说 transactions.resend.com 这种。

56:42

Yeah.

对。

56:43

Updates, notifications, whatever you want to use it.

更新、通知，你想拿它做什么都行。

56:46

And do I have more dots if I want like a dot b dot c dot recent.com.

那我要是想弄成 a 点 b 点 c 点 recent.com，还能加更多点吗？

56:50

You could do it, but we just see like doing, I don't know, five subdomains is like, yeah,

你可以这么做，但我们看到的是，弄，我也不知道，五个 subdomains 就有点，对，

56:58

not really practical, but like people do, yeah, like some special ones, like we do that

不太实际，但大家确实会这么做，对，比如一些特殊的，我们就会这么做。

57:04

at recent for our own emails, like marketing has a different stream than the transactional

在 Recent，就我们自己的邮件来说，比如 marketing 和 transactional 是两条不同的流。

57:09

ones.

对，连投资人那类也不一样，比如我用 confidential.recent.com 来发那些邮件。

57:10

And yeah, even the investor one has a different, like I use confidential dot recent.com to send

我都不知道还能这样，对，这是个好建议。

57:15

those emails.

这是个好做法。

57:16

I didn't even know you could, yeah, and that is a good advice.

所以最早是非常基于 IP 的，现在则非常基于 domain。

57:20

It's a good practice.

你知道，我，好吧，我遇到过太多糟糕的 IT，比如我不得不单独买一个

57:21

So the first was very IP based now, it's very domain based.

57:24

You know, I have, okay, I have received so many bad it, like I have to buy a separate

57:28

one.

一。

57:29

So because we use AI dot engineer in latent dot space, it's all like we're TLDs and nobody

所以因为我们在 latent dot space 里用 AI dot engineer，感觉就像我们是 TLDs，没人需要认可这个，对吧？

57:33

has to recognize this, right?

是啊。

57:34

Yeah.

所以有人告诉我要给所有东西都弄 dot coms。

57:35

So I was told to get dot coms for everything.

然后也许这会提高宜居性。

57:37

And then maybe that improves the livability.

问题是，现在，比如说，就拿 Coca-Cola 来说，对吧？

57:39

The problem with that is that now, like let's say, let's take Coca-Cola, right?

Coca-Cola 有一个 domain。

57:44

Coca-Cola has a domain.

57:46

Now your Coca-Cola like, oh, yeah, let's buy another domain that looks like our domain.

现在你们的 Coca-Cola 就像，哦对，咱们再买一个看起来像我们 domain 的 domain。

57:51

It looks like fishing.

这看起来像钓鱼。

57:52

It looks exactly like fishing.

这看起来完全就是钓鱼。

57:54

So the chances of, is this damage is way worse.

所以这个概率嘛，这种损害是不是严重多了。

57:57

Is this still a thing when you're buying the dot com, like I would get it the other way,

你买 dot com 的时候这还算个事儿吗，像我就会反过来买，

58:01

right?

对吧？

58:02

You're latent space.com, I buy latent dot space, it looks like fishing, right?

你是 latent space.com，我买 latent dot space，这看起来像钓鱼，对吧？

58:06

Yeah.

对。

58:07

The dot com still, man, you would be surprised, like there's just, like the things I see on

dot com 还在呢，老兄，你会惊讶的，就……我每天看到的东西，简直离谱。

58:14

a day-to-day basis, this is like crazy.

嗯。

58:17

Yeah.

所以我听到的意思是，别来，这是值得的。

58:18

So what I'm hearing is like, don't come, it's worth it.

嗯。

58:20

Yeah.

那都属于 email 那类业务，显然其中一部分是——你只是在 CharGBT 之前才刚起步。

58:21

That's all the sort of business of email, obviously some part of it is, you just got started

嗯。

58:26

before CharGBT.

58:27

Yeah.

58:28

And then you were a leading player that people liked, and then obviously, organically, you pop

然后你成了一个大家喜欢的头部玩家，然后很明显，靠自然增长，

58:32

up.

你就冒出来了。

58:33

You didn't do anything special for AEO.

你在 AEO 上没做什么特别的事。

58:35

What did you do for AEO?

那你为 AEO 做了什么？

58:37

I would say we were not the leaders, because we coming up, like in 2023, like very few people

我会说我们不是领先者，因为我们当时才刚起来，像 2023 年，几乎没多少人

58:44

knew about recent, and we were like, so behind, Twilio, Postmark, Mayo gun, those companies

知道 recent，我们就像，落后太多了，Twilio、Postmark、Mayo gun，那些公司

58:53

invested in SEO for 10 plus years.

已经在 SEO 上投入了 10 多年。

58:56

So for us to get ranked for anything was like almost like, it was uphill battle, like

所以对我们来说，想在任何东西上拿到排名，几乎就像，是一场硬仗，就像

59:02

so, so difficult.

太难了，真的太难了。

59:04

So then what we did was like, okay, let's focus on the docs, for sure.

所以后来我们做的就是，行，那我们就专注在 docs 上，这肯定要的。

59:08

And the docs, it was very intentional about like, okay, if we're doing a docs on, it has

而且 docs 这块我们是非常有意识去做的，比如，好，如果我们要做一个 docs，它同样是有 long tail 的，就像，哦，这里有一篇 docs，不只是讲 Python，而是讲你怎么把它和 Django 搭起来。

59:13

long tail, again, like, oh, here's a docs on, not just Python, like how you set it up with

对。

59:19

Django.

怎么用某某搭配某某。

59:20

Yeah.

没错。

59:21

How to blah with blah.

59:22

Exactly.

59:23

Short articles focus on traditional SEO, traditional SEO, nothing fancy there.

短文章就专注传统 SEO，传统 SEO，没什么花哨的。

59:28

And then when models started to come, then I was like, okay, maybe I got to change the

后来 models 开始出来的时候，我就想，好吧，也许我得改改页面内容，因为如果人们在提问，那我就得做些问答式的、提问回答格式的内容。

59:33

content of the page, because if people are asking questions, then I need to like create

所以就很多 FAQ。

59:38

content that is in action and answer, asking answer type of format.

现在几乎每个最近做的页面都有 FAQ，你知道吧，比如你去 pricing page，或者就像，对啊，pricing 几乎总是标配一个 FAQ。

59:43

So a lot of FAQs.

对。

59:45

Now almost every page that recent has an FAQ, you know, like you'll go to pricing page

59:51

or like, yeah, of course pricing almost always has an FAQ standard.

59:56

Yeah.

59:57

Yeah.

对。

59:58

But like you do FAQs in like all sorts of, yeah, I think maybe security also, yeah.

但就像你在各种地方都会做 FAQ 那样，对，我觉得可能 security 方面也是，对。

60:02

So like you see it's like a pattern of like all pages have that, because that's how people

所以就像你看到的，这有点像一种模式，所有页面都有这个，因为大家就是这样

60:08

are asking is recent socks to complain and then just have the answer right there.

在问最近的 socks 要抱怨什么，然后就直接把答案放在那儿。

60:12

And then, okay, now we have those things in, okay, now we got to do LLMS.TXT.

然后，好，现在这些东西都加进去了，好，现在得做 LLMS.TXT。

60:19

So we added LLMS.TXT on our docs, thankfully because of mentalify that we use great product

所以我们在 docs 里加了 LLMS.TXT，多亏了我们用的 mentalify，很棒的产品

60:26

and we're like, okay, how we take this to the marketing site too.

然后我们就想，好，怎么把这个也弄到 marketing site 上。

60:29

So then LLMS.TXT for every marketing page, same with the docs hatch, mark down versions

所以后来就是每个 marketing page 都有 LLMS.TXT，docs hatch 也一样，markdown 版本

60:36

of the docs.

文档的。

60:37

I'm going to do mark down versions of every single page on the website, on the marketing

我要给网站上的每一个页面都做 markdown 版本，营销网站上的每一个页面。

60:42

website.

然后这就是——

60:43

And that's-

是用 content sniffing，还是直接在后面加 .md？

60:44

Using a content sniffing or you just append like dot md.

嗯。

60:46

Yeah.

直接加就行。

60:47

Just append.

所以如果你去那个，对，security 页面，加 .md 就能用。

60:49

So if you go to the, yeah, security page, dot md it will work.

60:53

Yeah.

嗯。

60:54

But also if you do content negotiation with that page, it will render on that page like

但另外，如果你对那个页面做 content negotiation，它就会在那个页面上 render 出 markdown 版本，而不用去那个……对，对听众来说，这就是那种事情之一。

61:00

the mark down version without going to the, yeah, which for listeners, this is one

我不确定这重不重要。

61:04

of those things.

嗯哼。

61:05

I'm not sure it matters.

我真的不确定。

61:06

Mm-hmm.

嗯。

61:07

I'm really not sure.

61:08

Yeah.

61:09

I mean, you're winning really well.

我是说，你确实赢得很漂亮。

61:10

So, yeah.

所以，对。

61:11

I think one example that changed and maybe it doesn't help with distribution, but helps

我觉得有个例子变了，可能对 distribution 没帮助，但对

61:16

with accuracy.

accuracy 有帮助。

61:17

So go to the pricing page.

所以去定价页看看。

61:19

I remember this was like a random Sunday and I was just like doing competitor analysis

我记得那好像是个随机的周日，我当时就在做竞品分析

61:25

and I was just asking like, okay, if you send 1.2 million emails using recent and you

然后我就问，比如，如果你用 recent 发 120 万封邮件，而且你

61:32

did that with X and Y and Z competitor, who is the cheapest and who is the most expensive

用 X、Y、Z 这些竞品来做，谁最便宜、谁最贵

61:39

and it got completely wrong.

而且它完全搞错了。

61:41

It got the recent one like so bad and it actually said like very bad things about us like,

它把最近那个搞得特别糟，还真说了我们一堆坏话，像是：

61:46

oh, recent pricing is very obscure.

“哦，最近的定价非常不透明。”

61:49

I was like, what fuck, like, it's not obscure at all.

我当时就想，搞什么鬼啊，这根本没那么难懂啊。

61:52

So it was like super frustrating.

所以就特别让人抓狂。

61:54

So you made a calculator?

所以你做了一个计算器？

61:55

Yeah.

对。

61:56

So the problem was agents don't really drag a JavaScript widget.

所以问题在于 agents 其实没法真的拖动一个 JavaScript widget。

62:02

Okay.

好。

62:03

So then what I did was the .mt version of this page, which has a tremendous amount of

所以后来我做的就是这一页的 .mt 版本，顺便说一句，这个页面的

62:11

access by the way.

access 量非常非常大。

62:13

Like now.

就像现在这样。

62:14

Can you get another mix on this?

这个能再给我做一个 mix 吗？

62:15

Yeah.

嗯。

62:16

Yeah.

嗯。

62:17

Okay.

好。

62:18

Not through the regular analytics spots too because like we cannot track that.

也不是通过常规的 analytics 入口，因为那种我们追踪不了。

62:23

Yeah.

对。

62:24

It's a server side.

那是在 server side。

62:25

So then like we got to go to data dog to figure that out.

所以后来我们得去 data dog 里查才能弄明白。

62:27

Okay.

好。

62:28

And man, it was really crazy.

然后天哪，那真的太疯狂了。

62:30

And then I made that change.

然后我就做了那个改动。

62:31

I pushed that and then I asked again, like a few days later and then he got the numbers

我把它 push 上去了，然后大概几天后我又问了一次，然后他就拿到那些数字了。

62:36

right.

对。

62:37

I see.

我明白了。

62:38

So I'm like, okay.

所以我就想，好吧。

62:39

This is a way to live update and fix stuff in your city.

这是一种实时更新并修复你所在城市里各种东西的方式。

62:42

Exactly.

没错。

62:43

So that was one matter.

所以那是一件事。

62:44

Right.

对。

62:45

Okay.

好。

62:46

But it does mean that you don't have a single supposed to truth.

但这确实意味着你没有一个单一的所谓真相。

62:47

Like you have to update the MD and then you also have to update the.

比如你得更新 MD，然后你还得更新那个。

62:49

That sucks.

这太糟了。

62:50

But again, if I wanted the happy path and the easy path, like, I wouldn't be doing you

但话说回来，如果我想要 happy path 和 easy path，我就不会干，你

62:57

know.

懂的。

62:58

I'm like, that's the price of like, well, you know, if you're good at reacts, you can

我就想，这就是代价，嗯，你懂的，如果你擅长 reacts，你就可以

63:03

write a custom renderer.

写一个 custom renderer。

63:04

Yeah.

是啊。

63:05

Yeah.

嗯。

63:06

You know how to do this.

你知道怎么弄这个。

63:07

Yeah.

嗯。

63:08

What else?

还有呢？

63:09

Yeah.

嗯。

63:10

Yeah.

嗯。

63:11

Yeah.

嗯。

63:12

Yeah.

嗯。

63:13

I mean, I'm sure you have it somewhere in some food or something or whatever.

我是说，我敢肯定你在某种食物或者什么东西里有它，反正就那样。

63:16

Yeah.

是啊。

63:17

Just do slash migrate.

直接跑 /migrate 就行了。

63:18

I can give you some.

我可以给你一些。

63:19

By the way, 404 pages.

顺便说一下，404 pages。

63:20

I always think we should guess, you know, take the slug.

我一直觉得我们应该猜一猜，你懂的，拿 slug 去试。

63:23

Yeah.

是啊。

63:24

There's neighbors.

还有邻居。

63:25

Oh, yeah.

哦，对。

63:26

Okay.

好。

63:27

Man.

天哪。

63:28

I think there's a side.

我觉得还有另一面。

63:29

In case you ever break a URL, like, it's just auto fix itself, right?

万一你把 URL 弄坏了，它就会自动修好，对吧？

63:34

You don't even don't just serve the 404.

你甚至都不是只返回 404。

63:36

My intention with those is like, oh, let me try to do something that's not exactly like

我做那些的意图就是，哦，让我试着做点不完全是那种

63:47

comparison page.

对比页面。

63:48

Yeah.

嗯。

63:49

But I don't think it was a great idea because our competitors actually built the comparison

但我觉得这不是个好主意，因为我们的竞争对手其实做了那些对比页面，而且它们被推荐得比我们的还多。

63:55

pages and they get recommended more than ours.

嗯。

63:57

Yeah.

所以我们想做的是，比如，如果你去 postmark，就会看到，像是，好吧，我们来演示一下。

63:58

So we're trying to do, like, if you go to postmark, for example, it's like, okay, let's

首先，就像，嘿，如果你想 migrate，这里有个 code converter。

64:03

demonstrate.

这是个例子。

64:04

First, like, hey, if you want to migrate, here's a code converter.

64:07

It's an example.

64:08

Yeah.

嗯。

64:09

So if you just put any code block there.

所以如果你随便把一个 code block 放进去。

64:12

Whoa.

哇。

64:13

Why?

为什么？

64:14

Why, dude?

为什么啊，老兄？

64:15

Why?

为什么？

64:16

That's what we do.

我们就是这么干的。

64:17

That's what we do.

我们就是这么干的。

64:18

That's what we do.

我们就是这么做的。

64:19

This is so extra.

这也太夸张了。

64:24

There's a cube on the learning page.

学习页面上有一个 cube。

64:27

Everything's extra.

所有东西都太夸张了。

64:30

But the adoption is one that I pushed this last weekend.

但 adoption 这件事是我上周末推进的。

64:35

Like, now we have a graph that compares like our adoption with every different competitor.

比如，现在我们有个图表，能把我们的 adoption 和每个不同的竞争对手做对比。

64:42

And you see one thing, like, again, this was a lesson that I learned with the pricing

然后你会看到一件事，就像，再说一次，这是我从 pricing 页面学到的教训。

64:47

page.

64:49

They're not going to be executed.

它们不会被执行的。

64:51

So I spell out the same data in text so that LLM's can actually read that thing.

所以我把同样的数据用文本详细写出来，这样LLM才能真正读取那个东西。

64:58

So I think they're getting somewhat better.

所以我觉得它们变得稍微好一些了。

65:02

I see a lot of agent take use.

我看到很多agent的使用。

65:05

Okay.

好的。

65:06

Load it.

加载它。

65:07

Load website.

加载网站。

65:08

Click on image.

点击图片。

65:09

Interpret image.

解读图像。

65:10

Oh, it's slower.

哦，这个更慢。

65:11

But with computer use, they are getting much better.

但有了 computer use，它们变得好多了。

65:13

I agree.

我同意。

65:15

You're just comparing stuff.

你只是在比较东西。

65:17

They don't do that, right?

它们不会那样做，对吧？

65:18

I think they're starting to go more towards computer use.

我觉得它们开始更多转向 computer use 了。

65:21

I've been doing a lot of compare this and this and the browse of website is pretty computer

我一直在做很多比较这个和这个的事情，浏览网站这块也挺 computer 的。

65:26

use.

用。

65:27

Okay.

好。

65:28

Yeah.

嗯。

65:29

I noticed that you don't have, like, you know, another classic ayo tactic is like, here's

我注意到你没有，就像，你知道，另一个经典的 ayo 套路就是，这样：

65:34

how to connect our thing with clock code.

怎么把我们的东西和 clock code 连接起来。

65:36

Here's how to connect our thing with codex.

这里是怎么把我们的东西和 codex 连接起来。

65:38

Oh, yeah.

哦，对。

65:39

Oh, okay.

哦，好。

65:40

Yeah.

嗯。

65:42

Like, there are some things that are interesting, like if you go to the legal page, so recent.com slash

比如说，有些东西挺有意思，比如你去法律页面，就是 recent.com/legal，随便点开其中一个，哦，你总结一下，或者让它解释复杂的地方。

65:48

legal, anyone of these, like, oh, you summarize or ask to explain complexity.

我明白了。

65:54

I see.

这真是太不可思议了。

65:55

That's just an extraordinary thing.

是啊。

65:56

Yeah.

对。

65:57

Yeah.

嗯。

65:58

Yeah.

65:59

Again, not sure that matters.

再说，我不确定那是否重要。

66:00

You mentioned a lot about recency for ayo compared to traditional SEO.

你提到很多关于 ayo 相比传统 SEO 的 recency。

66:04

So.

所以。

66:05

Yeah.

嗯。

66:06

Yeah.

嗯。

66:07

But the thing that I was super curious about in the beginning, like, there was no reason

但一开始我特别好奇的是，就像，根本没有任何理由

66:13

for any of the models to recommend us because we're so behind in SEO.

让这些 models 里的任何一个推荐我们，因为我们在 SEO 上太落后了。

66:17

We don't have the typical glossary page or whatever, like, or articles that are so generic

我们没有那种典型的术语表页面之类的，或者说那种特别泛泛的文章

66:23

that were just like SEO heavy.

那些就是特别偏 SEO 的。

66:26

And I was trying to uncover, like, what is that?

我就想搞清楚，那到底是什么？

66:29

And I could totally tell, like, at least with opus 4.5, there was a recency bias in terms

我完全能看出来，至少用 opus 4.5 的时候，在……方面有 recency bias

66:36

of the technology and where they were fetching data.

也就是技术和他们抓取数据的地方。

66:39

I know recently, like, they said, like, maybe raplet is not necessarily something that

我知道最近他们会说，可能 raplet 未必还是大家会推荐的东西，

66:46

people suggest anymore, but definitely, like, newer tools, they started popping more.

但肯定的是，更新的工具开始更频繁地冒出来了。

66:52

And recent was one of those.

而且因为关于这个工具有很多，怎么说呢，organic content，然后我们就……

66:54

And because there was a lot of, I don't know, organic content about the tool, then we

66:59

would be getting recommended more.

就会得到更多推荐。

67:01

So the advice here is just, like, try to get as more people in more places to talk about

所以这里的建议就是，尽量让更多地方、更多的人来聊

67:05

the product as possible because it definitely influences.

这个产品，越多越好，因为这肯定有影响。

67:09

Obviously, I think that is a class of it.

显然，我觉得那算是其中一类。

67:11

I guess, did you ever engage SEO consultants?

我想问一下，你们有没有找过 SEO 顾问？

67:14

Are they useful or are they with some money?

他们有用吗，还是说就是要花点钱？

67:18

We never really did that.

我们其实从来没这么做过。

67:19

You seem like you're just doing everything you're saying.

你看起来就是把你说的每件事都在做。

67:21

You're a 20-20-year-old tweet of what you're talking about.

你简直就是一条2020年的推文，说的就是你现在聊的这些。

67:26

I think you were one of the first ones to pick it up, you know.

我觉得你是最早注意到这一点的人之一，你知道吧。

67:30

I mean, I think people generally, the state of things, because we did an AEO or GEO episode

我是说，我觉得大家普遍来说，当时的情况是，因为我们那会儿做过一期关于这些东西的 AEO 或 GEO 节目，

67:35

at the time around this stuff, and most people were like, just do what normal SEO does,

当时围绕这些东西，大多数人就说，照正常 SEO 那套做就行，

67:42

and don't block AI chatbots.

别去屏蔽 AI chatbots。

67:43

I think, and then most tools, like, profound at the time, they would just, like, track your

我觉得，然后当时大多数工具，比如 profound，基本上就是追踪你的 mentions，

67:48

mentions and then, you know, let you do experiments and all those things.

然后，你知道的，让你做实验之类的。

67:51

All those things are part for the course and normal.

这些都是司空见惯、很正常的事。

67:53

I'm just trying to, obviously, people are pinging me because I'm also doing AEO experiments

我只是想说，很明显，大家一直给我发消息，是因为我也在做 AEO 实验

67:59

on, like, what to do.

比如，研究该怎么做。

68:01

I have unfair advantage because I'm a content business.

我有不公平优势，因为我是做内容业务的。

68:03

Yeah.

是啊。

68:04

I pump up content.

我会大量产出内容。

68:05

Whereas you actually have a product.

而你其实是有产品的。

68:07

I mean, I guess how important is it to you, too?

我是说，我也想知道，这对你来说有多重要？

68:12

Like, you're very winning, but how much do you keep a track of how well our docs use?

比如，你非常擅长取胜，但你有多关注我们的 docs 用得怎么样？

68:18

How often are we recommended compared to competitors?

和竞争对手相比，我们被推荐的频率有多高？

68:20

Like, man, I think it's so crucial.

哎，说真的，我觉得这太关键了。

68:24

Because before we, if we're, like, a little bit unsure about which tool to use, we would

因为以前，如果我们有点不确定该用哪个工具，我们会

68:29

ask Google and then you would get to, like, those comparison pages, or you would get

去问 Google，然后你会看到那些对比页面，或者你会

68:35

to, like, the first three responses, and now you just get one response out of the gate,

看到前三个结果，而现在你一上来就只得到一个回答，

68:40

right?

对吧？

68:41

If you're asking which tool should I use?

如果你在问“我该用哪个工具？”

68:43

And, yeah, people just want to go with the one that is the most frictionless.

而且，是啊，大家就是想选那个最省事的。

68:48

Yeah.

对。

68:49

Which, I think, free tear is very important.

而我觉得，free tier 非常重要。

68:51

And I think Michael Grinich from Work OS is trying to promote this idea of off-MD,

而且我觉得 Work OS 的 Michael Grinich 正在推广这个 off-MD 的想法，

68:56

right?

对吧？

68:57

I don't know if you've seen this.

不知道你有没有见过这个。

68:59

The idea that you, it's very similar to Netlify as well, where you don't need an account

这个想法是，你，它跟 Netlify 也很像，就是你不需要账号

69:03

to do the basic action.

就能完成基本操作。

69:05

So it's, like, unauthenticated usage of your product, which is very dangerous, obviously

所以这就像是，对你的产品进行 unauthenticated usage，显然这非常危险。

69:09

because it's prone to abuse, but it is the easiest, like, no sign up.

因为它很容易被滥用，但这是最简单的，就，不用注册。

69:14

Yeah, that's the thing that I'm trying, yeah, I'm debating doing that or not.

对，那就是我想做的，对，我在纠结要不要那么做。

69:19

Do we need a provisioning API?

我们需要一个 provisioning API 吗？

69:22

And I keep going back and forth on this, like, okay, call this a provisioning and claims,

这个我一直来回纠结，就，好吧，叫 provisioning and claims，对吧？

69:27

right?

或者 provisioning and claims。

69:28

Or provisioning and claims.

但然后 free tier，对，就，我不知道是不是跟我们行业有关，但滥用很多。

69:30

But then free tear, yeah, like, I don't know if it's something about our industry, but

69:34

there's a lot of abuse.

69:35

Yeah, you just have abuse, yeah.

是啊，你就是会有滥用，是啊。

69:36

Just have a lot of abuse.

就是会有很多滥用。

69:37

Yeah, okay.

是啊，好。

69:38

Yeah, I mean, I think those are the main tactics.

是啊，我是说，我觉得那些就是主要策略。

69:41

I think the other thing that I would just tell listeners to think about is literally hook

我觉得另一件我想让听众思考的事，真的就是去 hook

69:48

your planker up to your analytics and your code base, right?

把你的 planker 接到你的 analytics 和你的 code base 上，对吧？

69:53

Like, you need to, like, fully complete the loop and then, like, basically auto research,

就像，你需要，像，把整个 loop 完整闭环，然后，像，基本上就是 auto research，

69:57

right?

对吧？

69:58

We decide, like, so what I have now is my cycle.

我们会决定，比如，我现在有的是我的 cycle。

70:00

I have both, I have a lane space CMO agent and I have a AIE CMO agent, right?

我两个都有，我有一个 lane space CMO agent，还有一个 AIE CMO agent，对吧？

70:05

And they spawn other agents or their job is to, like, research ideas, tricky track of

它们会 spawn 其他 agents，或者它们的活儿就是，比如，研究想法、跟踪我们试过的东西。

70:10

things that we've tried.

然后我会拍板批准，比如，我们想实验的那些事情。

70:11

Well, and then I sign off on, like, things that we want to experiment.

我们有一个实验窗口。

70:14

We have an experiment window.

我们就去做实验。

70:16

We go and do the experiments.

哦，真的吗？

70:17

Oh, really?

70:18

And then those are other agents that are monitored by them.

然后那些就是由它们监控的其他 agents。

70:21

And so I would say, that's kind of like the auto research loop for doing this stuff.

所以我会说，这差不多就是做这类事情的 auto research loop。

70:27

I think then you also need to give it tools.

我觉得接下来你还得给它 tools。

70:29

So like, I sign off an AHS, I sign off a data for SEO.

比如说，我会签核一个 AHS，签核一个 data for SEO。

70:32

And then most recently, I made my own AIEO tracker, which you can make.

然后最近，我自己做了一个 AIEO tracker，你也可以做。

70:37

Yep.

对。

70:38

And then, yeah, you just want to make the number go up.

然后，对，你就是想让这个数字往上涨。

70:40

Yeah.

对。

70:41

For sure.

当然。

70:42

And it's like, you don't know, right?

而且就像，你也不知道，对吧？

70:44

Like, you're throwing things at the wall.

就像，你只能各种乱试，看什么能成。

70:45

That's why, like, yeah, the thing on the legal page does that matter?

所以，就像，对啊，法律页面上那个东西，真的重要吗？

70:48

I have no idea.

我完全不知道。

70:49

But I'll keep throwing things at the wall and seeing what I think, like, Reddit doesn't

但我会继续各种乱试，看看我自己怎么想，比如说，Reddit 根本不重要。

70:54

matter.

对。

70:55

Yeah.

70:56

Yeah.

嗯。

70:57

Yeah.

嗯。

70:58

Yeah.

嗯。

70:59

Not anymore.

现在不一样了。

71:00

Twitter doesn't matter as much.

Twitter 没那么重要了。

71:01

Yeah.

嗯。

71:02

It's weird.

挺奇怪的。

71:03

Right.

对吧。

71:04

It's so crazy.

这也太疯狂了。

71:05

Do you have a tracker?

你有 tracker 吗？

71:06

No.

没有。

71:07

Like, we use all of those profounds of the world, they're all their products.

就，世界上那些 profound 我们全都在用，都是他们家的产品。

71:11

So we tried them all.

所以我们全都试过了。

71:12

Like, people ask me, like, oh, should I use God or profound or this other one?

就，有人问我，哎，我该用 God 还是 profound 还是另一个？

71:17

I'm like, yeah, we have four subscriptions.

我就说，对啊，我们有四个订阅。

71:20

You know, like, that's what it takes now.

你知道吧，就，现在就得这样。

71:22

Like, you, you got to see which one works best.

就像，你，你得看看哪个效果最好。

71:25

What I miss from those tools is, like, give me the action.

我从那些工具里缺的是，就像，直接给我行动方案。

71:28

And I think that the problem, they don't give you, like, what to do, because they're also

我觉得问题在于，它们不告诉你，比如，该做什么，因为它们自己也在试图搞清楚需要发生什么，对吧？

71:33

trying to figure out what needs to happen, right?

就像，哦，在 Wikipedia 上写篇文章，就像，是啊。

71:36

Like, oh, write an article on Wikipedia, like, yeah.

对。

71:39

Yeah.

我得说，像，老派的 SEO 工具，比如 Moz 和所有那一类，它们实际上在这些事情上会更有自己的看法。

71:40

I would say, like, the old school SEO tools, like Moz and all those kinds, they would actually

71:43

have more opinionation around these things.

71:46

Right now, the AOTOs, mostly, from what I hear, again, I'm not an expert, we should probably

现在，AOTOs 大部分，就我听到的来说——再说一次，我不是专家，我们可能该

71:50

talk about profound, is that they just are reporting tools.

聊聊 profound——就是说它们其实只是 reporting tools。

71:54

Right.

对。

71:55

I think it's reporting tools.

我觉得就是 reporting tools。

71:56

And that's fine.

这也没问题。

71:57

Yeah.

是啊。

71:58

It's a big business.

这是个大生意。

71:59

Atrefs is a Singapore business.

Atrefs 是一家新加坡的公司。

72:00

And apparently they do, like, I don't know, three, four hundred million dollars.

而且显然他们还真有，大概，呃，我也说不准，三四亿美元吧。

72:02

Yeah.

嗯。

72:03

Totally.

完全没错。

72:04

Uh-huh.

嗯哼。

72:05

The other parts of your business, like, we're going to talk about like the parenting side

你业务的其他部分，比如说，我们待会儿会聊到像育儿这块，还有个人这块。

72:09

and personal side.

我就是特别想说，好吧，你，你看起来特别松弛，你有50个人在全球各地远程工作。

72:10

I just really want to be like, okay, you, you see very chill, you have 50 people remote

72:16

all around the world.

72:17

You are worrying about MCP, worrying about AEO.

你在担心 MCP，担心 AEO。

72:21

That cannot be the way you're closing all these big customers, right?

这总不能是你拿下所有这些大客户的方式吧？

72:24

Like, where's your fancy sales guy with a steak dinner and a golf course?

就像，你那个能靠牛排晚餐和高尔夫球场搞定客户的高级销售呢？

72:28

We're just starting that motion.

我们才刚开始这套销售打法。

72:29

Yeah.

对。

72:30

And I think we are very lucky that we're able to go this far without having that motion.

而且我觉得我们特别幸运，能在没有这套打法的情况下走到这么远。

72:36

I've worked at an enterprise-only company in the past.

我以前在一家只做企业级客户的公司工作过。

72:40

So I knew, what does that take?

所以我知道，那需要什么？

72:42

And the steak ginners, like, all the apparatus around, like, that sales org, enablements,

还有那些 steak ginners，像是周围整套的配套体系，比如那个 sales org、enablements，有所有这些必需的东西。

72:49

there's all these things that are needed.

我知道，好吧，等时机合适，我们就得把那个开关打开，但最近我一直对这件事非常谨慎，是分层地做，而不是像在 20 的时候那样一下子来；我特别……可能因为我们远程办公，但我觉得我们对文化非常保护，就像，我们要确保不会一下子把文化破坏掉。

72:52

I knew that, okay, when the time is right, we're going to have to turn on that key, but

所以，不，我们试着找出合适的人。所以我们现在一年只有一个，这一直有帮助，一直，嗯，进展得超级好。

72:57

I've been very cautious about doing that a recent, like in different layers instead of

73:02

just like, at 20, I'm very, maybe it's because we are remote, but I think we are very protective

73:09

about the culture, like, let's make sure we don't damage the culture at once.

73:13

So no, let's try to figure out people who fit.

73:16

So we just have one a year now, and it's been helping, it's been, yeah, going super

73:21

on.

on。

73:22

Yeah.

嗯。

73:23

Good enough?

够好了吗？

73:24

Yeah.

嗯。

73:25

I would say, I think, you know, once you have, like, you know, you double your AR because

我会说，我觉得，你知道，一旦你，比如说，你知道，因为几笔单子把 AR 翻倍了，你就会觉得，好吧，这主意不错，因为人们愿意为一些东西多付你钱，你就会想，行吧，你知道，我跟你之间有协议，你的 software 就装在我的本地环境里，或者在我的 VPC 里，随便吧，

73:28

of, like, a couple deals, you're like, okay, like, this is a good idea, because people

73:32

want to pay you more for things that you're like, okay, well, you know, I have, like, the

73:37

agreement with you, I have, like, your software on my premises or in my VPC, whatever,

73:43

right?

对吧？

73:44

Yeah.

嗯。

73:45

So for sure.

那肯定的。

73:46

And what I'm learning too is like, you're building a different product.

而且我也在慢慢体会到，你是在打造一个不同的产品。

73:50

So that's how you need to see it.

所以你得这么看。

73:51

It's not only I'm building a different machine internally as a company, as I'm designing

这不只是我在公司内部打造一台不同的机器；当我在设计

73:56

the company, I'm designing, okay, which department who sits where how each person is, is

这家公司时，我在设计，好，哪个部门、谁坐在哪儿、每个人处在什么位置，

74:02

better equipped to do their job.

才更能胜任自己的工作。

74:03

No, like, for an enterprise, it's actually a different product.

不是，比如说，对于企业来说，这其实是另一个不同的产品。

74:07

Like, what do they care is very different from what $20 users care.

比如，他们关心的事情和那些花20美元的普通用户关心的完全不一样。

74:13

And if I just add, you know, work OS and SSO and, and SCAM and audit logs, it's not going

如果我只是加上，你知道，work OS和SSO，还有SCAM和audit logs，这肯定不够。

74:20

to be enough.

所以，这不只是把那些东西放到定价页面上，这肯定是一个过程。

74:21

So it's, it's not just putting those items on the, on the pricing page to, it's a journey,

在我们进入个人方面的话题之前，关于这个过程你还有什么想补充的吗？

74:28

for sure.

74:29

Any other things you want to reflect on the, the journey before we go into the, on the

74:33

personal side?

74:34

No, I think, yeah, we covered a lot as well.

不，我觉得，对，我们也聊了不少。

74:37

Oh, okay, then I'll talk a little bit about my use case, right?

哦，好，那我稍微讲一下我的 use case，对吧？

74:40

I'm a content creator.

我是做内容创作的。

74:41

Nice.

不错。

74:42

The bulk of your business, I assume, is transactional email.

我猜，你业务的大头是 transactional email 吧。

74:44

I probably look at that by doing it.

我大概是通过实际去做来了解这个的。

74:46

You have marketing emails, you have inbound emails, I'm sure less, like, sort of less outside

你有 marketing emails，也有 inbound emails，我确定会少一些，像是，业务之外会少一些。

74:51

of the business.

74:52

I pay, subtract, something like $10,000 a year, to send my emails, and that seems too

我每年要付，减去，差不多 $10,000 来发邮件，这看起来太高了。

74:58

high.

是啊。

74:59

Yeah.

所以我在想，为什么，为什么 marketing email 和 transactional email 这么不一样？

75:00

So I think, like, why, why is marketing email versus transactional email so different?

这是个好问题，因为你的 business model 也不一样，对吧？

75:07

It's a great question because your business model is also different, right?

确实。

75:10

It is.

确实。

75:11

It is.

而且我们也更贵。

75:12

And we are more expensive, too.

75:14

Yeah.

嗯。

75:15

Yeah.

嗯。

75:16

Then I was like, I can't do this.

然后我就想，这我可做不了。

75:17

Yeah, the industry has this pattern of, like, charging per contact.

对，这个行业有这种模式，就是按联系人收费。

75:21

That's the first thing.

这是第一点。

75:22

So the way they try to prices around the value you create in each contact.

所以他们定价的方式，是围绕你在每个联系人身上创造的价值来定。

75:28

But now, it's crazy, like, the AI companies that we serve, they grow so fast, like,

但现在很疯狂，比如我们服务的那些 AI 公司，它们增长得太快了，就像，

75:35

mistro users recent, okay?

mistro 用户最近，好吧？

75:36

So they're, like, the amount of users that they add is just insane.

所以他们就是，新增用户的数量简直离谱。

75:40

So, like, the business model of contact-based pricing is kind of broken, and we are realizing

所以就是，contact-based pricing 这种商业模式有点崩了，而我们也意识到

75:45

that and trying to fix that, too.

这一点，也在试着把它修好。

75:48

But that's how it had always been.

但一直以来都是这样。

75:50

And there's something about, like, when you send a transactional email, it was just, like,

而且就是，发 transactional email 的时候，它其实就像，

75:54

put a message in a queue and get out of the queue.

把一条消息放进 queue，然后从 queue 里出来。

75:57

With marketing email, you're building that engine of, like, batches and intervals between

而 marketing email 的话，你是在搭那样一个引擎，就是 batches，以及

76:03

batches, analyzing, increasing the speed, decreasing the speed as the sending is going, maybe

batches 之间的 intervals，做分析，发送过程中加快速度、放慢速度，可能

76:09

blocking.

blocking。

76:10

So...

所以……

76:11

Is there a max rate you should send emails at a time?

一次发邮件有没有一个最大发送速率？

76:13

Yeah, for sure.

有，肯定的。

76:14

Or domain.

还是按域名？

76:15

Okay.

好。

76:16

Specifically.

具体来说。

76:17

Yeah.

嗯。

76:18

Okay.

好吧。

76:19

So, like, let's say I have 100k emails to send out.

那么，比如说，我有 10 万封邮件要发出去。

76:20

Yeah.

嗯。

76:21

If you give me 10 minutes.

如果你给我 10 分钟。

76:22

Yeah.

嗯。

76:23

And, for example, now, with the 4.3 million users that we have, it takes more than a day.

而且，比如现在，我们有 430 万用户，发完要花一天多。

76:28

Oh, what?

啊？什么？

76:29

No, that's unacceptable.

不行，这不能接受。

76:30

Yeah.

嗯。

76:31

For sure.

当然。

76:32

So then I pay you more to get in a different queue.

所以我就多付你钱，好进另一个 queue。

76:34

Yeah.

嗯。

76:35

Okay.

好。

76:36

So you have to be, like, always trying to figure out, okay, we need more at the door,

所以你就得，呃，一直琢磨着，好，我们门口需要更多，

76:39

so...

所以……

76:40

How does that work?

那这是怎么运作的？

76:41

Yeah.

嗯。

76:42

Oh, that's interesting.

哦，这挺有意思的。

76:43

It's, like, really, it's for...

就是，真的，它是为了……

76:44

It's like talking...

就像在说话……

76:45

Yeah.

嗯。

76:46

Oh.

哦。

76:47

Exactly.

没错。

76:48

Where's my reset button?

我的 reset button 在哪儿？

76:49

Yeah, exactly.

对，就是这样。

76:50

Yeah.

是啊。

76:51

You also, like, you went into YC, you started this company as a young father.

而且你也，就，你进了 YC，你是作为年轻爸爸创办了这家公司的。

76:53

And I think, like, part of things, like, there are a lot of no kids, people in SF.

而且我觉得，就，某种程度上，SF 有很多没孩子的人。

77:00

And I think that's part of the consideration.

而且我觉得这也是考虑因素之一。

77:02

I think a lot of people are listening who are, like, I can't be a founder.

我觉得很多在听的人会想，我当不了创始人。

77:05

Like, I have this great idea, but, like, I got a young family.

就，我有个很棒的想法，但是，我有个小家庭。

77:08

How do you manage?

你是怎么兼顾的？

77:09

You got a professor with...

你有个教授……

77:10

You left SF for...

你离开 SF 是为了……

77:12

Yeah.

对。

77:13

Kids.

孩子。

77:14

Yeah.

对。

77:15

So this is the recent news, like, I'm spending more time in Brazil now because we're expecting

所以这是最近的新闻，就是说，我现在在 Brazil 待的时间更多了，因为我们快要有

77:19

our second kid.

我们的第二个孩子了。

77:21

And I think, at first man, it was crazy, because I came to SF.

而且我觉得，一开始，老兄，那太疯狂了，因为我来了 SF。

77:26

My daughter was one year old, and every YC founder around me had no idea, like, yeah, what

我女儿当时一岁，我身边每个 YC founder 都完全没概念，就那种，嗯，到底

77:34

does it mean to have kids?

有孩子意味着什么？

77:36

And...

而且……

77:37

What is this reproduction?

这生孩子到底是怎么回事？

77:39

You don't even...

你甚至都……

77:40

Don't even have to use B.

甚至都不用 B。

77:41

Like, what are you doing?

就，你在干嘛？

77:42

Married.

结婚了。

77:43

Like, totally.

就是，完全是这样。

77:44

And also, they weren't married at all.

而且，他们根本就没结婚。

77:46

And it's crazy how...

而且离谱的是……

77:49

I think, again, I always try to look at those things.

我觉得，再说一次，我一直试着去看那些东西。

77:53

At first, you see them as the advantage, like, oh, man, all of these guys, they went from

一开始，你会把它们看成优势，就像，天哪，这些人，他们从

77:58

like...

就像……

77:59

They came from MIT, all these amazing schools, that was not the case for me.

他们来自 MIT，这些很厉害的学校，而我不是这种情况。

78:04

Oh, they don't...

哦，他们不……

78:05

They are not married.

他们没有结婚。

78:07

That's not the case for me.

我的情况不是这样。

78:08

They don't have kids.

他们没有孩子。

78:09

That's not the case for me.

我的情况不是这样。

78:10

But still, going for that program, I'm like, man, I just want to build the best company

但话说回来，去参加那个项目的时候，我心想：兄弟，我就想在这一批里做出最好的公司。

78:16

on this batch.

不是第四名、第五名那种，而是 250 个里的第一名。

78:17

Not number four and number five, like, number one out of the 250.

好。

78:21

Okay.

78:22

Which cohort?

哪一届？

78:23

Winter 2023.

Winter 2023。

78:24

Okay.

好。

78:25

You realize, in the middle of that, is that there's no winner in YC, that there's no such

你在那过程中会意识到，YC 里没有赢家，也没有那种名气。

78:31

fame.

对。

78:32

Yeah.

是啊。

78:33

Yeah.

Mod you raise。

78:34

Mod you raise.

78:35

But, like...

但是，就像……

78:36

I'm pulling out...

我在往外拿……

78:37

Winter 23 now.

现在是 Winter 23。

78:38

I see who I know.

我看到我认识谁了。

78:39

Yeah.

嗯。

78:40

And especially what you understand is that...

尤其是，你理解到的是……

78:43

Winter...

Winter……

78:44

It's like an infinite game, right?

这就像一场无限游戏，对吧？

78:48

Like, even if you had an amazing batch and you raised the biggest round, it doesn't matter.

就是，就算你有一个特别牛的 batch，融到了最大的一轮，那也不重要。

78:54

Because, like, life goes on and it continues.

因为，就是，生活还得继续，它会一直往前走。

78:56

And many of the people that I met there, they are not here anymore.

而且我在那里认识的很多人，现在都已经不在这里了。

79:00

Yeah.

是啊。

79:01

They don't break ups, seven pivots, and they're dead, right?

他们没搞什么分手、七次 pivot，然后就死掉了，对吧？

79:06

But we are here.

但我们还在这里。

79:07

So, I was very much like, no, no, we're even the domain when we bought recent.com.

所以我当时就是那种，不不，我们买 recent.com 的时候，连 domain 都是我们的。

79:12

Like, no, it's going to be an amazing website with an amazing.com, like, it's not going

就像，不，它会是一个超棒的网站，配上一个超棒的 amazing.com，就像，它不会……

79:16

to be a.io or whatever, right?

成为 a.io 之类的，对吧？

79:19

But doing that with kids was so freaking hard, especially because I wanted to enjoy my

但带着孩子做那件事真的太难了，尤其是我还想好好享受和大女儿在一起的时光。

79:25

first daughter.

这我认真想过。

79:27

And I really thought about this.

就是，在 YC 期间有那么一个时刻。

79:29

Like, there was one point during YC.

所以，这个故事，怎么说呢，正好对应我接下来要说的。

79:32

So, this story, like, translates to what I'm going to say.

我来参加这个活动，PG 也在那儿。

79:35

I came to this event and PG was there.

我就想，哦天哪，program，万物之父，神在这儿，对吧？

79:39

So I'm like, oh, man, program, the father of everything, God is here, right?

79:45

So, everyone is a pre-intention to PG.

所以，每个人对 PG 都有一个 pre-intention。

79:47

And then he says, you know what?

然后他说，你知道吗？

79:49

We always tell you, don't do this thing, we tell you that as partners.

我们总是告诉你，别做这件事，我们作为合伙人这么告诉你。

79:54

And then you end up doing the thing and then months later you come to us and say, oh,

结果你还是做了那件事，然后几个月后你跑来找我们说，哦，

79:58

we're doing this or not doing this.

我们要做这个还是不做这个。

80:00

So he just said that.

所以他就这么说了。

80:02

So then I came to my partners and I asked like, okay, what are the things that I'm doing

然后我就去找我的合伙人们，我问，好吧，我现在在做的哪些事

80:06

now that I shouldn't be doing that you think it's wrong?

是我不该做的、而且你觉得是错的？

80:09

And then one partner, not my main partner, but another partner told me, like, you just

然后有一个合伙人，不是我的主要合伙人，而是另一个合伙人，跟我说，你就是太在意设计了。

80:13

worry too much about design.

你太在意这个了。

80:15

You care too much about that.

如果你不那么在意这个，你会走得远得多。

80:16

If you didn't care about that, you will be much further ahead.

这话让我特别难受，因为它是真的。

80:20

And that hurt me so bad because it was true.

比如我 pitch 的时候，我的 slides 很漂亮，但那根本不重要。

80:24

Like, my slides, when I was like pitching, it was like a beautiful slides and it doesn't

我就想，是啊，现在我知道这不重要了。

80:28

matter.

80:29

I'm like, yeah, now I know it doesn't matter.

80:31

But like, that hurt really bad.

但是说真的，那真的很伤人。

80:34

And then a few days later I was reflecting on that and I'm like, yeah, but if I were to

然后几天后我回想这件事，我就想，是啊，但如果我要

80:39

build a product that had like this identity and only 1,000 people used, I prefer doing

做一个有这种 identity 的产品，而且只有 1,000 个人用，我宁愿

80:49

that than having a product with 10,000 people that it's a product that I don't want to

做这个，也不愿有一个 10,000 人用的产品，而那个产品是我不想

80:54

go that I don't use myself.

去用、连我自己都不用的产品。

80:56

So you know what?

所以你知道吗？

80:57

I'm going to follow what I believe is right.

我要坚持我认为对的东西。

81:00

And I think in this industry, again, going back to the industry, everyone had terrible

而且我觉得在这个行业里，再说回这个行业，每个人都曾有过很糟糕的

81:06

products.

产品。

81:07

There was not enough like designing fusing to them.

当时在设计、在和他们融合这些方面，做得还不够。

81:10

So I felt that was a different shader and we needed a different shader in a such a

所以我觉得那是一个不同的 shader，而且我们需要一个不同的 shader，在这样一个

81:14

crowded market.

拥挤的市场里。

81:15

So I go back to that for my kid too, like, okay, 10 years from now I have, I'm with

所以我也会为了我的孩子回到那个想法上，就像，好吧，10 年后，我和

81:22

my daughter and my daughter wants to be with me so much and we play all the time.

我女儿在一起，而且我女儿特别想和我待在一起，我们一直都在玩。

81:29

And we are a 10 million AR company.

而且我们是一家 1000 万美元的 AR 公司。

81:32

I would rather live in that reality than live in a reality where we're a 100 million dollar

我宁愿活在那样一个现实里，也不愿活在一个我们是 1 亿美元……

81:37

company.

公司。

81:38

And my daughter doesn't want to play with me.

而且我女儿不想跟我玩。

81:40

It's not interested in me and we're not that close.

她对我没兴趣，我们也没那么亲近。

81:44

So that's the decision I made.

所以这就是我做的决定。

81:46

Maybe it's right.

也许是对的。

81:47

Maybe it's wrong.

也许是错的。

81:48

I have no idea.

我完全不知道。

81:49

I go extremely hard.

我会拼得特别狠。

81:51

I have a lot of ambition that's not going to stop me or anything, but like setting up

我有很多抱负，这不会让我停下来或怎么样，但就比如说把这些优先级排好、自己去琢磨什么对我行得通，我觉得这很重要。

81:58

those priorities and figuring out for yourself, like, what works for me, I think it was important.

所以是啊，比以往任何时候都更是这样，她对我来说是最重要的人，我会把她放在很优先的位置。

82:03

So yeah, more than ever, like, she's the most important thing for me and I prioritize

那你怎么把这个和你的抱负平衡起来，对吧？

82:09

her a lot.

就比如说，我也有这种情况，只是程度小一些。

82:11

How do you balance that with your ambition, right?

我没有孩子，但是啊，我有很多想做成的事，我还有个家庭。

82:13

Like, I also have this in a smaller way.

我该花时间陪他们，可当我陪他们的时候，心思又不在。

82:15

I don't have kids, but yeah, I have a lot of ambitious things and I have a family.

82:19

That I should spend time with and when I spend time with them, I'm not present.

82:23

I'm looking at my phone and I'm like saying, like, I should be reviewing Astra GPT-6, like

我看着手机，我就在说，就，我应该去评测 Astra GPT-6，就

82:30

it's just like, you know, AGI, no, it's not right, like, but I really feel like I should

就感觉，你知道吧，AGI，不对，这样不对，就，但我真的觉得我应该

82:36

be working and, you know, and like at the same time, like, I feel guilty and so, like,

在工作，而且，你知道吧，就，同时呢，就，我又觉得内疚，所以，就，

82:41

I'm unhappy in both areas.

我在两边都不开心。

82:43

Totally.

完全同意。

82:44

And that's how I feel.

我就是这种感觉。

82:45

I feel like I'm a terrible dad, I'm a terrible husband, I'm a terrible CEO and like, you're failing

我觉得我是个糟糕的爸爸，我是个糟糕的丈夫，我是个糟糕的 CEO，而且就，你在

82:52

in every single area of your life, I feel that way, exactly like that.

生活的每一个方面都失败，我就是这种感觉，完全就是这样。

82:56

But every day, I try to be a little bit more present, you know, and it's especially when

但每一天，我都试着更专注当下一点，你知道的，尤其是当你在家工作的时候，就像，哥们，你离开办公室，然后现在你是爸爸了，然后现在又不是了，哦，现在你得当丈夫，要关注你老婆，然后你脑子却在想，哦，天哪，我怎么搞这个企业，我得敲定这笔交易。

83:02

you're working from home, like, man, you leave your office and then now you're a dad, and

我也觉得自己一直在循环。

83:09

then now you're not anymore, oh, now you've got to be a husband and be attention to your

是啊。

83:13

wife, like, and your mind is like, oh, man, how I do this enterprise, like, I got to close

我们在同一条船上吗？

83:19

this deal.

83:20

I feel like I cycle all the time too.

83:22

Yeah.

83:23

Where are the same boat?

83:24

Yeah.

是啊。

83:25

Well, never easy.

嗯，从来都不容易。

83:26

And I think like it's just naming the thing for your category is important.

而且我觉得，给你的 category 命名这件事很重要。

83:29

And people don't really think about category creation enough, I think.

我觉得，大家对 category creation 的思考真的还不够多。

83:33

This is something, where do you do something like that Netlify tried to do and failed with

这事儿吧，Netlify 就试过类似的做法，结果在 GemStack 上失败了？

83:37

GemStack?

而且我觉得，真的应该有更多人去尝试做 category creation，因为一旦成功了，整个品类就都是他们的了。

83:38

And I think that more people should really actually try to do a category creation because

83:42

when they're successful, then they own the whole thing.

83:45

So which we did for this, recent, you know, it is the next generation Sangrid, let's call

所以这个我们最近做了，你知道，就是下一代 Sangrid，我们就叫

83:51

it, right?

它这个，对吧？

83:52

Or like, AI native Sangrid, something like that.

或者说，AI native 的 Sangrid，差不多这样。

83:55

I think that, you know, and you know, this is like me being like, probably too honest.

我觉得，你知道，而且你知道，这就像是我可能太坦诚了。

84:00

That is not category creation yet.

那还不算 category creation。

84:02

But it is category dominance, you know, like you own your category, that's good.

但这是 category dominance，你知道，就像你拥有自己的 category，这挺好。

84:06

But like, it is not clear enough yet.

但就是还不够清楚。

84:09

What is difference?

区别是什么？

84:10

So my classic example is another guest that we've had, Darmesh Shah from HubSpot.

所以我常举的经典例子，是我们请过的另一位嘉宾，来自 HubSpot 的 Darmesh Shah。

84:14

Yeah.

嗯。

84:15

Anyone who wants to think about category creation, go and read up about inbound marketing versus

任何想认真思考 category creation 的人，都可以去读一读 inbound marketing 与

84:20

outbound marketing, and how they created the category of inbound.

outbound marketing 的对比，以及他们是如何创造出 inbound 这个 category 的。

84:23

The ideal phrase of category creation is once you've heard the category, the two words

category creation 的理想表述是，一旦你听到这个 category，那

84:29

that your company represents, you don't want anything else.

代表你公司的两个词，你就不想要别的了。

84:33

There's no long essay, there's no long manifesto, there's no talk, there's just two words.

没有长篇大论，没有冗长的宣言，没有空谈，只有两个词。

84:38

Maybe some examples.

也许举几个例子吧。

84:39

But your audience can take it from there.

但你的听众可以从那儿接着往下。

84:43

It's very hard.

这很难。

84:44

Man, I've seen France trying to do that like his VIX from it's a webhooks as a service.

老兄，我见过 France 试着那么做，就像他的 VIX 来自一个 webhooks as a service 一样。

84:52

Yeah.

对。

84:53

They took my name.

他们用了我的名字。

84:54

Yeah.

对。

84:55

Man.

老兄。

84:56

It's just so freaking hard, right?

这真是难得要命，对吧？

84:58

I think they're doing okay, right?

我觉得他们做得还行，对吧？

84:59

Yeah.

是啊。

85:00

I don't know if they're still around.

我不知道他们现在还在不在。

85:01

There was a webhooks company, and then there was like an anti-webhooks company where

曾经有一家做 webhooks 的公司，然后还有一家像是反 webhooks 的公司，那家公司的做法是

85:05

you would take like, we don't do webhooks, we just help you publish.

你大概会说，我们不做 webhooks，我们只是帮你 publish。

85:09

It's just JSON file that you pull.

就是一个你去 pull 的 JSON file。

85:12

Oh, wow.

哦，哇。

85:13

You know, it kind of reverses the push versus pull the data flow.

你知道，这有点像把 data flow 的 push 和 pull 反过来了。

85:17

Yeah, this is again, COVID era dev tools, that was the most exciting thing there at the

是啊，这又是 COVID 时期的 dev tools，当时那是最让人兴奋的东西了。

85:21

time.

是啊。

85:22

Yeah.

酷。

85:23

Cool.

嗯，能叙叙旧总是很开心，也恭喜你取得的这一切成功。

85:24

Well, always a pleasure to catch up and congrats on all the success.

我还是很好奇，你到底把你所在的品类主导到了什么程度。

85:30

I still am very curious how much you have dominated your category.

我觉得你就是选得非常好，执行得也非常好，而且还能兼顾家庭，所以真的很佩服。

85:35

I think you just chose it very well and you executed it very well and you're managing

85:39

your family as well, so lots of respect.

85:41

Thanks so much.

非常感谢。

85:42

Yeah.

是的。

85:43

It was great chatting with you guys.

和你们聊天很愉快。

85:44

Thanks a lot.

非常感谢。

Designing for Agents — Zeno Rocha, Resend

Play Queue