本集简介
双语字幕
仅展示文本字幕，不包含中文音频；想边听边看，请使用 Bayt 播客 App。
我们在公司里提倡一种理念。
We start a mantra in the company.
我们想做一支爵士乐队，而不是行进乐队。
We wanna be a jazz band, not a marching band.
我喜欢这个说法。
I love that.
这就像一种自我反思的价值观。
It's like a self reflecting value thing.
它让我自己保持平衡。
It's auto equilibrium with me.
所以，就是这样。
So it's like, okay.
我是属于爵士乐队类型的人，所以我们一直在招更多像我这样的人。
I'm a jazz band person, so we've been hiring more and more jazz band person.
另外，过去两年半人工智能发展起来了，这时候更需要爵士乐队类型的人大放异彩。
Also, AI happened in the past two two and a half years, so more jazz band people can really shine during this moment.
好的。
Okay.
我们以前是经理模式。
So we had manager mode.
我们现在是创始人模式。
We have founder mode.
现在又出现了一种新模式。
There's this new mode.
我们该叫它什么好呢？
What should we call it?
叫爵士模式，倒也不错。
Jazz mode, maybe.
这名字其实挺合适。
That's not bad, actually.
今天到场的各位，有Notion的联合创始人兼首席执行官赵一鸣。
Everybody, today we have Ivan Zhao, the co founder and CEO of Notion.
我更愿意把他看作是重新创办人。
I like to think of him as the re founder.
他很久以前就创办了这家公司。
He started the company a long time ago.
那是2015年的事。
In 2015.
当时他还没找到产品与市场的契合点。
He didn't have product market fit.
他和联合创始人跑到京都——谁也没想到的地方——重新起步，把公司彻底重建了一遍。
He and his co founder moved to Kyoto, of all places, and kind of restarted the company and rebuilt it.
之后，他们终于找到了产品与市场的契合点，公司一下子起飞了。
After that, they got product market fit and it really took off.
到了2023年，他围绕生成式AI，在坎昆的一次外出会议中，又一次把公司从头来过。
He did the same thing in 2023 around generative AI in an off-site down in Cancun and literally restarted the company.
在我看来，他就是SaaS公司如何转型为AI公司的典范。
And in my mind, he's the canonical example of how a SaaS company can move and become an AI company.
这真是太有意思了。
So that's super interesting.
第二个有趣的地方，在于听他如何打造这家公司。
The second thing that's interesting is just hearing how he builds the company.
他在组织公司、薪酬设计和日常运营上，都完全是AI原生的思维。
He's very AI native in the way he organizes the company, the way he compensates people, the way he builds today.
这和几周前我采访的杰克·多西等几位CEO的做法很相似。
And it rhymes a lot with some other CEOs that have interviewed like Jack Dorsey a couple weeks ago.
希望你会喜欢。
I hope you like it.
对了，你觉得我这身打扮怎么样？
What do you think of my outfit, by the way?
我觉得你这裤子可以再松一点。
I think your outfit, it's your pants could be looser.
松一点？
Looser?
是的
Yeah.
好的
Okay.
是裤腿那里还是整体上
On the bottom or just In in
整体上
general.
好的
Okay.
我喜欢你的运动鞋。
I like your sneakers.
我不太清楚这是什么牌子。
I don't know what they are.
好的
Okay.
它们是意大利的，Premiata品牌。
They're Italian, Premiata.
挺酷的。
That's cool.
大家都说你特别有品位。
You're known as a very fashionable fellow.
你现在在纽约一般去哪儿买衣服？
Where where do you shop these days here in New York?
纽约这儿附近有一家不错的店，在苏豪区，叫La Casaun。
New York has a good store near here, in, like, Soho, Traveller called La Casaun.
不是那种大牌设计师款，更偏向欧洲风格，穿起来很舒服。
Not like a designer designer, but more like a European, easy to wear.
是的。
Yeah.
我喜欢那种感觉。
I like the vibe.
谢谢。
Thank you.
我最近打扮得挺讲究的，伊万。
I got up my game, Ivan.
或者直接学点
Or just channel some
挺不错的，咱们去逛街吧。
nice Let's go shopping.
你想去逛街吗？
You wanna go shopping?
就在附近。
It's right around here.
咱们去逛街吧。
Let's go shopping.
逛完之后我们再走。
Afterwards, we can go.
好的。
Okay.
我想聊点CEO的事，不只是时尚。
I wanna talk about CEO stuff, not just fashion.
我做Sequoia这一块已经好几年了，CEO圈子里变化可真不小。
I've been doing this Sequoia thing for a couple of years and a lot has changed in CEO dom.
之前出现了经理模式和创始人模式的区分，现在创始人模式还在，又冒出一种新模式。
Like the whole manager mode, founder mode thing happened, and now it's founder mode and some new mode seems to be happening.
咱们回头看看保罗·格雷厄姆那篇关于创始人模式的备忘录刚出来的时候。
Let's just reflect back when Paul Graham's founder mode memo came out.
我个人觉得。
My my person
那是什么？
What's that?
那是2020年的事了。
It's 2020 shoot.
对吧？
Right?
是2021年吗？
2021?
不是。
No.
不是。
No.
其实比那要晚一些。
It's like it was later than that.
大概是2024年。
It's like 2024.
我当时的反应其实是松了一口气，因为我本来就是那种特立独行的创始人类型，总有人劝我别这么干。
My reaction, by the way, was relief because I was a quirky kind of founder mode type of person, always being everyone's trying to talk me out of doing that.
你当时是什么反应？
What was your reaction?
说实话，我当时没什么特别的反应。
I didn't have too much of reaction, to be honest.
对。
Yeah.
现在呢，杰克·多尔西正在重新设计公司，把人工智能放在核心位置，布莱恩·阿姆斯特朗也在这么做。
Nowadays, you know, Jack Dorsey kind of redesigning his company to put AI in the middle and Brian Armstrong doing the same thing.
我觉得你们公司也在朝这个方向走。
My sense is you're kind of heading this way as well.
能聊聊你们的历程吗？公司是怎么一步步走到今天的？
Can you talk about your journey there, your company's journey?
有什么新进展？
What's new?
你对团队现在是怎么看的？
How are you thinking about the organization?
它在发生哪些变化？
How's it changing?
是的。
Yeah.
我这个过程其实有个小插曲，就是一度陷入创始人模式，总想亲力亲为，那时候Notion还很小，我们根本没学会放手。
My journey actually I have a small detour of the founder mode, which is like try to delegate, trying to grow your Notion has been very small as a company in the early days.
我们其实2019年、2020年初就已经找到产品市场契合点了。
Like, we had product market fit 2019, early twenty twenty.
2019年的时候，我们团队还不到十个人。
We're sub 10 ish people, 2019.
因为团队小，公司反而能盈利。
And company would be become profitable because we're small.
对吧？
Right?
你们早期是拿了风投，还是自己养活自己的？
Did you raise venture in the early days, or you bootstrapped
我们是自己养活自己的。
it?
我们确实拿了风险投资。
We raised venture.
真的吗？
You did?
哦，那一路可真不容易。
Okay, that one's been tough.
确实很难。
It was tough.
头五年，全是早期投资款在撑着，一撑就是五年。
Like, first five years, just sea stage money dragged us for five years.
是啊。
Yeah.
HubSpot 当初也是这样。
HubSpot was like that.
我们花了好长时间才找到产品市场契合点。
Took us a long time to get to product market.
真的特别难。
It's very tough.
对。
Yeah.
那时候的难，跟现在完全不一样。
It's a different type of tough compared to now.
那时候根本谈不上产品市场匹配。
Back then, there's no product market fit.
那就是绝望。
It's despair.
一片漆黑。
It's just black.
根本不知道路在哪儿。
You don't know where to go.
对吧？
Right?
从一开始我们就很小，但也因此尝到了甜头：团队小，利润高，融资也容易，生活节奏更轻松，所以我们一直刻意控制团队规模。
So we have been very small from the get go, and we sort of taste the sweet of it, because it's your small, you're profitable, easier to raise money, your default life, and we've been resisting growing team size.
疫情期间，很多公司都突然扩张了。
And during COVID, a lot of companies really blew up.
你还记得吧。
Remember that.
对吧。
Right?
是啊。
Yep.
你肯定也注意到了。
I'm sure you did.
我们确实也快速增长了，正好赶上了产品与市场的契合点。
We relatively yeah, blew up You got to product market fit at the right time.
我们也是差不多那时候找到了产品与市场的契合点。
We got to product market fit around the same time.
很多人都说你必须扩张。
There's a lot of advice that you gotta grow.
所以我们确实扩张了。
So we did grow.
我
I
明白了。
see.
但跟其他公司比，我们在人员规模上还是尽量控制增长。
Relative to other companies, still resistant to grow a lot in terms of headcount.
但就在那时，我渐渐明白：你得学会放权，得引入专业的管理团队。
But that's the moment that I sort of learned, oh, you need to delegate, you need to have introduce some professional manager management.
为什么？
Why?
谁跟你说的？
Who is telling you that?
这是个好问题。
It's a good question.
谁跟你说的？
Who tell us?
一些投资者。
Some investors.
你没跟我提过这个。
You didn't tell me that.
我不这么认为。
I don't think so.
它
It
在投资圈里，这算是普遍共识。
was conventional wisdom amongst the investor class.
我觉得这事有利有弊。
I would say there's a pro and con for that.
说实话，我之前一直抗拒这么做，所以没组建销售团队。
Like, we sort of I resisted that for a long time, so we did not build a sales team.
现在回头看，我早该早点成立销售团队了，因为我总想着从第一性原理出发，创造一种全新的产品驱动增长模式。
In retrospect, I should have started sales team earlier because I thought I can first principle the whole way, like create new kind of PLG, some kind of brand new motion.
这是个错误。
That's a mistake.
至于产品营销工程师这些职能，倒不能算错，只是我选择了更传统的方式来搭建。
The other part is product marketing engineer, those functions sort of, I wouldn't call mistake, but would opt in the traditional way of building it.
那个阶段，你们有没有请过有经验的人？
Did you hire, been there, done that, people in that phase?
有的。
We did.
现在还有留着的吗？
And do you have any of them still?
很多人至今还在公司。
Many of them are still with us.
是的。
Yeah.
我们的CTO已经任职三年了。
Our CTO, it's last three years.
在那之前，那个阶段还有一位CTO。
Before that, there's another CTO during that phase.
但我觉得，Notion的一些核心基因是由传统SaaS高管塑造的。
But I would say some part of Notion DNA are defined by the classic SaaS executive.
而过去三年，随着我们转向AI驱动开发，情况发生了变化。
Now the same change in the past three years since we're building with AI.
我们是最早采用语言模型技术的产品团队之一，开始意识到用这种新技术构建产品，很多旧规则都被打破了。
Like, we're one of earlier adopter of language modeling or product, and you start thinking things are breaking, building with this new technology.
过去那种清晰的角色分工，现在不太适用了。
The classic boundary between roles doesn't quite work.
我常打的一个比方，至今依然适用：开发传统软件就像造桥。
The analogy I like to use back then, it still is, is like building classic software is like engineering bridge.
只要你能设计出来，通常就能做出来。
If you can design it, you usually can build it.
挺有规律的。
Fairly predictable.
产品经理跟客户沟通，然后交给设计师，再传给工程师。
Like, product manager talk to customers, pass to designer, pass to engineers.
用语言模型做产品，那时候是这样，现在也差不多。
Building with language model back then is somewhat still is.
就像酿啤酒一样。
It's like brewing beer.
很多事情你根本没法准确预测。
You can't truly predict the things.
没错。
Right.
根本的问题在于
The underlying thing
这个比喻很贴切。
Good analogy.
你没法跟东部说，嘿。
You cannot tell the East, hey.
嘿，各位。
Hey, guys.
得往那个风味方向多靠拢。
Go go go towards that flavor profile more.
这事儿办不到。
You can't do that.
是的。
Yeah.
对吧？
Right?
你只能把最顶尖的人派进去，看看模型和科技能带来什么。
You sort of have to throw your best people in there, see what the model see what technology gives you.
所以这并不是以客户为中心。
So it's not customer first.
更像是先拿这项技术来做实验。
It's more like experiment with this technology first.
虽然过去一年左右，模型进步了很多，但核心还是技术驱动开发，而不是客户驱动开发。
Although the last year or so, model gets so much better now, but still, the spirit is technology first driven development rather than customer driven first development.
这迫使我们至少改变了产品的打造方式。
That really forced us to change the at least how we build product.
设计师才是最有生产力的角色。
Designer is the most productive thing.
我们的设计师、工程师和产品人员都混在一起，一起做评估、体验，摸索新东西。
Our designer and engineer and product people all just sit in the same bucket, then work with evals, work with experiences and figuring new things.
你们公司是其中一种角色占主导，另一种被边缘化了，还是现在大家都差不多了？
Your company has one of those personas kinda dominated and the other one's fallen aside, or is this your they all kinda do the same thing now.
我们总是从早期开始，那时候公司的本质还很模糊。
We always start by like, our DNA in the early days is always blurry.
我们是最早招聘能写代码的设计师的公司之一。
Like, we are one of the first to hire designer who can code.
我本人就是个能写代码的设计师。
I'm a designer who can code.
对吧？
Right?
所以我们公司从一开始，就是按照我和我联合创始人的样子来打造的。
So we we build a company based on my image and my company and my co founder's image.
这类人才特别突出。
So those profile really shines.
能写代码的设计师、有产品直觉的工程师、不怕亲自动手试错的产品经理。
Designer who can code, engineer who have strong product sense, PMs don't not afraid to roll up sleeve and try things.
比如，我们公司里消耗算力最多的人之一，就是这种类型。
Like, one of our heaviest token consumption person in the company is a for example.
这类人特别出彩，我们在Notion里有不少这样的人。
Those people really shines, and we have quite a few of them in Notion.
明白了。
Okay.
你看了多西的资料，也看了阿姆斯特朗的。
So you read the Dorsey stuff, you read the Armstrong stuff.
我的理解是，他们想的还是那种大三角形的组织架构。
My summary of it is they think of like a big triangle org chart.
信息在上下左右之间传递特别低效。
It's very inefficient flow of information up and down and across it.
这种模式已经存在几千年了，是时候重新想想了。
It's been around for thousands of years and it's time to rethink it.
简单说，就是在中间插进AI，慢慢把更多决策权交出去，让系统自己做决定，同时用合适的方式给它提供背景信息。
Basically stick AI in the middle and start delegating more and more over time and letting the system make decisions and feeding it context in good ways.
这大概就是我对他们做法的总体看法。
That's kind of my high level summary of it.
你觉得这两个人在做什么？
What's your take on what those two are doing?
那你现在做的是不是也差不多？
And are you doing basically the same thing?
某种程度上，我们是给其他公司提供一个现成的解决方案。
In some sense, we're a green product for other companies to do that.
那你自己呢？
But yourself?
核心理念其实差不多，至少我是这么理解的。
The high premise is similar, at least my interpretation of it.
现在有太多知识型工作其实就是填表跑流程。
It's so much knowledge work is paper pushing.
过去要让软件在人与人之间传递信息特别难，因为知识型工作本身很模糊，必须有人在中间协调传递。
In the past, it's really hard require software to push information from one person to another, but so much of a knowledge work is fuzzy, so human need to be in the middle to pushing information from one to the other.
结果发现，语言模型特别擅长两件事。
And turns out language model can do really good at two things.
第一件是写程序，帮你设计出最理想的流程，让信息和决策在公司里顺畅流动。
One thing is to write the software to design the perfect pipes for your company to pushing information and decisions.
第二点是，在两个决策点之间的空档期，实际完成这种小规模的决策。
Second is actually do this kind of mini decision making in the stop between one decision point to another decision point.
对吧？
Right?
过去很多公司里，很多人只是在中间传递信息和决策，以维持大规模运营，这部分其实可以重新思考。
So a lot of used to companies, a lot of people just in the middle to pushing information decision around to sustain a business at scale, that could be rethink.
好的。
K.
我常打的比方，你也提到过，就像在钢铁出现之前，建筑不可能超过五六层高。
The analogy I like to use what you've been using somewhat is, like, before steel, buildings cannot be higher than five, six floors.
纽约市的建筑大多是六层左右，用的是铁、石头和砖块。
And New York City are largely six floor tall, iron and stone, and bricks.
有了钢铁之后，建筑的信息承载结构和承重能力就能高得多。
After steel, there's information or weight structure can go much higher.
有时候，语言模型加上软件，就起到了这样的作用。
Sometimes language model plus software is that.
这是组织的钢铁。
It's the steel for organisations.
有了钢铁之后，组织会变成什么样子？
What does the organisation look like now that we have the steel?
我觉得大楼变高了，但组织会怎样呢？
I think I know buildings got taller.
组织会发生什么变化？
What happens to organisations?
我觉得大家都在摸索中。
I think everybody's trying to figure out.
比如多西的架构就分三层，或者三种人。
Like, Dorsey's structure is like three layers or something, or three type of people.
对吧？
Right?
是的。
Yeah.
围绕AI系统有三类人，主要职责是训练它，层级非常扁平。
Three types of people that sit around the AI system that mostly train it, extremely flat.
我觉得Brian Armstrong说是五层，每个人都是既当教练又当球员，我觉得这挺
And I think Brian Armstrong said it's five layers, and everyone's a player coach, which I think is pretty
有意思的。
interesting.
我觉得所有公司的做法都是对的。
I think all the spirits are right.
我觉得不应该这样，我认为应该是有层级的。
I don't think and I believe it should be a hierarchical.
比如，我觉得过去有人尝试过扁平化的组织结构。
Like, I think people tried in the past, like, flat organizational structure.
我觉得那不符合人性。
I don't think that's part of human nature.
人的本性就是喜欢分等级的。
Human nature are hierarchical.
那你觉得，有了AI之后，这种情况现在有改变吗？组织结构会不会变得更扁平？
And do you think that's changed now we have AI, that it can be flatter?
还是你觉得这根本就是胡说八道？
Or you're like, that's BS?
当然可以更扁平了。
Oh, definitely can be flatter.
我们现在正变得越来越扁平。
Like, we are getting flatter and flatter.
明白了。
Okay.
你直接管理多少人？
How many direct reports do you have?
我管七八个人。
I have seven, eight.
那一般来说，平均是多少？
And generally, what is it?
从公司整体来看，大概每个人管七到九个人。
The company wise, probably around seven, eight, nine people per person.
很多人现在管十五个、二十个，甚至二十五个。
A lot of people are going to fifteen, twenty, 25.
是的。
Yeah.
我们很多顶尖员工，现在管二十个、二十多个。
Many of our best people, twenty, twenty Got
明白了。
it.
你们公司是怎么调整的？
How have you changed your company?
你们怎么给员工发薪酬？
How you compensate people?
你们是怎么搭建组织架构的？
How you build your org?
你们怎么招聘人？
How you hire people?
就在过去这一年，随着这些变化发生。
Just in the last year as all this stuff has developed.
是的。
Yeah.
两年前，我们开始招聘不同类型的工程师。
Two years ago, we start to hire different shape of engineers.
好的。
K.
不再只看重经验，而是看重人的自主性、活力、乐观和好奇心。
No longer optimized for people's experience, optimized for people's agency, energy, optimism, curiosity.
我们目前用的一个大致框架是，比如说，
A rough framework we're using, like, okay.
公司里真正需要的是什么样的人才？
What is the talent in the company?
一个人的能力乘以他的品味或价值观，再乘以他的主动性，才真正构成他的潜力。
A talent equal to this person's capability slash experience times this person's taste or value system times this person's agency will.
对吧？
Right?
像语言模型这样的技术变革，就跟谷歌让每个人都能获取信息一样。
What language model changes like just like Google, everybody can access information.
语言模型让每个人都能成为不错的写作者和程序员。
Language model allow everybody to be a pretty good writer and programmer.
所以，能力变得普遍化、大众化了。
So capability got normalized, democratized.
但品味依然很重要。
Taste becomes still important.
你的价值体系是什么？
Like, what's your value system?
你希望为这个世界带来什么？
What do you wanna bring to the world?
你打算往哪个方向走？
Which direction do you wanna go?
这个方向不容易改变。
That you cannot change as easily.
你愿意付出多少努力？
And you will, how hard do you work?
这一点你是改不了的。
This you cannot change.
所以我们现在更应该聚焦在后者上。
So we optimize for the latter to today.
明白了。
Okay.
这意味着在工程和科技领域，我们更倾向于招聘年轻人。
That means in the engineering science, we're hired for much younger people.
公司内部通常叫他们 IC，刚毕业不久的初级工程师。
Internally, call it like, ICs, early, out of school ICs.
这跟我在面试的大多数首席执行官们的做法完全相反。
This is the opposite of the way most of the CEOs I'm interviewing are going.
挺有意思的。
It's interesting.
不，我觉得更多人其实很兴奋。
No, I think more people are thrilled.
他们说，有了这些工具，初级工程师的效率会高很多。
They say that like a junior engineer powered by this stuff's more productive.
如果你请到一个熟悉整个系统、资历深的资深工程师，他们一旦进入状态，能干一百个小时。
If you get a senior engineer that understands the whole system, been there a while, and they're pilled, they're like a hundred hours.
对。
Yes.
所以，招人就招特别初级的和特别资深的。
So hire for super junior, super senior.
哦，有意思。
Oh, interesting.
所以这种结构更像是哑铃形。
So the shape, it's not it's more like the Barbell.
哑铃状的结构。
Barbell shape.
对。
Yeah.
对吧？
Right?
因为资深工程师能把握方向，而目前语言模型里自动分配的信息还是太多了。
Because the senior provide taste, there's still a lot of the auto distribution information out in language model.
在语言模型里做架构设计，这方面普遍做得不好。
Architecturing in language models is known very bad with that.
所以他们得指明方向。
So they need to provide a direction.
差不多就是这样。
It's almost like this.
一个优秀的工程师可以同时管理四到六个编码代理。
A good engineer managing four to six coding agent at once.
对吧？
Right?
但一个资深的架构师只需要带两到三个初级工程师，每个人再带两三个，总人数就上去了。
But a good really senior architectural engineer can managing two to three junior interns or engineer, they each manage two or three, then that's the number is higher.
对吧？
Right?
他们还能培养下一代。
And they can train the next generation.
我觉得这种结构可能比一堆资深人员直接管编码代理更高效。
I think that might be a more optimal shape than a bunch of senior people manage coding agents.
明白了。
Okay.
我猜你还在招聘设计师和产品经理，对吧？
And you're I assume you're still hiring designers and still hiring PMs.
这个比例在变吗？
Is the ratio changing?
还是你们直接招人来全包了？
Or are you just hiring people do the whole thing?
设计师和产品经理的定义是不一样的。
The definition of designer PM are different.
我们是有意在招那些能做产品经理的设计师。
Like, we're intentionally hire designer who work as PM.
反过来，通常是那些能做产品经理的设计师。
PM can usually, the other way usually, it's designer who can work as PM.
产品经理想做设计师很难，因为涉及视觉方面。
It's harder for PM to work as designer because it's the the visual It's
这更像是天赋。
a genetic thing.
视觉设计这种手艺，得花好多年才能练出来。
It's more like a visual craft takes years to train.
是的
Yeah.
对吧
Right?
但是哪里
But where
设计师想跟客户交流、获取反馈呢？
designers wanna talk to customers and get feedback?
是的，这种情况很少见，但别对设计师刻板印象。
Yeah, that's rare, but don't stereotype designers, though.
确实有这样的人。
They do exist.
明白了。
Okay.
我看了你们网站，你们在招聘初级和高级工程师。
So, I look at your website, you're hiring an engineer, junior, senior.
你们也在招聘产品经理和设计师吗？还是说你们把这两个角色合在一起了，我都不太清楚该怎么称呼。
Are you hiring PMs and designers, or do you have some pot I don't know how to combine those two words together.
我们其实有个职位叫产品设计师，面试的时候我们会综合评估。
Well, our PM designer, by the way we interview them, we evaluate them.
我们最近在和两三位设计负责人沟通。
Like, we're in talk with two to three design leaders.
我们把他们当成产品经理来看待，本质上是一样的。
They're kind of we value them as PMs, basically.
你能不能自己推动事情往前走？
Are you how in can you drive things yourself?
光有好点子是不够的。
It's not just come up with good ideas.
除了设计能力，你平时跟客户交流得多吗？
How much are you interviewing customers besides you're good with craft, design craft?
有些东西跟潮流有关，比如很长一段时间大家都讲用户体验。
Some of this stuff is like the fashions change, like for the longest time it was experience.
那时候大家常说的是‘潜力’。
Then the word everyone used was slope.
我们招人看潜力，现在更看重品味、好奇心和主动性。
We hire for slope, and now it's kind of taste and curiosity and agency.
这其实也算潜力吧。
That's kind of slope though.
这就像……
That's like the
算是有点潜力的意思。
It's slope ish.
那具体有什么不同呢？
How is it different?
在我看来，潜力就是聪明有干劲。
Slope to me is just horsepower smart.
智力当然是第一位的，而且是多维度的。
Well, intelligence is first intelligence, multiple dimensions.
就像是
Like
GPU速度啊，你那个SAP系统跑得怎么样？
GPU speed, like how'd you do on your friggin' SAP?
但有很多聪明人就是懒。
But there's a plenty smart people are lazy.
这没错。
That's for sure.
你觉得这是高潜力还是低潜力？
Do you call that high slope or low slope?
威尔：我也不确定。
Will: I don't know.
我一直以来都是挑那些学得快、特别聪明的人。
I just always bucketed in to hire the people who learn fast and are very smart.
现在我觉得重点变成了‘品味’，你早就一直在谈品味了。
And I think it's changed to the word taste comes up a lot and you're, you've been talking about taste for a long time.
展开剩余字幕（还有 480 条）
整个行业都在谈这个。
The industry talks.
怎么突然就成热门话题了？
Why is that all of a sudden a thing?
这个词怎么现在到处都是？
Why is that word everywhere now?
因为语言模型里没有品位这回事。
Because taste is not in language models.
对。
Yeah.
如果你假设经过很长很长的时间，也许会有，但眼下还没有。
Taste is in in if you assume that a long, long time, maybe it is, but right now it's not.
对。
Yeah.
品位更多是根植于你的价值体系。
And taste is more rooted based on your value system.
你喝气泡水还是普通水？
Are you drinking sparkling water or still water?
你喝健怡可乐还是冰茶？
Are you drinking Diet Coke or iced tea?
每个人喜欢的不一样。
It's to each of their own.
没有对错之分。
There's no right answers.
你们其他人可不只是搞工程的。
The rest of you, you've got more than just engineering.
这里还有不少HubSpot的人，这很好。
You've a bunch of HubSpot people here, which is great.
咱们单说营销这块儿。
But let's just take marketing.
在营销领域，标准也变了吗？
Is the criteria changed there too?
我们实际上重组了市场部门。
We actually changed our marketing department.
现在不再设首席营销官这个职位了。
We no longer have a CMO organization.
这是今年才改的。
This is this year.
去年是这样，今年也是这样。
Last year, this year.
是的。
Yeah.
我们把原来的首席营销官团队拆成了内容叙事小组，没错。
And we break up our CMO organization into storytelling Yep.
这样更贴近产品，因为产品变化实在太快了。
Which is closer to product, closer to because the product is changing so fast.
传统的营销方式根本跟不上节奏。
It's like you just classic marketing can't keep up.
我们以前从没这么快交付过产品。
We haven't shipped this fast before.
对吧？
Right?
所以就让这部分直接贴近产品，跟社交媒体上大家讨论的地方直接连通，另一部分呢，就是支持销售和市场推广团队。
So let's just have that sit next to product, directly connect to social where people are discussion are happening, and then the other part is, like, serving the sales go to market function.
比如需求生成之类的。
Like, demand gen.
线索生成。
Lead gen.
线索生成。
Lead gen.
所以我们真的非常专注于这一块。
So it's like we really, really focus on that.
所以与其让信息在CMO那里转一圈，不如直接搞定怎么同时服务好产品和市场。
So you'd rather information round trip to a CMO than figure out how to serve both the product side and the gold market.
别再这样了。
No more.
让两边各自去摸索。
Just like let both sides
好的。
Okay.
自己理清楚。
Figure out themselves.
好的。
Okay.
不是靠自己，而是更去中心化一些。
Not with themselves, but more decentralized.
还有没有其他传统的营销职能包含在内？
And are there any other of the classic marketing functions in there?
这两个应该就是主要的了。
It's sort of those are the two.
就是这两个。
Those are the two.
设计师那边，还有讲故事方面的设计和创意部分。
The designers, there are also designer and creative part on the storytelling side.
比如社区吗？
Like community?
对。
Yeah.
社区也算在内。
Community is in there.
社区是整个生态或组织的一部分，像Notion的全球粉丝群体就是最重要的社区之一，我们要通过这些人来推动市场。
It's a community with as part of the community ecosystem or organization, which is startup where it's one of the most important communities, Notion fans around the world, and figure out how does to move our market through those people.
这些人就是这一类。
That's the bucket of people.
标准有变化吗？
And has the criteria changed?
销售、市场，还有其他那些职能呢？
Sales, marketing, all the different other functions?
你们对招聘的看法有变化吗？
Have you changed the way you think about the hire?
大多数公司面试时都会有个打勾的表格，
Like, most companies have a form when you interview someone has a Yes.
用来
Form to
其实就在刚才我说的工程师和设计师那段，我们是两年前改的。
actually in the middle of I just described engineer, Engineering designer, it's the we changed two years ago.
对。
Yes.
市场这块是大概六个月前改的。
Go to markets marketing would change, you know, six months ish ago.
销售这边现在正在改。
Sales were changing right now.
那你们招销售的时候，候选人必须得懂AI吗？
And and when you hire a sales rep, do they have to be AI AI built?
是的。
Yes.
我们其实改了第一轮面试的方式。
They're we actually changed our first interview.
不再看简历了，你得给我们做个东西出来。
No longer look at your resume, build something for us.
发个Notion链接过来，或者任何你亲手做的作品，我们只看你实际做出的东西。
Send the Notion link, send something that you built, we look at what you built.
那薪酬方面呢？
How about compensation?
这个有变化吗？
Is that changing?
我觉得以前在HubSpot，大家的生产力都远超平均水平。
My sense is we had an environment where, at least in HubSpot, we people are much more productive than the average person.
现在更是如此了。
Now it's even more so.
对。
Yeah.
你有重新考虑过吗？
Have you rethought?
我们也在调整这一点。
We are changing that as well.
明白了。
Okay.
他们需要更加注重实际能力。
So they need to be extremely lot more meritocracy.
你不能只是泛泛而谈。
You can peanut butter things.
我觉得SaaS时代就像把果酱随便抹开一样。
I think the Saas era is kinda like playbook peanut butter.
现在大家都觉得这是个挺平静的时期。
Everybody it's like a pretty peaceful era.
但现在已经是战时了。
Right now, it's a wartime.
完全对。
Totally.
这么一来，反而让人想问，我们以前做SaaS的时候到底在干嘛？
Actually, it makes the SaaS feels like, what are we doing back then?
很长一段时间里，根本没什么变化。
Nothing changed for the longest time.
一点都没变。
Like Nothing.
真的什么都没变。
It's nothing changed.
对吧？
Right?
就像我们当时那种想法
Like the notion we had
我离开HubSpot的部分原因就是，我们当时在打造一个平台。
And some of the reason I left HubSpot, it's like, okay, we're building this platform.
我们需要开发更多应用。
We need to build more apps.
我们需要持续开放这个平台。
We need to keep opening the platform.
我们需要招更多销售，就这么一直往前冲，直到几年前。
We need to hire more reps and just like, keep going until a couple years ago.
现在完全不一样了。
Now it's completely
变了。
changed.
你不可能再回来当运营了。
Well, you can't well, come back to become operator.
所以我才加入了红杉资本。
That's why I joined Sequoia.
我当时就想，天啊，这里有太多令人兴奋的事在发生。
I'm like, woah, there's so much exciting things going on.
当疫情爆发、市场剧变时，那更像是整个世界在改变，而不是人们在构建东西或技术在革新。
When we hit prior market when COVID happened, it's more like, world changing, but not like as people building things and technology changing.
你更喜欢战时还是平时？
Do you prefer wartime or peacetime?
战时有意思多了。
Wartime's way more fun.
这有什么区别呢？
It's like What's the difference?
你做法上有什么不同？
What do you do differently?
你会感觉更有活力。
You feel more alive.
好的。
Okay.
但你用的方法有什么不同呢？
But what methods do you use that are different?
我得照顾好自己的身体。
I need to take care of my body.
我每天都得锻炼。
I need to exercise every day.
我真没想到你会这么说，不过行吧。
I didn't think you were gonna say that, but okay.
是的。
Yeah.
感觉更灵活一些。
It's like it's more fluid.
你是更偏向自上而下，还是自下而上？
Are you more top down versus bottoms up?
我觉得现在公司对Notion的定位特别合适，因为我能非常顺畅地在任何方向上切换。
I think the company has a good spot Notion right now is a good spot with me in the sense that I can dance around really smoothly either direction.
我身边有很多值得信赖的人，他们能自下而上地推动事情。
There's enough people I trust really well, they can bottom up.
这一直让我很意外，毕竟我们公司有上千号人。
And it's constantly surprised me, like, we're a company of thousand people.
公司里大概有五六十个前创始人。
We have about fifty, sixty founders in the company ex founders.
他们真的能带头把事情做起来。
It's like they really can lead things.
对吧？
Right?
他们也欢迎我自上而下地介入任何我感兴趣、能带来新价值的领域，而且从不觉得我越界。
And and they welcome me to top down into jumping to any part that I can work on, I'm interested in, where I can provide you new value, and they don't feel territorial.
现在这种情况挺好的。
This is in a good place.
那些感觉不对的人，要么已经离开了Notion，那就让他们走吧。
Like, people who didn't feel that way, either no longer at Notion, let them go.
所以目前公司和我之间维持着一种挺特别的平衡。
So the company and me in a pretty interesting equilibrium at the moment.
我采访过Rippling的帕克·康拉德，他和很多创始人很像。
I interviewed Parker Conrad from Rippling, and he's similar as a lot of founders.
我注意到，HubSpot越做越大，但关键事务在各个部门之间却很难打通。
And the way it kind of strikes me is like the bigger HubSpot got almost all the important things cut across all those organizations.
在他们公司里，想找一个能真正推动跨部门事务的人实在太难了。
It was just very hard to find someone who could really drive something across the org, I suppose, in their org.
我们没有这么做。
We did not do that.
我后悔没在这方面多做点。
Regret not doing more of that.
我真正记得的是，疫情后我们开始扩张，后来试着按传统方式继续扩大，但那时我意识到，这根本不是我们的风格。
That's my real remember, I say right after COVID, scale, and then we try to scale more traditionally, I realized that's not us.
对吧？
Right?
那时候，也就是三年前，我们公司开始定下一个信条。
At that point, this was three years ago, we start a mantra in the company.
我们要做一支爵士乐队，而不是行进乐队。
We wanna be a jazz band, not a marching band.
我喜欢这个说法。
I love that.
我常拿交响乐团和爵士乐队作比较。
I I do the orchestra versus jazz band.
我不太喜欢行进乐队那种风格。
I like the marching band.
没错。
Yeah.
我意识到，我根本做不了行进乐队那种人。
Like, I realized I cannot be a marching band person.
这根本就不是我的风格。
That's just not me.
如果我把所有事都交出去，大家都会想：那我该做什么？
I will feel like if everybody I delegate everything, everybody do think, what do I do?
我觉得我做不到这样。
I feel like I can't do that.
这其实是一种自我反思的价值观。
So it's like a self reflecting value thing.
这对我来说是一种自我平衡。
It's auto equilibrium with me.
所以呢，就是这样。
So it's like, okay.
我是那种适合爵士乐队的人，所以我们一直在招更多像我这样的人。
I'm a jazz band person, so we've been hiring more and more jazz band person.
而且过去两年半人工智能崛起了，这时候更需要能即兴发挥的人才。
Also, AI happened in the past two two and a half years, so more jazz band people can really shine during this moment.
好的。
Okay.
我们之前是经理模式。
So we had manager mode.
我们还有创始人模式。
We have founder mode.
现在又出现了一种新模式。
There's this new mode.
这该叫什么名字呢？
What should we call it?
爵士模式？挺不错的。
Jazz mode, maybe.
这说法其实不坏。
That's not bad, actually.
因为爵士模式还是有一定结构的。
Because Jazz mode has some structure.
对吧？
Right?
你不想只是希望别人来参与。
You don't want to be like you want other people to contribute.
你希望大家都玩得开心，而不是只有你自己开心，或者你一个人说了算。
You don't want you want people to have fun, not just you having fun or dictating everything.
好。
Okay.
咱们试试看。
Let's give that a go.
咱们在X上发一下，看看反响如何。
Let's push that out on on X and see if it sticks.
我的意思是，
I mean,
确实是。
it's Yeah.
挺有意思的
Pretty
因为前几天我发了个调查，问大家这属于哪种风格。
because I I put a survey out the other day, like, what mode is this?
我当时在想，这是多西风格吗？
And I was like, is it Dorsey mode?
是原生风格吗？
Is it native mode?
我还问，是不是其他风格？
And I said, is it other mode?
结果有个人回复说，这是德佩切风格。
And somebody put in there, it's Depeche mode.
德佩切，德佩切，那都是乐队名。
Depeche, Depeche, That's all band.
德佩切。
Depeche.
是的
Yeah.
我们是在开玩笑。
Were making a joke.
但我喜欢爵士模式。
But I like jazz mode.
我觉得这个主意不错。
Think that's a good idea.
关于这个新世界，我还想问一个问题。
One other question on this new world.
我平时接触很多初创公司的创始人。
I work with all these startup founders.
做规划真不容易。
The planning's tricky.
我参加过你在红杉的路演，你当时有详细的计划，还有损益表。
And I sat through your pitch at Sequoia and you had a plan and you had a P and L.
我是说，从内部来看，计划方面到底是什么情况？
Like, what is it actually like internally, planning wise?
嗯，我们团队有自己的计划，跟现在大多数公司差不多。
Well, group has our plan, which is like most companies at this point.
如果这套计划能行得通的话，那我们团队的计划就有效。
If it works, if group has our plan.
对吧？
Right?
我觉得财务规划至少能让我知道，自己是不是在正轨上，大概跑得多快。
So I would say the financial planning is still at least I find useful to give you a you know you're on track, maybe you know how fast you're running.
对吧？
Right?
在我眼里，这就算一个财务计划了。
That's like a financial plan in my mind.
就是想知道，我现在是跑在七、八、一、九哪个阶段？
It's like, oh, am I running seven, eight, one, nine?
那到底是多少呢？
What number that is?
但产品战略？根本就没啥计划。
But product strategy is, oh, there's no freaking plan.
完全没计划。
No plan.
因为市场和技术变得太快了。
Because what changes the market, the technology is so fast.
不是六个月的事。
It's not in the six months.
也不是三个月。
It's not three months.
是每周都在变。
It's week by week.
所以你得随时调整这套东西。
So this is you have to jazz this thing.
财务方面，得往前推。
Financial things, march that.
对吧？
Right?
所以我们至少在财务上保持保守，偏向保守，同时也要敢于冒险。
Then at least we try to be conservative, mid to conservative with financial things and take risks.
得留点余地，成本成了另一个要考虑的因素。
There's a buffer, like, cost becomes another dimension.
如果你只是做软件，根本不用操心成本。
Like, if you're just building software, don't think about cost.
没错。
No.
但现在，至少我得认真考虑成本问题了。
Now you have to like, at least I have to really think about cost.
我能问你一下这个吗？
Can I ask you about that?
很多初创公司的毛利都很低。
Like, a lot of these startups' gross margins are awful.
像HubSpot这样的公司，毛利能达到86%左右。
And, like, I companies like HubSpot were, like, 86% gross margins.
这在SaaS行业算是常态。
That's sort of SaaS normal.
你们的成本和毛利现在怎么样？
What are happening to your costs and your gross margins?
你觉得像你们这样的公司，最终会稳定在什么水平？
And where do you think it settles out for companies like yours?
我没法给你一个确切的答案。
I don't have an answer for you.
我猜你们的毛利应该变差了，不过也许你们没怎么增加人手。
I'm assuming your gross margins have gotten worse, but maybe you didn't hire as much headcount.
这样的话，运营利润率可能反而好一些。
So maybe operating margins are better.
那你愿意让毛利跌到什么地步？
And how willing are you to be like gross margins to get radically worse?
你必须得接受。
Oh, you have to.
没错，就是这样。
You're Yeah.
愿意承受。
Down for
那你现在在做什么？
Like, what are you doing?
那你还没进入战斗状态吧？
Then you're not in the war mode.
对吧？
Right?
我觉得Notion面向的是知识型工作，可能和编程类产品的发展曲线不一样。
I think Notion is for knowledge work, which might have a different curve than coding products.
做编程类产品，你得用最前沿的模型，越智能越好。
Coding product, you run frontier models, usually the more the smarter the better.
但知识类工作就不一样了，其实我们已经能看到这种趋势了。
And knowledge work, you you can we already start to see this.
你完全可以满足于第二梯队的模型，比如中国产的、开源权重的那些。
You sort of can settle with the the second tier, the Chinese models, the open weight ones.
它们没那么聪明。
They're not super smart.
比如公司里那些填表、整理文档的工作，或者我咖啡洒在地毯上这种事。
Like, a lot of this kind of paper pushing information in your company, or I spill coffee on the carpet.
让HR团队下周派人来收拾一下就行了吧？
Can HR team come someone to tidy this up next week?
这种事根本用不着Opus这种高级模型来开工单。
That doesn't require opus to file tickets.
对吧？
Right?
所以这些问题正是Notion最擅长解决的。
So that's a lot of problem we are best at solving at Notion.
因此，它的增长曲线会跟现在最大的产品市场契合点——也就是编程助手——完全不同。
So that will have a different growth margin profile than the biggest product market fit today, which is a coding agent.
是的。
Yes.
明白了。
Okay.
你这思路跟我想法不谋而合，像是在重新创业。
You're kind of in my mind, king of refoundings.
对。
Yeah.
你已经这么干过两次了，至少Claude觉得你干过两次。
You've done it, I think twice, at least Claude thinks you've done it twice.
第一次是你和联合创始人搬去京都，把自己关在一个很小的房间里，重新打造了一次。
Your first one, you and your co founder moved to Kyoto, locked yourself in a very small room and rebuilt.
也没那么小，不过确实如此。
It's not that small, but yeah.
好的。
Okay.
最近，大家自然都了解这个故事了。
You know, more recently, obviously, everybody knows the story.
你去了墨西哥，提前拿到了模型，就这样转型成了一家AI公司。
You went to Mexico and you had early access to the models and you turned into an AI company.
谈谈这些重新创业的经历吧，当时你心里是怎么想的？
Talk about the different refoundings and what was your mindset at the time?
很多早期创业者都需要重新出发，很多成熟的SaaS公司也迫切需要转型。
There's lots of early stage founders who need to refound, and there's a lot of bigger SaaS companies that definitely need to refound.
对。
Yeah.
我觉得这个想法很好，顶尖的艺术家都会不断重塑自己。
I like as an idea, as the best artists reinvent themselves.
迈尔斯·戴维斯至少有三副面孔。
Miles Davis has three at least three faces.
对吧？
Right?
毕加索也是，两到三副面孔。
Picasso, two to three faces.
迈克尔·乔丹第二次复出，可能没那么厉害。
Michael Jordan, the second one, probably not as good.
你觉得呢？
No?
但真正最厉害的，还是史蒂夫·乔布斯，重新创业的王者。
But the the best one truly is Steve Jobs, the king of refunding.
Notion 最初的版本，我们花了四五年才找到产品市场契合点。
The first version of Notion is, like, it took us four, five years to find product market fit.
那段时间里，我们渐渐摸清了想做的产品方向，但钱快用完了。
In the middle of that, we sort of find the shape of the product we wanna build, but we'll be running out of money building it.
所以我和联合创始人萨姆决定，把所有人都遣散了。
So I co founder Sam and I said, let's just lay off everybody.
就剩我们两个人继续干。
Just go by the two of us.
这就是我们在日本的经历。
That's the Japan story.
我们是在日本重新打造了Notion。
We rebuilt Notion from Japan.
裁员这件事，做起来有多难？
How hard was that, the laying off everyone?
确实不容易。
It's not easy.
但有时候，你的直觉会告诉你，必须这么做。
But sometimes your body just tells you you have to do it.
就像你知道，如果不这么做，就真的没救了。
Like, you know you're dead if you don't do it.
你知道吗，那种感觉就像熊猫一样慵懒又平静
You know, there's like a panda energy just like
你干嘛偏偏选京都这么个地方？
Why the hell did you go to Kyoto of all places?
其实我们从来没去过日本。
Well, actually, we never been to Japan.
当你跟同事道别时，那些一起打拼的人，说到底，团队也没多大。
Because when you just say goodbye to your friends, all those people you work with not Notion, by all means, it's not big.
也就五个人左右。
Five people ish.
士气低得不行。
Morale so low.
对吧？
Right?
所以西蒙就说，咱们不如去个地方，换个心情。
So Simon and was saying, like, let's go somewhere that would just change the mood a little bit.
我们从来没去过日本。
And we never been to Japan.
一开始我们定了东京的民宿，结果发现房子都特别小。
So initially, picked Tokyo on Airbnb, realized that that the house are so small.
我很喜欢西蒙，但真不想和他挤在那么小的公寓里。
I don't I love Simon, but I don't wanna cram into a small apartment with him.
京都的公寓就大得多，因为二战时没被炸过。
And Kyoto, the apartment was much bigger because they didn't get burned on the World War two.
他们没被轰炸。
They weren't bombed.
他们没被轰炸。
They weren't bombed.
那座城市被保留了下来。
They spared that city.
那里有太多珍贵的东西了。
Too much treasures.
而且房子大得多，价格还更便宜。
And the house is much bigger, so and cheaper.
所以我们才选了京都。
So we that's why we picked Kyoto.
我们在旧金山有住的地方，也有办公室。
And we ran our place in SF and our office in SF.
这么做之后，第一次实现了现金流正向。
Actually become cash flow positive doing that.
这是头一回。
That's the first time.
但一旦落地，就开始疯狂写代码，吃点东西，再接着写代码。
But once you land, you start coding, just those kind of rapid coding, eating, coding, eating.
这种生活方式其实特别自由。
It's actually really liberating.
我在日本住了好几年。
I lived in Japan for a couple of years.
日本有没有什么特别的地方，启发了你重新创业或者打造这款产品？
Was there something about Japan that was inspirational in your refounding or the way you built the product?
那里有一种独特的美学，跟Notion的感觉很搭。
There's a certain aesthetic there that sort of rhymes with Notion.
还是说，你们在哪儿都行？
Or could you have been in anywhere?
你们去泰国不行吗？
Could you have been in Thailand?
我们一般都喜欢夏天暖和的地方。
Well, usually, we like summer warm.
西蒙喜欢温暖的气候。
Simon likes warm climate.
最好还能游泳。
Ideally, he can swim.
日本只是刚好我们俩都对那儿感兴趣。
Japan just happened to be we're both curious about it.
但你一旦去了京都，我就说不好了。
But once you go to Kyoto, I don't know.
你 probably 去过京都吧。
You've probably been to Kyoto.
那是一座非常特别的城市。
It's a very special city.
确实特别。
Very special.
对。
Yeah.
特别有启发性。
Very inspirational.
确实很有启发性。
Very inspirational.
那些东西都是上百年的老物件了，这家店都有一百五十年历史了。
Those things are hundreds everything oh, this shop, 150 years old.
对吧？
Right?
我来打开那扇门。
I'll open that door.
神社
The shrine
所以你去过泰国的普吉岛之类的地方，那里风景很美。
will So you're there for eight done a beautiful place in Thailand, Phuket, whatever.
你觉得即使换别的地方，你也能成功吗？还是说那个地点本身有什么特别的魔力？
Do you think you still would have pulled it off or was there something magical about the location that really worked?
大多数事情，其实都是我们讲给自己听的故事。
Most of the things are a story we tell ourselves.
当然。
Of course.
但故事才是真实的真相。
But story is the true story.
我们对自己讲的故事，就像说京都是个特别的地方。
The story we tell ourselves is like Kyoto is a special place.
如果你能在任何地方重新开始，那一定是在京都。
If you can pull off anywhere, you can pull off from reborn in Kyoto.
而且那里真的很鼓舞人心。
And it's it's also inspiring.
你看看周围的环境就知道了。
Just look at the surrounding.
这里的工艺水平，可能是亚洲之冠。
The craft is probably craft capital of Asia.
对吧？
Right?
所以只要你安心下来就好。
So just make you okay.
Notion 本质上是个工具，是为人类设计的，而那些使用者正在打造刀具、陶瓷杯、榻榻米坐垫——他们才是其他工具的使用者
If Notion fundamentally is tool, it's a tool for humans, and those people are building knives, ceramic cups, tatami seats, like, they're the tools other
工具是为一个核心服务的。
tools It's for a central.
这是一个工艺中心。
It's a craft central.
你怎么可能不被这种打造更好软件的氛围所激励呢？
And it's like, how do we how can you not be inspired by building a better software
你还没回答我的问题。
haven't answered my question.
什么问题？
Which is?
假如你去了菲律宾的某个地方，那儿也很美，你还能做成这样吗？还是说，身处这样一个以工艺为核心的地方，有什么特别重要的地方？
If you had gone to whatever, let's pick someplace in The Philippines, and it was pretty, would you have pulled this off, or was there something exceptionally important about being immersed in such a craft centric place?
我以为我做到了，但说实话，这个故事才更动人。
I thought I pulled it off, but yeah, but the story is the better story.
好吧。
Fine.
好的。
Okay.
所以你这次彻底重建，可不容易。
So your big refounding, not easy.
之后也没马上找到产品市场契合点，到现在还是没完全找到。
Didn't immediately hit product market fit after, still not
没错，没那么容易，花了一段时间。
an easy No, took a while.
是的。
Yeah.
之后呢？
After
那段时间。
that.
因为重建花了一年半的时间。
Cause it took a year and a half to rebuild
从零开始。
from scratch.
那时候你为什么不干脆放弃，重新开一家公司呢？
Why didn't you just quit and start a new company at that point?
我们从不为了创业而创业。
We never start a company for starting a company.
你真正想做的，是打造一些东西。
You want to build something.
从去年上大学起，我就一直着迷于这个工具型产品的想法。
Like I got obsessed about this notion shaped tool problem since last year college.
西蒙和我都在这个‘思维工具’的圈子里，这个传统可以追溯到感恩债务领域，最早的一批湾区计算先驱，这条脉络一直延续至今，直接源自他们。
And Simon and I were both in this kind of tools for thought community, which is you can trace the lineage back to grateful debt areas, the first generation computing pioneers around the Bay Area, like, the lineage traced way back, directly back to there.
我想延续这条脉络。
I want to continue that lineage.
所以对我来说，这并不是
So to me, it's not a
你那时候也没别的想法在脑子里转悠吧。
wasn't like you had another idea kicking around.
这可是你一生的事业。
Is your life's work.
这个想法太大了，我根本吃不透。
And this is, like, bigger than I can ever chew on this idea.
所以我根本不会去干别的。
So I wouldn't do anything else.
他也在那个圈子里，对这个想法同样着迷。
He's in that community, he's obsessed with this idea too.
如果我们不做这个，迟早也会去搞另一个公司，做同样的事。
So if we don't work on this, we'll work on another company doing the same thing.
明白了。
Okay.
我回头再聊重建的事，不过先说说格雷ateful Dead，你研究的那些六十年代重要的计算机科学家。
I'm gonna get back to the refounding, but just on the Grateful Dead, You study very important computer scientists from the sixties.
你肯定是史蒂夫·乔布斯的粉丝吧。
You're a student of Steve Jobs, I'm sure.
当年感恩而死乐队和他们的音乐有种匠人精神，公司和乔布斯本人更是如此。
There was sort of a craft ethos about the Grateful Dead and the music back then, but certainly about the company, certainly about jobs.
这种精神现在还存在吗？
Is it still here?
如今硅谷看起来更偏向商业化，而不是匠人精神了。
It seems like it's a lot more commercial than craft today in Silicon Valley.
我们是不是已经失去了它？
Have we lost that?
我觉得大多数人早就没有了。
I think most people don't have it.
大多数公司也都失去了。
Most company don't have it.
我同意。
I agree.
像史蒂夫·乔布斯这样的人，我敬仰他们，因为他们是人性与科技的交汇点。
Like Steve Jobs, I look up to him or people like him as like interception of humanity with technology.
但硅谷这边，科技更多是一种动手和玩票的文化。
Versus Silicon Valley, at least tech is largely a technology and tinker culture.
而这种玩票文化，根本不在乎自己的历史，甚至都不算科学。
And tinker culture doesn't even know the history of it's not even science.
就是纯粹的捣鼓。
It's tinkering.
对吧？
Right?
就是有人喜欢在车库里捣鼓东西，搞出个能用的玩意儿，碰巧符合市场需求，就慢慢做成了生意。
It's like people like to build things in their garage, and something works, then maybe has a product market fit and grow a business out of this.
不过说句公道话，沃兹和沃兹当年也是在车库里捣鼓的。
To be fair, Waz and Waz was tinkering in the garage.
对吧？
Right?
是的。
Yeah.
但你得亲自去认识一些人，才能体会到其中的人文层面。
But you require to meet someone to to bring the human side of it.
对吧？
Right?
我觉得极客文化的问题在于，你根本不知道前人走过哪些路。
I think tinker culture, the risk of us, like, you don't know what comes before you.
而科学至少会让你尊重它的历史。
Science, at least, you you respect the history.
比如，如果我们问大家道格拉斯·恩格尔巴特或者艾伦·凯是谁，大部分科技圈的人都一无所知。
Like, if we ask people who Douglas Engelbauer or Alan Kay, those computing pioneers are, most people in tech have no idea.
说实话，我之所以知道他们，是因为我曾经给雷·奥兹西工作过。
Honestly, the only reason I know them is I worked for this guy Ray Ozzi.
雷·奥兹西，就是那个
Ray Ozzi, like
当然了，你知道的，他是笔记系统的开创者，而且一直对这两个人着迷，整天挂在嘴边。
Well, you would know, of course, because he's the father of notes, but he was obsessed with those two and talked about them constantly.
我知道，雷亚兹。
I know, Riazi.
他以前还当过微软的首席技术官，对吧？
It's like Microsoft CTO for a while too, right?
不过，这条传承线现在已经断了。
So that lineage is broken, by the way.
那些记得七八十年代的人，现在大多都不在科技圈了。
The people who remember those eighties seventies and eighties, they were sort of people don't yeah.
他们现在早就不再从事科技行业了。
They no longer work in tech today.
所以科技行业根本不知道自己的过去。
So tech is like industry doesn't know its past.
如果你不了解自己的过去，就不懂历史，而历史就是人类本身。
If you don't know its past, you don't know history, which is humanity.
我觉得你的视角就是盯着眼前的东西，看竞争对手在做什么，但如果你能引入人性的维度，那你背后就有一整套其他领域的资源。
I think your your your point of view is, like, what's in front of you, what your competitor is doing versus if you bring humanity, you have all other discipline around you.
你背后还有整个历史作为支撑。
You have all the history behind you.
你能借鉴的好东西多得是。
There's way more good stuff you can steal from.
真希望更多科技创业者能从侧面、从背后去汲取灵感。
I wish more tech founders steal from aside or from behind.
我觉得他们确实会从中受益。
I think they'd benefit from it too.
再说回重新奠基这件事，很多创始人现在都在原地打转。
Back to the refounding, lots of founders are going sideways.
很多2021年成立的公司，现在都停滞不前了。
Lots of companies that started in 2021 are going sideways.
他们就十来个人。
They've got 10 people.
他们可能只是刚好收支平衡。
It's like maybe they're cashflow breakeven.
对于那些长期原地打转、拿了风投、还在找产品市场契合点的创始人，你有什么建议？难道每个人都该狠下心来，收缩规模，搬去京都重新开始吗？
What's your advice to somebody who's a founder has been going sideways for a long time, have ventured dollars in like you and working on product market fit, like, should everybody just be like, rip the band aid off and shrink it and move to Kyoto and and refound?
如果身体在提醒你的话。
If their body tell them so.
对我来说，就是一种直觉，你就是能感觉到。
Like, I, for me, it's like, there's you just feel it.
你得来点彻底的改变。
You have to do something drastic.
那时候西蒙正好拔了智齿之类的。
And Simon, during that time, was he broke his he got his wisdom teeth taken out or something.
我就一个人待着。
So I'm by myself.
我记得我当时在我前女友的工作室里。
I was like, I remember I was at my ex girlfriend's studio.
我就觉得内心有个声音在说，别无选择。
I just, like, an inner urge is like, there's no way out.
我必须这么做。
I have to do this.
等你落地日本的那一刻，就会感觉特别解脱。
And it's like, then you'd feel liberated once you land in Japan.
我觉得你的建议其实是，应该有更多人这么做，多听身体的信号，勇敢一点冒险。
I guess your advice would be is probably more people should do it, and they should listen to their inner body and be more risk seeking on that.
因为他们在风投支持的初创公司里很容易陷进去，久久出不来。
Because they can get stuck for a long time in a venture backed startup.
他们可能整整十五年职业生涯都在做一件没有起色的事。
They can spend fifteen years of their career on something going sideways.
我觉得现在就是做这件事最好的时机，因为市场机会正大开。
I think there's no better time than now to do that, because the market dynamic is wide open.
对，我同意。
Yep, I agree.
你第二次创业的时候，公司已经有一千名员工了。
Okay, you re founded a second time when you had a thousand employees.
那时候大概只有五百人左右。
Back then it was probably 500.
对。
Yeah.
明白了。
Okay.
五百名员工。
500 employees.
完全是另一回事了。
Different animal.
我不用再讲你讲过上千遍的那些故事了，毕竟大家都知道来龙去脉，但幕后到底是什么情况？
Without telling the same story you've told a thousand times, because I think everyone knows the lore, what was it like kind of behind the scenes?
你当时是什么感受？
What did you feel?
你身体的感觉也是一样的吗？
Was it the same thing your body felt it like?
那到底发生了什么？
What was the story?
第二次是在墨西哥的坎昆。
So the second time was in Mexico, Cancun.
对。
Right.
当时正好在外面。
Happened to be out somewhere.
对。
Right.
你总是在挺不错的地方。
You're always in a nice place.
我们当时提前拿到了GPT-4的模型。
We just got early access GPT-four model.
感觉整个世界都停了下来。
It feels like the world is stopping.
你第一次用GPT-3的时候，感觉还行。
Like, you first interaction with the GPT-three is fine.
对我而言，GPT-4简直像一场宗教体验。
GPT-four is a religious experience for me.
真的，太震撼了。
It's like, holy shit.
这是一种全身都受到冲击的宗教级体验。
It's a it's a full body religious experience.
你必须得做点什么，不能错过它。
Like, you have to do something with this.
就是这个，彻底改变了所有东西。
This is what changed everything.
不管你做什么，如果不抓住这个，其他都毫无意义。
Anything you do, if you don't do this, you will be meaningless.
我觉得公司一直挺坚定的，虽然当时很多人怀疑，毕竟三年前正是加密货币借贷的热潮期。
I think the company has been fairly there's be people doubt because remember that around three years ago, that's the the crypto era lending.
对吧？
Right?
很多人都在问，这不会又是另一个加密项目吧？
A lot of people, is this another crypto thing?
当时确实有不少人持怀疑态度。
There were a lot of doubters.
有些人后来离开了公司，但至少在早期，我和西蒙的信念非常明确——我们非做不可。
Some of them, obviously, no longer at the company, but the conviction is so clear, at least for me and Simon in early days, like, we just have to do this.
你们在重建整个业务的时候，有没有先退两步，好为了往前迈十步？
Did you take two steps back in the business to go 10 steps forward?
也就是说，重建期间，收入真的下滑了吗？
Like, did revenue really slow while you're rebuilding everything?
我们的第一个产品其实很快就带来了明显的收入增长。
Our first product actually has a meaningful revenue boost.
我们其实是在ChatGPT出现前两周推出了第一个AI产品。
We launched our first AI product actually two weeks before ChatGPT happened.
那是个AI写作工具。
It's like AI writing product.
这个产品带来了明显的增长。
That has a boost.
但之后，我们从2022年底开始一心想要打造一个智能代理产品。
But after that, we've been pretty much wanna build an agent product in 2022, end of twenty twenty two.
那可真是个艰难的过程。
And that's a slog.
花了我们一年半的时间。
That took us a year and a half
那时候根本没跑起来。
to really It didn't really work back then.
完全没成功。
Didn't work at all.
我觉得我和西蒙的问题在于，我们总是想得太靠前了。
Like, Simon I would say my problem and Simon's problem, we might be living too much into the future.
对。
Yep.
新技术一来，就给你当头一棒。
It bites you with new technology.
我们啥都试过，Tropit 还给我们搭了个模型。
Like, we try everything, and Tropit built a model for us.
我们试过 OpenAI 的微调，也试过别的模型。
We work OpenAI fine tuning or another model.
可全都不管用。
Just none of them work.
这一年半简直白熬了，士气低得不行。
That would like, year and a half just go with no error, and morale is definitely low.
直到一年半后，才终于有点起色。
Only start to work a year and a half.
收入转折点也差不多是那时候，就是AI产品开始有起色的时候，对吧？
Revenue inflection is around the same time, right, when the AI product
变得更好了。
got better.
是的。
Yeah.
但那确实是一段特别艰难的时期。
It's like but that was a was a pretty painful period.
在Notion的发展史上，那是你经历过的最痛苦的阶段吗？
Was that the most painful period in Notion history for you?
那种痛苦不太一样。
Different type of pain.
那最痛苦的是什么时候？
What was the most painful?
就是最初那两三年，产品还没找到市场契合点的时候，那种绝望感。
Just when those two, three, early days, pre product market fit is despair.
就是那种感觉
Was just like
绝望？
Despair?
你真用这个词吗？
Is that the word you use?
明白了。
Okay.
你根本不知道问题出在哪儿。
You don't know where it's coming from.
也不知道该往哪儿走。
You don't know where to go.
对吧？
Right?
那时候增长停滞，我们反复重建AI的基础架构。
When your growth rate low, like, during that period of time, were trying to rebuilding our AI foundation multiple times.
是啊。
It's yeah.
这可不怎么样。
That's not great.
现在简直是处处着火。
Right now, just everything on fire.
这是一种不一样的痛苦。
It's a different type of pain.
不错。
Nice.
很多五百人规模的公司，尤其是SaaS公司，都在搞AI，但真正重新打基础的却不多。
So a lot of people, 500 person companies, a lot of SaaS companies are doing the AI thing, but certainly haven't refounded.
有什么建议吗？
Advice?
先从你的产品开始。
Start with your product.
我们就把它当成酿啤酒，而不是建桥梁来对待。
We'll just talk about treat it like brewing beer and not building bridges.
归根结底，大多数科技公司靠的还是产品。
Because that's end of the day, most tech companies is the product.
对吧？
Right?
所以要从产品入手，这样创始人才能亲身体验这种做事方式。
So start with the product because then the founder has to be really hands on to experience that way of building things.
非创始人领导的公司能做到吗？
Can non founder led companies pull it off?
我不确定。
I don't know.
这会很难。
It would be tough.
我觉得只有创始人才有底气去推动改变。
I think founder has the moral high ground to to change things.
人们对创始人更宽容，也更容易原谅。
People are more tolerant and forgiving with founders.
对吧？
Right?
我觉得这个人得学会用语言模型，真正去感受它。
I think the person need to be use language modeling in just know how to feel it.
你得亲自体会AGI，体会AI。
Like, you gotta feel the AGI, feel the AI.
是的。
Yeah.
光靠看书是不行的。
You can't read about it.
光看YouTube也无济于事。
You can't watch YouTube.
你得亲手去搭建。
You gotta build.
关于 Bayt 播客
Bayt 提供中文+原文双语音频和字幕，帮助你打破语言障碍，轻松听懂全球优质播客。