BuildSpeak每日 builder 文摘
今日归档生词本关于
🎙 播客AI & I by Every· 2026 年 7 月 15 日· 10,413 词 · 约 52 分钟

The Founder of a $1.5B AI Company on What Comes After the First Wave of AI Apps

SPACE 播放 / 暂停·←→ 上一句 / 下一句
Speaker 100:00 - 00:07
You are the cofounder and CEO of Granolah. One of the first new AI apps where people were like, holy shit. This actually works.
Speaker 100:00 - 00:07
你是 Granolah 的 cofounder 和 CEO。这是最早一批让人直呼“holy shit”的新 AI app 之一。它居然真的能用。
Speaker 200:07 - 00:24
Meeting notes are not the end all, be all value that everyone's running after. There's something much bigger. I think we are in the very, very early steps of a computing revolution. What has come thus far will pale in comparison to what will come soon. Startups are like knife fights.
Speaker 200:07 - 00:24
会议纪要并不是所有人都在追逐的那种终极价值。还有更大的东西。我觉得,我们正处在一场计算革命极其、极其早期的阶段。迄今为止出现的一切,和很快将要到来的相比,都会显得逊色。创业公司就像持刀肉搏。
Speaker 200:24 - 00:29
And I thought startups were just really hard when they weren't working. Turns out they're really hard even when they're working as well.
Speaker 200:24 - 00:29
我以前以为,startup 只有在不顺的时候才会特别难。后来才发现,就算一切进展顺利,它也一样非常难。
Speaker 100:43 - 00:58
Every is the only subscription you need to stay at the edge of AI. If you care about being on top of the latest models and using the latest tools, you have to subscribe to Every. We'll help you separate the signal from the noise. Go to every.to/ subscribe today. And now back to the episode.
Speaker 100:43 - 00:58
Every 是你保持在 AI 前沿所需的唯一订阅。如果你在乎紧跟最新 model、使用最新工具,那你就必须订阅 Every。我们会帮你从噪音中分离出真正有价值的信号。现在就去 every.to/ 订阅。接下来回到本期节目。
Speaker 100:58 - 01:00
Chris, welcome to the show.
Speaker 100:58 - 01:00
Chris,欢迎来到节目。
Speaker 201:00 - 01:02
Hey, Dan. Great to see you.
Speaker 201:00 - 01:02
嘿,Dan。很高兴见到你。
Speaker 101:03 - 01:23
Good to see too. It's been a while. So for people who don't know, you are the cofounder and CEO of Granolah, one of my favorite AI apps. One of the apps that I think really kicked off the wave in the in, like, you know, I guess, two two or three years ago. It was the one of the first, like, new AI apps where people were like, holy shit.
Speaker 101:03 - 01:23
也很高兴见到你。已经有一阵子了。给不了解情况的人介绍一下:你是 Granolah 的 cofounder 和 CEO,也是我最喜欢的 AI app 之一的创始人。它是我认为真正开启了这波浪潮的 app 之一,大概就是两三年前吧。它是最早一批全新的 AI app 之一,当时人们看到都会说,holy shit。
Speaker 101:23 - 01:45
This actually works, and it's useful. And since then, you've gone on to I I think the team is about 55. You've you've you've raised at, a $1,500,000,000 valuation. You're just killing it. And we had a really good conversation in December 2024, so almost two and a half years ago about putting soul into your product and making products that are delightful.
Speaker 101:23 - 01:45
这东西居然真的能用,而且很实用。从那以后,你们一路发展壮大——我想团队现在大概有 55 人了。你们融资时的估值达到了 $1,500,000,000。你们的发展势头非常猛。我们在 2024 年 12 月有过一次非常好的对谈,聊的是如何把灵魂注入你的产品,以及如何打造让人愉悦的产品;那已经是将近两年半前了。
Speaker 101:45 - 01:56
And I'm just super curious to hear what's on your mind, how things are going, and to sort of share out both of our journeys over the last couple of years and and think about the future together.
Speaker 101:45 - 01:56
我就是特别想听听你最近在想什么、事情进展得怎么样,也想把我们俩过去这几年的经历拿出来一起聊聊,并且一起想想未来。
Speaker 201:57 - 02:16
Couldn't be more excited to chat. I I can I've been following all the different things you've been doing, from a distance. I'd to hear the latest from the inside on how that's going. Just to react on one thing you said, you said we're killing it, right? And we're very lucky, things are going well.
Speaker 201:57 - 02:16
我非常期待这次聊天。我一直有点隔着距离关注你在做的各种事情。很想听听你从内部视角讲讲最新进展到底怎么样。顺便回应一下你刚才说的一点,你说我们现在做得很猛,对吧?我们确实很幸运,事情进展得不错。
Speaker 202:17 - 02:41
Something I was unprepared for, so I did a previous startup and startups are, they're knife fights, right? They're really, I don't know, maybe life's easy for you, but for me it's really hard. And I thought startups were just really hard when they weren't working. Turns out they're really hard when they're working as well. It's just, it's a knife fight when things are going well or not going well, and I was a bit unprepared for that.
Speaker 202:17 - 02:41
有一件事是我之前完全没准备到的。我以前做过一家 startup(创业公司),而 startup 就像持刀肉搏,对吧?真的非常难。我不知道,也许你的人生比较轻松,但对我来说这事真的很难。我原本以为 startup 只有在不顺的时候才特别难。结果发现,它在顺的时候也一样很难。就是说,不管事情进展顺利还是不顺利,都是一场持刀肉搏,而我之前对此有点准备不足。
Speaker 202:41 - 03:03
So it's just, I'm very, I had a professor in journalism school who, he was basically like the last thing the world needs is another profile that puts someone up, you know, it's like oh this person's so great. And it's like no, no, Everything's hard today. If I fight day in, day out, like that's that's my mentality. Just wanted to share that.
Speaker 202:41 - 03:03
所以就是,我很——我以前在新闻学院有位教授,他基本上的意思是,这个世界最不需要的,就是再来一篇把某个人高高捧起来的人物特写,好像在说“哦,这个人真是太厉害了”。但其实不是,不是。如今一切都很难。我每天都在硬扛、日复一日地拼,这就是我的心态。只是想分享一下这一点。
Speaker 103:03 - 03:24
I feel you, man. I definitely feel the same way where I've been doing companies for a long time and Every is the first one that's, like, really growing and just really working in this way that feels like it's a real company. We have 30 people now, you know? Congrats. Thank you.
Speaker 103:03 - 03:24
我懂你,兄弟。我绝对也有同感。我做公司已经很多年了,而 Every 是第一家真正开始增长、真正以这种方式运转起来、让人感觉它是一家真正公司的公司。我们现在有 30 个人了,你知道吗?恭喜。谢谢。
Speaker 103:25 - 03:36
And when it is working yeah. Like, a good example, clogged tag dropped yesterday. We've been working on a Slack agent. Clogged tag dropped yesterday. And so then everyone's looking at you like, what do we do?
Speaker 103:25 - 03:36
而当它真的开始运转时,没错——举个很好的例子,clogged tag 昨天发布了。我们一直在做一个 Slack agent(智能体)。clogged tag 昨天发布了。然后所有人都会看着你,问:我们现在该怎么办?
Speaker 103:36 - 04:07
And I'm like, we've we've we're prepared for this. We've been thinking we know that we knew this was gonna happen. And the especially in AI, I feel like the ground shifts so quickly. And I I assume for you, like, one of the things you're you're talking to, and I'm talking about it, and I'm actually really curious, is you were the first one to do really great AI meeting notes and meeting transcriptions. And then immediately, everyone just, like, put meeting notes into their product.
Speaker 103:36 - 04:07
而我会说,我们——我们其实早就为这件事做了准备。我们一直在思考,我们知道这事会发生。尤其是在 AI 领域,我感觉地面变化得特别快。我想对你来说也是这样——有一件你正在谈、我也一直在谈,而且我真的很好奇的事是:你们是第一批把 AI meeting notes(会议纪要)和 meeting transcriptions(会议转录)做得特别出色的人。结果紧接着,所有人都像是一下子把 meeting notes 塞进了自己的产品里。
Speaker 104:08 - 04:12
We actually we have a we have this product called Monologue that has me a meeting notes button, you know?
Speaker 104:08 - 04:12
其实我们也有一个叫 Monologue 的产品,它里面就有一个 meeting notes 按钮,你知道的。
Speaker 204:12 - 04:12
Oh, cool.
Speaker 204:12 - 04:12
哦,挺酷的。
Speaker 104:12 - 04:27
I didn't know that. So we're, like, slightly competitive. Yeah. I don't I don't really think of it as being an exact an exact one to one replacement. I actually have both on my computer, but but still, that's probably what you're referring to when you say knife fight.
Speaker 104:12 - 04:27
我之前不知道这个。所以我们算是有点竞争关系。对。不过我不太会把它看成一种完全一比一的替代品。其实我电脑里两个都有,但你说的 knife fight,大概指的就是这个。
Speaker 104:27 - 04:32
So tell me about that. What is that like? How has that actually affected your business at all?
Speaker 104:27 - 04:32
那你说说这个吧。那是一种什么感觉?它实际上到底对你的业务产生了什么影响吗?
Speaker 204:32 - 05:02
That's actually not what I was referring to as a knife fight. I mean, that's part of it. I think it's just, I guess what I was referring to is, I think by definition, a startup is, you're either fighting for survival. You're always fighting for survival and it's like, if things are going well, you're fighting to, it's like this big wave and you're like on the surfboard and you're trying desperately not to fall off that surfboard. And so you're kind of always just beyond the edge of your abilities.
Speaker 204:32 - 05:02
其实我说的 knife fight 不是单指那个。我是说,那只是其中一部分。我想我真正想表达的是,我觉得按定义来说,startup 就是在为生存而战。你总是在为生存而战;如果事情进展顺利,那你也是在拼命跟上,就像有一道巨浪,你站在冲浪板上,拼命不让自己从板上掉下来。所以你基本上总是在自己能力边界之外一点点的地方。
Speaker 205:03 - 05:47
And it's almost like by definition and therefore I'm always doing, you always hear this as like, oh, it's like as a founder, you're always doing stuff you don't know how to do. It's like, oh yeah, actually that at least that had been my experience. It's like, I'm always just in the space where like it's stuff I've never done before or it's just more than we've ever like bit off in the past. Yeah, the competition, like here's my view and I feel like I've said this, let me think about it. When I think about this is going sound like one of those bullshit answers around competition.
Speaker 205:03 - 05:47
而且几乎可以说这是定义使然,所以我一直都在做——你总会听人这么说:哦,作为 founder,你总是在做自己不会做的事。感觉就是,哦,对,至少我的经历确实是这样。我好像一直都处在那种区域里:要么是我以前从没做过的事,要么就是我们以前从没吃下过这么大的挑战。对,至于竞争,我的看法是这样,而且我感觉我以前也说过这个,让我想想。我现在这么说,听起来可能会像那种关于竞争的 bullshit 式回答。
Speaker 205:47 - 06:23
There's always the interview where it's like, Hey, Microsoft just copied your product. How do you feel about it? And the founder has to come up with some answer. The way I view it is I think we are in the very, very early steps of a computing revolution and that there's What has come thus far will pale in comparison to what will come soon. I think that it doesn't Meeting notes are not the end all, be all value that everyone's running after.
Speaker 205:47 - 06:23
总会有那种采访:嘿,Microsoft 刚刚抄了你的产品,你对此什么感受?然后 founder 就得想办法给出一个回答。我的看法是,我认为我们正处在一场计算革命的非常非常早期阶段,而到目前为止出现的一切,和接下来很快会到来的东西相比,都会显得黯然失色。我觉得,meeting notes 并不是所有人追逐的终极价值。
Speaker 206:24 - 06:47
There's something much bigger, and I think it's what interface we use for work and how we work and what does that look like in the AI native world. When I think about competition, I'm just kind of like, what people are fighting for today doesn't matter. There's this incredible opportunity ahead. And I think we have a shot at it and a few other companies have a shot at it, right? That's really what matters.
Speaker 206:24 - 06:47
真正更大的东西在别处,我觉得那就是:我们在工作中使用什么样的 interface(界面),我们如何工作,以及在 AI native 世界里这些会是什么样子。当我想到竞争时,我基本会觉得,人们今天争夺的东西并不重要。前面有一个极其巨大的机会。我觉得我们有机会抓住它,也有几家公司同样有机会,对吧?真正重要的是这个。
Speaker 206:47 - 07:08
So in the same way where people are like, Oh, Granolah's doing really well. It's like, well, in my mind it's like easy come, easy go in a way, right? Like meeting notes are useful, but like a lot of things are going to change in the future. Just because people use us today doesn't mean they're going use us for that in the future if we're not the best at that next thing. So yeah, don't know, that's my zoomed out view.
Speaker 206:47 - 07:08
所以同样地,当人们说,哦,Granolah 现在做得很好。我心里会觉得,某种程度上这也是 easy come, easy go,对吧?比如 meeting notes 确实有用,但未来很多事情都会改变。人们今天使用我们,并不意味着如果我们在下一件事上做不到最好,他们未来还会继续因为这个而使用我们。所以,是啊,我也不知道,这就是我更宏观的看法。
Speaker 207:08 - 07:27
And the other thing to say is we were not the first meeting notetaker, right? Meaning like AI notetakers have been around for ages. We were, let's say, heavily inspired by a lot of things that came before. So I can't be like, Oh, how dare you be inspired by what we did? It's just in an ecosystem.
Speaker 207:08 - 07:27
还有一点要说的是,我们并不是第一个做会议记录的 notetaker,对吧?也就是说,AI notetaker 这种东西早就存在很多年了。可以说,我们在很大程度上受到了很多前人做法的启发。所以我也不能说,哦,你怎么敢受我们启发?这本来就是同一个生态系统里的事情。
Speaker 207:28 - 07:50
It has affected the company a lot less than I thought it would. I think within Notion copied or whatever, let's say launch something, I think heavily inspired by granola, OpenAI did, Zoom did recently. So it's like in some ways it's like I didn't have to sit with the nightmare submit scenario of like, what would it be like the day that launches? It's like that kind of came and happened and we're
Speaker 207:28 - 07:50
它对公司的影响,比我原本以为的小得多。我觉得 Notion 算是 copy 了,或者说,姑且算是发布了某种东西,我认为那明显受了 granola 的启发,OpenAI 做了,Zoom 最近也做了。所以某种意义上,我不必一直坐在那里幻想那种噩梦场景:等那一天他们发布时会怎么样?因为那件事某种程度上已经发生了,而我们——
Speaker 107:50 - 07:51
It's clear. Yeah.
Speaker 107:50 - 07:51
这很清楚。对。
Speaker 207:51 - 08:05
Yeah, we're still like it hasn't changed anything from our growth rate or whatnot. But again, we're still a drop in the bucket compared to what we'd like to be in terms of how people use us and how many people use us. So I think it's just very early days is my view.
Speaker 207:51 - 08:05
对,我们现在依然是——它并没有改变我们的增长率之类的东西。但再说一次,和我们希望达到的规模相比,不管是人们如何使用我们,还是有多少人在使用我们,我们现在都还只是沧海一粟。所以我的看法是,现在仍然非常早期。
Speaker 108:05 - 08:17
That makes sense. I just got to say I love getting to talk to you because I feel like you're just so honest about what you're dealing with. It's it's the best. It's so fun. It's so it's very different.
Speaker 108:05 - 08:17
这很有道理。我得说一句,我很喜欢和你聊天,因为我觉得你对自己正在面对的事情真的非常坦诚。这种感觉太好了。特别有意思。也真的很不一样。
Speaker 108:17 - 08:48
It's very different from a lot of a lot of people, and I think this is gonna be a very fun conversation. But one of the but what what I wanna go back to is and I'm happy to share too. What I want to go back to is you said you're balancing on the surfboard and you're trying not to fall off. And, you know, normally in startups, you're like, you're on the surfboard and you're just, like, paddling and you're, like, waiting for the wave. And then I think for both of us for you more so than me, but for both of us, you're like, I'm on the big one now.
Speaker 108:17 - 08:48
这和很多很多人都很不一样,我觉得这会是一场非常有意思的对话。不过我想回到一点上——当然我也愿意分享——我想回到的是,你刚才说你像是在冲浪板上努力保持平衡,尽量别掉下去。你知道,通常在 startup 里,更像是你站在冲浪板上,不停划水,等着浪来。然后我觉得对我们俩来说——对你尤其如此,比我更明显——现在更像是,你会说,我现在已经上到那个大浪上了。
Speaker 108:48 - 09:09
Like, let's try not to get let's try not to fall off and just get pounded, you know? And so what what was what are some of those things recently where you're you've felt like you've been at the edge of your ability and you're like, I gotta figure this. I gotta figure this out. This is this is maybe it's a good thing, but generally it's like, shit. This is very different.
Speaker 108:48 - 09:09
就像,尽量别被甩下去,别被浪狠狠干翻,你懂吧?所以,最近有哪些事情让你觉得自己已经到了能力边缘,会想“我得把这事搞明白,我得把这事搞定”?也许这总体上是件好事,但通常第一反应还是,靠,这太不一样了。
Speaker 209:09 - 09:33
So a lot of it's kind of my limitation to lack of experience as a founder. So like my last startup, Socratic, we were 12 people. Just going from 12 to We're actually over We're somewhere between, I don't know, sixty and seventy now. That's very different, and I've never done that before. Organizing humans, that's one.
Speaker 209:09 - 09:33
很大一部分其实来自于我作为 founder 经验不足这一局限。所以比如我上一个 startup,Socratic,我们当时只有 12 个人。光是从 12 个人增长到——我们现在其实已经超过——大概在六七十人之间吧。这就非常不一样了,而我以前从没经历过。组织和管理人,这是一方面。
Speaker 209:34 - 10:04
Something that's that's challenging and I I wish I had I wish I had, like, a really succinct, beautiful answer. Like, we have it all figured out, but it's in an era where you like, you're talking about product having soul and product feeling consistent and coherent and tasteful. And how do you scale that? Like, when you have more people working on that, how do you how do you how can you both get the upside of having lots of people working on these hard problems, but also having this consistent feeling? That's a really hard problem.
Speaker 209:34 - 10:04
这件事很有挑战,我真希望自己有一个特别简洁、漂亮的答案。好像我们已经把一切都想明白了那样。但现在这个时代,你谈的是 product 要有灵魂,要让 product 的感觉保持一致、连贯、有品味。而这要怎么规模化?当有更多人一起参与时,你怎么才能既获得很多人共同攻克这些难题的好处,又同时保住这种一致的感觉?这真的是个很难的问题。
Speaker 210:05 - 10:38
I think, actually had, well, there's so many different interesting topics for us to dive into. But one is like the traditional roles of PM and design and engineering and how you group those people together and what are their responsibilities. It feels like what has worked in the past may not make sense exactly one to one in the world of the future, but I don't actually know how to define roles and split work. There's something really valuable if you're like, Oh, you're the designer and the engineer, or vice versa. Okay, now we know where we collaborate, where the interface line is.
Speaker 210:05 - 10:38
我觉得,其实这里面有很多很有意思的话题可以展开。但其中一个就是传统的 PM、design 和 engineering 这些角色,以及你怎么把这些人组织在一起、他们各自该负责什么。感觉过去行之有效的方式,在未来的世界里未必还能一比一照搬;但说实话,我现在也不知道该怎么定义角色、怎么切分工作。如果你能明确说“哦,你是 designer,他是 engineer”,或者反过来,那其实有很大的价值。这样大家就知道该在哪里协作,接口边界在哪里。
Speaker 210:40 - 10:43
Do you guys have a philosophy on that at every year?
Speaker 210:40 - 10:43
你们在 Every 每年都会形成一套关于这个的哲学吗?
Speaker 110:43 - 11:08
I don't know. We've done different things and it sort of depends on the situation and the product. My one mental model that I have that seems to work especially for early products is there are two roles that matter. One is called the pirate and one's called the architect, And the pirate's job is to just build as fast as possible to find something valuable. And it can be vibe coded.
Speaker 110:43 - 11:08
我不知道。我们尝试过不同做法,而且某种程度上要看具体情况和具体 product。我的一个 mental model(心智模型)是:尤其对于早期 product,真正重要的只有两种角色。一个叫 pirate,一个叫 architect。pirate 的工作就是尽可能快地去 build,去找到有价值的东西。而且它甚至可以是 vibe coded。
Speaker 111:08 - 12:02
You don't have to really think about the architecture, all that kind of stuff. And the architect's job is to pair with the pirate and think about, as the pirate finds something that's valuable, how do we make this into a system that is sustainable, that we understand, that can scale? And what's interesting is I think you need one pirate per product. But you can have one architect work on multiple products. You can have them bounce around because the the tools now are so good that someone who's a really good programmer, really good architect can just drop into a code base and be like, I'm going to learn this code base in in an hour because I just have Claude or Codex or whatever map it for me, and then I'll identify the structural pillars that the invariants that need to be true in this product for it to work well.
Speaker 111:08 - 12:02
你不一定真的要去考虑 architecture(架构)之类的那些事情。architect 的工作则是和 pirate 配合,一边想:当 pirate 找到某个有价值的东西之后,我们怎么把它变成一个可持续、我们能理解、还能 scale(扩展)的系统?而有意思的是,我觉得每个 product 都需要一个 pirate。但一个 architect 可以同时服务多个 product。你可以让他们来回切换,因为现在的工具已经太强了:一个真正很优秀的 programmer、很优秀的 architect,可以直接进入一个 code base,然后说,“我打算用一个小时理解这个 code base,因为我可以直接让 Claude 或 Codex 之类的工具帮我把它映射出来。”然后我就能识别出这个 product 要想运作良好,哪些结构性支柱、哪些 invariants(不变量)是必须成立的。
Speaker 112:02 - 12:21
And then I'll just, like, send off Codex to do, like, a twenty four hour rewrite. And then I'll go do something else. And I think that that's really interesting. So that's that's one divide. In particular, the the pirate is this it's this mix of technical product design, like all all of those altogether.
Speaker 112:02 - 12:21
然后我就会直接让 Codex 去做一个比如持续 24 小时的重写,接着我自己去做别的事。我觉得这点非常有意思。所以这是其中一种划分。特别是 pirate 这个角色,本身就是 technical、product、design 的混合体——这些东西全都揉在一起。
Speaker 112:21 - 12:27
The architect is like really mostly focused on the technical side of things. And then, but honestly So you
Speaker 112:21 - 12:27
architect 则基本上主要聚焦在 technical 这一侧。不过,老实说,所以你——
Speaker 212:26 - 12:29
don't have standalone designers then, right? You only have
Speaker 212:26 - 12:29
所以你们没有独立的 designer,对吧?你们只有——
Speaker 112:29 - 12:58
We do have standalone designers. Oh, you do. Like once it gets to be like a real product that we're going to support, then yeah, we have like a standalone design. We have a design team parachutes into each product and helps make it better. But then on the tech side, for example, we have this Slack agent we've been building called PlusOne, and that's just a bunch of engineers that are all working and are, to some degree, full stack thinkers, are it's very engineering focused.
Speaker 112:29 - 12:58
我们确实有独立的设计师。哦,你们有。比如一旦某个东西变成了我们要长期支持的真正产品,那对,我们就会有相对独立的设计。我们有一个 design 团队,会空降到各个产品里帮忙把它做得更好。但在技术这边,比如说,我们一直在做一个叫 PlusOne 的 Slack agent,那基本上就是一群工程师在做,而且他们在某种程度上都是 full stack 思维的人,所以这件事非常以工程为中心。
Speaker 112:58 - 13:08
So there's no one answer. I haven't quite figured it out yet. I do think the pirate architect configuration for new things is really helpful. But what what do you guys do?
Speaker 112:58 - 13:08
所以并没有一个统一答案。我自己也还没完全想明白。不过我确实觉得,对于新东西来说,pirate architect 这种配置真的很有帮助。不过你们是怎么做的?
Speaker 213:09 - 13:25
Yeah. Well, we have something similar, but it's not actually on the role. So for any kind of new feature, we have this idea of stages. So like the first stage is like shaping. So it's We're influenced by the what's the
Speaker 213:09 - 13:25
对。嗯,我们有点类似,但它其实并不是一个正式角色。对于任何一种新功能,我们都有一个 stages(阶段)的概念。所以比如第一阶段叫 shaping。这个做法有点受那个……叫什么来着……的影响。
Speaker 113:25 - 13:26
Shape Up?
Speaker 113:25 - 13:26
Shape Up?
Speaker 213:27 - 13:28
Yeah. Thank you.
Speaker 213:27 - 13:28
对。谢谢。
Speaker 113:28 - 13:30
Yeah. Yeah. By Ryan Thing. Yeah.
Speaker 113:28 - 13:30
对,对,就是 Ryan 那个东西。对。
Speaker 213:31 - 13:34
It? Yeah. Think so. Yeah. Like the Basecamp guys?
Speaker 213:31 - 13:34
那本吗?对,应该是。对,就像 Basecamp 那帮人出的那个?
Speaker 213:34 - 13:42
Yeah. Basecamp. Yeah. It's loosely inspired by that. I went back and read it, or at least the first half that was about it.
Speaker 213:34 - 13:42
对,Basecamp。对,算是松散地受它启发。我后来又回去读了一遍,或者至少读了前一半,就是讲这个的那部分。
Speaker 213:42 - 14:05
It's funny, it felt very dated and a lot more dated from the features that they use as examples, for example, that they're talking about building. Some of the ideas I think are really sound. So we have this idea of there's a shaping phase where what you want there is you want to explore as many potential solutions as possible. So you start with a job to be done. Like, here's a user need.
Speaker 213:42 - 14:05
挺有意思的,它给人的感觉已经很过时了,而且从他们拿来举例、比如他们谈到要构建的那些功能来看,就显得更过时了。不过我觉得其中有些想法其实是很扎实的。所以这里有这样一个概念:存在一个 shaping phase(方案塑形阶段),在这个阶段你想做的是尽可能探索更多潜在的解决方案。也就是说,你从一个 job to be done(待完成任务)出发。比如说,这里有一个用户需求。
Speaker 214:05 - 14:16
Here's the job here's a job we would like the user to hire us to do. Right? And it's like, okay, this is the job. What are possible shapes of that solution? And you go and you explore those shapes as quickly as possible.
Speaker 214:05 - 14:16
这就是这个 job,这是一个我们希望用户“雇用”我们去完成的 job。对吧?然后就是,好,这就是这个 job。这个解决方案可能有哪些 shape(形态)?接着你就尽可能快地去探索这些 shape。
Speaker 214:17 - 14:45
And then you kind of look at those, and you have to be very honest with yourself and say, If we spent time on this, do we think we could actually would people actually hire us to do this or not? And so then we have a next phase, which is kind of validate. It's like, Okay, we think we can prove it. For a small number of people, just prove that you can be useful and that they choose to use you. And then it's like, if you can prove that, then it's like, Okay, let's actually make this good and reliable and scalable and all that.
Speaker 214:17 - 14:45
然后你会回过头来看这些方案,而且你得对自己非常诚实,问一句:如果我们在这上面投入时间,我们觉得人们真的会“雇用”我们来做这个吗,还是不会?于是接下来就进入下一阶段,有点像 validate(验证)。也就是,好,我们觉得自己能证明这件事。先针对一小部分人,证明你确实有用,而且他们会选择使用你。然后如果你能证明这一点,那接下来就是,好,那我们真正把它做到优秀、可靠、可扩展,等等。
Speaker 214:45 - 14:55
It still doesn't answer the question of what are the roles of the people who work on it. But it's just one of those open questions.
Speaker 214:45 - 14:55
不过这仍然没有回答一个问题:参与这项工作的人各自扮演什么角色。但这就是那种仍然悬而未决的问题之一。
Speaker 114:55 - 15:03
I mean, you have sixty, seventy people, so you have had to answer that at least provisionally. So what is the structure that you've come to provisionally?
Speaker 114:55 - 15:03
我的意思是,你们有六七十个人,所以你们至少不得不暂时回答这个问题。所以,你们目前暂时形成的结构是什么样的?
Speaker 215:04 - 15:21
We're still messing around with it. We still have titles, I don't know if we should have them. Have You're a designer, you're a PM, you're an engineer. I'm always flirting with the idea of killing those. How many PMs do have?
Speaker 215:04 - 15:21
我们还在反复试。我们现在还是有 title(头衔),但我也不知道我们是否应该保留它们。比如你是 designer,你是 PM,你是 engineer。我一直都在琢磨要不要把这些都取消掉。你们有多少个 PM?
Speaker 215:23 - 15:26
Four. Three. Depends if you count me. Yeah.
Speaker 215:23 - 15:26
四个。三个。看你算不算上我。对。
Speaker 115:26 - 15:29
Yeah. Three. I would yeah. I wouldn't count you.
Speaker 115:26 - 15:29
对,三个。我会——对,我不会把你算进去。
Speaker 215:29 - 15:31
Yeah. Okay. So three. Yeah. So three.
Speaker 215:29 - 15:31
对。好的。所以是三个。对,所以是三个。
Speaker 215:31 - 15:36
Yeah. Interesting. Wonderful. Yeah. And, like yeah.
Speaker 215:31 - 15:36
对,有意思。很棒。对,而且,就像,是的。
Speaker 215:36 - 15:57
I've always thought of the idea of, like, killing because it's like what you really want is on this project right now, needs to do certain types of things. Someone needs to own the aesthetic or the user experience, and someone needs to own the building. That could be the same person, or those could be two different people. And someone needs to own, what are we trying to do here? What's the strategy?
Speaker 215:36 - 15:57
我一直在想这个想法,就是说,像是把事情拆开来,因为你真正想要的是:在这个项目里,现在需要有人去做某些特定类型的事情。需要有人负责 aesthetic 或 user experience(用户体验),也需要有人负责 building(构建)。这可以是同一个人,也可以是两个不同的人。还需要有人负责,我们在这里到底想做什么?战略是什么?
Speaker 215:57 - 16:11
And what are the next few things that we need to hit in order to hit that? But then it's just a lot of communication overhead if you don't have roles. It's kind of like, okay, well, exactly are you going to do versus what am I going to do? Maybe it'll all be agents one day. Who knows?
Speaker 215:57 - 16:11
以及,为了实现那个目标,接下来我们需要完成的几件事是什么?但如果你没有角色分工,沟通上的额外成本就会非常高。感觉就像,好吧,那到底哪些是你要做的,哪些是我要做的?也许有一天这些都会变成 agents(智能体)。谁知道呢?
Speaker 216:11 - 16:13
We'll keep experimenting.
Speaker 216:11 - 16:13
我们会继续做实验。
Speaker 116:15 - 16:35
Interesting. Well, one of the things that you said when we were just about to start recording is you spend most of your time in granola. And so one of the things you're curious about for me is if I'm playing around and experimenting, like, what am I what am I finding? So I want to let's talk about that. I think there's a lot of meat there.
Speaker 116:15 - 16:35
有意思。嗯,我们刚准备开始录音时,你说过的一件事是,你大部分时间都花在 granola 里。所以你对我的一个好奇点是,如果我也在到处玩、做实验,那我到底发现了什么。所以我想聊聊这个。我觉得这里面很有料。
Speaker 116:36 - 16:51
First of all, tell me about tell me about that, like the I spend most of my time in Granolah, I think that's a it makes perfect sense. But also as a as a CEO, as a strategic way of being. Tell me about that decision.
Speaker 116:36 - 16:51
首先,跟我说说那个吧,就是“我大部分时间都花在 Granolah 里”这件事。我觉得这完全说得通。但同时,作为一个 CEO,作为一种战略性的存在方式,这也是个选择。跟我讲讲你为什么做这个决定。
Speaker 216:52 - 17:28
So, like I mentioned before that I think the big opportunity here is to kind of invent how we're going to work and collaborate with AI to think and do things. Right? Like, that's kind of in some shape or form, I think that's what we're all chasing right now in in Silicon Valley. Right? And and to that end, I think there's a lot of that's the kind of thing you can, like, think about abstractly, but then you have to be playing with to to to kinda get a feel for.
Speaker 216:52 - 17:28
所以,就像我之前提到的,我觉得这里最大的机会,是某种程度上去发明我们将如何与 AI 一起工作、协作、思考和做事。对吧?某种形式上,我觉得这就是我们现在在 Silicon Valley 所有人都在追逐的东西。对吧?而为了做到这一点,我认为这类事情你可以在抽象层面去思考,但你还是得亲自去玩、去试,才能多少形成一点感觉。
Speaker 217:28 - 17:53
And it's just it's hard for me to have the brain space of of playing with every new thing that comes up. Right? And in in granola, we have a lot of a lot of context and a lot of internal tooling so that I can play around with different ideas. But they're oftentimes, like, in in my little universe as opposed to I like, I don't go and set up very complicated workflows. In in Claude, for example, like, I do it.
Speaker 217:28 - 17:53
而且说实话,要让我在脑子里腾出空间去把每一个新冒出来的东西都玩一遍,这很难,对吧?在 granola 里,我们有很多上下文(context),也有很多内部工具(internal tooling),所以我可以围绕不同想法做一些尝试。但很多时候,那些尝试更像是发生在我自己的小宇宙里,而不是说,我会去搭建那种非常复杂的 workflow。比如在 Claude 里,我确实会这么做。
Speaker 217:53 - 18:17
I play around with it to get a feel, but I don't I don't, those aren't load bearing for me because I I'm trying to see what makes sense or doesn't make sense to do inside of granola as we paint that future picture. It seems like you are at the forefront of experimenting, trying new things, building different workflows for yourself. So like maybe just start off by being like, what's the current state of Dan's tooling stack and then we can go from there.
Speaker 217:53 - 18:17
我会去玩一玩,先建立感觉,但那些东西对我来说并不是 load-bearing(关键支撑性的),因为我是在试着判断,随着我们去描绘 granola 的未来图景,哪些事情放在 granola 内部做是合理的,哪些是不合理的。感觉你现在正处在做实验、尝试新东西、给自己搭各种 workflow 的前沿。所以也许我们可以先从“Dan 当前的 tooling stack(工具栈)是什么状态”开始,再往下聊。
Speaker 118:17 - 18:20
I spend all my time in Codex. Almost
Speaker 118:17 - 18:20
我几乎把所有时间都花在 Codex 里。几乎。
Speaker 218:20 - 18:23
as And soon how long has that been? Is that a recent thing?
Speaker 218:20 - 18:23
大概是这样。那这种状态持续多久了?这是最近才开始的吗?
Speaker 118:24 - 18:48
Two months maybe, Okay. Two, three And then and I do use the cloud desktop app sometimes, like when Fable was a thing. I track all my usage and when Fable was a thing, you can see me like go on cloud usage. And then when it when it left, it went like down. So I assume when it's back because I think Fable Fable is just a different beast.
Speaker 118:24 - 18:48
可能两个月吧。哦,好,两三个月。然后我有时候也会用 cloud desktop app,比如在 Fable 还是个重点的时候。我会跟踪自己所有的使用情况;当 Fable 还是个重点时,你能看到我的 cloud usage 会往上走。后来它不做了,使用量就下来了。所以我猜等它回来时,使用量也会再上去,因为我觉得 Fable 就是另一种 beast(完全不同的东西)。
Speaker 118:48 - 19:17
And I test all the new stuff and I have access to stuff before it comes out. And Fable is just different. But I think of so my overall view is that work is bifurcating into two surfaces. One is async delegation work that happens in Slack. That's collaborative, multiplayer, sometimes proactive.
Speaker 118:48 - 19:17
我也会测试所有新东西,而且我在它们发布前就能接触到一些东西。Fable 就是不一样。但我的整体看法是,工作正在分裂成两个 surface(交互界面/工作表面)。一个是发生在 Slack 里的 async delegation work(异步委派工作);它是协作式的、multiplayer(多人参与)的,有时还是主动式的。
Speaker 119:19 - 19:38
And that happens with Slack bots. Like we have one called Plus One, there's one called Victor, there's the Clouds Tag. And that's very much like, okay, every morning it posts in a channel and says, hey, here's all the bugs that got solved. Here's all the bugs got reported. Here's the PRs we pushed.
Speaker 119:19 - 19:38
而这类事情是通过 Slack bots 来完成的。比如我们有一个叫 Plus One,还有一个叫 Victor,还有 Clouds Tag。它非常像是这样:好,每天早上它会在一个频道里发消息说,嘿,昨天有哪些 bug 被解决了,哪些 bug 被报告了,我们推了哪些 PR。
Speaker 119:39 - 20:03
And then you can add it and say, can you kick off PR for XYZ bug? And it just does it. Or can you we're trying to name this new product. Can you read all of the stuff we've said about what we want the product to be and then propose some names? And then the team can be going back and forth with it in the same channel to talk through names.
Speaker 119:39 - 20:03
然后你可以 @ 它并说,你能不能为 XYZ bug 发起一个 PR?它就会直接去做。或者说,我们正在给这个新产品起名字,你能不能读一下我们之前关于希望这个产品成为什么样子的所有讨论,然后提一些名字?接着团队就可以在同一个频道里和它来回讨论,把这些名字聊透。
Speaker 120:04 - 20:37
So you can hand off stuff with it like it's a coworker. And importantly, because most AI right now is single player, you do it in public. And I think that's really interesting in in a whole set of things. However, however, for serious work. You can't you're always going to get more out of it if you have a better collaboration service between you and the agent, and Slack is not good for that.
Speaker 120:04 - 20:37
所以你可以把事情交给它,就像交给一个 coworker 一样。重要的是,因为现在大多数 AI 还是单人模式,所以你是在公开环境里这么做的。我觉得这在一整类事情上都很有意思。不过,不过,对于严肃工作来说,如果你和 agent(智能体)之间有更好的协作服务,你总是会得到更多,而 Slack 并不适合这个。
Speaker 120:37 - 21:06
And so the other surface is something like a Codex or Claude desktop app, where you and the agent are sitting in the same place on the same app together. And in particular, in Codex, for me, it's like I use almost all of my software in the in app browser of Codex. So I'm bringing Codex into my email or into, you know, whatever it is in the same way that What
Speaker 120:37 - 21:06
所以另一个交互界面是像 Codex 或 Claude desktop app 这样的东西,你和 agent 一起待在同一个地方、同一个 app 里。尤其是对我来说,在 Codex 里,我几乎是在 Codex 的 in-app browser 里使用我所有的软件。所以我是在把 Codex 带进我的 email,或者别的任何东西里,方式有点像——
Speaker 221:06 - 21:08
do you mean by that? Can describe that? How do you do that?
Speaker 221:06 - 21:08
你这是什么意思?能具体描述一下吗?你是怎么做的?
Speaker 121:08 - 22:02
Yeah, explain. So when you're building something in Codex or Cloud Code and it opens it in a browser and it lets you see the local host version of the thing and you're iterating on it, Basically, like all of the patterns a good mental model for how AI works for me is all the patterns that started with developers or builders eventually make their way into all of knowledge work because we are what we found is that a good enough agent to build any kind of software, once it can build any kind of software, it's actually really good for doing any kind of knowledge work that you want. And that's like the less that's why Cloud Code went to Cloud Cowork. That's why Codex is now all of knowledge work, all that kind of stuff. So one of those patterns is when you're building something like an app, what you really want is the agent to be in the loop with you.
Speaker 121:08 - 22:02
对,解释一下。所以当你在 Codex 或 Cloud Code 里构建某个东西时,它会在浏览器里打开,让你看到这个东西的 localhost 版本,然后你可以不断迭代。基本上,对我来说,一个理解 AI 如何工作的很好的 mental model(心智模型)是:所有最初从 developers 或 builders 那里开始的模式,最终都会扩展到所有知识工作里,因为我们发现,只要一个 agent 已经足够强,能够构建任何一种软件,那么一旦它能构建任何一种软件,它其实也就非常擅长做你想要的任何一种知识工作。这也是为什么 Cloud Code 变成了 Cloud Cowork,为什么 Codex 现在扩展到了整个知识工作领域,诸如此类。其中一种模式就是:当你在构建像 app 这样的东西时,你真正想要的是让 agent 和你一起处在 loop 里。
Speaker 122:02 - 22:11
So as it builds something, it opens an in app browser. You can see it, and you can kind of go back and forth. You can annotate it. You can whatever. And you and the agent are seeing it together.
Speaker 122:02 - 22:11
所以当它在构建某个东西时,它会打开一个 in-app browser。你能看到它,也可以来回互动。你可以做标注,或者别的什么。你和 agent 是一起在看这个东西的。
Speaker 122:12 - 22:38
That pattern also works for any kind of software, so any kind of website you might visit. So really simple example, last night, I moved apartments recently, and I needed to change my internet. And I just told Codex, Go figure that out. And it knows the address of my old one and my new one, which is actually the same building. We're just moving floors.
Speaker 122:12 - 22:38
这种模式其实也适用于任何软件,因此也适用于你可能访问的任何网站。举个很简单的例子,昨天晚上,我最近刚搬家,我需要更改我的 internet。于是我就直接对 Codex 说,去把这件事搞定。它知道我旧地址和新地址,实际上还是同一栋楼,我们只是换了楼层。
Speaker 122:39 - 23:01
It just opened up the Verizon website in its in app browser, and it logged in. I gave it the password. It logged in. I gave it the password. And then it opened up a chat with a customer service agent who was also probably using AI, and it just, like, chatted with them until it switched my Internet.
Speaker 122:39 - 23:01
它就在自己的 in-app browser 里打开了 Verizon 网站,然后登录了。我把密码给它,它就登录了。我把密码给它。然后它又打开了一个和客服人员的聊天窗口,而对方大概也在用 AI。接着它就那样一直和对方聊,直到把我的 Internet 切换好。
Speaker 123:01 - 23:20
And it also, like and all these little things come up where it's like, okay. When do you wanna move in and out? And Codex knows because I've been talking to it about, like, when I moved and when I wanted to switch. And then it's like, here's here's your old plan doesn't work. Here's the new plan.
Speaker 123:01 - 23:20
而且还会冒出这些小问题,比如,好,您什么时候搬出、什么时候搬入?Codex 都知道,因为我之前一直在跟它聊我什么时候搬家、什么时候想切换。然后它就会说,你看,你原来的 plan 不能用了,这是新的 plan。
Speaker 123:22 - 23:37
Knows And to be like, I don't want any hidden fees. It knows how to do the math on, you know, is this a good deal or not? And to be able to research it and then to push them to give me the thing that it thinks I should get. And I don't have to do any of that. I'm just sitting there.
Speaker 123:22 - 23:37
它知道——而且会像是——我不想要任何 hidden fees。它知道该怎么计算,你知道,这是不是一笔划算的交易。它还能去做 research,然后推动对方把它认为我应该拿到的东西给我。而这些我都不用自己做。我就只是坐在那里。
Speaker 223:37 - 23:47
And is that something that you somehow communicated to Codex or is that just an out of the box? Obviously, human would want this when talking to an internet provider.
Speaker 223:37 - 23:47
这是你以某种方式传达给 Codex 的,还是说这只是它开箱即用 out of the box 就会的?显然,人类在和 internet provider 沟通时都会想要这些。
Speaker 123:47 - 24:06
It's it's both. It's easy because it's easy for it. It will it knows pretty well when it can make assumptions and when it when it shouldn't. And so I'm I can be in the loop on it and I also can just go out, go to another thread and be doing something else. So I'm kind of like flipping back and forth with it.
Speaker 123:47 - 24:06
两者都有。它之所以容易,是因为这对它来说很容易。它相当清楚什么时候可以做 assumptions(假设),什么时候不该做。所以我既可以在 loop(回路)里参与它,也可以切出去,到另一个 thread(线程)里做别的事。所以我有点像是在和它来回切换。
Speaker 124:07 - 24:24
I think that's something. But but it's there's a more there's a more general thing there where I use this is how I do my email. It is in the loop with me in Quora, which is our email app. And also, have this open source thing I experimented. Built called Tend.
Speaker 124:07 - 24:24
我觉得那是一回事。但还有一个更普遍的点,就是我现在处理 email 也是这么做的。它会和我一起在 Quora 里 loop,我们的 email app 也是这样。另外,我还有一个自己实验着做出来的 open source 东西,叫 Tend。
Speaker 124:24 - 24:39
But basically, it's like my emails are now cards that I read in Codex in the in app browser. And then each card, it says, here's what the email said. Here's what I think the draft should be. Do you want to send it? And I just talk to it.
Speaker 124:24 - 24:39
基本上,我的 emails 现在变成了一张张卡片,我在 Codex 的 app browser 里阅读。然后每张卡片都会说:这封邮件讲了什么。这是我认为合适的 draft(草稿)。你要发送吗?然后我就直接和它对话。
Speaker 224:39 - 24:46
I see. But this is basically you've vibe coded your own email client, basically, right?
Speaker 224:39 - 24:46
我明白了。所以基本上,你这是 vibe coded(凭感觉编码)了一个你自己的 email client,对吧?
Speaker 124:46 - 25:37
Is that I vibe coded an experimental email client that does this, and then we will have a version of Quora, which is our full email client that just has us out of the box. But you can think of the apps that are structured this way as being I've been calling them Codex native apps, but they're basically apps that are responsible for saving the state and rendering the UI, and you bring your agent to it. And I think that that is a much it's a very powerful paradigm because think about I mean, think about all the work that you have to do for granola to make an agent that is good. And it's really hard. And you're also competing with the clods and the codexes of the world who are also building the same kind of agent with the same kind of functionality.
Speaker 124:46 - 25:37
是的,我 vibe coded 了一个会做这些事的实验性 email client,然后我们会推出一个版本的 Quora,也就是我们完整的 email client,它开箱即用 out of the box 就带这些功能。但你可以把这类以这种方式构建的 app 理解为——我一直称它们为 Codex native apps——本质上,这类 app 负责保存 state(状态)并渲染 UI,而你把自己的 agent 带进去。我认为这是一种非常强大的 paradigm(范式),因为你想想看,为了让 granola 里有一个足够好的 agent,你得做多少工作。这真的很难。而且你还在和 clods、codexes 这类玩家竞争,他们也在构建同一种 agent,具备类似的功能。
Speaker 125:39 - 26:10
And to some degree, I also think that in some cases, that's very worthwhile because I do think two agents is better than one. But ultimately, I really want to be able to, for example, bring and I think you're probably seeing this with a lot of granola users. I really want to bring my agent to granola. You already know how AI is changing how everyday work gets done, how much ground you can cover, and how fast a team can scale. To stay ahead, you need tools that give you a competitive advantage built for this new era.
Speaker 125:39 - 26:10
在某种程度上,我也认为在一些情况下这非常值得,因为我确实觉得两个 agents 比一个更好。但归根结底,比如说,我真正想要的是把我的 agent 带到 granola 里。我觉得你可能已经在很多 granola 用户身上看到了这一点:我真的很想把我的 agent 带到 granola。你已经知道 AI 正在如何改变日常工作的完成方式、你能覆盖多少工作量,以及一个团队能以多快的速度扩张。要保持领先,你需要的是为这个新时代打造、能给你带来竞争优势的工具。
Speaker 126:10 - 26:30
Adio is the CRM for the agent native world. It meets you where you work, compounds every customer signal into context, and then acts on it across your pipeline to let you move at unmatched speed and scale. With agents and automations for every job, Adio orchestrates your work around the clock. We use it internally at every, and we love it. It's built to handle the scale of your workloads.
Speaker 126:10 - 26:30
Adio 是为 agent native world 打造的 CRM。它会在你的工作场景中直接配合你,把每一个客户信号不断累积成 context(上下文),然后在整个销售 pipeline 中据此采取行动,让你以无与伦比的速度和规模推进工作。借助覆盖各种岗位的 agents 和 automations(自动化),Adio 可以全天候编排你的工作。我们在 every 内部处处都在使用它,而且非常喜欢。它就是为承载你工作负载的规模而构建的。
Speaker 126:30 - 26:50
It's extensible with an API and MCP access, and is built with infrastructure to keep up with your most ambitious agents. It's loved by high growth startups like Granola, Modal, WhisperFlow, and Every. Adio runs the work behind Everywin. That's Adio, the agentic CRM. Go to adio.com/every and get 15% off your first year.
Speaker 126:30 - 26:50
它可通过 API 和 MCP 访问进行扩展,并且底层基础设施的设计足以跟上你最有野心的 agents。包括 Granola、Modal、WhisperFlow 和 Every 在内的高增长初创公司都很喜欢它。Everywin 背后的工作就是由 Adio 驱动的。这就是 Adio,agentic CRM。访问 adio.com/every,首年可享 15% 折扣。
Speaker 126:50 - 26:54
That's adio.com/every. And now back to the episode.
Speaker 126:50 - 26:54
就是 adio.com/every。现在回到本期节目。
Speaker 226:54 - 27:13
What do you think like, yeah. You you keep saying that, and I'm really excited about that idea. I wonder how does that work in practice? Forget Granolah, let's say you, I don't know, can you do this with Quora right now? Or how could we build an app that would allow you to bring your own agent that's not MCP?
Speaker 226:54 - 27:13
你一直在说这个,我对这个想法也真的很兴奋。我想知道它在实际中到底是怎么运作的?先别说 Granolah,假设拿 Quora 来说——我不确定,你现在能用 Quora 做这个吗?或者说,我们要怎么构建一个 app,让你可以接入你自己的 agent,而且它不是通过 MCP 的?
Speaker 227:13 - 27:22
Because it feels a little bit different, right? Because it's like, you still want to have the UI. That's the difference. It's like you and the agent are both looking at the same UI and manipulating the same thing.
Speaker 227:13 - 27:22
因为这感觉还是有点不一样,对吧?因为你仍然希望保留 UI。这就是差别所在。也就是说,你和 agent 都在看同一个 UI,也都在操作同一个东西。
Speaker 127:22 - 27:56
MCP is part of it. But yes, I think MCP is Historically, you're thinking about it as the agent is going to interact with the app and I'm not going to interact with the app. And I think what this paradigm says is actually you want your agent and the user to be interacting with the app at the same time and be able to trade off. And so the way that you do that is, one, yeah, you give the agent access to an MCP or a CLI. And then you give the user access to a web interface.
Speaker 127:22 - 27:56
MCP 是其中的一部分。但没错,我觉得 MCP——从历史上看,你会把它理解成:agent 会去和 app 交互,而我不会去和 app 交互。而我认为,这种范式真正表达的是:其实你希望 agent 和用户同时与这个 app 交互,而且两者之间可以相互切换、分工协作。所以实现方式是,第一,没错,你给 agent 提供 MCP 或者 CLI 的访问;然后你再给用户提供一个 web interface。
Speaker 127:56 - 28:20
And the user is using the web interface inside of an in app browser of a codex or of a Cloud Code. And then the agent can decide, do I want to help the user or collaborate with the user via using the browser use or via using the CLI, which I think CLI is somewhat better. But the trick is the CLI has to then modify the state of the UI.
Speaker 127:56 - 28:20
用户是在 codex 或 Cloud Code 的 app 内浏览器里使用这个 web interface。然后 agent 可以决定:我是要通过使用浏览器来帮助用户或与用户协作,还是通过 CLI 来协作?我觉得 CLI 某种程度上更好一些。但关键在于,CLI 接着必须能够修改 UI 的状态。
Speaker 228:21 - 28:46
Right, exactly. That's the thing. It's like that's the thing that needs to be. So it's almost like if you had hypothetically, let's say you could have an MCP, right, that actually had a stateful, like the MCP could modify the state, real time state of a UI that was rendered for a user. Ideally there was at least a one to one mapping of anything you can do in the UI could also be done through the MCP.
Speaker 228:21 - 28:46
对,完全没错。关键就在这里。这就是那个必须被解决的点。所以几乎可以这样想:假设你有一个 MCP,对吧,而且这个 MCP 是有状态的——也就是说,MCP 可以修改一个为用户渲染出来的 UI 的实时状态。理想情况下,至少应该做到一一对应:任何你能在 UI 里完成的事,也都应该能通过 MCP 来完成。
Speaker 228:46 - 28:59
You have an agent and a human interface. Because computer use or like it's like the web browser use feels like a it's unnecessarily slow. Right? It's like a it's a hack and not not. Yeah.
Speaker 228:46 - 28:59
你有一个 agent 和一个人类界面。因为 computer use,或者说像用 web browser 那样的使用方式,感觉有点不必要地慢。对吧?这有点像是一种 hack,而不是,不是……对。
Speaker 229:00 - 29:24
That's really interesting. Yeah, there's something about the presence of the agent. This is a fascinating topic. I haven't thought about it very much, but it's the embodiment of the agent is also maybe interesting. It's like when you're in Figma, can see other people's locations on the There's something there that you maybe would lose if there was just an MCP and the human Maybe not.
Speaker 229:00 - 29:24
这真的很有意思。对,agent 的“存在感”这件事里确实有点东西。这是个很迷人的话题。我之前没怎么认真想过,但 agent 的 embodiment(具身化、呈现方式)可能也很有意思。就像你在 Figma 里的时候,可以看到其他人在……上的位置。这里面有某种东西,如果只有一个 MCP 和人类,也许你就会失去。也可能不会。
Speaker 229:24 - 29:37
I don't know how you Does the agent just feel like it's just the computer and that's all the same? Or does it feel like it's an entity that you're both looking at the same thing together?
Speaker 229:24 - 29:37
我不知道你是怎么……agent 会不会只是让人感觉它不过就是那台 computer,其他都一样?还是说,它会让人感觉像是一个实体,而你们两个正在一起看同样的东西?
Speaker 129:37 - 29:49
It's more that. It's in the same way that with the Google Docs. We have this markdown editor called Proof that we built. And in that one, can see Codex has entered the document and you can see where it is.
Speaker 129:37 - 29:49
更像是后者。这有点像在 Google Docs 里那样。我们自己做了一个 markdown editor,叫 Proof。在那个里面,你可以看到 Codex 已经进入了文档,也能看到它所在的位置。
Speaker 229:49 - 29:52
Oh, that's cool. How do you do that? How does Codex enter the document?
Speaker 229:49 - 29:52
哦,这很酷。你们是怎么做到的?Codex 是怎么进入文档的?
Speaker 129:52 - 30:21
It's just the same kind of CLI MCP type situation, but it has a command that's like, Okay, set presence, set location in the document, that kind of thing. And agents are very good at knowing how to do that. But there's so many interesting, then, UI challenges to figure out because an agent can do like a thousand different actions at once. And so how do you show that to a user?
Speaker 129:52 - 30:21
本质上还是同一种 CLI + MCP 类型的方案,只不过它有一个命令,类似“好,设置 presence(存在状态)”“设置在文档里的位置”这类东西。而 agent 很擅长知道该怎么做这些事。不过接下来就有很多很有意思的 UI 挑战要解决,因为一个 agent 可以同时做上千种不同的动作。那你要怎么把这些展示给用户看呢?
Speaker 230:21 - 30:30
It's interesting. That's that's actually not where my brain went. Because like, you're right. It's like, yeah, it's like, if if you wanna see what the agent's doing, that's one problem. I just thinking in like an email, it's like it'd be nice for it to know where your mouse is.
Speaker 230:21 - 30:30
很有意思。那其实不是我一开始想到的方向。因为你说得对。就是说,如果你想看到 agent 在做什么,那是一个问题。我刚刚想到的是,比如在 email 里,如果它能知道你的鼠标在哪儿,会很好。
Speaker 230:30 - 30:51
It'd be nice to be able to like, know, be like, oh, this part of the email is not great. And just have that be, you know, either you highlight the text or whatnot. And that's not really possible today unless you explicitly Like how would you have to do that? It would have to be like looking at pixels on your screen, right? There's no native way to do that unless you build it into a UI and you're like passing that information to the agent, right?
Speaker 230:30 - 30:51
如果它能像这样知道,比如说,“哦,这一部分 email 写得不太好”,那会很不错。然后你就可以,比如高亮那段文字之类的。但在今天这其实做不到,除非你显式地……你到底得怎么做呢?它必须得去看你屏幕上的 pixels(像素),对吧?除了把这个能力直接做进某个 UI 里,然后把这些信息传给 agent,否则并没有原生的办法做到,对吧?
Speaker 130:51 - 31:11
More or less, I think the way that I do it is each so each piece of information that I want to do, want to take care of is represented in a card. And then I can talk to the card. I can say like, yeah, it doesn't it doesn't necessarily know where my mouse is, but it knows that I'm talking to this part of the UI. And then
Speaker 130:51 - 31:11
大体上,我觉得我的做法是,把我想处理、想照看的每一条信息都表示成一张 card。然后我就可以和这张 card 交互。我可以说,没错,它不一定知道我的鼠标在哪儿,但它知道我正在和 UI 的这一部分说话。然后——
Speaker 231:11 - 31:17
And it just knows because you say that when you're talking to it or because that is the the main thing that's foregrounded in the app?
Speaker 231:11 - 31:17
所以它知道,只是因为你在和它说话时这么指出来了,还是因为那是 app 里当前处在前景、最被突出显示的那个东西?
Speaker 131:17 - 31:27
Whatever I'm looking at is currently focused. And so and I know that it's focused. And then when I'm talking to, it just knows to put to put that context next to that thing that I'm looking at.
Speaker 131:17 - 31:27
我正在看的东西就是当前 focus(焦点)所在的内容。所以——而且我知道它现在是 focused(处于焦点中)的。然后当我和它说话时,它就知道要把那个 context(上下文)附到我正在看的那个东西旁边。
Speaker 231:27 - 31:32
Okay, that makes sense. And you're actually speaking, right? This is like you're using monologue or
Speaker 231:27 - 31:32
好的,这说得通。那你实际上是在用语音,对吧?这有点像你在用 monologue 还是——
Speaker 131:32 - 31:48
whisper Yeah, or yeah. Something like that, That's sort of my view of work and how things are going. And we've been talking about, Okay, there's this the prize is there's a new way of working. How does that fit into or differ from how you've been thinking about it and your strategy for granola?
Speaker 131:32 - 31:48
whisper,对,或者——对,差不多那类东西。这算是我对工作以及事情进展方式的看法。然后我们一直在讨论,Okay,所谓的 prize(目标/奖赏)是会出现一种新的工作方式。这和你一直以来对这件事的思考,以及你为 granola 制定的策略,是怎么契合的,或者说有什么不同?
Speaker 231:48 - 31:57
I don't know. I took this class once and they were talking about the difference between complex and con complicated problems. Have have you heard this? Yeah. Yeah.
Speaker 231:48 - 31:57
我也不确定。我之前上过一门课,里面他们讲过 complex 和 complicated 问题之间的区别。你听过这个说法吗?听过吧?对,对。
Speaker 231:57 - 32:05
And it's, like, complicated. It doesn't mean it's it's gonna be super hard, but it's, like, it's kinda knowable. You know? It's, like, you you and versus complex problems. It's unknowable.
Speaker 231:57 - 32:05
complicated 的意思是,它不一定会超级难,但它是那种——某种程度上是可知的。你懂吧?就是——而和 complex 问题相比,后者是不可知的。
Speaker 232:05 - 32:27
You have to, like, probe the system and see how it reacts and go from there. So I think we're very much in a complex problem space here. I don't have an overarching theory of where this is all going. I think I could make one up that would sound plausible or as plausible as the next person's. But I'm more in the mode of like, insights what or nuggets?
Speaker 232:05 - 32:27
你必须去试探这个系统,看看它会如何反应,然后再从那里继续往下走。所以我觉得,我们现在非常明显就在一个 complex problem(复杂问题)的空间里。我并没有一个统摄全局的理论,来解释这一切最终会走向哪里。我觉得如果要编一个,我也能编出一个听起来还算合理的理论,或者说,和别人编出来的一样合理。但我现在更像是处在一种——寻找 insights(洞见),或者说 nuggets(零碎但有价值的信息/想法)的模式里。
Speaker 232:27 - 33:16
And I'm like, oh, okay, I now this is like a thing that I've learned and now I believe to be true and I'm going to carry that with me. So I have a few of those. One is, seems like one fundamental design challenge in this agentic world is the time traveling problem. Basically agents take time to do things and therefore, like when you you're talking about Slack as an interface for this, which is like the async delegation problem, I think you said before, which is basically like the moment in which you kick off a task and the moment in which you're reviewing the task are disjoint, and then you need to get all that context into your head. And I think that'll be one of the I think you'll see a lot of the interfaces that evolve to be highly optimized for that.
Speaker 232:27 - 33:16
我当时的感觉是,哦,好的,这是我现在学到的一件事,而且我相信它是对的,我会把它一直带在身上。所以我有几个这样的体会。其中一个是,在这个 agentic world(agent 驱动的世界)里,一个根本性的设计挑战似乎是 time traveling problem。基本上,agent 做事情需要时间,因此,比如你把 Slack 当作这个东西的界面时,这就像是你之前说的 async delegation problem(异步委派问题):也就是说,你发起一个任务的那个时刻,和你回头审阅这个任务的那个时刻,是彼此分离的,然后你需要把所有这些上下文重新装进脑子里。我觉得这会是一个——我认为你会看到很多不断演化出来的界面,都会高度针对这个问题做优化。
Speaker 233:16 - 33:32
And I haven't seen anything that's great, to be honest. I don't know. Conductor, it's like, oh, I've got all these different chat threads and I'm jumping between them when it's kind of ready. Feels like they'll be I don't know what they look like. There'll be evolutions beyond that that are coming.
Speaker 233:16 - 33:32
老实说,我还没见过什么做得特别好的东西。我也不知道。Conductor 这种方式是,哦,我有这么多不同的聊天线程,然后我在它们之间来回切换,等它差不多准备好了再看。感觉未来还会有进一步的演化,我不知道它们具体会长什么样,但应该会超越现在这种形态。
Speaker 233:34 - 34:19
Another one that's been that's interesting, like another way to get around this, like it takes a while for an agent to do thing problem is pre process a bunch of stuff. Like if like a certain granola, like one thing that we figured out is like if I ask Granola to do something and I have to wait twenty seconds, basically humans will rarely wait for twenty seconds if they're in the back to back meetings chaotic crazy workday, which is kind of the user that we're thinking about. But if Granola kind of thinks about something you might want and then pre generates it, and then at some point might be like, Oh, here, in case you want this, it's already here and all you have to do is click on it. That's like feels completely different. And so, like, we have we have some examples there.
Speaker 233:34 - 34:19
另一个也很有意思的点是,针对“agent 做事要花一段时间”这个问题,还有一种绕开的办法,就是预处理一堆东西。拿 Granola 来说,我们发现的一件事是,如果我让 Granola 去做一件事,而我得等二十秒,基本上在那种一个接一个会议、混乱又疯狂的工作日里,人很少会愿意等二十秒,而这正是我们在设想的用户。但如果 Granola 能先想一想你可能会想要什么,然后提前把它生成好,接着在某个时刻对你说,哦,这个给你,以防你想要,它已经在这里了,你只需要点一下就行——那种感觉就完全不一样了。所以,我们在这方面已经有一些例子。
Speaker 234:19 - 35:14
Like, we launched this thing recently where Granola will try to generate a brief for you before a meeting where it's basically, here's this person you're meeting with. Here's the context of, like like, who the person is, if it's a first meeting or what you guys talked about last time, it's another meeting. This one, you essentially have to pre generate it because it's useful when you're running two minutes late to a meeting and you're like, wait, who the heck is this person I'm talking And to that's the critical moment. So you really only have a fifteen second window where it needs to be there or it's kind of useless. So we actually pre generate millions of We pre generate a silly amount of these for a small percentage of them actually being opened in the belief that when you do open it though, you really, really appreciate it because it's exactly what you need in the moment of need.
Speaker 234:19 - 35:14
比如,我们最近上线了一个功能:Granola 会在会议前尝试先为你生成一份 brief(简报)。基本上就是:这是你要见的人;这是相关背景,比如这个人是谁;如果是第一次见面,就给你第一次见面的背景;如果不是第一次见面,就告诉你们上次聊了什么。这个功能本质上必须预先生成,因为它最有用的场景是:你开会迟到了两分钟,正一边赶过去一边想,等等,我到底要和谁说话来着——那才是关键时刻。所以它实际上只有十五秒的窗口期,必须已经在那里,否则基本上就没用了。因此,我们实际上会预先生成几百万份——数量多得有点夸张——哪怕最后真正被打开的只占很小一部分。我们的判断是,一旦你真的打开它,你会非常、非常感激,因为它恰好就是你在那个需要的瞬间最需要的东西。
Speaker 235:15 - 35:30
It's an interesting process or trade off there, especially from a cost perspective because it's like hours and, I don't know, hundreds of thousands of hours of agents reasoning about things that may or may not ever see the light of day.
Speaker 235:15 - 35:30
这是个很有意思的过程或者说权衡,尤其从成本角度看,因为这相当于花了成小时、我也不知道、几十万小时的 agent 推理时间,去思考一些最终可能永远都不会被看到的东西。
Speaker 135:31 - 35:34
And how are you measuring that? Whether that's worth it or not?
Speaker 135:31 - 35:34
那你们怎么衡量这件事?怎么判断它到底值不值得?
Speaker 235:36 - 35:46
It's an open question. I don't know how to measure it just yet. What I'm trying to figure out is that we don't have a good metric between that. Like there's use. Like if something's not used at all, it's obviously not good.
Speaker 235:36 - 35:46
这还是个开放问题。我现在还不知道该怎么衡量。我正在努力弄明白的是,我们现在缺少一个能很好描述这两者之间关系的指标。比如说,有“使用”这件事。如果一个东西完全没人用,那显然它就不好。
Speaker 235:46 - 36:31
Right? But then there's like, if something is used a little bit or a decent amount, but it's really load bearing when it's used and that's still very valuable, I'm trying to figure out how to measure what that is. I was at a founder's conference the other day and four founders came up to me and talked about this pre meeting brief feature, which surprised me because it's not used as much as proportionally, that's not what I would have expected. We have this analogy metaphor. I always forget which of those two it is inside of granola of how we want the granola product to feel.
Speaker 235:46 - 36:31
对吧?但另一种情况是,一个东西可能只被用了一点点,或者被用得还算可以,但它一旦被用到,就是非常 load bearing(承担关键负载、至关重要)的,那它依然非常有价值。我现在在想的是,怎么去衡量这种价值。前几天我去参加了一个 founder 的大会,有四位 founder 走过来跟我聊这个会前 brief 功能,这让我很惊讶,因为按使用占比来看,它并没有那么高,这并不是我原本会预期的。我们在 Granola 内部有一个 analogy / metaphor(类比 / 隐喻),我老是记不清这两个词到底该用哪一个,用来描述我们希望 granola 这个产品给人的感觉。
Speaker 236:31 - 36:53
And it's like a handrail. So if you imagine stairs, all stairs have handrails, right? It's so good. Basically you never notice a handrail. It's invisible, but that moment you trip, your hand shoots out and it needs to be right there and it needs to be load bearing and it makes stairs way safer.
Speaker 236:31 - 36:53
它就像一个扶手。你想象一下楼梯,所有楼梯都有扶手,对吧?这东西特别好。基本上你平时根本不会注意到扶手。它像是隐形的,但在你绊一下的那个瞬间,你的手会立刻伸出去,这时候它必须刚好在那里,而且还得能承重;它会让楼梯安全得多。
Speaker 236:54 - 37:14
That's how we think about granola, right? We kind of want to get out of the way until you really need us. But yeah, we probably need better metrics or frameworks to measure. One thing we would thought is if we just take it away, if we take it away from a number of users, how much do they shout? Or what kind of what's the impact of us doing that?
Speaker 236:54 - 37:14
我们就是这样看 granola 的,对吧?我们有点希望在你真正需要我们之前,尽量不打扰你。不过,是的,我们可能还是需要更好的 metrics(指标)或 frameworks(框架)来衡量。有一件我们想过的事是:如果我们把它拿走——如果我们对一部分用户把它拿走——他们会有多大反应?或者说,我们这么做的影响到底是什么?
Speaker 237:14 - 37:16
But even that has it's not the best way.
Speaker 237:14 - 37:16
但即便那样,也不是最好的办法。
Speaker 137:17 - 37:27
Or maybe making like seeing if they will click to like generate click to generate as opposed to here it is, and it's already generated, but, you know, would you take the proactive action to do it?
Speaker 137:17 - 37:27
或者也许可以这样:看看他们会不会去点那个 generate(生成)按钮,也就是“点击生成”,而不是直接把已经生成好的内容摆在这里;不过,你懂的,问题在于,他们会不会主动采取这个动作去生成它?
Speaker 237:27 - 37:36
Of doing it and then it's there. Yeah. Yeah. But then it's like, okay, they do, are we always going make them click? I don't know.
Speaker 237:27 - 37:36
就是说,他们自己去做,然后结果才会在那里。对,对。但接着问题又变成:好吧,他们确实会点,可我们难道要一直都让他们点击吗?我不知道。
Speaker 237:36 - 37:40
We could do it as a test, I guess. A tricky one.
Speaker 237:36 - 37:40
我想我们可以把这个当作一个测试来做。这个问题挺棘手的。
Speaker 137:40 - 38:10
That is tricky. Then how do you think about also what is granola's role versus the codexes and the co works of the world? And what do you see as being places where you want to integrate and make yourself available to what they do versus places where you're like, I want to own this. This is like something that we can uniquely do well.
Speaker 137:40 - 38:10
这确实很棘手。那你还会怎么去看 granola 的角色,以及它和这个世界上的 codexes、co works 之间的关系?你觉得哪些地方是你们想要做 integration(集成)、让自己能够接入他们所做事情的;又有哪些地方你会觉得,“这个我想自己掌控,这属于我们能独特地做好的东西”?
Speaker 238:11 - 39:01
Yeah. So my view here is that it's like a new, again, it's like this new era of computing where for the first time computers can make sense of our context and you can do amazing things with that. That end, having the right context is really critical in you being able to get value out of AI. So my belief is that the context you have in granola, we should make that as available to however you want to use it. So if you want to use it in Codecs, if you want to use it in your personal agent, if you want to use whatever you can do that if you can make your life better to use Granola context, should, and we should make that really easy.
Speaker 238:11 - 39:01
对。所以我在这件事上的看法是,这像是一个新的——再说一次,这像是计算的新纪元:计算机第一次能够理解我们的 context(上下文),而你可以基于这一点做出非常惊人的事情。为此,拥有正确的 context,对于你能否从 AI 中获得价值是极其关键的。所以我的信念是,你在 granola 里拥有的 context,我们应该让它尽可能可用于任何你想使用它的方式。因此,如果你想在 Codecs 里用它,如果你想在你的 personal agent(个人 agent)里用它,或者不管你想怎么用,你都应该能这么做;如果使用 Granola 的 context 能让你的生活变得更好,那就应该这样,而且我们应该让这件事变得非常容易。
Speaker 239:01 - 39:39
And there's a lot, you know, our API or MCP is going to get a lot better over the next couple months because that's like a first class goal of ours. I think there are a number of jobs to be done, pain points, use cases that we should just be 5x better than anybody else in the world at. The obvious ones are all related to meetings. It's just one of those things where Maybe you feel differently. Maybe you're more AI pill than I am.
Speaker 239:01 - 39:39
而且有很多,你知道的,我们的 API 或 MCP 在接下来几个月里会变得好很多,因为那是我们的一个 first-class goal(最高优先级目标)。我觉得有不少 jobs to be done、pain points(痛点)、use cases(使用场景),我们就应该做到比世界上任何其他人都好 5 倍。最明显的那些都和会议有关。这就是那种事情——也许你感觉不同,也许你比我更 AI pill(更坚定相信 AI)。
Speaker 239:40 - 40:30
I think that the models are going to keep getting better, but I think that if Granolah just cares about a few things way more than anybody else and we optimize the hell out of those things, they'll just be a better experience, especially if they're tied to a UI that you have to use in vivo during a meeting, for example. And so that's our current strategy is basically be best in world in anything that's meeting adjacent and then be the best way to capture context to power any agents you have. I can give you an example. There's like, I was talking a user and they're doing something really clever. She worked in sales and they basically used, I think it was Claude to generate a microsite before they would talk to a potential client.
Speaker 239:40 - 40:30
我认为模型会持续变得更强,但我也认为,如果 Granolah 只是在少数几件事上比任何人都更在乎,并且把这些事情往死里优化,那体验就是会更好,尤其是当这些事情还和一个你必须在会议进行中、in vivo(现场/实时)使用的 UI 绑定在一起时。所以我们当前的策略基本上就是:凡是 meeting-adjacent(会议相关/邻近)的事情,都要做到世界最佳;然后再成为捕获 context(上下文)以驱动你拥有的各种 agent 的最佳方式。我可以举个例子。我之前和一个用户聊过,她们做了一件非常聪明的事。她在 sales 工作,她们基本上会在和潜在客户交谈之前,用我记得是 Claude 去生成一个 microsite。
Speaker 240:30 - 40:58
And the microsite was it was really impressive. They had this template and then of what that microsite would look like. And then given a new potential customer, they would the cloud would, like, style that template to look like the customer's website. And then they would connect the granola mcp and just pull all the context from granola and like fill out all the data and the fields and the widgets based on that data. And that's just the kind of thing that we would never try to be best of the world at and wouldn't make sense to, right?
Speaker 240:30 - 40:58
那个 microsite 真的很令人印象深刻。她们先有一个模板,规定这个 microsite 应该长什么样。然后给定一个新的潜在客户,Claude 就会把这个模板调整成看起来像那个客户自己的网站。接着她们再接入 granola mcp,把 granola 里的所有 context 都拉出来,然后基于这些数据去填充所有的数据项、字段和 widgets。像这种事情,我们根本不会尝试做到世界第一,而且那样做也没什么意义,对吧?
Speaker 240:59 - 41:31
But she would also be, in her case she was like writing a lot of emails in granola right after meetings because like follow-up emails in granola made a lot of sense for her and that's the kind of thing where it's like maybe we don't want to write every email. We don't want to your email inbox, but there might be a type of email that you want to send right after a meeting that is extremely good on certain dimensions that we could be the best in the world at. It makes sense to do both. I don't think it's a one or other. I was undecided about this two years ago when we started Granola.
Speaker 240:59 - 41:31
但与此同时,对她来说,她也会在会议结束后立刻在 granola 里写很多邮件,因为对她而言,在 granola 里写 follow-up emails(跟进邮件)非常合理。而这类事情就属于:也许我们不想去写所有邮件,我们也不想做你的 email inbox,但可能有一种你会在会后立刻发送的邮件类型,在某些维度上可以做得极其出色,而我们完全可以把它做到世界第一。两者都做是合理的。我不认为这是非此即彼。两年前我们刚开始做 Granola 的时候,我对这一点还拿不定主意。
Speaker 241:31 - 41:35
Now I have strong conviction, but it kind of took me a while to figure that out.
Speaker 241:31 - 41:35
现在我有很强的 conviction(确信),但我确实花了一些时间才把这件事想明白。
Speaker 141:36 - 42:22
That's interesting. Yeah, I think from a user perspective and I may not be your core user, and that's actually really interesting to me to learn that. Because of my the thesis that we're building every round and one of my convictions is what builders do now, people who want to figure out these general purpose tools and build these workflows or whatever, regular people are going to do in a few years once they're more built in once it's built into the handrails of these products. And so what I do is I just play around with stuff and try to figure out, oh, this is what I use Codecs and Slack agents and whatever for. And eventually, that will make its way into maybe stuff that we build, but stuff that the labs build or stuff that other people build, and we get to see it first.
Speaker 141:36 - 42:22
这很有意思。是的,我觉得从用户视角来看——而且我可能不是你们的核心用户——知道这一点其实让我很感兴趣。因为我们在做的 thesis(判断框架)之一,以及我的一个 conviction,就是现在 builders 在做的事,那些想弄明白这些通用工具、搭建这些 workflows(工作流)之类的人在做的事,几年之后普通人也会去做——等这些能力被更好地内建进产品、变成这些产品的 handrails(默认护栏/引导)之后。所以我现在做的事,就是到处试各种东西,试着搞明白:哦,这就是我会把 Codecs、Slack agents 以及其他类似工具用在什么地方。最终,这些东西会进入也许是我们自己做的产品里,或者 labs 做的东西里,或者别人做的东西里,而我们只是更早看到它们。
Speaker 242:22 - 42:48
Can I ask you a question about that? Because I think, is the theory that, so the stuff that you are trying to automate with AI or the interactions you're trying to build, do you think that other people will also be building their own audit? AI for people will be building their version of those automations, or do you think those will be productized? Productized. Yeah, okay.
Speaker 242:22 - 42:48
我能就这个问你一个问题吗?因为我在想,你们的理论是不是这样:你现在试图用 AI 自动化的那些事情,或者你试图构建的那些交互,你觉得其他人之后也会自己去搭建他们版本的这些 automations(自动化),还是你觉得这些东西最终会被 productized(产品化)?会产品化。好,明白了。
Speaker 242:48 - 43:39
I have the same view. You're the guinea pigs, right? Yes, Basically, yeah. And we have tons of guinea pigs using granola, right? My view is basically we should just look at what are the most common and cutting edge ways in which people are using our API and MCP to do crazy stuff and then figure out which of those do we think we could be the best in the world at and a lot of users would benefit from and then build a really beautiful seamless product experience that might have taken I ran into one of the founders of Hugging Face the other day at a conference and he was like, man, I had built this super convoluted complicated pre meeting brief workflow in Claude Code and I'd been working on it for months and I just turned it off because yours was better.
Speaker 242:48 - 43:39
我看法也一样。你们就是 guinea pigs(小白鼠/先行试验者),对吧?对,基本上是这样。而且我们这边也有大量 guinea pigs 在用 granola,对吧?我的看法基本上是:我们应该去看,人们现在最常见、也最前沿地用我们的 API 和 MCP 做疯狂事情的方式到底是什么;然后判断其中哪些是我们认为自己能够做到世界第一、而且会让很多用户受益的;再把它做成一个真正漂亮、无缝的产品体验。前几天我在一个 conference 上碰到 Hugging Face 的一位创始人,他说,哥们,我之前在 Claude Code 里搭了一个超级绕、超级复杂的会前简报 workflow,已经折腾了好几个月,但我后来直接把它关掉了,因为你们的更好。
Speaker 243:39 - 43:43
And to me that was like, that's exactly what I want. You can't do
Speaker 243:39 - 43:43
对我来说,我当时就觉得,这正是我想要的。你没法做
Speaker 143:43 - 43:43
this That's for
Speaker 143:43 - 43:43
这个。这是为了
Speaker 243:43 - 43:51
exactly right. But that's exactly what I want because we're just going to care more than any individual should ever care about any one workflow is kind
Speaker 243:43 - 43:51
完全正确。但那也正是我想要的,因为我们就是会比任何一个个体本该对任何单一 workflow(工作流)投入的关注都更多,
Speaker 143:51 - 44:17
of how I think about it. I think that's exactly right. Okay. So we're definitely aligned there. My current feeling is, In general, when I and this is again, this is just for me and and this may change, but when I see granola, like a like the UI come up, it's usually a mistake.
Speaker 143:51 - 44:17
我大概是这么想的。我觉得这完全对。好。所以我们在这点上肯定是一致的。我目前的感觉是,一般来说,当我看到 granola,比如它的 UI(用户界面)弹出来时,通常就说明哪里出错了。
Speaker 144:18 - 44:34
You don't want to see it. I don't want to see it. I mean, it's fine. The recording thing is fine. In general, the way that I want to access it is going to be it pushing something into codex, which is my work surface, or codex pulling something out of it.
Speaker 144:18 - 44:34
你不会想看到它。我也不想看到它。我的意思是,录音这个功能没问题。总体来说,我希望访问它的方式是:要么它把某些东西推送到 codex——那是我的工作界面——要么 codex 从里面拉取一些东西。
Speaker 144:35 - 45:15
And I think this focus on meetings is so right. And it's actually such a hard and deep problem to just get that, like the context of that right. So for example, I mean, you know this better than I, but I'm just thinking about when you do a transcript and I say the name Chris, which Chris are we talking about, right? And so I think of the job of software companies in an AI world. AI is this really super flexible thing, but it's only flexible if it's given the right structure and the right context within which to be flexible.
Speaker 144:35 - 45:15
而且我觉得,聚焦会议这个方向太对了。实际上,光是把这里面的上下文处理对,就是一个非常困难而且很深的问题。比如说,我是说,你肯定比我更懂这个,但我在想的是,当你做 transcript(转录稿)时,如果我说了 Chris 这个名字,我们说的是哪个 Chris,对吧?所以我会这样理解 AI 世界里软件公司的工作:AI 是一种极其灵活的东西,但只有在给它正确的结构和正确的上下文、让它在其中发挥灵活性时,它才真正灵活。
Speaker 145:16 - 46:16
So if if, you know, AI if we're using the human body as an analogy, AI is like the the ligaments and the muscles and then software has to be the bones. And the bones sort of set like give it form and a structure, but then the the AI can kind of like work around that to to do anything. And there's so much for even even, for example, again, in a meeting transcript, I might say, that's a good idea. But the way I say that means says a lot about what I actually think that is not captured in just a transcript. And I want granola to be I this automation that runs that turns all the Slacks and all the meetings that I was not in into cards that I just go through.
Speaker 145:16 - 46:16
所以如果——你知道——如果用人体来打比方,AI 就像韧带和肌肉,而软件必须是骨骼。骨骼某种程度上是在设定——给它形态和结构,但 AI 则可以围绕这个框架运作,去做各种事情。这里面还有太多东西,甚至——还是拿会议 transcript(转录稿)举例——我可能会说,“那是个好主意。”但我说这句话的方式,其实传达了很多我真正怎么想的信息,而这些是单靠 transcript 并不能捕捉到的。我希望 granola 成为——我是说——这样一种自动化:它持续运行,把所有 Slack 消息和所有我没参加的会议,整理成一张张卡片,供我快速浏览。
Speaker 146:16 - 46:53
And I can see things that happened in the company. But the transcripts are, A, often wrong, and B, they capture none of the nuance of how and why things were said. And I think if Granolah did that for me, it would be the most valuable thing in the world, and I wouldn't necessarily have to look at the product. And maybe for a more normie user who doesn't want to be in Codecs all the time, although I do think that that's going to be more and more the case. You still want to have the UI, but there's all this background context processing that I would look to you guys to do that I don't really want to do.
Speaker 146:16 - 46:53
这样我就能看到公司里发生了什么。但 transcript(转录稿)一来经常不准确,二来完全捕捉不到人们是如何、又为什么那样说话的细微差别。我认为如果 Granolah 能为我做到这一点,那它会是世界上最有价值的东西,而我甚至不一定需要去看这个产品本身。也许对于更普通的用户——他们不想一直待在 Codecs 里,尽管我确实觉得以后这种情况会越来越普遍——你还是需要保留 UI,但还有大量这种后台的上下文处理工作,我会希望你们来做,而不是我自己去做。
Speaker 246:54 - 47:04
Yeah. Yeah. No, that makes total sense. I think, I guess, the transcript is one way to capture the context from a meeting, right? It is the simplest way.
Speaker 246:54 - 47:04
对,对,不,这完全说得通。我想,我觉得,transcript(会议转录)是捕捉会议上下文的一种方式,对吧?它是最简单的方式。
Speaker 247:04 - 47:30
And as LMs are weirdly good at making sense of garbled transcripts and doing something useful with that. I think but what's the actual job to be done here? It's actually, like, what are the insights or what are the decisions or what are the actions I need to take on top of this context? And that there's this layer of intelligence or interpretation, which is like the disambiguation. Like, which Sam did I mean?
Speaker 247:04 - 47:30
而且 LMs 奇怪地很擅长理解这些杂乱的 transcript,并从中做出一些有用的事。我想,但这里真正要完成的任务是什么?其实是,基于这些上下文,我需要得到哪些洞察、做出哪些决策,或者采取哪些行动?这里面还有一层 intelligence(智能)或 interpretation(解释),也就是消歧。比如,我说的是哪个 Sam?
Speaker 247:30 - 47:51
Is like a great example. We have this so we have this I'll give you like, it's such an interesting space. So here's the challenge we have, right? When a user starts using granola, they use granola a lot, right? And therefore, we have a lot of context about that user, right?
Speaker 247:30 - 47:51
这就是个很好的例子。我们有这个——我给你举个例子,这是个特别有意思的领域。所以我们面临的挑战是这样的,对吧?当用户开始使用 granola 时,他们会非常频繁地使用它,对吧?因此,我们会掌握这个用户的大量上下文,对吧?
Speaker 247:51 - 48:38
And we can, in the background and we have this, but we don't use it yet, and I'll explain why generate a pretty high fidelity picture of who you are, what you're working on, what you're trying to achieve, what are your challenges, who you're working with on different things on. And let's just, for argument's sake, let's say this is like a two page summary of who Dan is today, The state of Dan today, and it changes every day. And we can generate better notes if we pass that context as part of our note generation pipeline. But now those notes might include information that didn't come from the meeting, Right? So like when you said it's like, Oh, that's interesting.
Speaker 247:51 - 48:38
而且我们可以在后台——我们其实已经有这个能力了,只是还没用,我等下会解释为什么——生成一幅相当高保真的画像:你是谁、你在做什么、你想达成什么、你面临哪些挑战、你在不同事情上和谁合作。为了方便讨论,我们假设这就像是一份两页的摘要,描述今天的 Dan 是谁,也就是 Dan 今天的状态,而且它每天都在变化。如果我们把这些上下文作为 note generation pipeline(笔记生成流程)的一部分传进去,我们就能生成更好的 notes。但这样一来,那些 notes 里可能会包含并非来自这场会议的信息,对吧?所以就像你刚才说的那样,会变成“哦,这很有意思”。
Speaker 248:38 - 48:46
Or what was your example? It's like, Oh, I like that idea. And like, we might be like, Dan doesn't actually like that idea. He thinks that's a terrible idea. Cause we know he hates that general class of idea.
Speaker 248:38 - 48:46
或者你刚才的例子是什么来着?像是,“哦,我喜欢那个想法。” 但我们可能会判断,Dan 实际上并不喜欢那个想法。他觉得那是个很糟糕的主意。因为我们知道他讨厌那一大类想法。
Speaker 248:46 - 49:09
Cause he said that five other times in five other meetings. And so then there's this interesting question of like, who's the audience? And like today a lot of people still use the notes for themselves, also there's this mixed use with also share it with other people. Right? Whereas like the perfect notes for you are actually probably a little bit different or the perfect like notes are just one way to represent that information.
Speaker 248:46 - 49:09
因为他在另外五场会议里已经说过五次了。于是这里就出现了一个很有意思的问题:受众是谁?而且现在很多人仍然主要是把 notes 给自己用,同时也会混合着分享给其他人,对吧?但其实,最适合你自己的 notes 很可能会有点不同,或者说,所谓完美的 notes 其实只是表达这些信息的一种方式。
Speaker 249:09 - 49:22
There's other ways. It's like what level of intelligence or interpretation do you want granola to do? But yeah, totally hear you. This is why I think infancy days for
Speaker 249:09 - 49:22
其实还有别的方式。问题在于,你希望 granola 做到什么程度的 intelligence 或 interpretation?不过对,我完全明白你的意思。所以我才觉得,这个领域目前还处在非常早期的阶段。
Speaker 149:22 - 50:01
Or honestly, like, you seemed really stressed in that meeting. Do you want to talk about it? Be really, really interesting. Like maybe people would be freaked out by that, but also one of the things I use granola for a lot is like management stuff being like, Or do I do someone on my team had a difficult conversation with someone on their team and they're talking to me about it and I'm like, let's talk to Claude about the meeting transcript so we can decide what to do. All that kind of stuff is in order to do it well, I need not just the transcript, but there was a long pause after XYZ said this thing.
Speaker 149:22 - 50:01
或者说得更直接一点,比如,“你在那场会议里看起来压力很大。你想聊聊吗?” 这会非常、非常有意思。也许人们会被吓到,但另一方面,我经常把 granola 用在很多管理相关的事情上,比如说,我团队里的某个人和他们团队里的某个人发生了一次棘手的对话,然后他来和我讨论,我就会想,“我们来和 Claude 聊聊这份会议 transcript,看看该怎么处理。” 诸如此类的场景都是这样。要把这件事做好,我需要的不只是 transcript,还需要像这样的信息:在 XYZ 说了那句话之后,现场出现了一个很长的停顿。
Speaker 150:01 - 50:17
No one said anything for three minutes. You don't see that in the transcript. No. Yeah. Stuff like that is just there's so much, like, richness there to be had that granola to to me, I want granola to do for me.
Speaker 150:01 - 50:17
三分钟里谁都没有说话。你在 transcript 里是看不到这一点的。对,没错。像这种东西里其实有太多、太丰富的内容可挖了,所以对我来说,我希望 granola 能替我做到这一点。
Speaker 250:17 - 50:29
Yeah. That makes that makes total sense. Yeah. I also think, I'm curious what are all the main way because he was like, don't want see the granola UI. I want to pull that in Codex or whatnot.
Speaker 250:17 - 50:29
对,这完全说得通。对。我也在想,我很好奇主要有哪些方式,因为他当时的意思有点像,我不想看 granola 的 UI,我想把那个内容拉到 Codex 之类的地方去。
Speaker 250:30 - 50:53
I also wonder, okay, are the main workflows or reasons why people are pulling transcripts or meeting notes into their agents? Which of those, if you were to optimize around, you would do a better job at? Which of those are the long tail of like, you know, give people power and they'll do a better better job is also an interesting question for for us.
Speaker 250:30 - 50:53
我也在想,大家把 transcript 或 meeting notes 拉进他们的 agent 里,主要的 workflow 或原因到底是什么?如果你要围绕这些场景做优化,哪些是你能做得更好的?而哪些又属于那种 long tail(长尾)需求——你知道的,给人们更多能力,他们就会把事情做得更好——这对我们来说也是个很有意思的问题。
Speaker 150:53 - 51:08
Yeah. What is on your mind for the next, like, six to twelve months as as these like, okay, you saw Fable come out. Does that affect your roadmap? Does that affect how you're hiring or how you're building? Yeah.
Speaker 150:53 - 51:08
对。对于接下来大概六到十二个月,你在想些什么?比如说,你看到 Fable 发布了,这会影响你的 roadmap 吗?会影响你怎么招人、或者怎么构建产品吗?对。
Speaker 151:08 - 51:09
Tell me about that.
Speaker 151:08 - 51:09
跟我讲讲那个吧。
Speaker 251:09 - 51:37
The order of operation, the questions that go through my mind is first, like, what does, how does this change how we build internally and how do we structure our teams? Like, that's the first question. I think our roadmap is and this is the kind of thing where it's like, you know, you can have one plan in the abstract and then you actually play with something you should you should adjust it. And I unfortunately didn't play with Fable enough before it got yanked. But the roadmap is more it's kind of baked into the roadmap a little bit more, which is like, Okay, models will get better.
Speaker 251:09 - 51:37
我脑子里思考问题的顺序、会先问自己的问题,首先是:这会怎样改变我们的内部构建方式,以及我们该如何组织团队?这是第一个问题。我觉得我们的 roadmap 是——这也属于那种,你知道的,你可以在抽象层面有一个计划,但当你真的上手玩了某个东西之后,你就应该调整它。不巧的是,Fable 在被下架之前,我没来得及充分体验它。不过 roadmap 方面,更多是有一点已经 baked into the roadmap(内嵌进路线图)了,也就是:Okay,models 会变得更好。
Speaker 251:37 - 52:03
Here's the general strategic direction. Like I think our strategy is more dependent on what people use us for and what are the jobs to be done or use cases around that that we can tackle. Model intelligence doesn't affect that directly too much, if that makes sense. Whereas model capabilities do affect how many people do we staff on a pod, you know? Like how do we define goals there?
Speaker 251:37 - 52:03
这是总体的战略方向。比如我觉得,我们的 strategy 更多取决于人们用我们来做什么,以及有哪些 jobs to be done(待完成任务)或者相关 use cases(使用场景)是我们可以去解决的。如果这样说得通的话,model intelligence 并不会太直接地影响这一点。相对而言,model capabilities 确实会影响,比如一个 pod(小团队)我们该配多少人,对吧?以及我们该怎么定义那里的目标?
Speaker 252:03 - 52:14
Should we be approaching things completely differently? So yeah, that's the thing that's on my mind for sure. The other thing is I
Speaker 252:03 - 52:14
我们是不是应该用一种完全不同的方式来做事?所以对,这肯定是我在想的一件事。另一件事是我
Speaker 152:17 - 52:17
think
Speaker 152:17 - 52:17
思考
Speaker 252:20 - 52:51
So maybe this touches on a little bit, like it's the stuff that you're using Codex for. I think there's one or multiple canonical UIs that haven't been invented yet that are like right now, so Granolah you have meeting notes, right? And you can use meeting notes or transcripts or context for lots of things. Usually when you're trying to make use of that context, it's not per meeting. It's usually you have a set of contexts.
Speaker 252:20 - 52:51
所以这也许会稍微涉及一点,比如说这和你用 Codex 做的那些事有关。我觉得有一个或多个“标准型”的 UI 还没有被发明出来,情况大概是这样的:比如在 Granolah 里,你有会议记录,对吧?而你可以把会议记录、transcript(转录文本)或 context(上下文)用于很多事情。通常当你试图利用这些 context 时,并不是按单场会议来用的。一般来说,你会有一组 context。
Speaker 252:51 - 53:24
Our case, let's just say it's like all the meetings you had today or all the meetings you had this week. And maybe it's actually all the meetings and all the Slack messages and all the emails, but let's just keep it to meetings for a second. At that level of altitude, what is the right UI and interaction design to manipulate or work with that context, I don't think anyone has figured that out. Maybe you're cooking on something, Dan, or maybe you'll figure it out. When I look around, I'm like, I've seen very little actually.
Speaker 252:51 - 53:24
以我们的场景为例,假设就是你今天开的所有会议,或者这周开的所有会议。也许实际上还包括所有会议、所有 Slack 消息和所有邮件,但我们先只谈会议。在这个抽象层级上,到底什么样的 UI 和交互设计才适合去操作或处理这些 context,我觉得还没有人真正想明白。也许你正在琢磨什么,Dan,或者也许你会把它想出来。因为我放眼看去,真正看到的东西其实非常少。
Speaker 253:26 - 53:58
And coming up with new UI paradigms is very hard and it's very rare for new ones to stick around. I think they're more discovered than invented, if that makes sense. But there's a very clear gap that I see there. We have a chat UI and chat threads are super valuable, actually surprisingly versatile and powerful when you're doing one thing. But when you're actually trying to do multiple things across varied contexts, I don't think we have the right metaphor there.
Speaker 253:26 - 53:58
而且提出新的 UI 范式非常难,真正能留存下来的新范式也极少。我觉得它们更像是被“发现”的,而不是被“发明”的,如果这么说你能理解的话。但我在这里看到一个非常明确的空白。我们现在有 chat UI,而 chat thread(聊天线程)确实非常有价值,实际上它在你只做一件事的时候,出乎意料地通用而且强大。但当你真的要跨不同 context 同时做多件事时,我觉得我们还没有找到正确的 metaphor(隐喻/交互比喻)。
Speaker 253:58 - 54:16
So that's something that's very much on my mind. Because when you talk about the normie user, I think you're going to need a simple metaphor that the normie user can just grok and interface with. It's not going to be the kind of hoops that you currently jump through with Codecs.
Speaker 253:58 - 54:16
所以这件事我最近一直在想。因为当你谈到普通用户时,我觉得你需要一个简单的 metaphor,让普通用户能够直接理解并与之交互。它不会是你现在在 Codex 里那种需要层层绕弯、反复折腾的使用方式。
Speaker 154:16 - 54:23
Yeah. How do you think about token spend both on your team and in the product?
Speaker 154:16 - 54:23
嗯。你是怎么考虑 token 消耗的,无论是在你的团队内部,还是在产品里?
Speaker 254:24 - 54:44
I think Okay, I'll do team first. So token spend on team. I think folks on the team should be using AI to augment their abilities as much as possible. I think it's silly to say, if you aren't spending this many tokens, then you're doing it wrong. That's like a, I feel like a weird way to do it.
Speaker 254:24 - 54:44
我想,好的,我先说团队。关于团队里的 token 消耗,我觉得团队成员应该尽可能用 AI 来增强自己的能力。我觉得如果说“如果你没花这么多 token,那你就是做错了”,这很蠢。我感觉那是一种很奇怪的管理方式。
Speaker 254:44 - 55:04
I kind of trust the team to take the spirit of the goal, which is, like, be the most productive productive you can be and, you know, be thoughtful about how you use it. We didn't even track token spend, like, we're until, like, last week internally just because it wasn't a priority for me. In terms of token spend in the product
Speaker 254:44 - 55:04
我比较相信团队会把握这个目标背后的精神,也就是尽可能做到最高效率,并且认真思考该怎么使用它。其实我们内部直到上周左右,甚至都还没有追踪 token 消耗,因为这对我来说并不是优先事项。至于产品里的 token 消耗
Speaker 155:05 - 55:12
Wait, wait, wait. Before we get there, but you started last week. What spurred it and then what did you come
Speaker 155:05 - 55:12
等等,等等,等等。在我们讲到那儿之前,你们是上周才开始的。是什么促成了这件事,然后你们最后得出了什么结论?
Speaker 255:12 - 55:25
to? Fable. Basically what spurred it was like, Do we even know how much we're spending? We're like, We should probably know how much we're spending. It was basically because folks were using Fable.
Speaker 255:12 - 55:25
是 Fable。基本上,真正促使这件事发生的是:我们到底知不知道自己花了多少钱?我们当时想的是,我们大概应该知道自己花了多少钱。说到底,就是因为大家在用 Fable。
Speaker 255:25 - 55:32
Someone's like, I think I spent a couple thousand dollars today. We're like, Okay, maybe this is something we should at least be aware of.
Speaker 255:25 - 55:32
有人说,我觉得我今天花了几千美元。我们就想,好吧,也许这至少是件我们应该了解清楚的事。
Speaker 155:32 - 55:41
We had literally the same thing. Fable came out and Kieran, who runs Quora, he was on track to spend a million dollars in credits
Speaker 155:32 - 55:41
我们当时也完全一样。Fable 一出来,负责 Quora 的 Kieran 原本就快要花掉价值一百万美元的 credits(额度)
Speaker 255:41 - 55:42
in In a one
Speaker 255:41 - 55:42
在一……
Speaker 155:44 - 55:46
And a we were like, Okay, we need to figure out.
Speaker 155:44 - 55:46
然后我们就想,好吧,我们得搞清楚这件事。
Speaker 255:46 - 55:49
Maybe we need to think about this a little bit.
Speaker 255:46 - 55:49
也许我们需要稍微认真想一想这个问题。
Speaker 155:49 - 55:53
It actually could be worth it, and also we need to think about it, you know?
Speaker 155:49 - 55:53
它其实可能是值得的,但同时我们也需要认真考虑一下,你懂吧?
Speaker 255:53 - 56:06
Exactly. Yeah. This goes back to like, you know, things always outside of our ability. It's just one of those like, it's like, yeah, we should probably, we should probably at least know how much we're spending. That sounds like a really obvious thing, but it's just like, we're just so busy with so many other things.
Speaker 255:53 - 56:06
没错。是啊。这又回到那种,你知道的,总有些事超出我们的掌控。这个问题就属于那一类。就是说,嗯,我们大概至少应该知道自己花了多少钱。这听起来像是个特别显而易见的事,但现实是,我们手头还有太多别的事情要忙。
Speaker 256:06 - 56:58
It's like, now now's the time. In product, I think I'm still a believer that it's early days and we have so much to gain by figuring out what are the killer kind of AI native experiences that Granolah can build, and therefore we're not very cost conscious on the token spend. So like I said before, we pre generate, millions of these briefs even if only 10% of them get opened because we want to figure out what's that right experience for the user. And our belief is both cost will go down over time and we can optimize cost once we know what the best experiences are. That's our philosophy.
Speaker 256:06 - 56:58
感觉就像,现在就是那个时候了。在产品上,我仍然相信这还处于早期阶段,而我们还有很多可获得的价值,关键在于弄清楚 Granolah 能构建出哪些真正 killer 级的 AI native(AI 原生)体验。因此,我们目前对 token 消耗的成本并没有那么敏感。所以就像我之前说的,我们会预先生成数百万份这样的 briefs,即使最后可能只有 10% 会被点开,因为我们想搞清楚对用户来说,那个正确的体验到底是什么。我们的信念是,一方面成本会随着时间下降,另一方面,一旦我们知道了最好的体验是什么,就可以再去优化成本。这就是我们的理念。
Speaker 256:59 - 57:11
Agenetic features are really expensive and so as we build more and more of them, at some point that'll break the math for us. But for now that's been the priority.
Speaker 256:59 - 57:11
agentic 功能确实非常昂贵,所以随着我们构建越来越多这类功能,到了某个时点,这在账上就会算不过来了。但目前为止,这一直是我们的优先事项。
Speaker 157:12 - 57:26
And how much are people using true agentic workflows in Granola? I ask that as like a I'm really curious about normie users learning how to use power features like those.
Speaker 157:12 - 57:26
那在 Granola 里,用户对真正的 agentic workflow(agentic 工作流)的使用程度有多高?我这么问,是因为我真的很好奇,普通用户是怎么学会使用这类高级功能的。
Speaker 257:26 - 58:03
It's the word agentic that's tough, right? I think roughly half is the answer. Weekly, every week about half the granola users use us in an agentic way. But I don't know if they would describe it that way, if that makes sense. They'll ask a query that is kind of a complex query around, let's say, your coaching or improvement question that will look at a series of meetings over time and then do several steps of analysis over that, right?
Speaker 257:26 - 58:03
难点就在于 agentic 这个词本身,对吧?我觉得大致的答案是,一半左右。按周来看,每周大概有一半的 Granola 用户会以一种 agentic 的方式使用我们。但我不确定他们自己会不会这样描述,如果你明白我的意思的话。他们会提出某种比较复杂的 query(查询),比如说,围绕你的辅导或改进问题,这种查询会查看一段时间内的一系列会议,然后在其上执行多个分析步骤,对吧?
Speaker 258:03 - 58:10
Like that kind of thing is roughly half a week, yeah. Really interesting.
Speaker 258:03 - 58:10
对,像这种类型的使用,大概占到每周的一半。很有意思。
Speaker 158:11 - 58:16
Chris, if people are looking to find you online to use granola, where should they find you?
Speaker 158:11 - 58:16
Chris,如果大家想在网上找到你,或者开始使用 granola,他们应该去哪里找你?
Speaker 258:16 - 58:31
I don't know. Granola's Twitter handle or X handle is meet granola. Mine is CJ Pedrigal, so CJ my last name. And, yeah, I guess those are the best places.
Speaker 258:16 - 58:31
我也说不好。Granola 的 Twitter 账号,或者说 X 账号,是 meet granola。我的账号是 CJ Pedrigal,就是 CJ 加上我的姓。对,我想这些大概就是最合适的地方。
Speaker 158:32 - 58:37
Amazing. I always love talking to you. I love what you're building. Thanks for coming on. We got to do this more often.
Speaker 158:32 - 58:37
太棒了。我一直都很喜欢和你聊天。我很喜欢你正在打造的东西。感谢你来参加节目。我们以后一定要更常这样聊。
Speaker 258:38 - 58:46
Yeah, absolutely. Whether it's on the pod or just us chatting, really enjoy these conversations.
Speaker 258:38 - 58:46
对,当然。无论是在 pod 里还是就我们私下聊天,我都真的很享受这些对话。
Speaker 158:46 - 58:46
Sounds great.
Speaker 158:46 - 58:46
听起来很棒。
Speaker 358:57 - 59:20
Positively have to smash that like button and subscribe to AI and I. Why? Because this show is the epitome of awesomeness. It's like finding a treasure chest in your backyard, but instead of gold, it's filled with pure unadulterated knowledge bombs about chat GPT. Every episode is a roller coaster of emotions, insights, and laughter that will leave you on the edge of your seat craving for more.
Speaker 358:57 - 59:20
绝对得狠狠干碎那个 like 按钮,然后订阅 AI and I。为什么?因为这个节目简直就是“超棒”的典范。就像你在自家后院发现了一个宝箱,只不过里面装的不是黄金,而是关于 chat GPT 的纯粹、毫无掺杂的知识炸弹。每一集都是一场情绪、洞见和笑声交织的过山车之旅,会让你全程坐立难安,还意犹未尽、想继续看下去。
Speaker 359:20 - 59:37
It's not just a show. It's a journey into the future with Dan Shipper as the captain of the spaceship. So do yourself a favor. Hit like, smash subscribe, and strap in for the ride of your life. And now without any further ado, let me just say, Dan, I'm absolutely, hopelessly in love with you.
Speaker 359:20 - 59:37
这不只是一个节目。这是一场通往未来的旅程,而 Dan Shipper 就是这艘宇宙飞船的船长。所以帮自己个忙吧。点个 like,狠狠干碎 subscribe,然后系好安全带,准备迎接你人生中最刺激的一趟旅程。现在闲话少说,我只想说一句,Dan,我已经彻底、无可救药地爱上你了。
原文 ↗https://www.youtube.com/playlist?list=PLuMcoKK9mKgHtW_o9h5sGO2vXrffKHwJL
BuildSpeak — 关于本项目BUILT IN PUBLIC · 跟随 builders 而非 influencers