The Every Podcast - SaaS 末日是金矿:专访 Figma 的 Matt Colyer 封面

SaaS 末日是金矿:专访 Figma 的 Matt Colyer

The SaaS Apocalypse Is a Goldmine With Figma’s Matt Colyer

本集简介

“软件末日”——即人工智能将使软件即服务(SaaS)变得过时的恐慌——并未让 Figma 的 Matt Colyer 感到不安。作为公司面向开发者的产品管理总监,他已自行构建智能体两年,且购买的软件服务比以往任何时候都多。 除了论证 AI 是 SaaS 公司的“金矿”外,Colyer 还与 Dan Shipper 在《AI & I》节目中探讨了为何优秀的设计需要“钻石型”流程:首先发散,尽可能生成大量创意,然后收敛至最佳方案。对话是线性的,这使其适合对单一设计进行迭代,却不利于生成大量选项。Figma 新推出的画布内智能体正是对此的首次尝试。 他们还深入探讨了为何 AI 设计工具需要摆脱文本框的束缚、Figma 的 MCP 服务器如何打通代码与设计之间的闭环,以及为何“审查”已成为 AI 辅助产品开发中的最大瓶颈。 如果您觉得本期节目有趣,请点赞、订阅、评论并分享! 想了解更多来自 Dan Shipper 的内容: 订阅 Every:https://every.to/subscribe 在 X 上关注他:https://twitter.com/danshipper 时间戳: 1:03 - 引言 2:15 - 为何“软件末日”叙事本末倒置 5:27 - Matt 的邮件智能体起源故事 13:21 - 发散与收敛的设计思维 17:39 - Figma 的 MCP 服务器 19:45 - 为何设计智能体需要个性化 22:09 - 每个问题本质上都是上下文问题 25:12 - Apple 和 Google 作为上下文领域的霸主 28:18 - 为何审查成为新的瓶颈 本期节目中提到的资源链接: Matt Colyer 在 X 上:https://x.com/mcolyer Figma:https://figma.com Figma MCP 服务器:https://www.figma.com/blog/introducing-figma-mcp-server/

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。

Speaker 0

这可以说是SaaSpocalypse,或者换个说法,软件的下一个时代。

The SaaSpocalypse or the next era of software, if you will.

Speaker 0

我对此非常兴奋,我觉得Figma和其他很多SaaS公司也一样。

I'm really excited about it, and I think Figma and a lot of other SaaSpocalypse are too.

Speaker 0

我长期从事开发者工具领域的工作,大概五到十年前,全球开发者人数的估算大概是两千万、三千万,或者四千万左右。

Because, like, I've worked in developer tools for a long time, and maybe five, ten years ago, the estimate of, like, number of developers worldwide was, like, 25,000,000, 30,000,000, 40,000,000, give or take.

Speaker 0

我觉得这个时代最令人兴奋的地方在于,开发者人数可能会达到十亿,甚至更多。

I think what's most exciting about this time is that I think it's gonna be, like, a billion, maybe even more than that.

Speaker 0

我们正经历一个前所未有的时期,产品开发和科技普及都在加速推进。

I think there's this incredible time that we're moving through of product development and really the democratization of technology.

Speaker 0

科技。

Technology.

Speaker 0

最终的结果就是,世界上软件的数量大幅增加了。

I think the end result is that there is dramatically more software out there in the world.

Speaker 0

如果你身处这个领域,那就意味着一座金矿正在向你敞开。

If you're in that space, it means it's a gold mine.

Speaker 0

对吧?

Right?

Speaker 1

要跟上AI的前沿,Every是你唯一需要的订阅服务。

Every is the only subscription you need to stay at the edge of AI.

Speaker 1

如果你关心最新模型和工具,那就必须订阅Every,才能从信息洪流中抓住真正有价值的内容。

If you care about being on top of the latest models and using the latest tools, you have to subscribe to Every to separate out the signal from the noise.

Speaker 1

去 every.com

Go to every.

Speaker 1

今天就订阅吧。

Tosubscribe today.

Speaker 1

马特,欢迎来到节目。

Matt, welcome to the show.

Speaker 0

谢谢邀请,丹。

Thanks for having me, Dan.

Speaker 1

对于还不了解你的人,你是Figma负责开发者产品的高级产品经理。

So for people who don't know you, you are the director of product management for developers at Figma.

Speaker 1

我想先聊聊大家最关心的那个大问题。

And I want to start with, I think, the big question on everyone's mind.

Speaker 1

我得说一句,我大概两个月前买了一堆Figma的订阅,就是因为现在到处都在谈什么‘SaaS末日’。

I should probably say I bought a bunch of Figma, like, probably like two months ago, something like that, because there's this whole SaaSpocalypse narrative.

Speaker 1

我特别想跟你深入聊聊AI和产品管理,你肯定有很多亲身经历可以分享。

And what I want to get in detail or get get into with you, I think you have a lot of stuff to share about AI, AI and product management, all the stuff that you've been doing yourself.

Speaker 1

不过我也想先问问,AI时代,SaaS工具会变成什么样?

But also, I'd love to I'd love to start with what is going to happen to SaaS tools in AI?

Speaker 1

我觉得Figma是个特别有意思的例子,总有人觉得‘我以后不用Figma了’。

And I think Figma is a really interesting example where there are always people who are like, oh, I I don't have to use Figma anymore.

Speaker 1

你们不是刚在产品里推出了一个智能助手吗?

There are, you know, you guys just launched an agent, like, in your product.

Speaker 1

你们还有Figma MCP,对吧?

You also have Figma MCP.

Speaker 1

所以,从Figma刚起步时还没AI,到现在成了大产品、AI无处不在,这个转变过程你是怎么看的?

So if you're transitioning from so if you're transitioning from a world where you'd you know, there was no AI when Figma started to now you're a big scaled product and now there is AI.

Speaker 1

那这到底是怎么运作的?你们是怎么考虑要不要让产品支持智能代理的?

Like, how does that work and how are you how are you thinking about do we open the product up to agents?

Speaker 1

我们是该自己开发一个代理?

Do we build our own agent?

Speaker 1

哪些做法是有效的?

What's working?

Speaker 1

哪些行不通?

What's not?

Speaker 1

诸如此类的问题。

All that kind of stuff.

Speaker 1

我对这些特别感兴趣。

I'm really interested in that.

Speaker 0

是的。

Yeah.

Speaker 0

我今天很想聊聊这个。

I'd love to talk about that today.

Speaker 0

对我来说,这有几个不同的角度。

I think for me, it comes from a couple of different angles.

Speaker 0

我觉得首先一点是,所谓的SaaSpocalypse,或者说软件的下一个时代,换个积极的说法,可能就是这样。

I think the first thing is like the SaaSpocalypse or like, you know, the next era of software, if you will, maybe is more of positive framing on that.

Speaker 0

我特别兴奋,因为我一直在做开发者工具,时间挺长了。

I'm really excited about it because like I think like, I've worked in developer tools for a long time.

Speaker 0

大概五年前、十年前,全球开发者人数估计是两千五百万、三千万,或者四千万左右。

And like, you know, maybe five, ten years ago, the estimate of like number of developers worldwide was like, I don't know, 25,000,000, 30,000,000, 40,000,000 give or take.

Speaker 0

我觉得现在最令人兴奋的是,我觉得未来可能会达到十亿人。

I think what's most exciting about this time is that like, I think it's gonna be like a billion.

Speaker 0

对吧?

Right?

Speaker 0

甚至可能更多。

Like, maybe even more than that.

Speaker 0

对吧?

Right?

Speaker 0

所以我们正处在一个产品开发和科技普及的绝佳时期。

And so like, I think there's this incredible time that we're moving through of product development and really the democratization of technology.

Speaker 0

现在到处都在讲自研软件这类说法,不过这些咱们以后再聊。

Mean, there's a lot of catchphrases around like homegrown software and like we can talk about some of that stuff later.

Speaker 0

但归根结底,世界上软件的数量比以往多太多了。

But I think the end result is that there is dramatically more software out there in the world.

Speaker 0

所以回到你提到的SaaSpocalypse,对那些已经有成熟产品的公司来说,这意味着一个金矿。

And so kinda coming back to your point about like the SaaSpocalypse and like what does that mean for companies that have an established product is like if you're in that space, like it it means it's a gold mine.

Speaker 0

对吧?

Right?

Speaker 0

到处都是机会,我对此特别看好。

That that there's all this opportunity out there and that I'm really excited about it.

Speaker 0

Figma和其他很多SaaS公司也都是这么想的。

And I think, you know, Figma and a lot of other SaaS businesses are too.

Speaker 0

另外,网上一些负面讨论主要在说:要是能靠‘ vibe coding ’写出所有应用怎么办?

And so I think the other part, you know, kind of some more of the negative sentiments of the discussions you see online is around like, well, what if I could just vibe code every app?

Speaker 0

对吧?

Right?

Speaker 0

我觉得现在特别有意思的是,不知为什么,今年一月成了这个趋势真正成为主流叙事的转折点。

And I think what's really interesting about this time is like, for whatever reason, like, January of this year was like the point at which it, like, became the larger narrative.

Speaker 0

我其实早就开始做这些事了,大概有一年半到两年了吧。

Like, I've been doing this stuff for, like, probably eighteen months or two years.

Speaker 0

那时候我就觉得,对,咱们该把所有东西都做出来,但我觉得整个世界在今年一月突然就开窍了。

So, like, I was already like, yeah, let's go build everything, but I feel like the whole world's cut up in January of this year.

Speaker 0

我当时就想,对,咱们该把所有东西都做出来。

I was like, yeah, let's go build everything.

Speaker 0

现在大家都这么想了。

And, like, people are.

Speaker 0

我很期待看到接下来会发生什么,因为就我个人来说,最让人兴奋的就是把第一个版本做出来的时候。

And I am excited to see what happens because, like, I know my own personal journey through that is, like, it's really fun to build the initial version of it.

Speaker 0

对吧?

Right?

Speaker 0

其实两年前我就自己搞过一个智能代理,第一个就是邮件处理的。

And, like, I actually built one of my own agents two years ago, and the very first one was, an email agent.

Speaker 0

我后来回头看了看,它是怎么起步的。

And I had to look back as, like, how it started.

Speaker 0

说白了,就是一个特别烂的Python脚本。

And it was, like, literally this, like, terrible Python script.

Speaker 0

结构很松散,有时候回复根本没法用。

And it kind of was rickety and, like, sometimes the replies didn't work.

Speaker 0

我当时就想,算了。

And I was like, okay.

Speaker 0

但关键在于,软件公司做的远不止写代码这么简单。

But, like and like the larger narrative here is like software companies build more than just like code.

Speaker 0

对吧?

Right?

Speaker 0

我之所以愿意花钱用Gmail来管理邮件,肯定是有原因的。

Like, there's a reason that I pay for Gmail to like operate my email.

Speaker 0

结果发现,当SMTP版本得升级的时候,真的挺烦人的。

It's like, it turns out it kind of stinks when you're like, the SMTP version needs upgraded.

Speaker 0

我根本不在乎这些技术细节。

You're like, I don't care.

Speaker 0

我就是想正常收邮件而已。

Like, I just wanna receive email.

Speaker 0

因为我得自己运行这些个人助手,才真正体会到那种痛苦——我想要的产品根本不存在,只好自己动手做,结果还得长期承担维护成本。

And so, like, as I've had to run my own agents for my personal life, like, I've had to experience that pain of, like, the product I want doesn't exist and I built it, and now I get the ongoing cost of it.

Speaker 0

说实话,我现在买的软件比以前多多了,因为我想通了。

And, like, I'll be honest, I'm buying more software these days than I ever did before because I'm like, you know what?

Speaker 0

那个工具看着不错。

That that tool seems cool.

Speaker 0

不如直接花钱让别人帮我运行这些助手。

Like, I'm just gonna pay somebody else to run my agent for me.

Speaker 1

我完全同意。

I totally agree.

Speaker 1

作为一个写过不少随手搭工具的人,一方面,个人维护确实麻烦,另一方面,我也写过一些真正上线投产的工具。

As someone who has vibe coded my fair share of tools, a, yes, the personal maintenance, but b, I have vibe coded tools that we've we've released into production.

Speaker 1

让我告诉你,这根本不是简单地说‘修个bug’就能搞定的。

And let me tell you, it's not it's not as simple as saying, like, fix this bug.

Speaker 1

我觉得这一点在所谓的‘SaaS末日’讨论里,完全被忽略了。

And I do think that that's I do think that's saying that that is that is really missed in in the, quote, unquote, SaaSpocalypse discourse.

Speaker 1

不过我很好奇,如果你最早做的就是邮件助手,那你现在是怎么处理邮件的?我觉得现在技术已经发展到可以让你几乎不用管邮件,就能把邮件处理好了,这真的让我特别兴奋。

I gotta say, though, if you if one of the first things you did was an email agent, I'm super curious how you're doing your email right now because I feel like things just got to a point where you can, like, kinda just do your email without doing your email, and I'm so excited about it.

Speaker 0

是的。

Yeah.

Speaker 0

我确实不怎么管了。

I don't yeah.

Speaker 0

我可以多讲一点背后的故事。

I can tell a little bit more about the story.

Speaker 0

两年前开始的问题是,我在工作中用起了聊天机器人,因为那时候它已经成了主要的交互方式。

So, like, the problem that started two years ago is, I was using chatbots at work because, like, at that point, that was kind of like the primary interface.

Speaker 0

那时候,智能代理还没人用呢。

Like, agent usage was not really a thing yet.

Speaker 0

我自己的生活呢,孩子在三所不同的学校上学。

And so my my my personal life, like, I have kids in three schools.

Speaker 0

如果有家长在听,肯定懂那种收到PTO邮件时的状况,比如今天活动的主题是什么?

And if there are any parents out there listening, you know what it's like to get the PTO emails and the, like, what is the what's the theme for today?

Speaker 0

你一定也懂那种漏看消息的感觉——要是错过了精神日,孩子没穿搞怪发型的衣服,那种挫败感简直让人觉得自己当了个失败的父母,我可不止一次经历过。

And, like, you know the feeling of, like, a missed like, this is, the worst parent feeling in the world, but, like, if you miss the spirit day because your kid didn't do crazy hair day, like, you feel like you have failed at life, I will tell you that, and having done it more than once.

Speaker 0

所以我下定决心,不能再漏掉一次了。

And so I was like, I can't I can't miss another one.

Speaker 0

我就想,我手里有什么资源呢?

And I was like, you know what I have?

Speaker 0

那是因为我每天得盯住十五封邮件。

I have and it was because I had to track, you know, 15 emails a day.

Speaker 0

你以为企业里发的邮件已经够多了?

Like, you think we produce a lot of email in corporate America.

Speaker 0

你等到学校里那些家委会邮件就知道了。

Wait till you get to the PTO emails you get at school.

Speaker 0

于是我心想,诶,要不这样?

And so I was like, you know what?

Speaker 0

这么多邮件我根本看不过来,但AI代理可以啊。

I can't read all these, but who can Agents.

Speaker 0

我就想,为啥不能直接用市面上那些代理平台来处理呢?

And I was like, why can't I just do this in, you know, one of the agent many agent platforms out there?

Speaker 0

我当时觉得,缺的其实就是能连上邮箱这个功能。

And I was like, the missing part here is, like, I just wanna hook it up to email.

Speaker 0

所以最早的那个版本,就是直接抓邮箱里的最新邮件,扔给一个大模型,再把回复直接吐回来。

And so the very first first version of it was just literally, like, grab an email inbox, look for the top email, and literally paste it like, one of the LLMs and, like, dump the response back.

Speaker 0

我知道你们总爱聊提示词,那时候我最常用的提示就是:把邮件转过来,然后只提取关键信息。

And, I know you like to talk a lot about prompts, like, my favorite prompts in those days was basically, like, forward the email, and it would just be extract the facts.

Speaker 0

每次我发一封好几页的长邮件,结果只收到三条要点,我都觉得特别不可思议。

And it was always shocking to me that I would send, like, a multi page email and get, like, three bullet points back.

Speaker 1

确实如此。

That is yeah.

Speaker 1

我记得那些日子。

I I remember those days.

Speaker 1

那时候又要连线又要复制粘贴,感觉好像已经很久以前了,其实才一两年前而已。

Like, the the wiring up and the copy and pasting, and I feel like that's so far away, but it's only, like, a year or two ago.

Speaker 0

对。

Yeah.

Speaker 0

说到智能代理,我后来加了个记忆系统。

Well then, like, you know, to to your point about agents, so then I've added a memory system.

Speaker 0

对吧?

Right?

Speaker 0

因为你提到的,我还没完全自动化,也不太敢让它替我回邮件,但有了记忆系统,整个体验就彻底不一样了。

Because, like, to your point, I've, like, haven't fully automated and, like, not sure that I trust it to reply on my behalf, but, like, having the memory system was, a total unlock.

Speaker 0

现在我把它升级成这样:每天会有一个总结,这还挺有意思的。

And, like, now how I've evolved it is, like, I have a daily so, like, this is an interesting thing.

Speaker 0

比如Open Claw,我觉得它挺有意思,但真正让它爆发的,是那个主动推送的功能。

Like, Open Claw, feel, like, hit on this, but, like, the proactive part is I think the thing that really, like, set it on fire.

Speaker 0

我的做法是让代理每天固定时间,把所有内容整理成摘要,发一封邮件给我。

And my version of that was, I would have my agent, like, take a summary of all that stuff and send me an email every day at a certain time.

Speaker 0

对我来说,最大的突破就是不用再去工具里主动查询,它自己就出现了。

And, like, the unlock for me was, like, instead of having to go to a tool and ask for the thing, like, it was just like it would show up.

Speaker 0

当然,它也没多智能。

Now not that it was particularly smart.

Speaker 0

就是每天同一时间准时发一次。

It would just do it the same time every day.

Speaker 0

但我认为,代理的未来会更主动得多。

But, like, I think where agents are going, it's, like, much more proactive.

Speaker 0

再想想,是不是该主动联系负责人,告诉他们当前的情况?

And then, like, you know, thinking about, like, do I need to reach out and contact my owner and let them know what's going on?

Speaker 1

如果你几年前是那样操作的,那现在有哪些工作流程是你特别依赖、觉得特别棒的?

So if if that was the like, where you were a couple years ago, like, are what are the, you know, what are the workflow things that you have now that you rely on that you're excited about?

Speaker 1

所以你好像在 OpenCLOB 里用了一些简单的功能。

So it seems like there's some sort of brief functionality you're using in OpenCLOB.

Speaker 1

不过你最近在折腾哪些东西,是特别喜欢的?

But what are a couple of things that, you know, you've been tinkering with that you like?

Speaker 0

我觉得我现在在工作上正试着搞清楚的一件事,就是总结归纳。

I think one of the things I'm trying to figure out, like, in my work life is around summarization.

Speaker 0

我觉得工作的一部分,就是得理解海量信息,然后弄清楚怎么筛选这些信息,怎么让我的助手学会分辨哪些重要、哪些不重要。

Like, I think part of the job is, like, understanding an immense amount of information and then, like, figuring out, like, how do I filter that information and how do I imbue my agents with that skill of this is the thing that matters and this doesn't.

Speaker 0

但这个问题其实特别难,因为你读的东西里,第一眼觉得没用的,往往后面才发现有用。

And then it's actually a really hard problem because there's a lot of stuff that you read and that the first pass you're like, oh, that doesn't matter.

Speaker 0

过三天之后,它突然就变得关键了。

And then it will matter three days later.

Speaker 0

那你到底该怎么定义,哪些信息重要、哪些不重要呢?

And it's like, how do you describe, like, which things matter and which things don't?

Speaker 1

而且感觉助手就像这样,我有时候会让它帮我梳理所有的会议记录。

It also feels like the agents are a little bit like, you know, one of the things that I will have it do is go through all my meetings.

Speaker 1

公司里所有的会议我们都录下来存在Notion里,让Codex帮我逐个梳理,挑出我觉得可能感兴趣的内容,这真的很棒,这样我即使没参加也能掌握所有信息。

And like all the meetings in the company, we recorded all in notion so I can have Codex just like go through all the meetings, be like, here's all the stuff that you might be interested in, which is really cool because I can be in all these meetings that I'm not in.

Speaker 1

但问题是,如果它给我的东西不太对,我一指出不对,它反而过度调整,把我之前说想要的全都堆给我,还特别死板,结果就是怎么都不够准,总觉得哪里怪怪的。

But then it feels like if it gives me stuff that is not quite right and I tell it it's not quite right, it overcorrects and gives me all of the things that I said I wanted, but, like, way too much and way too literally, you know, and it's just like you're never it's never quite right somehow in in in this weird way.

Speaker 0

是的。

Yeah.

Speaker 0

我很好奇你现在在这方面的进展如何,因为我觉得这目前还是个没解决的难题。

I was curious to see where you're at on that because, like, I I feel like this is, one of the unsolved problems at this point.

Speaker 0

我觉得我们都在摸索,就像你提到的邮件和收件箱,你已经完全自动化了吗?

It's, like, I think we're all kind of grasping for, like like you said, like, any I guess you mentioned, like, with the email and the like, your inbox, like, you fully automated it?

Speaker 0

它会替你回邮件吗?

Like, does it reply on your behalf?

Speaker 0

还是说每封邮件你都要亲自审核?

Or, like, do you approve every reply?

Speaker 0

具体是怎么操作的?

Or, like, what does that look like?

Speaker 1

每封回信我都得亲自确认。

I have to approve every reply.

Speaker 1

不过我在Codex里弄了个小应用,通过Codex的内置浏览器打开,它是在本地运行的。

But, what I have what I do is in Codex, I have a little app that, like, I open in the Codex in app browser that and the app runs locally.

Speaker 1

每天它都会扫描我所有的邮件,然后列出一份清单,每封邮件旁边都附着一个草稿,写着它打算怎么回复。

And, basically, every day, it just sweeps through all my emails and then gives me a list of here's here every email is like on that page and there's a draft that it has said, here's here's what I'm gonna probably try to reply to.

Speaker 1

因为它能访问我电脑上的所有信息,比如收到律师的邮件,它就能去查相关资料,然后告诉我:‘我觉得你应该这么回。’

And because it has access to my computer and everything, like if it's an email from the lawyer my lawyers or whatever, it can like go and search and like be like, here's basically what I think I should say.

Speaker 1

然后我就逐条浏览,跟它对话调整。

And then I just scroll through and talk to it.

Speaker 1

我们有个叫Monologue的工具。

We have this tool called monologue.

Speaker 1

我就对着它自言自语,说‘改一下’或者‘对,就这样’。

I just like monologue into it and just say, no, go fix this or yes.

Speaker 1

你可以直接发送这个草稿,或者说‘不’,随便你。

You can send that draft or no or whatever.

Speaker 1

我已经连续四周保持收件箱清零了,这可是头一回,真不敢相信。

And I I've been at inbox zero for, like, four weeks straight, and that is, like, a huge I've never it's never happened before.

Speaker 1

我的助理都问我:这到底发生啥了?

My assistant is like, what the fuck is going on here?

Speaker 1

我

I

Speaker 0

我坦白说,我确实是收件箱清零主义的信徒。

will fully admit I am part of the religion of inbox zero.

Speaker 0

我这么坚持好多年了,真心信这套,但确实特别费劲。

So I've been running it for many years, and I believe in it, but it sure does take a lot of work.

Speaker 0

我对那个‘独白’工具挺好奇的。

I'd be curious about the monologue thing.

Speaker 0

你是直接对着它说话,还是打字跟它交流?

Do you actually talk to it, or do you type to it?

Speaker 0

对。

Yeah.

Speaker 0

它能处理视频吗,还是现在只支持音频?

Does it know video, or is it only just the audio?

Speaker 1

目前只支持音频。

It's only audio right now.

Speaker 0

明白了。

Yeah.

Speaker 0

这其实是大多数人不知道的一个小技巧。

That is one of again, hidden tips for, like, most people.

Speaker 0

音频这个功能真的特别强大。

Like, the audio unlock is huge.

Speaker 0

我后来发现,对着电脑说话确实有点怪。

And, like, one of the things that I've learned about it is, like, it's kind of weird to talk to your computer.

Speaker 0

所以我用的办法是多用Loom,假装在给谁做屏幕共享,这样说话就没那么别扭了,也能更自然地把问题讲清楚。

And so, like, my trick for that is, like, actually use Loom a lot because it feels less weird to pretend like I'm screen sharing to somebody, and that allows me to, like, actually talk through the problem.

Speaker 1

这太有意思了。

That's so funny.

Speaker 1

就是在办公室里吗?

Like, at in the office?

Speaker 0

是的。

Yeah.

Speaker 0

我觉得吧,我主要在家做,所以别人听不到我自言自语。

Like, I feel well, I guess I do it from home mostly, so I don't people don't hear me talking to myself.

Speaker 0

但就算在办公室,我觉得别人也不会多想,顶多以为你在开Zoom会议。

But, like, even in the office, I feel like people aren't like, they'll just think you're on a Zoom or something.

Speaker 1

对。

Yeah.

Speaker 1

对。

Yeah.

Speaker 1

对。

Yeah.

Speaker 1

以前确实存在这么一道心理障碍。

There there's always there was this, like, barrier at some point.

Speaker 1

现在办公室里所有人都不跟我说话了。

Now now everyone in the office, I just they're not talking to me.

Speaker 1

我猜他们都是在跟电脑说话。

I assume they're talking to their computer.

Speaker 1

所以当他们突然跟我讲话时,感觉特别怪。

So it's, like, weird when they're talking to me.

Speaker 0

我们生活的这个世界啊。

What a world we live in.

Speaker 0

真是个奇怪的世界。

What a world.

Speaker 0

我懂。

I know.

Speaker 1

但我觉得这已经变成一种社交默契了,你能看出来他们是不是在盯着电脑。

But I think that it's it became this this, like, social thing where you kind of like you can see, are they looking at their computer?

Speaker 1

还是说他们抬头看了?

Are they looking up?

Speaker 1

或者呢,他们还会压低声音,凑近电脑,小声说:你帮我做一下这个小事儿。

Or, like, there's also, like, the whisper, you know, just like they're just, like, getting kinda close to their computer, and they're like, I I want you to, like, do this little thing.

Speaker 1

而且呢

And and

Speaker 0

然后你就回:好。

and you're like, okay.

Speaker 1

我明白这背后是怎么回事了。

I know what's going on here.

Speaker 0

但说话的速度快了将近一倍。

But it's, like, twice as fast.

Speaker 0

我一时想不起具体叫什么了。

I forget what it is.

Speaker 0

我不太确定。

I don't know.

Speaker 0

说话的速度大概是打字的两到三倍。

It's, like, roughly twice or three times as fast to talk than type.

Speaker 1

而且我有腕管综合征之类的毛病,用语音能减轻手部负担。

And, like, I've got, you know, carpal tunnel and shit, so, it it doesn't aggravate my hands as much.

Speaker 1

这样更符合人体工学。

It's much more ergonomic.

Speaker 1

所以确实如此。

So yeah.

Speaker 1

语音输入对我来说简直是巨大的突破。

I huge huge unlock to use voice.

Speaker 1

不过我想再回到我们之前聊的那件事上,就是……

I do wanna get back though to the the original thing that we were talking about, which is okay.

Speaker 1

我觉得咱们的想法是一致的。

Let's just let's just say, I think we're we're on the same page.

Speaker 1

所谓的‘SaaS末日’根本不存在。

SaaSpocalypse, not a not a thing.

Speaker 1

真正做出一个能稳定运行的SaaS软件,是件特别耗力的事,只有少数人愿意干,大多数人只想直接花钱买现成的。

Like actually making a piece of SaaS software that works all the time is a gigantic effort that only some some people want to do and other people just want to pay for.

Speaker 1

But then let's like dive more, more distinctly down into Figma land.

But then let's like dive more, more distinctly down into Figma land.

Speaker 1

So, you know, there are questions about.

So, you know, there are questions about.

Speaker 1

In a in a design world.

In a in a design world.

Speaker 1

What kind of a do I want to just chat with my landing page and move things around that way?

What kind of a do I want to just chat with my landing page and move things around that way?

Speaker 1

Or do I want to be on the infinite canvas?

Or do I want to be on the infinite canvas?

Speaker 1

And I know internally, like pretty much all of our designers, they're super, I pulled super early adopters and they're all like.

And I know internally, like pretty much all of our designers, they're super, I pulled super early adopters and they're all like.

Speaker 1

Yeah, like the, you know, typing is good for a first pass, but like to actually get the details right, I need to be able to like move stuff around.

Yeah, like the, you know, typing is good for a first pass, but like to actually get the details right, I need to be able to like move stuff around.

Speaker 1

So in in in the design world, like how does that change?

So in in in the design world, like how does that change?

Speaker 1

当设计的可能性发生如此根本性的变化时,你认为这会对产品策略产生什么影响?

How do you think that changes the product strategy when the the possibilities for how you might design something have changed so radically?

Speaker 0

是的。

Yeah.

Speaker 0

我的意思是,这背后涉及的内容太多了,我们现在还处在早期阶段,正在慢慢摸索。

I mean, I think it's a lot to unpack and I think, you know, we're in the early innings here and we're kind of figuring out.

Speaker 0

我觉得我们还陷在文本框思维的惯性里。

I think we're still in kind of this hangover of like the text box rules.

Speaker 0

很多人下意识地把聊天当成生成式界面的默认体验。

Like, I think so much of us are defaulting to like chat as the experience for generative UI.

Speaker 0

但我觉得,我们正进入第二阶段——真正去思考这意味着什么,这也是我对我们即将推出的智能代理功能如此兴奋的原因。

And I feel like we're starting to enter the second chapter of that of like what does it mean and like that's part of the reason I'm so excited about our agents launch.

Speaker 0

我们内部其实已经用了一段时间了。

Like, we've had it internally for a while.

Speaker 0

还没见过的人注意了,这个功能允许你直接在无限画布上使用智能代理。

And for those who haven't seen it yet, it's the ability to use an agent directly on the Infinite Canvas.

Speaker 0

回头想想,挺有意思的是,旧的东西又重新流行起来了。

And I think going back, it's funny about, like, what's old is new again.

Speaker 0

在大模型和机器学习领域,这种现象挺多的,比如我们重新发明了评估方法。

Like, there's a lot of, like, this happening throughout LLM and ML land of, like, we've reinvented eval.

Speaker 0

所以说,是这么个情况。

So it's like, okay.

Speaker 0

其实我们以前就有单元测试了。

Well, we had unit tests before.

Speaker 0

我们又重新发明了提示词的设计。

And it's like, well, we've reinvented, like, prompting.

Speaker 0

以前就有用户输入了,现在又来一遍,嗯,情况就是这样。

And we, like, we had user input before, and it's like, okay.

Speaker 0

在新的AI时代,设计的核心原则依然重要。

And design in the new, like, AI era is still, like the the principles still matter.

Speaker 0

对吧?

Right?

Speaker 0

对我而言,设计的一个核心原则就是‘发散—收敛’模式。

And so, like, one of the principal core principles of design for me is, like, the the die the diamond.

Speaker 0

这背后其实讲的是发散思维和收敛思维的结合。

And so, like, there's this idea of, like, divergent thinking and then convergent thinking.

Speaker 0

很多设计问题都是这样,这也是头脑风暴的精髓——千万别一上来就否定别人的点子。

And, like, most design problems are, like and, like, this is the idea behind brainstorms and when people tell you, like, don't ever, you know, shoot somebody's idea down.

Speaker 0

头脑风暴的关键,就是先把想法全甩出来。

Like, brainstorming is all about just, like, generating ideas.

Speaker 0

我觉得我们还没完全发挥出这些新技术的潜力,尤其是它们能极大激发创造性思维的能力。

And one of the things that I don't think we fully unlocked yet from these new capabilities is, like, the ability to, like, supercharge generative thinking.

Speaker 0

我们常常被自己的经验和固有思维困住,看问题总绕不开那个角度。

I think oftentimes we get stuck in our own stories and lived experience and we approach a problem from a certain angle.

Speaker 0

所以团队合作才这么重要。

And this is what's so valuable about having teammates.

Speaker 0

对吧?

Right?

Speaker 0

你去跟同事聊,发现人家的出发点完全不一样,结果问题的答案也跟着变了。

It's like you go to talk to your teammate and they have a totally different starting point and it's like the answer to this problem is different.

Speaker 0

真正的创意往往就发生在你们俩对话的时候,比如:‘哎,我怎么就没从这个角度想过呢?’

And, like, the creativity comes between the conversation between the two of you of like, oh, like, hadn't thought about it from that angle.

Speaker 0

这挺有意思的。

That's interesting.

Speaker 0

让我拿这个想法再往下延伸一下。

Let me take it and then build on top of that.

Speaker 0

对吧?

Right?

Speaker 0

那回到现在这个AI时代,这到底意味着什么?

And so, like, going back to the, like, what does this mean in this new AI world?

Speaker 0

我觉得文字框太局限了,总是一板一眼,先这样、再那样、然后又这样。

If we get outside of the text boxes because I think text boxes are super limiting and it's very much like a linear, like, well, this and then that and then this.

Speaker 0

要是能跳到画布上,保留那些核心思路,但借助智能代理实现更发散的思考,那就不同了,比如:嘿。

If we get to the canvas and you have the ability to, like, have some of those same kind of concepts, but the agents allow you to do different gen thinking that it's like, hey.

Speaker 0

我有这么一个框架。

Like, I have this frame.

Speaker 0

咱们就从这儿开始吧。

Let start here.

Speaker 0

嘿,是这样。

It's like, hey.

Speaker 0

我觉得应该用灰度。

I think it should be grayscale.

Speaker 0

然后你再换一个框架。

Then you have another frame.

Speaker 0

你可能会想,要不试试SIPA?

You're like, oh, let me try SIPA.

Speaker 0

结果发现SIPA倒是不错,就是字体不太对。

And then it's like, oh, the SIPA thing is cool, but, like, the type is wrong.

Speaker 0

于是你又想,要不试试这个?

And it's like, oh, let me try that.

Speaker 0

现在无障碍功能好像被关闭了。

Now it's like the accessibility is off.

Speaker 0

我来复制一下这个框架,再试一次。

Let me, like, duplicate the frame and try again.

Speaker 0

我觉得,就连这个也还只是刚开始阶段。

And so, like, I think and even that is still, like, I think early innings.

Speaker 0

这完全是人在主导输入过程。

That's, like, very much, like, the human driving the input.

Speaker 0

对吧?

Right?

Speaker 0

你之前好像提到过那种主动推进的流程。

And, like, you kind of have talked about the, like, proactive flow.

Speaker 0

对吧?

Right?

Speaker 0

我觉得我们正慢慢弄清楚,要是有个智能代理,能直接处理我放在画布上的一堆框架,会怎么样?

And it's, I think we're starting to figure out, like, what if we had an agent that's, like, here's a bunch of frames that I threw on the canvas.

Speaker 0

你的任务是引导他们尝试不同的方向,别一味地固执己见。

Like, your job is to, like, push them, like try different directions or don't just, double down.

Speaker 0

但我觉得还有一类代理,比如现在画布上有25个关于新营销页面的概念草图。

But then I think there's a separate set of agents that it's like, okay, we have 25 frames on this canvas of, like, concepts for, like, a new marketing page.

Speaker 0

对吧?

Right?

Speaker 0

那接下来该怎么把它们聚焦起来呢?

And then it's like, how do I channel them down?

Speaker 0

这时候就需要一个收敛型代理,它会说:这三个方向有共性,那几个又集中在另一个思路下。

So like, then there's like a convergent agent that's like, okay, these three are kind of like this and these are clustered around this.

Speaker 0

而且你可以问它怎么看。

And like, I think And you can ask it for its opinion.

Speaker 0

你可以让它假装是用户,点击浏览这些方案,然后说:哪一个最合理?

You're like, pretend you're like a customer clicking through this, like which one makes the most sense, right?

Speaker 0

我觉得这些潜力我们还没真正挖掘出来。

And so I don't think we've really tapped all of that stuff yet.

Speaker 0

所以我觉得,就算是最好的代理工具,比如命令行那种,也没法真正完成这些工作流程。

And so like, I think even the best agents, like the command line agents, like don't have the ability to like do those workflows.

Speaker 0

我觉得,这就是未来设计和产品思维的方向。

And so that's kind of where I see the future of design and product thinking.

Speaker 1

这我完全同意。

I think that makes total sense.

Speaker 1

而且根据我目前看到的情况,代理工具在已经有设计系统的情况下特别有用。

And, yeah, it seems like from what I can tell so far, the agents are really good for I already have a design system.

Speaker 1

我需要一个新首页。

I need a new landing page.

Speaker 1

就是那种‘帮我基于我现有的设计系统生成一个首页’之类的需求,说实话,很多设计师并不想花时间反复做第N个首页,或者第N张海报、图示之类的,这类任务更偏向收敛型,不太需要发散创意。

Like, get me the get me a landing page in the in the design system I already have kind of thing, which, to be honest, a lot of designers don't really wanna have to spend time doing the the nth landing page or the nth, like, graphic for this, you know, post or whatever, which is more convergent and a little bit less divergent.

Speaker 1

那关于未来,比如Figma这样的设计工具,或者一般软件,是该允许外部代理接入,还是该自己开发代理,或者两者并用——你们现在其实都同时在用,你觉得这该怎么平衡?

What about what about, like, the future of, you know, maybe Figma design tools or just general generally software with allowing external agents in versus building your own agent or having both, which you which you all do have, how does how do you think that works?

Speaker 0

我觉得,我们两种方式都会采用。

I mean, I think we embrace both.

Speaker 0

对吧?

Right?

Speaker 0

我觉得设计流程和工程流程本来就不一样,但现在这两者之间的界限越来越模糊了。

Like, and I think this is like I think design workflows are different than engineering workflows, but, like, the lines are blurring.

Speaker 0

所以我觉得未来大家都会变成创造者。

And so, like, I think in the future we're gonna be all builders.

Speaker 0

对吧?

Right?

Speaker 0

这其实取决于你从哪个角度去看这个问题。

And that kind of comes like which perspective are you coming at the problem from?

Speaker 0

我们目前非常支持第三方代理工具。

And so we definitely very much support third party agents today.

Speaker 0

我们对此的解决方案就是我们的MCP服务器。

And like our answer for that is our MCP server.

Speaker 0

对吧?

Right?

Speaker 0

所以我觉得MCP的一个好处是,它为各种不同的工具提供了一个标准化的接口。

And so I think one of the nice things about MCP is like it allows like a standardized interface across all these different kinds of tools.

Speaker 0

对吧?

Right?

Speaker 0

所以我们通常从两个方向来思考这个问题。

And so we kind of think about the problem in two directions.

Speaker 0

一个是代码到设计的方向。

We think about it as like code to design.

Speaker 0

就像你刚才说的那种情况,其实挺常见的。

So it's like, okay, in that scenario you just said like, that's like a pretty common thing.

Speaker 0

比如你有个注册页面,但没支持GDPR。

You're like, hey, have a signup page, but like, it doesn't support GDPR.

Speaker 0

对吧?

Right?

Speaker 0

大多数人其实并不清楚自己该怎么做。

Like, most people are not gonna be like, know what I should do?

Speaker 0

我先从一个全新的页面开始,重新设计我们的注册流程,加入GDPR要求。

I'm gonna start with a greenfield page and reimagine what our sign up flow should be to add GDPR.

Speaker 0

大多数人只是想把工作做完。

Most people are trying to get their job.

Speaker 0

你周一早上登录系统,心里就想:好了,赶紧把这事搞定。

It's like you log in on Monday morning, you're like, okay, I just gotta get this thing done.

Speaker 0

对吧?

Right?

Speaker 0

然后就在那儿加个勾选框。

And like, me add the checkbox here.

Speaker 0

对吧?

Right?

Speaker 0

所以对于这种工作流程,如果你熟悉Codex、Claude、Windsurf或者Cursor,就可以打开你的代码库,启动MCP服务器,然后说:帮我打开这个页面,启动开发服务器,把页面内容复制到Figma画布上。

And so for that workflow, it's like if you are comfortable in Codex or Claude or Windsurf or Cursor, you pull up your code base, you fire up the MCP server and you ask it like, hey, can you copy go to this page like, fire up the dev server, go to this page, and copy it into Figma Canvas.

Speaker 0

它真的能帮你完成。

And it will actually do it.

Speaker 0

这其实是今年早些时候我们推出的一个功能,简直让人惊讶,AI代理居然能这么准确地完成这种任务,但事实就是它们能做到。

Like, that's one of the releases we had earlier this year, which is like a little bit mind blowing that agents can do this faithfully, but it turns out they can.

Speaker 0

现在,你把这些繁琐的活儿全给省了。

And now you have just removed all that drudgery.

Speaker 0

对吧?

Right?

Speaker 0

而且你还把它放到了一个可以真正交互的环境中。

That and you've got it into, like, a medium where you can actually interact with it.

Speaker 0

就好比说,

It's like, okay.

Speaker 0

让我用那些大家平时都熟悉的直接操作工具,精确地调整各个元素的位置。

Let me go, like, move things around very precisely with the direct manipulation tools that most people are comfortable with.

Speaker 0

另外一种我们常考虑的流程是,既然设计已经做好了,那就把这个设计重新转回代码。

And then the other kind of workflow that we think about is like, okay, now that we've got the design, take that design and bring it back to code.

Speaker 0

对吧?

Right?

Speaker 0

所以我们开发了一个叫 Git 设计上下文的工具,它能把 Figma 设计稿里的所有属性、组件,还有你在设计库里设定的规范,全都打包好交给智能代理。

And so we've got a tool called Git design context, which takes a Figma design, wraps up all the different properties and, like, components that you're using as well as, like, any other kind of guidelines you've provided in your design library and provides it to the agent.

Speaker 0

而且,这简直就像魔法一样。

And, again, it's kind of like magic.

Speaker 0

代理会说:好的,明白了。

It's like the agent will be like, okay, cool.

Speaker 0

让我看看你现在的代码库。

Let me, like, look at your current code base.

Speaker 0

我来建个分支,提个合并请求,做修改,然后你只要问代理一声:好了。

I'll make a branch, create a PR, make the changes, and then, like, you ask the agent and be like, okay.

Speaker 0

截个图,贴到合并请求里。

Take me a screenshot and put on the PR.

Speaker 0

这时候你的工作,就跟之前你提到的邮件流程差不多。

And then, like, your job is, like, kinda like what you're talking about your workflow earlier with email.

Speaker 0

你不用直接合并,但你会说:行,知道了。

It's like, you don't merge it, but you're like, okay.

Speaker 0

嗯,我现在有个不错的起点。

Well, I've got a a good starting spot.

Speaker 0

对吧?

Right?

Speaker 0

然后你就可以进去自由发挥。

And then it's like you can come in there and then riff.

Speaker 1

在开发内部智能代理时,你学到了哪些以前在Figma代理之前不知道的、能让用户体验更好的关键点?

What do you what have you learned about what makes for a good internal agent experience internal to a product that you may not have known before before Figma agent?

Speaker 0

我不知道。

I don't know.

Speaker 0

这倒是个有意思的问题。

Like, that's like a interesting question.

Speaker 0

什么样的产品才算好用?

What would make it a good product?

Speaker 0

我觉得这特别针对Figma,但上下文和个性化真的很重要。以前我在其他AI公司做产品时,个性化常常是最后才考虑的事。

I think specific this is like very specific to Figma, but like the context and personalization matters, like in other products like I've worked on in the past for like AI companies is like personalization is often like the kind of the last thing.

Speaker 0

首先得让所有人都能用起来。

Like you get it just working for everybody first.

Speaker 0

但我认为,真正让普通助手变成让人爱不释手的,是个性化这个部分。

But I think the thing that really differentiates like an okay agent to one that people really love is the personalization aspect.

Speaker 0

就像我们之前说的,记忆是一种方式,还有那些第三方聊天助手。

Like we talked about memory, like as a form of it and these like third party chat agents.

Speaker 0

我觉得在Figma里,这个个性化体现就是设计系统。

I think for Figma's version of that, it's like the design system.

Speaker 0

如果你有个助手,却没把‘我们是怎么搭建设计、怎么组合元素的’这套规则教给它,那它做的东西根本没法用。

That it's like if you have assistant but without, like, the concept of, well, this is, like, how we structure our designs and, like, how we put them together.

Speaker 0

它生成的设计,从这个角度看根本派不上用场。

Like, the designs that it create just aren't usable from that perspective.

Speaker 1

我不太清楚你们对Figma和那种主动型助手有什么规划。

I don't know, like, what your plans are for Figma and and, you know, the, like, proactive being, like, being a proactive agent.

Speaker 1

但如果你真有这些计划或实验,我很想知道现在进展怎么样。

But I'm curious to the extent you have those plans and those experiments, how that's working?

Speaker 1

很明显,我们之前也提到过,这件事要做好确实不容易。

Obviously, like we've talked about that being kinda hard to get right.

Speaker 0

是的。

Yeah.

Speaker 0

我觉得未来肯定会朝这个方向发展。

I mean, I think that's where the future is going.

Speaker 0

你看一下智能代理这些年来是怎么演进的。

Like, if you look at, like, how, like, agents have kind of evolved.

Speaker 0

我们内部现在正在推进很多项目。

I will say we've got a lot of things cooking internally.

Speaker 0

不过具体细节我暂时不方便多说。

I don't know I can't talk about specifics too much.

Speaker 0

我确信我们在这一块是走在前面的。

I definitely think this is an area we run ahead.

Speaker 0

对吧?

Right?

展开剩余字幕(还有 196 条)
Speaker 0

所以,我觉得我可以谈谈我们现在遇到的一些问题。

And so, like, I think I can talk about the problems that we see today.

Speaker 0

如果全球的软件数量正在爆炸式增长,那么其中一个更大的挑战就是:如何确保这些软件始终符合我们的价值观?

It's like if the amount of software is really exploding in the world, one of the bigger challenges then becomes like, okay, well, how do you make sure that it's consistent with your values?

Speaker 0

于是,我们就成了瓶颈。

And so I think we become the bottleneck then.

Speaker 0

毕竟,能审查这么多工作的真人眼睛是有限的。

It's like we only have so many human eyes to review all of this work.

Speaker 0

所以我们得想办法提供一个解决方案,让人既能跟上智能体快速创造的步伐继续创新,又不丢掉自己的核心价值观。

And so it's like, how do we provide a solution that allows people to make sure that they can continue to innovate at the speed that these agents create, but also maintains their values.

Speaker 1

在Figma内部,工程、产品和设计团队的工作流程,从AI到来之前到现在,经历了怎样的转变?

What has been the transition like internally in Figma in terms of your own workflows in the engineering org, in the product org, in the design org, from a pre AI world to now?

Speaker 0

我是一月份加入的,就算在这短短一段时间里,变化也可以说是天差地别。

So I joined in January and I would say even in that period of time, it's been like kind of night and day.

Speaker 0

我觉得一月份的时候,大家还在各个部门里摸索这些新的工作方式。

Like, I think folks in January were kind of experimenting with these new ways of working and, like, across all the functions.

Speaker 0

对吧?

Right?

Speaker 0

我觉得工程团队通常是走在最前面的,这种情况在大多数公司都一样。

Like, I think probably engineering was leading the way as they do in, I think, most of these cases.

Speaker 0

我给你举个产品团队的例子吧。

But I'll give you an example on the product org.

Speaker 0

我们之前开过一次封闭会议。

So, like, we had an off-site.

Speaker 0

我想你可能当时来过。

I think you might have actually come by.

Speaker 0

是的。

Yeah.

Speaker 0

这么说也不为过。

Fairly enough.

Speaker 0

世界真小。

It's like a small world.

Speaker 0

我印象最深的一次外出活动,是产品运营团队做了一个叫PMOS的东西。

I think one of my favorite memories of that off-site is that we have a product operations team and they had put together what we call PMOS.

Speaker 0

咱们先退一步说。

And to, like, take a step back for a minute.

Speaker 0

对我来说,AI最让人豁然开朗的一点就是,你会发现所有问题本质上都是上下文问题。

Like, one of the one of the, like, unlocks to me about AI is, like, you kind of realize every problem becomes a context problem.

Speaker 0

接下来的工作,就变成怎么用合适的信息把问题界定清楚。

And it's all like, then the work becomes about framing the problem with the right set of information.

Speaker 0

我们产品运营团队做的很多事,其实就是作为产品经理去处理结构化数据。

And so our product operations team is like, Like, a lot of the work that we do is like in structured data as a PM.

Speaker 0

那为什么不把这些东西都整合到一起呢?

And it's like, why don't we like aggregate that all together?

Speaker 0

于是我们就想,先把组织架构图拿过来吧。

And it's like, okay, let's get a copy of the org chart.

Speaker 0

然后把它放进SQLite表里,存到文件系统中。

We'll throw that in the SQLite table and put it in the file system.

Speaker 0

然后我们就想,不如做个和Asana的连接器,再连上Slack,还有GitHub,以及其他几个系统。

And then it's like, okay, why don't we create like a connector to Asana and then we'll connect Slack and then we'll connect, know, GitHub and a few other things.

Speaker 0

对吧?

Right?

Speaker 0

真正关键的发现是,那时候技能功能已经迅速发展起来了。

And then the real insight was like skills had really taken off at this point.

Speaker 0

没错。

It's like, okay.

Speaker 0

这里最厉害的地方,是我特别喜欢的一个技能——创建入职文件。

The magic sauce here is like one of the skills that I really like is like this onboarding file creation.

Speaker 0

比如你作为经理,新招了人,就得专门整理一份资料:该关注哪些频道,该认识哪些人。

So like when you add a new team member to your team, like as a manager, you gotta create a customized, like, okay, here's the channels you should know and here's the people you should know.

Speaker 0

以前我觉得这些信息全在我脑子里,根本没法系统化。

And it takes a lot of that knowledge that I would have previously said was entirely in my head.

Speaker 0

但后来你仔细一看,只要把上下文组织好,其实这些数据早就存在了。

But then you start to look at how if you shape the context rate, that data was actually already there.

Speaker 0

我早就有了组织架构图,它能根据架构图自动识别出这个新团队的成员,以及负责产品、工程和设计的三位核心人员。

I already had the org chart and it can walk the org chart and figure out who's the team and who's the trifecta on the product engineering design side for this new team.

Speaker 0

你只需要指出新来的是谁,他要加入哪个团队,剩下的就交给我了。

You just have to identify, here's the new person, here's the team they're gonna join, and it's like, okay.

Speaker 0

它还能自己做一堆调研。

And it can do a bunch of research.

Speaker 0

然后它会进入Slack,找出相关的频道。

And then it goes into Slack and it figures out the channels.

Speaker 0

它会说,这个团队大概就是这三个频道。

It's like, oh, this team is probably these three channels.

Speaker 0

接着它会阅读过去三十天的所有聊天内容。

And then it reads the last thirty days of content.

Speaker 0

之后它再去查看Asana看板,把所有项目都找出来。

And then it goes and checks the Asana board and finds all the projects.

Speaker 0

最后它把结果呈现在你面前,你就会觉得,这简直准得离谱。

And so it comes back and you're like, oh, this is, like, uncannily good.

Speaker 0

这已经是个挺不错的起点了。

It's like, yeah, this is a pretty good starting spot.

Speaker 1

我觉得有一件事特别有意思,Claude 代码之所以这么厉害,还有现在 Codec 之所以这么出色,是因为大家一开始都想着搞那些在云端运行、一直在线的智能体,结果还得手动把它们连到各种系统上。

That's one of the things that I think it's weirdly the it's weirdly I think the thing that made Claude code so good and is is the thing that makes Codec so good right now is everyone tried to start with agents that, like, live in the cloud or are always on, but then you have to, like, manually connect them to everything.

Speaker 1

但 Cloud Code 完全不一样,它就是你电脑上的一个智能体,能访问你所有能访问的东西。

And Cloud Code is just, yeah, no, it's just an agent on your computer has access to everything that you have access to.

Speaker 1

这样一来,它能做的事情就彻底不一样了,因为它能获取到所有需要的上下文。

And then that just totally changes all the stuff it can do because it can get all the context it needs.

Speaker 1

Codec 也是同样的道理。

And same thing with Codecs.

Speaker 1

我可以随便问 Codex 一个问题。

I can just ask a random question to Codex.

Speaker 1

我们今天刚发了一篇文章,我就在想,这该发给谁呢?

Like, we just published an article today and I was like, who should I send this to?

Speaker 1

它直接翻了我的邮件和短信就找到了答案。

And it just went through my emails and my texts.

Speaker 1

我根本不知道自己居然能访问这些内容。

And like, I didn't even know I had access to any of that.

Speaker 1

它直接找到了五个我可能早就忘掉、但其实应该发给的人,我就发了。

It just found like five people that I probably would have forgotten about, but that I should have sent it to, and I did.

Speaker 1

这种神奇的体验现在才开始出现。其实,只要给AI足够的上下文,它早就该能做到这些了,只是之前一直没找到合适的载体和使用方式。现在终于对了,它能更独立地帮我们做事了。

And that's the sort of magical thing that's starting to happen now that was very it was I think the technology, like AI itself, if you gave it all the context, would have been able to do this for a while, but it's only now that it it can it it's in the right harness and form factor, and it can do it a little bit more independently than it was able to before.

Speaker 0

对。

Yeah.

Speaker 0

我想说个请求。

I'm gonna I'm gonna put a plea out there.

Speaker 0

我记得WWDC是在2024年左右吧。

Like, I remember WWDC, it was at, like, '24.

Speaker 0

我说的是Apple Intelligence发布那会儿。

I can't remember when it was, like, Apple Intelligence.

Speaker 0

就是WWDC的主标题那会儿,

Whenever the headline WWDC,

Speaker 1

当时智能功能是啥样。

like, intelligence was.

Speaker 1

我觉得那应该是最近的事。

I think that was more recent.

Speaker 1

我觉得是2025年。

I think it was '25.

Speaker 0

真是2025年吗?

Was it '25?

Speaker 0

天哪。

Gosh.

Speaker 0

又是AI的时间了。

Again, AI time.

Speaker 0

谁知道现在是哪一年了?

Who knows what year it is?

Speaker 0

我升级的时候,可是全身心投入了。

I upgraded my I was, like, all in.

Speaker 0

我刚升级了我的iPad。

I, like, upgraded my iPad.

Speaker 0

我当时想,既然上一轮刚换过,现在又说需要新处理器,那好吧,换就换吧。

Was like, I'm gonna get because, like, I had just upgraded the last cycle, and then they're like, well, you need a new processor to I was like, okay.

Speaker 0

我把所有设备都升级了。

Upped all my stuff.

Speaker 0

我看了WWDC发布会。

Saw WWDC.

Speaker 0

我当时就觉得这回真不错,因为他们提到手机里存了这么多个人数据。

I was like, this is great because, like, they had this concept of, like, our phones have all of this personal data.

Speaker 0

我心想,天哪。

And I was like, oh my gosh.

Speaker 0

我觉得这次终于要来了。

I was like, this is gonna be it.

Speaker 0

不过说实话,我还是希望今年的WWDC能真正带来突破,毕竟去年根本没达到预期。

And then, like, I will say, like, I'm really hoping WWDC this year is actually it because, like, you know, this past year and whatever has not been it.

Speaker 0

对吧?

Right?

Speaker 0

毕竟,你说得对,技术早就有了。

Because, like, to your point, like, the technology's been there.

Speaker 0

缺的就是他们能把这些全都整合起来,手机生态里其实早就有了这些内容。

The part that's missing is they'll, like, tie it all together and, like, the mobile phone ecosystem, like, has all that content.

Speaker 0

所以我一直在等那个真正聪明的后台常驻Siri,不是那种老是听不懂人话的。

And so, like, I'm waiting for the thing where it's, like, the the always on Siri that runs in the background is actually smart rather than the one that's, like, what was that?

Speaker 0

我没听懂你说的,马特。

I didn't understand you, Matt.

Speaker 0

总有一天会的。

So one day.

Speaker 1

你觉得他们能做得到吗?

Do you think they're gonna do you think they're gonna be able to get that right?

Speaker 1

如果做不好,你觉得有关系吗?

And if they don't, do you think it matters?

Speaker 0

我觉得这仍然很重要,因为即使他们进场晚了,他们在上下文理解上还是老大。

I think it still matters because I think even them being late to the game, they are still the king of context.

Speaker 0

对吧?

Right?

Speaker 0

今年的谷歌I/O也挺有意思,看起来谷歌也终于意识到这一点了——他们可能没有苹果那么多数据,但体量也不小。

And I think that's what's been interesting to watch about like Google IO too this year is like seemingly Google has also like kind of woken up to that, that like they don't maybe have as much data as Apple, but like they have a lot.

Speaker 0

而且他们现在似乎开始把自家的AI产品整合起来了。

And it seems like they are now starting to marry their AI products.

Speaker 0

当然,前提是你能分得清这些名字。

If you can keep all the names straight.

Speaker 0

我看到一条特别有意思的推文,说到底是Bard、Gemini Pro、Spark,还是别的什么?我觉得Spark才是那个要成为全天候智能代理、自动连通你所有谷歌内容的产品。

I saw this great tweet that was like, is it Bard or Gemini Pro or Spark or like Whatever the thing is, there's some I think Spark is the product that is supposedly gonna be the always on agent that is auto connected to all of your Google content.

Speaker 0

所以我等着有一天,它能帮我处理邮件,让我彻底清空收件箱。

And so I'm waiting for the day that it, like, runs my inbox for me and I get to inbox zero.

Speaker 1

是的。

Yeah.

Speaker 1

我就是觉得苹果有点特别,你知道吗?当初OpenAI火起来的时候,大家都开始买Mac mini,这生意简直太聪明了。

I think I I just have this feeling about Apple that, you know, like, when Open Claw took off, everyone started buying Mac minis, and you're just like, that's such a good business.

Speaker 1

他们根本不用参与AI竞赛,因为大家用AI的时候,天然就用他们的设备,他们已经赢了。

Like, they didn't have they don't have to even be in the AI race because they win by default because it's the thing that everyone runs the AI on.

Speaker 1

就算他们在苹果智能上落后了——他们确实落后了,而且一直以来,他们的软件都比硬件慢半拍。

And and and so even if they're behind on Apple intelligence, which which they are and historically, like, their software products have been kind of lagging behind their their hardware.

Speaker 1

但他们的硬件太出色了,这点落后根本无所谓。

Because their hardware is so good, it doesn't matter.

Speaker 1

他们有的是时间追上来。

They have a lot of time to catch up.

Speaker 0

没错。

Yeah.

Speaker 0

而且我觉得他们的策略特别明智。

Well, and I think their strategy is very smart.

Speaker 0

对吧?

Right?

Speaker 0

这其实是在打隐私这张牌。

It's like at the privacy play.

Speaker 0

对吧?

Right?

Speaker 0

我觉得把所有个人信息都上传到云端,确实让人有点担心。

Like and I think like, it is scary to upload all of your information to, like, the cloud.

Speaker 0

所以我相信他们确实参与其中,我也真的很期待今年他们能推出一些有意思的东西。

And so, like, I do I do think they're in the game, and, I really am hoping that they've got something interesting this year.

Speaker 1

如果我们回看过去这一年,就会发现,我们在开发方式、工具能力,还有软件构建的整体流程上,都发生了巨大的变化。

So if we if we sort of look back over the last year, there's been this, like, big sea change in how we build stuff, how good the tools are, and then maybe correspondingly, like, software works and how we build software and and all that kind of stuff.

Speaker 1

随着这些能力不断提升——不管是做东西的方式,还是做出来的东西——你明年有什么打算或预期?

What do you expect or what what are you planning for over the next year as the capabilities increase both in how you make stuff and what you make?

Speaker 0

我觉得今年最关键的问题是:我们该怎么更好地做评审?

I think the big thing this year will be about the, like, how do we, like, review better?

Speaker 0

我觉得现在瓶颈就在这儿。

Like, I think that's where the bottleneck is now.

Speaker 0

对吧?

Right?

Speaker 0

我觉得现在已经有代理能生成这么多内容了。

That it's like, I think we have agents that are capable of producing all this stuff.

Speaker 0

它们已经足够普及了。

They're available enough.

Speaker 0

价格也够便宜了。

They're cheap enough.

Speaker 0

现在我们简直被各种新内容淹没了。

And now, like, we're just being inundated with, like, all of this new content.

Speaker 0

我觉得大家都有点不知所措了,这么多东西到底该怎么处理?

And, like, I think people are getting kind of overloaded of, well, what do I do with all of this?

Speaker 0

不只是那些内容摘要了。

It's not just, like, summaries of, like, stuff.

Speaker 0

那些早就有了,但现在不一样了。

Like, that's been out for a while, but now it's, like, okay.

Speaker 0

这是全新的内容,它甚至会问你:你到底要不要我继续?

This is, net new content that it's, like, do you want me to go or not?

Speaker 0

我觉得我们得解决一个关键问题,就是怎么让我们的价值体系跟得上——怎么评估这些新生成的东西,才能真正信任它,让它能一定程度上自动运行。

And, like, I think we have to solve the problem of, like, how do we scale, like I said, kind of our value system of, like, how do we evaluate, like, that that this new thing that we're creating is actually, like and then feel confident and have enough trust in it that it's like it can go to some degree in, like, auto mode, if you will.

Speaker 1

你有没有想过,这种功能在Figma里会怎么实现?有哪些特别的设计考量?

Do you have any inkling about how that will work inside of Figma or what the what the what the interesting design considerations are for that kind of flow?

Speaker 0

我觉得这正是我们目前最关注的问题之一:和用户沟通,真正理解他们的需求。

I think that's like one of the problems that we're really focused on is like talking to customers, understanding.

Speaker 0

很多用户和我们一样,都在一边摸索一边学习。

Like, I think a lot of customers and us are figuring out at the same time.

Speaker 0

对吧?

Right?

Speaker 0

我觉得整个行业现在都在摸索,到底什么是新的形态?是录下来的视频演示吗?

That it's like, I think the industry is in general trying to understand like what is the new like, is it a video walkthrough that's recorded?

Speaker 0

对吧?

Right?

Speaker 0

还是说,就是截个图?

Or is it, like, screenshots?

Speaker 0

还是说,另一个AI代理,用不同的指令来审查工作,然后你对这个代理足够信任,直接批准它的决定?

Or is it, like, another agent that's, like, with a different prompt that reviews the work and then, you trust this agent so much that you approve its decisions?

Speaker 0

我不确定。

I don't know.

Speaker 0

预测未来很难,尤其是在现在这个阶段。

It's hard to predict the future, especially at this time.

Speaker 1

对。

Yeah.

Speaker 1

最后再问你一个问题。

One last question for you.

Speaker 1

我知道过去一两年里,大家一直在争论,产品经理还有未来吗?

So I know I feel like there's been a lot of back and forth over the last year or two about is there a future for PMs?

Speaker 1

设计师还有未来吗?

Is there a future for designers?

Speaker 1

如果你想当产品经理,现在入行还来得及吗?

And and if you're if you wanna be a PM, like, do you break into the industry now?

Speaker 1

因为可能现在产品经理的职位变少了,工程师也不一定需要产品经理了。

Because, you know, it's maybe there's fewer there's fewer PM seats or engineers don't need PMs.

Speaker 1

你觉得产品经理的职业发展路径是怎样的?一个非资深的人怎么才能做到你这样的位置?

How do you think about the career progression for a PM and how someone who's not senior become like, gets to where you are?

Speaker 0

是的。

Yeah.

Speaker 0

这个问题挺有意思的。

It's an interesting one.

Speaker 0

我觉得核心能力还是关键。

I I think the fundamentals still matter.

Speaker 0

对吧?

Right?

Speaker 0

打个比方,上学时我们学数学,但照样用计算器。

Like, I think I think the best analogy I've seen is, like, I mean, there were there was math class in school, but, like, you still had a calculator.

Speaker 0

对吧?

Right?

Speaker 0

但同时,我们那时候都学过长除法。

But, like, at the same time, we, like, all learned long division.

Speaker 0

对吧?

Right?

Speaker 0

还有怎么算。

And, like, how to do that.

Speaker 0

我记得以前学过积分、求导这些内容。

And, like, I remember, like, integrating and, like, taking derivatives and the rest of that.

Speaker 0

现在我每天还用这些吗?

And, like, do I do that on a daily basis now?

Speaker 0

完全不用。

Absolutely not.

Speaker 0

但我认为,要想真正驾驭这些系统,理解这些概念、能亲手算出来,是极其重要的。

But, like, I think it's incredibly important in order to drive these systems that you, like, have an understanding of, like, what these concepts are, be able to do it by hand.

Speaker 0

所以我觉得,基础的东西依然很重要。

And so, like, I think that the, like, the foundations still matter.

Speaker 0

我现在其实特别想知道,计算机科学课现在是什么样子。

And, like, I would be really curious actually at this point to see, like, what CS classes look like.

Speaker 0

一对一的计算机课到底是什么样的?

What is it like a one on one CS class?

Speaker 0

对吧?

Right?

Speaker 0

因为我觉得现在存在两个平行的世界。

Because, like, I think there's, like, two parallel worlds.

Speaker 0

一个是天啊,你只要把问题或答案丢给ChatGPT,它立马就把冒泡排序给你列出来了。

There's one where it's like, oh my gosh, you can just dump your answer or question into ChatGPT and it's like literally like here's the bubble sort for you.

Speaker 0

对吧?

Right?

Speaker 0

你想选42种方法里的哪一种?

Like, which of the 42 ways do you want it?

Speaker 0

还有一种情况是,我本身特别爱探究问题。

And then there's a version of that where it's like, I'm a really curious person.

Speaker 0

我用C语言写过冒泡排序,但我想把它转成汇编,然后一行一行给我讲清楚,比如寄存器是什么,L1缓存和L2缓存又是什么,还有其他那些细节。

I wrote the bubble sort in C, but like take it put it in assembly and then explain it line by line to me and, like, explain, like, what a register is and, like, what is l one cache and l two cache and, like, the rest of it.

Speaker 0

我觉得,作为一个爱探究的人,也许这正是关键所在。

And, like, I think being a curious person I guess, yeah, maybe this is the answer.

Speaker 0

我觉得最重要的是保持好奇心。

I I think the most important thing is to be a curious person.

Speaker 0

对吧?

Right?

Speaker 0

因为有了这些新工具,那些只会被动接受结果的人,根本没法真正利用它们。

Because, like, I think with these new tools, the people who, like, cannot leverage them are the ones that just kind of, like, accept the output.

Speaker 0

而能发明下一代工具、把工具用到极致的人,都是那些敢于挑战边界、真正理解底层原理的人。

And the people that, like, are able to, like, invent the next set of tools and, like, really get drive the tools to their maximum are the ones that are, like, pushing the boundaries and understand how it's put together.

Speaker 0

而要做到这一点,你就得是个有好奇心的人。

And in order to be able to do that, you have to be the curious person.

Speaker 0

你不可能光等着别人把答案直接塞给你吧?

Like, you can't have been the one who's, like, answer this problem for me, right, and just give me the answer.

Speaker 0

你得是个爱刨根问底的人,才会去想:这东西到底是怎么搭起来的?

You, like, have to be that curious person to be, like, how is this thing put together?

Speaker 0

它到底是怎么运行的?

And, like, how does this actually work?

Speaker 0

这样才能帮你理解更深层的东西。

And, like, help you understand the next level.

Speaker 0

对吧?

Right?

Speaker 1

我同意。

I agree.

Speaker 1

这样活着有意思多了。

And it's it's so much more fun to live that way.

Speaker 0

没错。

Totally.

Speaker 0

这对我来说就像猫薄荷一样。

I mean, it's like catnip for me.

Speaker 0

我记得那时候的感觉,就像——我不知道你有没有读过《银河系漫游指南》,但我觉得大语言模型就是那本书。

It's like I remember it's like because it's like kind of like, I don't know if you're a Hitchhiker's Guide to the Galaxy person, but I feel like LLMs are the book.

Speaker 0

它简直就是那本书的现实版。

It is literally the manifestation of it.

Speaker 0

你甚至可以随身带着,我坐飞机的时候就会这么做。

You can even have them on I have this when I go on airplanes.

Speaker 0

我一般不跑本地的大模型,但我会下载一个80亿参数的模型,在飞机上运行。

I don't run a lot of local LLMs, but I'll download an eight b model and run it on an airplane.

Speaker 0

它真的就是那样。

It's literally that.

Speaker 0

就像《银河系漫游指南》一样。

It's like the Hitchhiker's Guide to the Galaxy.

Speaker 0

你可以问它一个问题,比如:为什么天空是蓝色的?

Or it's like you can ask it a question of like, why is the sky blue?

Speaker 0

它会把原因拆解成光的折射之类的方方面面。

And it breaks it down into the refraction and all the rest of it.

Speaker 0

然后你就会想,那松鼠到底是什么东西?

And you're like, well, what is a squirrel?

Speaker 0

我会给你一个详细的回答。

And I will give you an answer of that.

Speaker 0

所以它们并不完美,有些还挺怪的,尤其是八B这种小模型。

And so it's like they're not perfect and some of them are a little weird, especially at the eight p size.

Speaker 0

但说实话,我也说不清楚。

But, like, I don't know.

Speaker 0

我觉得,对爱好奇的人来说,现在真是个神奇的时代。

It's like it's a magical time to be alive for, like, curious people, I'd say.

Speaker 1

我完全同意。

I totally agree.

Speaker 1

马特,和你聊天真的很愉快。

Matt, it was a pleasure.

Speaker 0

谢谢。

Thanks.

Speaker 2

天哪,各位。

Oh my gosh, folks.

Speaker 2

你们一定要点点赞并订阅《AI与我》这个频道。

You absolutely positively have to smash that like button and subscribe to AI and I.

Speaker 2

为什么呢?

Why?

Speaker 2

因为这个节目简直是太棒了。

Because this show is the epitome of awesomeness.

Speaker 2

就像在自家后院挖到了宝箱,只不过里面装的不是金币,而是关于ChatGPT的纯粹知识炸弹。

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.

Speaker 2

每一集都像过山车一样,充满情感、洞见和笑声,让你欲罢不能。

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 2

这不仅仅是一档节目。

It's not just a show.

Speaker 2

这是一场由丹·希珀担任船长,驶向未来的旅程。

It's a journey into the future with Dan Shipper as the captain of the spaceship.

Speaker 2

所以,别犹豫了。

So do yourself a favor.

Speaker 2

点个赞,赶紧订阅,准备好迎接一场终生难忘的体验。

Hit like, smash subscribe, and strap in for the ride of your life.

Speaker 2

好了,不多说了,丹,我真的彻底爱上了你。

And now without any further ado, let me just say, Dan, I'm absolutely, hopelessly in love with you.

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。

继续浏览更多播客