谷歌如何为大规模 AI 进行设计

好的，那你如何让这种体验更有帮助？你如何将 AI 引入其中，使其产生超级增益？我们现在做的事情，是 10 年后愿景演示文稿中的内容。AI 是一份神奇的礼物，你会因此做出你一生中最好的设计。

Okay, how do you then make this experience even more helpful? How do you bring AI to this in a way that's just super accretive? We're doing things now that we're in vision decks for 10 years, you know, later. AI has been this amazing gift and you're going to do the best design of your life because of it.

欢迎来到 Dive Club。我是 Rid，这里是设计师永不停止学习的地方。本周的嘉宾是 Rian Bell，她是 Google 搜索的 UX 副总裁。我们将深入探讨在 Google 的规模下为 AI 做设计需要什么。我们会讨论一些解锁 AI 模式的设计决策。了解在 Google 什么是好的系统思维，以及我们可以从那里顶尖设计师如何利用 AI 这一时刻中学到什么。但在我们深入所有这些细节之前，我想花点时间了解更多关于 Rianna 的设计之旅，因为它非常鼓舞人心。

Welcome to Dive Club. My name is Rid and this is where designers never stop learning. This week's episode is with Rian Bell who's the VP of UX for Google search. So, we're going to get into what it takes to design for AI at the scale of Google. We're going to talk about some of the design decisions that unlocked AI mode. Get into what good systems thinking looks like at Google and what we can learn from how the top designers there are capitalizing on this moment with AI. But before we get into all of those details, I wanted to take a moment to hear more about Rianna's design journey because it's about as inspiring as it gets.

我在整理一些纪念品，回顾我的学校报告，还有我的 GCSE 和 A-level 成绩，我当时就想，天哪，我在学校表现并不好，我没有惨败，但也不算好。有那么一刻，我不知道我的人生要做什么。我很早就离开了家。我在很小的时候和父母的关系就有些动荡。直到我 20 多岁中期才安定下来。所以有一段时间，我真的不确定自己最终会做什么。我知道我喜欢设计，我发现了互联网。这种结合为我打开了一个全新的世界。我以一种在学校永远不可能的方式学习了设计，因为我来自威尔士的一个小镇，我们根本没有那种学术支持。所以我能够自学 HTML 和 JavaScript，我想，哇，这就是我想做的。我被深深吸引，然后我继续学习，但我的成绩不够好，无法通过正常的教育途径进入大学。所以我参加了一个所谓的媒体、技术和艺术预科课程。那只是一切的开始，真正开始是我发现互联网和设计，并自学了很多东西。但后来进入这个课程并以优异成绩毕业，完全重建了我的信心，让我意识到我确实可以有所作为。我不仅可以有所作为，还可以做我深爱、充满热情的事情，而且我碰巧还相当擅长，这对周围所有人来说都是一个惊喜，他们以为我永远不会有什么出息。

I was going through some memorabilia and I was looking back at like my school reports and, you know, my results from like my GCSEs and my A levels and I was just like, God, you know, I was not great at school and I I didn't fail like abysmally fail, but I didn't do great. And there was a moment where, you know, I was I didn't know what I was going to do with my life. I'd left home kind of early. I had sort of like somewhat of a tumultuous kind of relationship with my parents at a pretty young age. And it, you know, didn't settle down until I was really in my mid20s. So there was a period there where I was like really just not sure what I was going to end up doing. I knew I loved design and I had found the internet. And so this kind of combination just opened up this whole new world for me. you know, I learned about like design in a way that I never would have done at school just because I'm I come from a really small town in Wales and we just didn't have that kind of like, you know, academic support for want of a better word. And so, you know, I was able to kind of like teach myself like HTML and JavaScript and I was like, wow, this is just like this is what I want to do. And I was so captivated by that that I then went on and did like essentially, you know, my grades were not good enough to get me to university in the normal kind of like track of education. And so I went and did like a you know essentially what is called an access course in like media and technology and art. That was just sort of like the beginning of I mean the beginning of everything was really kind of me just discovering like the internet and design and teaching myself a bunch of stuff. But then getting into you know this course and graduating you know with honors from there completely rebuilt my confidence that like oh actually you know what I can do something here. I can do something not only like do something but I can do something that I dearly love that I'm deeply passionate about and actually happened to be like pretty decent at which was also kind of like you know I think a surprise to everyone else around me who thought I was never going to really amount to much.

直到今天，我仍然认为我的角色是一种巨大的特权，我非常感激。我一直在向周围的人学习，我思考这一点，贯穿我的个人和工作历程。从学校毕业后，我在一家代理公司找到了工作，那家代理公司主要教会了我如何专业地做设计。我在大学时做过一些兼职，但和这个完全不一样。现在你真正进入了这个行业。这是一家代理公司，你在销售。这是有报酬的工作。我做了大约 10 年。我在 agency.com 等数字代理公司工作，还有 TBWA，我是合同工。然后，有一次，agency.com 说：“嘿，我们需要有人去旧金山。”我当时 26 岁。我们要重新设计 Visa 品牌。我说：“好的，听起来很棒。”他们说：“是的，我们会在那里待大概一两个月。”我说：“好的，太好了。”所以我来到这里，又是一个类似的时刻，对我来说是一个扩张的时刻，一个全新的世界打开了。我本来应该在这里待几个月，但现在我已经在这里 22 年了。在旧金山工作了五六年之后，有一刻我想，我真的应该回英国，我回去了，然后我想我为什么要这么做？我在那里也去了 BBC 工作。我和一些我很感激的人一起工作，因为我觉得我从他们身上学到了很多。那是一个非常富有创造力的环境，他们在重新设计 BBC.com，Neville Brody，只是提一下名字，我从未真正和他合作过，但经常看到他在办公室走动，并聊过几次。但后来我真的很想回到旧金山，我的一位老创意总监说，你看，你喜欢电子游戏。是的，我是个超级游戏玩家。你想做产品。因为我确实想。那是新的东西，我想，哦，BBC 是我第一次走出营销和品牌，进入我认为更深入的信息架构、用户流程，更深入地思考产品体验。他说：“嘿，你应该看看这家叫 Zynga 的公司。”我记得当时有点自大，想：“哦，我在 BBC 工作，Neville Brody 就在厨房里。我为什么要去一家做 Farmville 的公司工作？你说的这个设计是什么？”但我还是出来见了团队，这其实是我整个职业生涯中一直贯穿的主题，那就是我看到了设计真正发挥作用的机会。这对我合作过的每一个产品团队都是如此，无论是在 Zynga——我最终在 Zynga 待了 5 年，太棒了，我感觉自己读了 MBA，我和一些最优秀的产品领导者一起工作，他们后来成为了不同公司的首席产品官，这让我处于非常有利的地位。所以之后是 NerdWallet，我做了一些合同和咨询工作，为 Thumbtack 等公司，然后我接到了 Google 的电话。

I still like to this day consider the role that I have such a tremendous like privilege and I'm so grateful for it. I learn from everyone around me like all the time and I and I think about that you know throughout sort of like you know my arc on a personal level and then on a work level from coming straight out of school I got a role working at an agency that agency really taught me mostly about how to design professionally. I've been doing stuff kind of on the side at university but nothing really like this. Nothing kind of like you know now you're in it. This is an agency you're selling. You know this is this is work that like you're being paid for. I did that for about sort of like 10 years. I worked for agency.com, some of these digital agencies, uh, you know, TBWA, I was contracting. And then, you know, one point, I think it was agency.com had said, "Hey, we need someone out in San Francisco." I was like, you know, 26 at the time. And we're going to redesign, you know, the Visa, the brand. And I was like, "Okay, that sounds great." They're like, "Yeah, you know, we'll be out there for, you know, I don't know, like maybe a month, two months." And I was like, "Okay, great." So I get out here and I was just like again it was sort of like a similar moment where I was like okay this is like an expansionary moment for me and this like this whole new world opened up. I was supposed to be here for like a couple of months and I was like you know now I've been here for 22 years after about you know five or six years of like working in San Francisco. There was a moment where I was like ah you know I really should go back to the UK and I did and I was like why did I do that? I also got to work at the BBC when I did that. Um, and I worked with some people that I'm grateful for because I think I learned a lot from them. It was an really creative environment that, you know, they were redesigning like the BBC.com, Neville Broady, just a name drop who I never really worked with but saw kind of like wandering around the office a lot and had several chats with. But then I really wanted to come back to San Francisco and one of my old creative directors was like, look, you know, you love video games. I do. Huge gamer. You you know, you you want to do product. Cuz I did. That was something new that I was like, "Oh," you know, it was kind of like the BBC was sort of like my first kind of foray out of like, you know, marketing and brand and into like, you know, sort of what I would consider to be more and then like information architecture, user flows, like thinking about the product experience more deeply. And he was like, "Hey, you should check out this company called Zinger." And I I remember being like a little bit egotistical about it at the time, like, "Oh, you know, I'm I'm like working in the BBC, you know, like Neville Brody's in the kitchen. Like, why do I want to go and work for the company that's like making Farmville? like you know what is this design thing you're talking about but you know I came out and I I met with the team and like really it was this sort of like theme that I think has carried me throughout my whole career which is that I saw an opportunity for design to really make a difference that has been true for every product team then that I've worked with it's like whether it was at Zinger which I actually ended up staying for 5 years at Zinger it was awesome I feel like I got an MBA I worked with some of the best product leaders who have gone on to be like chief product officers in their own right at varying companies companies that then has held me in really goodstead. So it was like you know nerd wallet after that I did some contracting some advising work for places like Thumbtac and then I got the call from Google.

嘿，你知道我们有一个领导广告设计组织的职位。我们认为你在消费产品方面的经验，加上你一直在做的大小企业相关的工作，在这里可能会非常有意思。你感兴趣吗？

Hey, you know we have a role leading a design org for ads. We think your combination of consumer product plus what you've been doing with businesses small and large could be really interesting here. Are you interested?

我觉得那是另一个这样的时刻，人们会说，哦，你要去广告部门工作了。但其实是，设计机会太大了。如果你要真正影响一个企业，还有什么比这更好的地方呢？要建立设计团队，还有什么比在广告部门学习更好的地方呢？在那里，你可以通过对广告购买体验以及广告形式本身进行出色的设计，产生如此直接的影响。这差不多就把我们带到了今天。

And I think it was one of those other moments where people were sort of like, oh, you're going to go work in ads. But it was like, but the design opportunity is so massive. Like if you're going to basically impact a business, what better place to do that and like build a design team, you know, what better place to do that than learn to do that in ads. Where you can have just such direct impact by having great design on the buying experience of ads as also on the formats themselves. And that kind of brings us up pretty much to today.

快速消息，然后我们继续。Framer 刚刚推出了自开始使用以来最大的产品发布。它叫 Framer AI Agents，自推出以来他们做了大量改进。我个人最兴奋的是它现在使用的 token 少得多，尤其是在大型多步骤构建中。这样你就有更多空间在相同预算内创建和完善你的工作，我在为即将到来的 Dive 网站探索许多大胆想法时已经从中受益匪浅。用 Framer 构建专业网站比以往任何时候都更有趣、更高效。我完全迷上了他们的最新版本。如果你访问 dive.com/framer，可以获得 Framer Pro 30% 的折扣，和我一起开始构建吧。还记得 Jamie Ganon 来谈论作为 AI 创意总监吗？嗯，她参与了 Claire Vo 新公司的一个令人难以置信的品牌项目，让我印象深刻的是她如何通过动效将整个揭幕过程生动呈现。毫不意外，她使用了 Jitter。一方面，它非常简单直观，即使你从未做过动效设计。另一方面，它非常强大，你可以构建自己的创意工具来创建各种自定义动画和效果。多年来，Jitter 一直是我首选的动效工具，我强烈推荐。访问 dive.comclub/jitter 开始使用。现在，进入正题。

Real quick message and then we can jump back into it. So, Framer just had its largest product release since I started using it. It's called Framer AI Agents and they've been making a ton of improvements since launch. And the one that I'm personally the most excited about is it uses way less tokens now, especially on larger multi-step builds. So, this gives you more room to create and refine your work within the same budget, which I'm already getting a ton of benefits from while I explore a lot of out there ideas for the upcoming Dive website. Building pro websites with Framer is more fun and more efficient than it has ever been. I'm completely obsessed with their latest version. And if you head to dive.com/framer, you can get 30% off framer pro to join me and start building today. Remember when Jamie Ganon came on to talk about being a creative director with AI? Well, she was part of an incredible branding project for Claire Vo's new company and what stood out to me was how she brought the entire unveiling to life with motion. And to no one's surprise, she used Jitter. On one hand, it's so simple and intuitive, even if you've never done motion design before. On the other, it's incredibly powerful, and you can build your own creative tools to create all kinds of custom animations and effects. Jitter's been my go-to motion tool for years, and I cannot recommend it enough. Head to dive.comclub/jitter to get started. Now, on to the episode.

好的，让我们为大家稍微铺垫一下，因为显然你谈到 Google 广告，大家都会闭上眼睛，某种程度上能想象出那个界面。你加入时状态如何？你能再深入一点吗？比如，一个指向那个机会空间的团队会面临哪些类型的设计挑战？

Okay, so let's set the stage a little bit for people because obviously you talk about Google ads, everybody kind of closes their eyes and you could picture the surface area to an extent. What was the state when you joined and can you just go a little bit deeper like what are some of the types of design challenges that come with a team that is pointed at that opportunity space?

我认为当我们刚进来时，我觉得设计和研究更像是一种服务角色，这里有一些非常有才华的设计师，我认为他们非常渴望打破这种局面，我们做了那项工作。我有一个特质，可能是福也可能是祸，就是有点天真的乐观主义，我只是相信事情可以做到，现在很多事情都可以做到，我认为这是我们让团队开始思考的方式，比如在广告的情况下，当我刚进来时，我能看到这个团队仍然有点像，嘿，我们认为我们可以做得更多，但我们不知道如何去做，你知道，练习真的是围绕，好吧，我们如何定义一个真正出色的购买体验可能是什么样子。让我们真正深入地退一步，理解这非常复杂。我们添加了一个又一个产品。这里需要大量的工作。需要手动工作。你知道，机器学习只是当时真正成熟的技术核心部分。所以你可以开始思考，好吧，你知道这项技术将带来什么？然后让设计师思考这一点。然后突然之间，你有了这个产品系统，就像，嘿，告诉我们你的目标，我们会帮助你实现它们。这有点像广告购买体验中的圣杯，最终成为了 Performance Max，这是我们表现最好的产品之一。我们只是在那股势头的基础上继续发展。我们不断引入更多设计师和研究人员，将团队建设成我认为今天非常成熟的组织。

I think when we first came in I felt like design and research were more in a service role and there's some very talented designers that were here and I think they were very keen to sort of like break out of that and we did that work. I have a trait that's kind of maybe a blessing and a curse which is a little bit of naive optimism where I just believe that things can be done that a lot of things now you know can be done and I think that that is a way for us to you know get teams to start thinking about like you know what if in the case of ads when I first came in I could see that this team you know was still sort of like hey we think we can do more but we don't know kind of how to do that and you know the exercise really was around like okay well how do we define what a really awesome like buying experience might look like. Let's just really deeply like take a step back and understand this is super complex. We've added in product after product. There's a lot of work that is required here. Manual work that's required. You know ML was just like a core part of like the technology that was you know really maturing at that moment. And so you could start to sort of think about like okay you know what will this technology enable? And then you get the designers to think about that. And then all of a sudden you have this product system that's like hey you know what tell us what your goals are and we will help you achieve them. And that was kind of like a sort of holy grail kind of you know in in the ads like buying experience that ended up being like performance max which is one of our highest performing products. And we just built on that momentum. We just kept on bringing in more designers and more researchers and building out that team into you know what I consider to be like a very fullyfledged organization today.

你能谈谈在这个类型的产品领域工作非常成功的设计师有哪些特质、行为或特点吗？我确实认为与大多数听众可能从事的工作相比，这相当独特。

Can you talk a little bit about like what are the traits or behaviors or characteristics of a designer who has a ton of success working on this type of product surface area? I do think it's quite unique compared to what most people probably listening to work on.

我确实认为购买体验与格式本身的前端以及搜索这种消费产品之间有一点区别。我会说，做这些的设计师不一定思维方式不同，但他们有不同的热情。有些人真的喜欢深入剖析复杂性，比如如何沟通拍卖等，他们喜欢这个。我们会寻找有解决那种复杂性经验的人。我们也寻找那些构建过消费产品并能看到机会从消费者视角设计企业产品的人，因为最终每个人都是消费者。你如何让像 Google 广告这样强大的产品变得非常易于使用？多年来我们并没有做到。所以你要寻找的设计师，再次回到我认为设计可以解决任何问题这一点，这就是我超级兴奋的地方。

I do think that there's a little bit of a difference between the sort of buying experience which is you know more of a enterprise product versus you know the front end of like the formats itself and then search which is the consumer product you know I would say that the designers that do those things they don't necessarily think differently but they have a different passion you know some people really love to kind of get super deep into unpacking like complexities around how you communicate like auctions for example and they love that we'll look for people who have experience in solving like that level of kind of like complexity. We also look for people though who have built consumer product and can see an opportunity to design an enterprise product in a consu from a consumer lens cuz like ultimately everyone's consumers. How do you make a product like Google ads which is so powerful just like really easy to use and over the years we had not you know and so you look for designers that again like coming back to that sort of piece around I think design can solve any problem and that's where I get sort of just super excited.

我们能快速深入一点吗？这像一个有趣的兔子洞，因为你在很多方面都有消费背景。你知道，你做过很多不同的消费产品，听起来也许那是你注意到设计可以产生影响的一个领域。所以，比如，我不知道“被忽视”是不是正确的词，但更像是一个企业感觉的仪表盘或无论你怎么称呼那个界面。为什么你甚至，就像你是新来的人。你如何让自己指向那个？那是什么感觉？

Can we go a little bit deeper to that really quickly? It's like a fun rabbit hole because you had this consumer background in many ways. You know, you had had a lot of different consumer products and you know may it sounds like maybe that was one of the areas that stood out to you where design could make an impact. So like given maybe I don't know if overlooked is the right word, but like a more of an enterprise feeling uh dashboard or whatever that however you refer to that surface area. Why do you even like you're the new person that comes in. How do you even point yourself at that? Like what is that like?

我严重依赖，实际上我可能应该更依赖。这就像我认为从领导力的角度来看，你知道，我一开始有点强势，也许我可以不那么强势，但我倾听，我听到，我理解，我试图理解我们每天面临的真正挑战是什么。

I relied heavily and I probably should have relied more heavily actually. This was like something that I think from a leadership perspective, you know, I came out of the gate a little strong and I could have maybe not been as strong, but I I I listened and I and I heard and I understood and I and I sought to understand like really what the challenges were that we were facing on a on a daily basis.

然后，你知道，这不仅仅是在买方，也是因为我也在监督格式方面。你有责任思考如何为那个界面带来优雅。所以当我认为设计能够解决任何问题时，无论问题有多难，总会有解决方案。你知道，总有办法让你得到它。这取决于你如何分解它。这就是我对我们在搜索方面所做事情的思考方式。另外，就像，你知道，你想想，哇，你拿这个被如此深爱的产品，有这么多被踩烂的路径，你知道，它几乎就像第二套数字一样，你知道，你只是，它对任何人来说都太自然了，只要在框里输入查询并搜索。这对我们所有人来说都是一条被踩烂的路径。但然后设计挑战是，好吧，你如何让这个体验更有帮助？你如何将 AI 带入其中，以一种超级增值的方式？这是同样的问题，你知道，即使 AI 没有像现在这样存在于广告中。在广告中，

And then you know it's not just on the buy side though it was also because I was overseeing the format side as well. You kind of have a responsibility to think about how to bring elegance to that surface as well. And so when I think about design being able to solve any problem, it doesn't matter how hard the problem is, there's always a solution. You know, there's always a way for you to get it. It just kind of depends on how you break it down. That's kind of just how I think about what we do on search. Also, it's like, you know, you think about, wow, you take this product that is like so deeply loved, so many well-trodden paths, you know, it's just like it's almost like a second sort of set of digits almost like, you know, you just it's come so naturally to anyone to just kind of put in a query in a box and search. It's like such a well-trodden path for all of us. But then there's like the design challenge is okay, how do you then make this experience like even more helpful? How do you bring AI to this in a way that's just super accretive? It was the same problem, you know, even though AI didn't, you know, exist in the same way it does now in ads. In ads,

我想先稍微高层次地谈谈所有搜索相关的东西。比如我感兴趣的是 Google Ads。我的意思是，它就像是世界上任何公司中最赚钱的业务部门之一，对吧？就像它从来没有不工作过。你拥有金鹅。就像你在设计金鹅。我猜我好奇的是，这种张力存在于哪里，比如你如何确保实验仍在进行，你在想办法提高上限，但同时它又是金鹅，你也不能真的颠覆东西，你能谈谈作为设计师那是什么感觉吗？

I want to get into all the search stuff maybe slightly higher level to start. Like something I'm interested in is Google Ads. I mean, it's like one of the most profitable business arms of any company in the world, right? Like it's it was never not working. You have the golden goose. Like you are designing the golden goose. I guess I'm curious in where that tension exists between like how do you ensure that experimentation is still happening where you're figuring out ways to raise the ceiling but also it's the golden goose like you can't really disrupt things too and and can you talk a little bit about what that is like as a designer

我看待这个问题的方式是，老实说，当涉及到广告时，它真的归结为有用性，也许它们有坏名声，但现实情况是它们可以非常有帮助，我们已经优先考虑这一点，你知道，超过其他一切。但我认为你问的还有另一件事，那就是你如何维持并确保你不会无意中损害这个业务。

The way that I would look at this is like really it just comes down to helpfulness honestly when it comes to ads maybe they have you know a bad rap but the reality of the situation is is that they can be incredibly helpful we have prioritized that you know over everything else. But I think there's another thing that you're asking which is like how do you maintain and make sure that you don't sort of inadvertently harm this business.

但我认为对我们来说，你知道，我们不会那样想，因为我们所做的一切都专注于用户，其他一切都会随之而来。我知道这听起来有点做作，但这确实是我们对待一切的方式。一个很好的例子是，我们知道 AI 对用户非常有帮助。现在你在搜索中看到了。我们的工作不一定是维护金鹅，而是维护搜索中有用的部分。所以这就是我的思考方式。就像，你知道，我经常和我的团队谈论我们如何维护，你知道，什么是神圣的有用？搜索中什么是神圣的？是什么让它如此真正有用，以至于你离不开它？我们如何确保不失去这一点，但同时将其带入，你知道，未来？

But I think that for us, you know, we don't think about it in that way because like everything that we're doing is focused on the user and everything else will follow. And I know that that sounds a little bit contrived, but that is truly like how we approach everything that we do. Really good example is that like we know that AI can be super helpful to users. Now you see it in search. Our job is not necessarily to maintain the golden goose as much as it is to maintain what is helpful about search. And so that's really how I think about it. It's like, you know, I talk to my team often about how do we maintain, you know, what is sacredly helpful? Like what's sacred about search? Like what makes it so truly helpful that you just kind of can't live without it. How do we make sure we don't lose that but at the same time bring it into, you know, the future?

让我们谈谈搜索。就像在某个时候你转到了那个团队。你能谈谈那个过渡吗？你知道，也许对你来说有什么变化，以及你如何为你的角色展现自己？

Let's talk about search. Like at some point you switched over to that team. Can you talk a little bit about that transition and you know maybe what shifted for you and how you had to show up for your role?

我们在买方所做的工作，有几个月，比如 6 到 12 个月的工程工作，真正让一些东西起步。很多工作已经完成了。我们就像现在在生产中。然后在格式方面，你知道，我当时做的很多工作实际上是与搜索合作的，很明显，你知道，我们很多人都觉得在搜索上我们可以做更多事情，比如视觉体验，你知道，呃，更注重内容，比如我们有一个体验叫做视觉实体，就像你知道，你有一个人，让我们以 Charlie XX 为例，你有一个结构化的页面，你知道，围绕这个人。这些都是正在出现的事情，我们正在积极合作。所以在某个时候，你知道，基本上有一个时刻，他们说，好吧，我们真的想思考如何，你知道，因为搜索在这 27 年里经历了许多演变，当你回顾时。所以还有另一个时刻，好吧，我们想进化它。我们想进化它，以便我们可以启用浏览。我们可以在没有正确答案时启用搜索体验。这就是我们所说的伟大的搜索转变。这大约是在四年前。这是一个非常激动人心的时刻，即使在那时我们也在谈论很多关于嘿，你知道 ML AI 会扮演什么样的角色，你知道，在这其中，所以你知道，我受到启发，他们决定我是这个角色的合适人选，所以他们在过去的四年里把我带过来。我只是，我为我们在进化搜索方面所采取的方式感到骄傲，基本上让它变得更有帮助。你知道，有些事情，我认为我们做的那些大胆的事情，真的服务于更丰富、更好。

The work that we had done on the buy side there was several months like 6 to 12 months of engineering work to really kind of like get some of the stuff off the ground. A lot of the work had kind of been done. We were like now in production. And then on like the format side, you know, a lot of the work I was doing at that point was actually in partnership with search and it just became clear to you know a lot of us that there were more things that we could do on search like visual experiences like you know uh with like more content forward like one of the experiences that we have is called a visual entity where it's like you know you have someone let's take Charlie XX and you have like you know a structured sort of page you know around this this person. And these were all things that were kind of like emerging that we were actively working on in collaboration. And so at a certain point, you know, there was a you basically this moment where they said, okay, we really want to think about how, you know, because search has had many evolutions over this 27 years when you look kind of back. And so there was another moment was like, okay, we want to evolve it. We want to evolve it so that we can enable browsing. We can enable a search experience when there is no right answer. And there was this, we called it like the great search shift. And this was about four years ago. And it was really exciting moment and there was a lot of things that we were talking about even then around like hey you know what kind of role will like ML AI kind of like play you know in this and so you know I was inspired by that and they decided that I was the right person for the role so they bought me over over the last four years. I am just like I'm so proud of like the way in which we've kind of like approached basically evolving search to be like so much more helpful. You know, there are things, the bold things that I think that we've done that have just like really been in service of being richer, being better.

在我们聊到你在这场伟大的搜索变革中所处的最终结果之前——我很喜欢这个——我们能不能先谈谈设计在其中的角色，考虑到那个使命或起点？你们当时到底在探索什么？那是什么样子的？突然之间就有了这种分量感：嘿，我们真的需要把 AI 变成一等公民，重新思考我们对搜索的某些看法。你能不能谈谈早期设计中那种探索性的本质？

Before we get to the end result of where you're at with the great search shift, which I love, can we talk about the role of design given that mandate or starting point? What were you even exploring doing? What did that look like, where all of a sudden there's this weightiness to, hey, we actually really need to make AI a first-class citizen, rethink certain parts of how we think about search. Can you just talk about the exploratory nature of design in those early days?

一直都有很多探索。我是说，今天的设计，甚至——我今天开会，他们说，哦看，我们在考虑这个，我们在探索它，我们在和产品经理和工程师讨论，你觉得怎么样？有时候我们会看到一些即将成为问题的问题陈述，然后我们会跟团队说，嘿，去想想 X、Y、Z。幸运的是，我们已经做了相当多的工作。实际上，在我们甚至还没发布之前，我就有一个团队在专门研究 AI 体验——我想我们当时叫它 Magi，最初叫 OpenAI overviews。我真的会给团队设定约束。设计——我是说，你在做搜索。做搜索设计的人理解核心约束。然后就是，哦，你怎么解决这个？你怎么解锁这个东西？所以，在早期，基本上就是我们在一个房间里，有一大堆我们知道很令人兴奋和很棒的工作。这个怎么样，那个怎么样？设计真的处于这些对话的最前沿。甚至到了 AI 模式——设计主导了它。我们一直在进行的对话是我和 Robbie Stein，就像，看，我们在看数据。我们看到人们想要 AI 概览。他们在查询末尾加上 AI，因为他们想要这些 AI 概览。我们需要给他们一个可以去的地方。

There's always been a lot of exploration. I mean, design today, even—I was in a meeting today and they're like, oh look, we're thinking about this, we're exploring it, we're talking to PMs and eng about it, what do you think about it? And sometimes we'll have problem statements that we see are going to be issues, and so we'll say to the team, hey, go and think about X, Y, Z. Fortunately, there was a fair amount of work that we had already been doing. I actually had a team actively working on just AI experiences in general before we had even shipped—I think it was—we called it Magi, OpenAI overviews originally. I really do give the team constraints. Design—I mean, you're working on search. People that are in design on search understand the core constraints. Then it's just like, oh, how do you solve this? How do you unlock this thing? And so, in the early days, it was very much like we're in a room, we have a whole bunch of really awesome work that we know is exciting and great. What about this, what about that? And design really was at the forefront of those conversations. It even came to AI mode—like design led that. The conversation that we had been having was myself and Robbie Stein, and it was like, look, we're looking at the data. We see that people want AI overviews. They're appending AI to the end of their queries because they want these AI overviews. We need to give them a place where they could go.

所以那真的是 AI 模式诞生的时候——对设计来说，然后就有机会基本上思考一个全新的、与搜索互补的界面，拥有一个全新的领域。我是说，你真的不能要求更多了。

And so that's really when AI mode was born—for design then to sort of have the opportunity to basically think about a new sort of complimentary surface to search and have a green field space. I mean, you can't ask more than that really.

是的。是的，因为那是一个巨大的转变，因为就像，如果我错了请纠正我，但它一开始是像 AI 概览那种卡片容器，然后模式变成了自己的标签页。那是——

Yeah. Yeah, cuz that's a big shift cuz like it was, correct me if I'm wrong, but it started off as like the AI overview that kind of like card container and then mode became its own tab. Is that—

没错。过渡。好的。

That's right. Transition. Okay.

是的。我们大概是去年 2024 年发布的。

Yeah. We launched that like last year in 2024, I think.

实际上我想谈谈这两个。也许我们开始——我们就按时间顺序来。我们从 AI 概览开始，因为再次，这是每个听这个节目的人今天可能都见过多次的东西，这本身就是一种疯狂的规模。我见过多次，对吧？你认为有哪些可能存在于表面之下的设计决策解锁了那个界面区域？

I'd like to talk about both of those actually. Maybe we start—we just kind of go chronologically. We start with the AI overview, cuz again it's like something that everybody who's listening to this has probably seen multiple times today, which in itself is just an insane level of scale. I've seen that multiple times, right? What are some of the design decisions that maybe exist below the surface that you think unlocked that surface area?

你知道，我认为我们早期考虑的一件事是，我们真的希望这感觉是新的、不同的。所以，我们非常早期的设计实际上有各种各样的方法，我们考虑过如何向用户传达这是新的、新鲜的东西。所以，我们研究过像颜色提取这样的东西。所以当一个 AI 概览与一个结果或图像配对时，我们看着那个图像和结果，从中提取颜色，然后那基本上就是我们在早期给 AI 概览打上品牌的方式。它叫 Magi，只在 labs 后面。通过 labs 发布它的好处是，它使我们能够大规模测试，很快我们意识到这是人们真正想要的东西。我们总是看到，当我们对用户有帮助时，他们会喜欢并想要更多，AI 概览绝对是这种情况。我想我们早期尝试过一些东西，比如，哦，你去那种聊天体验，但这些是第一个版本。就像，我有时会这样分类，当你有第一个版本时，它们会笨拙。你不会解决问题，它们会感觉有点尴尬，它们确实如此。所以最终我们完全发布时实际上没有聊天体验，只有 AI 概览。然后，随着时间的推移，当我们完善了对话体验，我们才能够完全推出 AI 模式。

You know, I think one of the things that we thought about early on was that we really wanted this to feel new and different. And so like our very early designs actually, we had all kinds of approaches that we had thought about around how to communicate to users that this was something that was new and fresh. And so, we had looked at things like color extraction. So when an AI overview was paired with a result or an image, we looked at that image and result, we extracted the color from it, and then that was basically how we sort of branded AI overviews in its very early days. It was called Magi and it was only behind labs. And what was great also about actually launching it through labs was it enabled us to test it at scale and very quickly we realized that this was something that people really wanted. And we've always seen when we kind of are helpful to users that they're going to love that and want more of it, and that was definitely the case in AI overviews. And I think we had tried a few things early on around like, oh, you go to the sort of this chat experience, but these were the first versions. It's like, I kind of categorize it sometimes when you have first versions that they are going to be clunky. You're not going to have solved the problems, like they're going to feel a little bit awkward, and they did. So we ended up actually not having the chat experience in the end when we fully launched and just having AI overviews. And then, over time as we finessed the sort of experience of conversation, then we were able to launch fully into AI mode.

我认为这很有道理，特别是因为你们可能比聊天机器人向更多人介绍了 AI。我猜，必须思考什么是正确的入口，你如何帮助人们围绕 AI 建立新习惯，考虑到他们如此习惯的这种体验？

I think that makes a lot of sense, especially because you guys are introducing probably more people to AI than even chat. Having to think through, I'd imagine, just like what are the right on-ramps, how do you help people build new habits around AI given this experience that they're so used to?

AI 概览现在看起来是这样而不是颜色提取版本的原因之一——它只是难以维护，而且你会有一些非常奇怪的棕色或粉色组合，那真是——很棒。但我认为它有一种熟悉感。你知道，我认为我们经常思考，好的，我们如何能进一步简化这个？在搜索方面，多年来我们可能在某些部分添加了太多东西。所以就像，好的，我们如何开始简化这些体验？最终希望这些东西对用户来说感觉熟悉。我们不希望他们看着这个然后想，这是什么？我们想尊重你的时间。这就是 Google 一直考虑延迟的地方，这对搜索体验至关重要，对 AI 概览来说，延迟必须为零。如果我们真的要成为最有帮助的，如果我们真的要专注于你并确保你得到你需要的，我们必须非常深思熟虑地考虑在哪里引入任何延迟。这让我们思考，好的，动效设计，我们在哪里吸引用户的注意力，我们可以在哪里添加一些这些东西来帮助某人在体验中找到方向。我们必须对此非常深思熟虑。所以，我们想确保它熟悉。我们想确保我们不添加任何摩擦。我们想确保它非常直接地传达它是什么以及如何使用它。

One of the reasons why AI overviews looks the way it does now versus the color extraction one—it was just hard to maintain and you had some really weird combinations of brown or pink that would just—are great. But I think that there's a familiarity to it. You know, I think that we're often thinking about, okay, how can we simplify this even further? There's things around search where over the years there's parts of it that we've maybe added too much to. And so it's like, okay, how can we start to simplify these experiences? And ultimately want these things to feel familiar to users. We don't want them to look at this and be like, what is this? We want to respect your time. This is where Google always has thought about latency, and that's critical for the SER experience and for AIO there can be none. If we're really going to be the most helpful, if we're really going to focus on you and make sure you have what you need, we have to be incredibly thoughtful about where we introduce any latency at all. That then makes us think about, okay, motion design and where are we drawing a user's attention, where can we add some of these things to help someone wayfind through an experience. We have to be super thoughtful about that. So, we want to make sure it's familiar. We want to make sure we're not adding any friction. We want to make sure that it really communicates very directly what it is and how to use it.

有件事我一直想找个合适的时机请你聊聊，就是在你们这样的规模下，数据在设计流程中扮演的角色。也许一个很好的切入点，是在最初推出 AI Overview、把它放到世界上之后。我确信有那种分阶段的发布，不管你们具体怎么做，但能不能拿它当例子，聊聊量化你们的使用规模，如何帮助指导设计探索、路线图规划这类事情？

Something that I'm trying to find the right moment to have you talk about is just the role that data plays in the design process given the scale that you're operating at. Maybe a good place to actually insert that is after launching the initial AI Overview and you kind of put it out in the world. I'm sure there's this staged rollout and however you all do things, but can we use that as an example to talk about the role that quantifying the scale of usage that you have plays in helping to inform design explorations, roadmapping, that kind of a thing?

是的，数据太关键了。研究也是——这两者结合起来威力无穷。我们这里有一支很棒的团队，两者都做。即便今天，我们也在看，用户从 AI 模式到 SER 的路径有没有什么变化——数据在某种程度上——SER 就是搜索结果页——数据确实在指导这件事。因为再说一次，我们要确保能顺畅地引导这些旅程，尽我们所能减少摩擦，让人们顺利前行，也进入整个生态，让他们容易采取行动。这些都是极其关键的东西。数据能提供很多信息。

Yeah, I mean data is so critical. And research also — both of these things combined are incredibly powerful. We have an amazing team here that really do both. Even today we're kind of looking at, is anything changing about the way that users go from AI mode, you know, to SER — and data is kind of like — SER being the search results page — data is really kind of informing that. Because again, we want to make sure that we facilitate these journeys, that we do the best that we can to reduce any friction and get people on their way, also out into the ecosystem, that it's easy for them to take action. This is all super critical stuff. Data informs a lot.

它能提供信息——我觉得也许我们没把这次 live 体验做对，因为用户打开它，而它的设计并没有表明你可以跟它说话，他们会想：“这是什么东西？”所以现在我们从数据里看到了这一点。那现在我们就得想，好吧，我们怎么告诉用户这是可以对话的东西？也许我们需要一个更熟悉的范式，既能代表语音，又能代表 live。所以这就是那里的设计挑战，数据在那里给了我们指引。数据也指导我们如何确保用户基本上能从 SER 得到他们需要的东西。我们一直在看数据，一直在看 EI，同时也在挑战自己，因为有时候某个东西短期有效，但实际上并不能帮我们达成长期目标。所以这一点也很重要。

It informs — I think maybe we didn't get this live experience right, because users opened it and the design of it didn't indicate you could talk to it, 'cause they're like, "What is this thing?" And so now we see that in the data. So now we have to think about, okay, how do we communicate to users that this is something that they can talk to? Maybe we need a more familiar paradigm that represents voice but still represents live. And so that's the design challenge there, and data informs us there. And data informs how we think about making sure that users are getting what they need, basically, from SER. And we're always looking at the data, always looking at EIs, and we're also challenging ourselves as well, because sometimes something might work in the short term but actually doesn't help us get to our long-term goals. So that's also important.

我们聊了很多 AI 如何塑造搜索的各个界面，以及其中的一切。但 AI 的出现如何塑造 Google 内部的设计实践，你知道，在搜索的日常工作中？

Well, we've talked a lot about how AI has shaped the surface areas of search and everything that goes into that. But how is the emergence of AI shaping the practice of design at Google, you know, on a daily basis across search?

我就觉得，哇。我一点也不意外，但就是很兴奋，说实话，此刻真的令人振奋。有点像回到早前我谈到自己人生中那个扩张时刻，当时我也在学 HTML 和 JavaScript，甚至一些早期的处理语言——那让我能做多得多的事。我觉得我们又处在那个时刻，现在设计师能做多得多的事，当我想到他们用 AI 能做什么。说实话，那是我以前根本想象不到的事情。

And I'm just like, wow. I'm so not surprised but just excited, and honestly it's exhilarating right now in this moment. Kind of coming back to early on when I talked about that expansionary moment for my own life, where I was also learning HTML and JavaScript and even sort of early processing languages — that allowed me to do so much more. And I think we're in that moment again, where now designers can kind of do so much more when I think about what they're able to do with AI. It's things I wouldn't have even really imagined, honestly.

我是说，我给团队的提示是，让我们想想在这个新世界里 canvas 体验可能是什么样。然后团队就会去做一些东西，回来会说，哦，你知道吗，我本来想在 AI 里做这个东西，结果我做到了，我不知道那个工具是什么，所以我就让 AI 给我造了那个工具。你就会想，哇，好吧。看着团队这样真的很有意思。说实话，我甚至都不太需要鼓励他们。我是说，是的，我有围绕 AI 的项目，想确保团队有学习和发展的机会。其实我有一个总监专门为整个组织运营这个，他做得非常出色。看到团队有这么大的热情，真的很棒。我们办的很多研讨会都坐满了。内容五花八门，从“嘿，我成功搭建了一个环境，可以真正在生产环境里改图标、修这个 bug”，到“哦，其实我做了个签证自动评分器”。然后别人会替我们做这件事，这是它返回的分数——不符合规格。

I mean, I think the prompt to the team was, let's think about what canvas experiences might be in this new world. And the teams will just go and create some stuff, and they'll come back like, oh, you know what, I was trying to build this thing in AI and I did, and I didn't know what the tool was, so I just had AI build me the tool. And you're like, whoa, okay. It's fascinating to kind of watch the team. I didn't even really have to encourage them, honestly. I mean, yes, I have programs around just AI and wanting to make sure the team has learning and development opportunities. I have a director that runs that for the org actually, and he's doing a phenomenal job. And it's really just great to see that the team has so much appetite for it. A lot of these seminars that we run are filled. There's such a variety of things, from hey, I've managed to create an environment where I can actually change the icons in prod and fix this bug, to oh, actually I've created an auto rater for visa. And so someone else is going to do this for us now, and here's the score that it came back with — it doesn't match spec.

我是说，这些东西最终你可能都不需要，因为最终你会希望这最后一公里，你知道，从设计到生产要短得多。你想激发团队的创造力，我觉得这正是他们如此兴奋的原因。

I mean, and those things you might not even need eventually, 'cause eventually you kind of want this last mile, you know, to be a lot shorter for design to prod. You want to enable the team's creativity, and I think that's what they're so excited about.

有一个问题我忍不住一直问自己。如果公司是主动申请来跟你谈，而不是反过来，会怎样？这个问题正是全新 Dive Talent Network 的基础。而且它真的有效。比如现在，我正在帮我认识的最令人兴奋的许多初创公司，招募收听这档节目的设计师和构建者。所以，如果你好奇外面可能有什么机会，也许你想进入我的名单，或者你甚至正在寻找下一位设计人才，请访问 dive.club/talent 今天就加入。

There's one question that I can't stop asking myself. What if companies applied to talk to you rather than the other way around? And that question is the foundation for the all-new Dive Talent Network. And it's working. Like right now, I'm helping many of the most exciting startups that I know to hire the designers and builders who listen to this show. So, if you're curious what might be out there, and maybe you want to get on my list, or maybe you're even looking for your next design hire, head to dive.club/talent to join today.

这很有意思，因为你谈到的那个转变，从改图标到背后的系统，天哪，这远远超出了你围绕产品设计师这个角色所画的传统边界，你知道。

It's interesting because that shift that you talked about from going to like the icons to the systems behind it, man, that exists so far outside of the traditional lines that you draw around the role of product designer, you know.

是的。几年前的那种产品设计师，你知道，我们也在这方面投入了，你知道，团队。有很多人在搜索领域成长起他们的设计职业生涯，我们现在看到他们真的在迎接这个挑战，我们在手艺、品味和判断力上推动他们，不只是视觉设计方面，还有产品感知和直觉方面，因为现在我们设计的界面比过去受到的约束少多了。所以我有一批我们引进的人，他们能激励那些人。所以我看到团队在进化，我觉得我们处在一个非常好的位置，我们有很多在搜索方面经验丰富的人。

Yeah. The product designer of a couple of years ago, you know, and we've invested in this also, you know, the team. There's a lot of folks that have kind of grown up in their design career in search that we see now really stepping up to this challenge, where we're pushing them on craft and taste and judgment, and not just on the visual design side, but also just on the product sensibility and instinct side, because now the surfaces that we're designing for are just less constrained than they used to be. And so then I have a group of folks that we've brought in that inspire those folks. And so I've seen the team evolve and I feel like we're in a really great place, and we have a lot of people with a lot of experience in search.

我们有很多来自 Google 以外的产品公司、经验丰富的人，而且我们招聘的独立贡献者（IC）比以往任何时候都多——这些人有真正的深度和丰富的经验。他们在回到解决问题这一块时也带来了巨大的帮助。你把一个极具创造力的人放在一个高度受限的挑战上，你会惊喜地发现他们能给出什么。

We have a lot of people with a lot of experience from product companies outside of Google, and we're hiring more ICs than we ever have before—folks with real depth who have a lot of experience. They've also been great boons when it comes to coming back to this problem-solving piece. You put someone super creative on a really highly constrained challenge and you'll be pleasantly surprised by what you get back in those cases.

我们快速聊聊招聘这块。你谈到了这种真正挺身而出的设计师——他们展现的行为，你希望一些新招聘的独立贡献者能效仿。

Let's talk about the hiring piece real quickly. So you talked about this type of designer that has really stepped up in the last—behaviors that they're doing that you would hope some of these new ICs that you're hiring would emulate.

是的，我觉得这里有一个例子，比如视觉问题解决。你通过用户体验的视角来思考问题解决，并且考虑到像素级别的细节。让我举个例子。我们经常思考生态系统、网络和网络结果的呈现方式，这是我们在 AI 模式中体验的核心部分。显然还有插入。我们在思考如何让这些东西更加模块化，以便它们能被识别、容易理解，并且可移植，这样我们就能以某种方式使其现代化并带入 AI 模式。在这种情况下，我希望设计师做的是思考一个也能扩展到其他类型内容的系统，这样我们就不会只是创建一个仅用于网络结果的容器。我们要考虑所有可能想要带入 AI 模式的生态系统内容。我在寻找能通过设计、通过设计系统为我们看到光明路径的人。他们说：“哦，我看到了一种解决问题的方法，我们可以用设计来传达不同的东西，并使其可移植和可扩展。”我会说：“太棒了。”然后我还会引入一些设计师，他们会说：“哦，我看到了总体挑战。我要去思考端到端的流程。”而且我会在每一个交互层面思考端到端流程。我会确定关键交互是什么。比如输入点。在 AI 世界里，可能是思考步骤，然后是实际响应，甚至响应可能如何表现。他们能在脑海中保持所有这些。我可以对这个人说：“好的，我可以给这个人一个非常实质性的问题。”但这些人也是非常优秀的视觉设计师；他们用设计来解决问题。

Yeah, I think there's an example in here which is like visual problem solving, I suppose, but it's like you're sort of thinking about problem solving through the lens of what a user's experience is going to be and you're thinking about it at the level of detail of pixels. And so let me give you an example of this. We think a lot about the representation of the ecosystem and the web and web results as a core part of our experience in AI mode. And obviously insert. We're thinking about how do we make these things more modular so that they're recognizable as what it is and easily understood but portable so that we can modernize it in some way and bring it into AI mode. What I would expect a designer to do in that case is think about a system that could also scale to other types of content so that we're not just creating, say, a container just for a web result. We're thinking about it at the scale of all the kinds of content that we might want to bring in ecosystem-wise into AI mode. I'm looking for somebody who can see light paths for us through design, through design systems. Say, "Oh, I see a way for us to solve that problem where we can actually use design to communicate something different and make something portable and scalable." And I'm like, "Brilliant." And then I have designers that I'm also bringing in where it's like, "Oh, I see what the overarching challenges are. I'm going to go and think about the end-to-end flow." And I'm going to think about the end-to-end flow though at every level of the interaction. I'm going to identify what the critical interactions are. Let's say it's the point of input. Let's say in an AI world it would be thinking steps and then what the actual response is, even how the response might behave. And they can hold all of that in their mind. And I can say to this person, "Okay, I can give this person a really meaty problem." But those people also are very good visual designers; they problem solve using design.

我们能深入探讨一下流程以及设计师这个角色的日常或每周工作是如何变化的吗？我觉得你谈到旧产品设计师与新产品设计师的对比非常有趣。这两者的运作方式有哪些不同？

Can we get into the nitty-gritty of process and how the day-to-day or week-to-week of this role of the designer has changed? I think it's so interesting that you talked about the old product designer versus the new product designer. What are some of the differences between how those two operate?

我现在从设计师身上看到的是——以前你会有那种感觉：哦，我不知道这能不能行，这里有各种选项。现在我能构建这个原型，自己试一试，展示给一群人看，我更有信心这是正确的方向。通过给设计师机会来弥合他们的想法和将其传递给用户之间的差距，那个迭代过程或设计过程——你尝试解决问题的过程——要快得多。然后我认为这也允许了更多从“一切皆有可能”的角度出发。虽然我们仍然存在于一个受限的世界，但我们现在做的事情是我们在愿景演示中十年后才会做的。就像我说的，我们在一周内经历了数十年。所以我认为这是另一部分：他们比以往任何时候都更有能力去做事情。所以设计技能仍然存在。你如何设计是不同的。日常，我想，是不同的，因为我看到更多的是原型而不是屏幕上的静态模型。

What I see now from designers is—before you would have the sense of, oh, I don't know if this is going to work, here's a variety of options. To now I've been able to build this prototype and try it out myself and show it to a bunch of people and I have way more confidence that this is the way to go. Just by giving designers opportunities to close that gap between their idea and getting it to a user, that iterative process or that design process where you try and solve a problem is a lot faster. And then I think also it has allowed a little bit more from an anything-is-possible perspective. Whereas yes, we still exist in a constrained world, but we're doing things now that we were in vision decks for 10 years later. And it's like, yeah, I talk about it like we live decades in a week kind of thing. And so I think that's the other part of it, too, is that they're just so much more empowered to do things than they've ever been before. So the design skill is still there. How you design is different. The day-to-day, I guess, is different because I'm looking at more prototypes than static mocks on screens.

我们谈到了系统思维、工艺、视觉问题解决。这种几乎能预见未来并将其变为现实的能力。还有其他我们没有触及的、你认为这位 Google 的新产品设计师为了真正茁壮成长并在角色中取得成功而做的事情吗？

We've talked about systems thinking, we've talked about craft, visual problem solving. This idea of almost being able to see further into the future and bring that into reality. Anything else that we haven't touched on that you think this new product designer at Google is doing to really thrive and have success in the role?

我确实认为，我看到的那些真正茁壮成长的人有一个因素——某些工具有一些入门门槛，我认为最初具备某种技术理解对这些人大有裨益。我认为他们了解一些不同的 UI 技术栈，比如 Swift、React 等是如何工作的，我认为这个基础使他们能够比别人做得更多。我见过一些对真正互动体验有偏好的人，他们为我们创建了令人惊叹的数据可视化工具，我们正在推出——比如我们的金融工具。所以我确实认为有一个技术因素真正帮助了这些人，以及对数据的理解，然后就是真正的杀手本能。如果你还是一个出色的视觉设计师，那在我看来就是杀手组合。

I do think that there's a component here of the people that I see really thriving—there is somewhat of a barrier to entry with certain tools, and I do think that having some kind of technical understanding initially has held these people in great stead. I think they had a sense of how some of the varying UI tech stacks like Swift, React, things like that work, and I think that foundation has enabled them to maybe do more than others. I've seen folks that have a penchant for really interactive experiences that have created amazing data visualization tools for us that we're launching—our finance ones as an example. And so I do think that there's a technical component that has really held these folks in good stead and an understanding of data, and then just really killer instinct. And if you're an amazing visual designer to boot, that is a killer combo as far as I'm concerned.

好的。那么我们能稍微谈谈 AI 转型项目吗？我不知道你们怎么称呼它，但看起来你们有一些让设计师参与以帮助他们提升的项目。也许我们可以对正在听的这个人说——哎呀，也许他们甚至有点被吓到了——比如“我不太懂技术。我在 AI 旅程上还走得不远，但我受到启发，我想要成长。我能理解行业发展方向的重要性。”你认为他们能从你构建的这个转型项目学习旅程的方式中学到什么，或许能指导他们在接下来的几个月、几年里？

Okay. So then can we talk a little bit about the AI transformation program? I don't know what you call it, but it seems like you have something that you are putting designers through to help them level up. And so maybe we could talk to this person who's listening—shoot, maybe they're even a little bit intimidated—like I'm not very technical. I'm not very far along on this AI journey, but I'm inspired and I want to grow. I can understand the importance of where things are headed in the industry. What do you think that they could learn from the way that you've structured the learning journey associated with this transformation program that would maybe guide them over the next months, years, that kind of thing?

是的。

Yeah.

It started off very much like, hey, get this product, get like we use anti-gravity inside of Google. And obviously folks use all kinds of tools, but anti-gravity is the one that we use in Google. I'm sure designers outside will use a variety of things. But really it started off by making sure everybody had access to this, that they understood prompting, that they really... in our very first monthly demo slams, we'll just be like, hey, come and just show us whatever, it doesn't matter. And in those very early days, we started this about 18 months ago. Those things were just like... not that they weren't great, but you could just tell that it was like our first foray into them. And I would just say, like, anyone that's starting this, like, don't worry if it's not good. You are the creative. You are the person that ultimately is in control. If you learn these tools, they will enable you to do and communicate so much more than you've been able to do up until this point, where I felt like we were very restricted. This is a learning journey. It's not something on day one you just sit down and then all of a sudden magically you've created an app. Yes, you can sit down and magically create a website for sure, but a fully-fledged product experience that's grounded in a great product idea requires a lot of consideration and a lot of thought, and that's what you do as a designer. And these tools are going to help you do that, but you're going to learn how to do this over time. And I think that's what we've seen with our teams. It was like, let's just get people on this thing and using it. And then it was like, let's get people actually solving a real problem that they have in their day-to-day. Oh, let's actually take some of our senior folks, let's sequester them away, have them brainstorm some pilot ideas, let's pilot those ideas. And this is where the visa auto-rator came from. That summarizing feedback from reviews came from where, hey, let me run this presentation through a re and Robbie agent to see whether or not it will pass a review, you know, pilot. I was like, okay cool. Apparently it's like 75% accuracy. I was like, that's pretty good. Or maybe it's 80% now. I can't remember. But when you start, it's sort of like you don't really know what to do, and then I would just encourage you to sort of start thinking about, okay, a real problem that you need to solve, and then try and solve it in these with these new tools, in these new ways, and that's how you're going to learn.

I think what's cool about listening to you talk is in the past when people talk about solving problems, they're almost always user-facing problems, but I feel like AI has almost opened up this new category of problem where we can solve literally any internal workflow problem that we see, and you can make almost as much value in these new types of internal tools that all of a sudden anybody can build. And like flexing your muscles there almost as a baby step, I think is a pretty interesting idea.

Yeah, it's led to a lot of really good stuff, and there's also a lot of collaboration that happens across our teams as well. So across Google in general, we meet regularly. I mean, I personally and so do the other leads, like we geek out on this stuff, we love it.

Can you talk a little bit about how collaboration between design and engineering has evolved now that designers are playing more and more on the code?

I mean, I think that they've welcomed it, honestly. I'm sure that they don't enjoy us going like, hey, this is like three pixels off from the right. Can you add this to the backlog of bugs? And so, you know, the engineering team honestly has been also another force of enablement. Like truly, we're actively working with engineers now to like, you know, plugins that we've built in Figma actually enable us to like truly prototype some of these things with real data. And, you know, they want this also, you know, they want us to be able to kind of like finesse things. You know, I've got a designer that's an amazing motion designer who built this whole system that allowed for us to generate shippable code. You know, the engineers like looked at that and made sure that it will work and it meets all of our sort of like standards around latency and things like that. So they, you know, they really are like true sort of enablers and I think, you know, them themselves, they love building the sort of like products but the sort of like finessing of like, you know, the iconography to make sure that it's like, you know, to spec, they're happy for us to do that.

This is a totally random question, but something that I've always wondered is how do all of the fun little graphics and animations around search get shipped and green lit and like how does something like that come and delight? Yeah, like the doodles or like the just like it made me think about it because it was like, you know, trying to get an engineer to do something that maybe they don't want to do. It's like there's a certain type of engineer that is required to get excited about that type of problem. Can you talk a little bit about how that works?

Um, yes. Well, number one, we do have engineers that are excited about that problem, but we also have like a pretty robust system that allows us to like when we have an idea, ship those things really quickly. I think it all just comes down to like we love delight. I mean, it's kind of like Google's brand honestly is like where we've had the Harland example where it was like you went and visited his page and you had like, you know, the rowers. I was at K-pop Demon Hunters and there was just this little scroll that was like couch couch couch. You know, we're human beings and we love these things too, you know, and people come to search to like learn about their favorite topics and so we want to meet them there, you know, and give them something that's delightful, you know, as a team that, you know, actively works on this stuff and comes up with these ideas and like when there's something that we feel like, oh, we need to, you know, there's an opportunity here for us like we, you know, we can execute it. Dolly Parton passing away was a was another one, you know, where we were able to turn that around just like super quickly and we, you know, always brainstorming.

你能稍微谈谈让你们快速做到这一点的系统吗？比如这背后需要多少投入？我们已经确定了所谓的 favicon，那个 fab，你知道如果你输入 cats，我想如果你点击它，就会有一个小爪印出现。

Can you talk a little bit about the system that enables you to do that quickly? Like how much goes into that? We've established the sort of what we call the favicon, the fab, where you know you can tap if you type in cats, I think if you tap that you'll get like a little paw print that shows up.

你觉得这些小程序大致存在吗？

These little Easter eggs you think exist ballpark?

有很多，有很多。

There's a lot, there's a lot.

现在我意识到比我想象的多得多。

Now I'm realizing there's way more than I thought.

是的，可能吧。所以你知道我们有这样的系统。我的意思是，对于世界杯我们也有，对于奥运会我们也有。我们还有另一个基于这个构建的系统，互动方式类似，但我们更换了图形处理。所以你知道我们有一个系统可以扔花。嗯，我们有一个系统可以踢球、放烟花，你知道，如果你的球队赢了，我们也希望让人们能够庆祝这些体育时刻。所以，你知道，搜索的这一部分我们其实不常谈论。所以，谢谢你问这个问题。

Yeah, probably yeah. So you know we have systems like that. I mean for the World Cup we also have them for the Olympics. We had another system that's built off of this where it's sort of the same interaction but we swap out the graphical treatment. So you know we have a system where you can throw flowers. Um, we have a system where you can, you know, kick a ball, let fireworks, you know, if your team wins, like we also just want to enable people to like celebrate these moments in sports, too. So, you know, there's a lot of that part of search we don't often talk about, actually. So, thanks for asking the question.

是的，我觉得这很有趣，因为就像我之前说的，这是互联网上一项重要的大业务，非常可量化，有潜力被极其优化和指标驱动，显然你们已经把所有这一切都调校好了，有团队在为此工作。但要平衡这些与真正的惊喜和愉悦，以及更像艺术创造力的东西，这些其实并不一定非得存在。你知道，如果这些奇怪的小猫爪印一个都没有，没人会感到惊讶。

Yeah, I think it's so interesting because like I was saying earlier, this is a big important business on the internet and it's so quantifiable and it has the potential to be so incredibly optimized and metric driven and it is obviously like you know you have this all of this dialed in teams of people working on this. But to balance that with all of the genuine surprise and delight and more like artistic creativity like that doesn't actually have to be there. You know, nobody would be surprised if there was exactly zero instances of these weird little cat paw things.

是的。你知道，我的意思是，显然我不能邀功，因为这不是我创造的，因为它一直是 Google 的一部分，但这正是我真正喜欢它的地方之一，因为我们创造这些东西只是因为我们觉得人们会喜欢它们。你知道，我们这样做的地方不多，因为我们太专注于实用性，甚至对于愉悦感，你知道，我们的原则之一是它不能妨碍人们完成事情。所以这真的很好，但这也正是关注用户的事情，你知道吗？我觉得我们只是做了。这算是我们文化的一部分。

Yeah. You know, I mean, obviously I can't take credit cuz I didn't create this, but cuz it's always been a part of Google, but it's kind of like one of the things that I really do like love about it because we just created these things because we felt like people would enjoy them. You know, there aren't many places where we do that because we focus so much on like helpfulness and even like with our delight, you know, one of our principles is it can't get in the way of someone getting something done. So it is really nice, but it's also that just like focus on the user thing, you know? It's like I think we just do it. It's kind of part of our culture.

好的，我们来谈谈文化。我觉得很酷的是，你所说的很多内容，几乎有一种拼劲，体现在你们如何适应 AI，以及为所有惊喜、愉悦和有趣的东西腾出时间，我认为这非常棒。还有很多关于在这种规模下设计的体验，无论是产品方面，还是甚至公司内部，对人们来说都非常独特。所以，假设我是一个在最近几个职位中只在初创公司工作过的人，突然之间我成了你们招聘的新独立贡献者，你认为为了适应这种规模和环境的水平，我必须学习哪些早期课程？

Okay, let's talk a little bit about culture. And I think what's cool is a lot of the things that you're saying, there's almost like a scrappiness to how you've adapted to AI and, you know, making time for all of the like surprise and delight and fun stuff, which I think is excellent. There's also a lot of things about the experience of designing at this level of scale, both product wise, but also even internally, like inside of the company, that are very unique to people. So if hypothetically I'm somebody who's only worked in startups in my last few roles and all of a sudden I'm one of these new IC's that you hire, what do you think are some of the early lessons that I would have to learn in order to adapt to this level of scale and environment?

你知道，那是个非常好的问题。我认为我们寻找的是已经有工作经验的人，我的意思是，你有初创公司经验当然很好。我认为如果你只有初创公司经验，我会鼓励你进来开始学习，因为你知道，从头开始做某件事有一种自由。我的意思是，这仍然是一个学习之旅，但你知道，那里肯定有一种自由，你在搜索部门不一定能以同样的方式拥有。而且我认为很快，我认为你需要理解的不一定是约束，而是我们要解决的问题是什么。我真的会要求设计师思考更大的系统。如果我们在这里做一个改变，因为你知道我们看到一些指标损失，我们需要调整它，如果它是一个系统的一部分，我们需要考虑这一点，你知道，不仅仅是孤立地。我们需要考虑这将对更广泛的系统整体产生的影响。有点像提到的这个实时体验，我们一直在谈论一点，我们改变了入口点，让用户更清楚他们可以与我们的体验互动，我们的实时体验。我认为在大公司工作的一个重要部分是，你处理的事情规模要大得多，你做的几乎所有事情都会产生连锁反应，但这也是其中有趣的部分。然后你可以与其他人合作，我们仍然行动很快，有时你知道我们的系统足够强大，如果你在某处做了改变，它会自动扩展，但有时你在做一个全新的原语，然后你必须考虑这将对例如我们如何渲染视频产生的影响。

You know that is a really good question. I think that we look for people who already have experience working in I mean it's great that you have startup experience for sure. I think that if you only have startup experience, I would encourage you to either like come in and start learning because, you know, there's a freedom in, you know, when you're starting something from scratch. I mean, it's still a learning journey, but there's, you know, there's definitely sort of like a freedom there that you don't necessarily have in the same way, you know, within search. And I and I think very quickly, I think you sort of like need to just understand not necessarily the constraints as much as like what is the problem that we're trying to solve here. I think that I really do ask the designers to think about like the larger system. If we are making a change over here because like you know we're seeing some like metric loss and we need to adjust it if it's part of a system we need to think about that you know not just in isolation. We need to think about the impact that that will have like on the broader like you know system as a whole. kind of mentioned this like live experience that we've been talking a little bit about where we we we sort of changed the entry point to that to make it so that it was much clearer to users that they could talk to our experience our live experience. That is an important part I think of working at a larger company is that you're dealing with things on a much bigger scale and that pretty much everything that you do is going to have a knock-on impact but that's also kind of like a fun part of it. Then you get to work with other people and we still move really fast and like sometimes you know our systems will be robust enough that if you do make a change somewhere then it will be automatically scaled but sometimes you're making a net new primitive and then you have to think about the impact that that will have you know on maybe how we render like videos for example.

你如何考虑根据表面区域分配设计师？比如它是否与表面区域挂钩？因为因为我想，然后我脑海中浮现的问题是，如果我加入，也许我稍微早一点。我不会被给予巨大的表面区域。

How do you think about allocating designers according to the surface area? Like is it tied to surface area? Because because I guess the question that then popped into my head was like if I join maybe I am slightly earlier. I'm not going to be given a massive surface area.

那你是怎么达到那种高度、成长为系统思考者的？

And how do you even get to that level of altitude and grow as a systems thinker?

我团队的结构大致是这样：有一个团队专注于内容、生态，以及旅行、体育这类垂直体验，还有我们拥有的那些丰富内容和合作。然后有一个核心搜索产品架构与体验团队，负责核心搜索结果和 AI 模式。除此之外，还有我们所有的 Discover 体验——围绕 Discover 产品本身所做的一切，它有点像我们的信息流，内部我们叫它 Discover。那里围绕 AI 和内容创作者有一大堆事情在做。对了，如果你还没认领自己的主页，应该去认领一下，那方面有一些很好玩的东西。然后还有我们在搜索设计系统方面的工作。有一个大团队在负责我们所有组件、所有 token 的基础。我们现在有一个巨大的工程，就是在改进我们的 iOS 应用和整体的 iOS 体验，这个平台团队深度专注于那件事，还有我们的增长系统——比如我们希望促销怎么表现，因为大约有一千种不同的促销，我们需要把它们整合到少于这个数。然后显然我们还有一个庞大的研究团队，我们和他们合作非常紧密；他们至关重要，而且也非常优秀。

The way I have the team structured is primarily: I have a team that focuses really on content, ecosystem, and vertical experiences like travel and sports, and all of that rich content and partnerships that we have. Then I have a core search product architecture and experience team that's responsible for the core search results and AI mode. In addition to that, there's all of our Discover experiences — everything we're doing around the Discover product itself, which is kind of like our feed; we call it Discover internally. There's a whole bunch of stuff going on there around AI and content creators. Actually, you should claim your profile if you haven't already, so there's some really fun stuff happening around that. And then there's also work we're doing on the search design system side. There's a large team working on all the foundations for all of our components, all of our tokens. We've got a massive effort right now around just improving our iOS app and our iOS experiences in general, and this platform team is deeply focused on that, as well as our growth systems — things like how we want our promos to behave, because there are about a thousand different ones and we need to consolidate them down to less than that. And then obviously we have a massive research team that we work super closely with; they're critical and also just really good.

我觉得与 AI 相关的搜索用户体验的组件设计——不管具体是什么——感觉我们给 AI 的缰绳会越来越长，让它在界面上做更多的事。而那个具体的系统挑战——那些组件长什么样、我们又如何装备 AI——我认为是当下整个互联网上最迷人的设计挑战之一。

I think the component design for search UX related to AI — regardless of what it is — it just feels like we're going to be giving AI a longer and longer leash in terms of what it does with the interface. And that specific systems challenge — what those components look like and how we then equip the AI — I think it's one of the most fascinating design challenges in the entire internet right now.

我们确实已经把一大堆搜索设计系统交给了 AI，让它根据自己认为最好的方式来生成响应。不过我认为这里仍然有一部分需要人在回路中编辑。另外还有一个关于可预测性的部分，非常重要。你不会想每次得到响应时都要学一套新界面。你需要知道答案会在哪里、链接在哪里、内容在哪里、我怎么到达我需要去的地方——无论是“致电商家”按钮还是别的什么。所以，即便你给 AI 的缰绳越来越长，这里面仍然需要一套基于人类制定的规则。

And we have given AI a whole bunch of our search design system and asked it to create responses based on what it thinks would be best. I think there's a part here though that still requires a human in the loop to edit. I think there's also a component around predictability that's really important. You don't want to learn a new interface every time you get a response. You need to know where the answer's going to be, where your links are, where the content is, where I can get to where I need to get to — whether it's the call a business button or what have you. And so there does need to be, still within the longer and longer leash you give AI, a set of human-based rules within this thing.

对。

Yeah.

然后我觉得真正酷的地方在于，它会进入类似“我要为这个设计一个小部件”的状态。比如，给我解释某个数学概念，它会生成一个图表让我——数学不是我的强项，所以我就不举例了，因为刚才那个例子不好。

And then I think where it's really cool is where it gets into, like, “I'm going to design a widget for this.” So, like, explain some mathematical concept to me and it will create a chart that allows me — math was not my strength, so I'm not going to give you an example because that was bad.

好。你能再具体一点吗？

Yeah. Can you get a little bit more specific there?

可以。我们也有一些科学类的例子，比如：需要多少气球才能让一辆自行车浮起来，或者承载一辆自行车，它会为这类问题生成交互式小部件。我们正在积极做这些东西。那些东西真的很好玩。

Yeah. I think we had some scientific ones as well, where it was like, how many balloons does it take to float a bicycle or carry a bicycle, and it'll create interactive widgets for those kinds of things. And we're actively working on that stuff. And that stuff's really fun.

甚至想想像 Google 这样的系统需要多少原语——因为显然你可以拥有所需原语数量的 5 倍，也就是理想数量，或者你是以最小重叠实现完全覆盖？我对这种东西可以变得特别极客。真的很酷。

Even thinking about how many primitives a system like Google needs — because obviously you can have 5x the amount of primitives that are required, like the ideal number, or do you have complete coverage with minimal overlap? I could get so nerdy about that kind of stuff. It's really cool.

嗯，我得说我们正在试着减少一些。

Well, I will say we're trying to reduce a few.

我相信。对。你怎么找到那个中间地带？这很有意思。

I'm sure. Right. How do you find that middle ground? It's interesting.

好，我们聊了很多当今世界重要的技能，以及你希望更多人在内部效仿的最优秀设计师的特质。我想简要谈谈反面。当你回顾自己作为设计师的历程，想想什么重要、什么推动了你，或者你多年来共事过的最优秀的人的特质，有没有某些我们历史上赋予很高价值、但你认为现在重要性正在下降的特征或技能？

Okay, so we talked a bunch about skills that matter in today's world and what the traits of the best designers are that you're hoping more people would emulate internally. I guess I kind of want to touch on the flip side briefly. When you think about your journey as a designer and what's mattered, what's propelled you, or maybe the traits of the best people that you've worked with over the years, are there certain characteristics or skills that we've attributed a lot of value to historically that you think have diminishing importance now?

没有什么是立刻浮现在我脑海、让我说“哦，设计师以前这么做，现在不做了”的。我想说，这更多不是关于设计流程——我觉得其中有些环节被简化了，比如 Figma 里的迭代流程，现在我可以直接创建原型。所以更多是：我绘制或设计的画布是什么？并没有单一的一件事，因为归根结底，我认为设计师的核心就是创意人，我们表达自己的方式可能会变，但技能本身比工具更大。所以我认为角色本身保持不变，工具在变。我做的事情可能会变，但那些硬技能——我能不能解决问题、能不能以系统的方式思考、能不能看着这个说“哦，我已经勾选了它需要解决的所有优先事项”——所有这些仍然非常重要。我想说，选择你想用来发挥创造力的工具。我不是说你必须用所有这些工具。我觉得你可以挑一两个你觉得会是最好的起点。我认为 AI 应该能帮你做到这一点，论坛也可以，同行也可以，你所在的大学也可以帮你选择正确的工具。我只想鼓励你坚持用一个并学会它。没人指望你懂所有这些，或者懂所有工具。

There's not anything that immediately comes to mind that I would say, oh, designers used to do this and now they don't anymore. I guess I would say it's less about the process of design, which I think maybe some of the bits get shortcut — the iterative process in Figma, as much as, okay, now I can create a prototype. So there's a little bit more of, what is the canvas that I'm drawing in or designing in? There isn't one thing, because ultimately at our core I think designers are just creatives, and the way in which we go about expressing ourselves might change, but the skill really is bigger than maybe the tools are. So I think the role itself stays the same; the tools change. So what I do might change, but the hard skills around can I problem solve, can I think about things in a systemic way, can I look at this and say, oh, I've checked all the boxes of the priorities of things it needs to solve — all of those things still really matter. I would say choose your tools that you want to use to be creative. I wouldn't say you need to use all of these tools. I think you can pick one or two maybe that you feel like are going to be the best starting point. And I think AI should be able to help you with that, and so can forums, so can your peers, and so can wherever you are in university — can help you sort of choose what the right tools are. And I would just encourage you to stick with one and learn that. No one expects you to know all of this stuff or know all of the tools.

好。

Okay.

那么，我们能对正在收听、可能感到有点不知所措的人说几句吗？因为感觉事情越来越多，越来越多，而且你甚至没有明确回答，你知道，我们可以放下什么？就像，不，不，继续捡起东西。现在就做这个。现在就做这个。做这个。现在，最后一件事，也许我们可以以此收尾，如果你假设，也许是个傻问题，但就像如果明天你有能力在 Google 搜索顶部放一个横幅，每个设计师都会看到，你会在那个横幅上放什么信息？

So, can we talk to the person who's listening who maybe is feeling even a little bit overwhelmed because it feels like it's just more and more and more, and the fact that you didn't even have a clear answer to, you know, what can we set down? It's like, no, no, just keep picking stuff up. Like, do this now. Do this now. Do this. Now, the last kind of thing that maybe we can ride off into the sunset on is if you hypothetically, maybe a silly question, but like if tomorrow you had the ability to place a banner at the top of Google search that every designer would see, what kind of message would you put on that banner?

搜索顶部的横幅，我会说，你知道，AI 一直是一份惊人的礼物，你会因为它做出你一生中最好的设计。

The banner at the top of search, I would just say like, you know, AI has like been this amazing gift and you're going to do the best design of your life because of it.

嗯，我想不出比这更鼓舞人心的结束语了。所以，你知道，你给了我们一个窗口，让我们看到我们都看到但不太了解如何运作以及发生了什么的世界。所以，我非常感谢你拉开帷幕。

Well, I can't think of a more inspiring note to end on than that. So, you know, you gave us a window into a world that we all see, but like we don't really know how that all works and and what's going on. And so, I super appreciated you pulling back the curtain.

完全。

Totally.

只是多听一点你的旅程。真的很鼓舞人心。

Just hearing a little bit more about your journey. It was really inspiring.

谢谢邀请我。是的，聊天很棒。

Thanks for having me on. Yeah, it's great to chat.

嘿，如果你喜欢这一集，别忘了每周四我和 Tommy Gio 在 Dive Radio 直播，我们深入探讨热门话题，并听取业内众多不可思议的设计师的意见。所以，如果你想把它加入日历，只需访问 dive.comclub/radio。我们周四下午 4 点东部时间见。

Hey, if you enjoyed this episode, don't forget that every Thursday I go live with Tommy Gio for Dive Radio, where we deep dive into trending topics and hear from a ton of incredible designers in the industry. So, if you want to get that on your calendar, just go to dive.comclub/radio. And I'll see you on Thursdays at 400 p.m. Eastern.