乔尼·艾维的奇特法拉利

我喜欢那个玻璃质感的换挡杆。

I like the glass key stick.

还有钥匙也是吗？

And the key as well?

钥匙设计真酷。

The key is very cool.

钥匙确实很酷。

The key is very cool.

没错。

Yeah.

他们聊得好像只要介绍视频里大家夸钥匙有多酷就够了，但其实这辆车真正重要的地方他们完全没提到。

They're talking like like, if if your introduction video is them agreeing on how how cool the key is, you've missed something important about

到底该是什么样子。

what the car

才是它该有的样子。

is supposed be.

当然。

Sure.

就像你说的，这辆车有很多细节确实很出色。

Like you said, there are a ton of details about this car that are very cool.

我完全同意。

No disagreement.

但它

But it's

不是法拉利？

not a Ferrari?

它不是法拉利。

It's not a Ferrari.

但如果你只听了这段音频，我建议你去看看那个视频。

But if you you if you were only just listening to that audio, I encourage you to go watch that video.

视频就在法拉利官方社交媒体账号上，顺便说一句，他们所有视频的评论区都在吐槽这辆车。

It's just on the main Ferrari social channel, which, by the way, all of their videos are getting roasted in the comments about how about the car.

我有点过意不去。

So I feel I feel bad.

但那个视频就在法拉利的官方社交频道上。

But, like, that video is just on the main Ferrari social channel.

你可以看看刘易斯和查尔斯是怎么努力表现出对这辆车的兴奋的。

So you can watch, like, Lewis and Charles, like, try to be excited about the car.

他们脸上的表情全写出来了。

Like, this is all over their faces.

是啊。

Yeah.

然后他们上了车，还真的在赛道上跑了几圈。

And then they're in the car, and, you know, they do, like, hot laps in the car.

看到真正的F1车手硬要装出这车快得惊人，特别好笑——毕竟他们平时开的可是F1赛车。

And it's very funny to see literal Formula one drivers have to pretend the car is shockingly fast because it's just like, they're Formula one drivers.

这车哪有他们平时开的那么快啊。

Like, it's not as fast as their their day job.

没错。

Right.

这速度要是真跟他们平时开的一样，那可就违法了。

It's not it's it's literally illegal for it to be as fast as their

他们的日常工作。

day jobs.

我也不太清楚，各位。

I don't know, guys.

现在市面上快车多的是。

Like, there's a lot of fast cars out there.

你还能看到他们开着更快的车在摩纳哥赛道上飞驰。

Like, you can watch them drive around Monaco in much faster cars.

我知道。

I know.

我这么说吧。

Here's what I'll say.

其实我对美国电动车这个说法不太认同。

I I actually disagree with you about The US EV thing.

我给你举个小例子吧。

And I'll give you one tiny example.

我们有一辆凯迪拉克Vista。

We have a Cadillac Vista.

这是款三排座的凯迪拉克电动车SUV。

It's the three row Cadillac EV SUV.

外观跟凯雷德一模一样，就像个迷你版凯雷德。

It looks just like a Escalade, like a baby Escalade.

大家都特别喜欢这车。

Everyone loves this car.

我们当时入手的价格特别划算。

Like, we got a great deal on it.

那时候正好是拜登政府快结束的时候，各种补贴政策都在放，经销商急着清库存，以为市场要崩了。

It's sort the end of the Biden administration where all the tech credits were flying, the dealers are trying to move all the cars because they thought the market was collapsed.

太好了。

Great.

我们买这辆车花的钱，和以前买吉普车一样多，因为当时这车优惠实在太大了。

We pay as much for that car as we did our old Jeep because it was such a great deal all the time.

大家都特别喜欢这车。

Everyone loves this car.

我们停好车，总有人过来问：这车是啥？

Like, we park it, and people are like, what's that?

我就说：这是电动车。

And, like, I'm like, it's electric.

他们就哦一声。

And they're like, oh.

但这其实不是我想说的重点。

And that's really not the point of the story.

重点是：不对。

The point of story is No.

但这才是故事的重点。

But that is the point of the story.

它看起来像一辆凯迪拉克凯雷德，只不过用的是电池，这才是关键。

The it looks like an Escalade, but it happens to have a battery is the is the correct answer.

因为油价涨得太离谱了。

Into it because gas prices are out of control.

对吧？

Right?

所以大家都说，哇，这太棒了。

And so, like, everyone's like, oh, that's great.

现在开电动车其实特别省钱。

And now you use EVs are really cheap.

但真正有意思的是，我们拥有这辆车已经快一年了。

But the real turn is we're about a year into having this car.

对吧？

Right?

贝基开了我的汽油车，那车我平时根本不碰，就停在车道上，纯属应急用的，比如需要启动发动机的时候。

And Becky drove my gas car, which I never drive, which, like, lives in the driveway for in my case of emergency, like, start engine.

对。

Sure.

有次我们碰巧得去两个不同的地方，这种事很少发生。

And we we just needed to be in different places for in one of the rare occasions, we needed to be in different places.

她就开了我的车去。

And she drove my car.

她回来后，我就问她：开起来感觉怎么样？

And she came back, and I like, was it fun?

杰夫，你觉得呢？

Jeff, fun?

她说：开完这辆大SUV之后，你的车简直像卡丁车一样。

She's like, your car is basically a go kart after driving this giant SUV.

我就问：你开得快吗？

And I was like, did you go fast?

她说，我也没开多快。

She's like, I didn't go fast.

事实上，我现在特别讨厌开油车，因为我爱开电动车，根本不想再回头了。

And in fact, I hate driving gas cars now because I love driving the EV, and I don't wanna go back.

这基本上也是所有汽车高管都会说的话。

And this is by and large what all the car executives will tell you.

对吧？

Right?

只要你真正体验过那种生活——再也不用去加油站，开车时感受那种瞬间加速和强劲扭矩，你就再也回不去了。

That once you, like, live the lifestyle and you never go to the gas station anymore and you drive the car and you feel the acceleration and the instant torque and the blah blah blah blah blah, you're not gonna go back to a gas car.

他们就是没法让更多人转投电动车。

They just can't get enough people into the EVs.

比如Rivian R2，这款即将停产的车，它的主要卖点是——Plug Wasim，Rivian软件负责人，我们之前也在那个播客里聊过他。

And so, like, the Rivian r two, the big argument for this car, which is on its way out, Plug was sim, who is the head of Rivian software, just on the coder we were talking on there too.

它基于全新的平台打造。

It's on on the new platform.

他的核心观点是，只要有一款主流车型，外观符合美国人喜欢的SUV样式，人们就会愿意上车，然后我再给他们一套远超燃油车的软件体验——毕竟这车就像一台电脑，能运行搭载AI技术的Rivian助手。

His whole argument is, like, once you have the mainstream car that looks like the SUV Americans wanna buy, they will go into the car, and then I will deliver them a software experience that's superior to the gas car because it's a little computer, and it can run Rivian assistant with the power of AI.

就像《Decoder》每一期节目一样，我们都在谈AI的力量。

And just like every episode of decoder, we talk to the power of AI.

是的。

Yeah.

我觉得其中三分之二说得对。

I think two thirds of that is correct.

但一旦涉及到语音助手驱动的用户界面，整个系统就彻底崩了。

And then you hit voice assistant powered user interface, and the whole car explodes.

全盘崩溃。

It all falls apart.

不过我觉得你真正点出的是，人们喜欢特斯拉，是因为特斯拉的软件确实出色。

But I I think you're what what you have really identified is people love Teslas because the software in Teslas is great.

对吧？

Right?

开特斯拉的那种软件体验，每年都能收到系统更新，整个体验连贯一致，这就是人们喜欢特斯拉的原因。

The software experience of being in a Tesla, of having that all be one experience of over the year updates, this is why people love their Teslas.

没错。

Sure.

其他大多数车企的做法是：车是电动的，但软件太差了。

Most other carmakers are like, it's electric, and it has bad software.

大多数人实际感受到的就是这种体验，他们并不想要其他那些东西。

And that's the experience most people are having, and they don't want the other stuff.

也许吧。

Maybe.

也许吧。

Maybe.

但关键是，这车毕竟是电动的。

But it's also it it's electric.

而且软件确实很烂。

It has bad software.

而且你不觉得它长得特别像宇宙飞船吗？

And isn't it cool how much it looks like a spaceship?

根本不是这么回事。

It's like, no.

对啊。

Yeah.

人们喜欢法拉利，是因为法拉利够拉风，兄弟。

People like Ferraris because Ferraris are hot, dude.

没错。

Like Yeah.

这辆车有很多特点。

This this car is many things.

但它一点都不性感。

It is not sexy.

法拉利Luce根本谈不上什么性感。

Like, there's nothing sexy about the Ferrari Luce.

对于那些觉得法拉利就应该造性感车的人来说，这其实是法拉利的一个基本事实，老兄。

For our that would like, a fundamental fact of Ferrari is that it makes sexy cars, dude.

就像说——

Like

乔尼·艾维最近做的什么东西是性感的？

What was the last thing, Jony Ive made that was sexy?

这是个很好的问题。

It's a terrific question.

当他伸手去转中控台那个旋钮的时候，我确实有点心动。

I will say, I did have feelings when he grabbed the handle on the the center console thing and turned it.

那才叫性感，老兄。

That's hot, dude.

这辆车有太多细节能让人觉得性感，但它本身并不性感。

There are so many little details about this car that are sexy, but it is not a sexy car.

我觉得iPhone 4才是最后一个让人觉得‘哇，真性感’的产品。

I would argue that the iPhone four was the last, like, oh, that's sexy.

很有魅力的物件。

Charming object.

但你知道，我说的是，你知道我对iPhone 4的看法。

But, you know, you mean, you know how I feel about the iPhone four.

苹果过去五到十年的设计，完全没有任何性感可言。

Like, Apple design for the past, I don't know, five years, a decade, has been totally sex free.

没错。

Yes.

当然。

Yeah.

毫无疑问。

For sure.

不过，确实如此。

But, like yeah.

我觉得，说这是乔尼·艾维被自己的设计理念冲昏了头，反而可能是最简洁的解释，说明这辆车为什么既有优点也有缺点。

And I I I think the idea that this is Jony Ive high on his own supply, I actually think is maybe the Occam's razor explanation of why this car is the way that it is, both good and bad.

因为他聊了很多，我觉得。

Because there are I think he he talked a lot.

他接受了克莱奥·亚伯拉姆的长时间采访，在她的YouTube频道里谈了很多，说不想放弃数字屏幕和这种交互式界面，但也不希望所有东西都变成触屏。

He did a long interview with Cleo Abram on her YouTube channel as part of this and talked a lot about not wanting to give up on digital screens and and that kind of interactive interface, but also not wanting to have everything be a touch screen.

这辆车里有很多细节，把模拟体验和数字体验结合得特别棒。

And there are a lot of details in this car that are really cool mixes of analog and digital experiences.

对吧？

Right?

比如那个钥匙就是个好例子。

Like, key is a good example.

把钥匙插进去一按，他就特别兴奋，说钥匙的颜色会移到挡把上，暗示车已经启动了。

You put the key in, you push it down, and he he got really excited about how the the color of the key sort of moves over to the gear shift as a way of communicating like the car is now on.

车上到处都有这种小细节，真的又酷又有意思。

And there there there are little details like that all over the car that are genuinely very cool and interesting.

如果你把法拉利看作汽车设计的先锋，然后这些理念慢慢渗透到凯美瑞这样的车里——我不确定是不是真这么发展，但这里面确实有让人兴奋的可能。

And I think if you look at Ferrari as sort of a a tip of the spear of car design and eventually this all trickles down into Camrys, I don't know if that's actually how it works, but there's something potentially exciting there.

但在这个案例中，这个缺口远小于所有这些因素的总和。

But the hole in this case is much less than the sum of all of those things.

我觉得这正是人们真正被吸引的地方。

And I think that's what people are really responding to.

对。

Yeah.

它必须看起来像一辆法拉利。

It needs to look like a Ferrari.

就像乔尼·艾维的设计风格，几乎是必然的。

Like, you know, the Jony Ive design aesthetic is it's inevitable.

正如我常说的，限制反而会成为产品的核心特征。

And the you know, as I always say, the limitations become the defining feature of the product.

但在这里，法拉利本不该有什么限制。

And here, it's like, Ferraris aren't supposed to have limitations.

你觉得，如果乔尼·艾维得设计一辆必须控制在凯美瑞价格水平的车，会不会更有意思？毕竟，这款车可是有史以来最贵的法拉利。

Do you think it would have been more interesting in that respect if Jony Ive had had to design a Camry that had to cost like a Camry as opposed to because in that case, this is the most expensive Ferrari ever.

这可是毫无保留，全力以赴。

Like, this is truly no holds barred.

想怎么设计就怎么设计。

Do whatever you want.

我觉得他确实有些地方做得很好，也有一些地方没做好。

And I think that there are clearly some ways in which he got it right and some ways in which he got it wrong.

但如果你跟乔尼说，你得做个售价三万五千美元的车，那这个项目可能会有趣得多。

But I think if you were like, Johnny, you have to make a car that is $35,000, it this might have been a vastly more interesting project.

那当然了。

Oh, without question.

乔尼·艾维的超能力，就是能在限制中突破限制。

Again, the limitations Jony Ive overcoming constraint limitation is the superpower.

没错。

Yeah.

他所有成功的产品，无一不是这样。

And it is true of every product he's ever made that has been a hit.

对吧？

Right?

初代iMac的限制在于CRT显示器，而他偏偏让这个部件成了产品的亮点。

The the constraint and limitation of the first iMac was the CRT display, and he made that thing the star of the show.

没错。

Yeah.

这种做法贯穿了他设计的每一款产品。

And that just goes on through all of the products.

AirPods必须内置天线。

The AirPods needed to have antennas in them.

结果我们就有了那对小柄。

Now we have little stems.

你每次都能看到他这么干。

Like, you you just see him do it every time.

他总能把让产品设计变得困难的限制，变成最亮眼的设计元素。

He makes the thing that makes the product design hard into the star of the show.

法拉利通常主打的是V12引擎，但这次你根本没有这个东西。

And with Ferrari, usually, star of the show is the v 12 engine, which you do not have in this case.

那你能怎么办呢？

So, like, what are you gonna do?

是的。

Yeah.

对吧？

Right?

我觉得丰田凯美瑞的限制条件反而会更有趣，这本来才是苹果造车的初衷。

Like, I I think the constraint of a Toyota Camry would have been much more interesting, and and that was the promise of the Apple car.

你知道什么样的车设计起来更有意思吗？

You know what a more interestingly designed car is?

就说现在吧，Slate卡车的预售已经开了，其实就是把电池和电机塞进一个箱子里。

Like, as we're speaking, preorders for the slate truck have opened up, which is just basically, like, an EV battery and motor in a box.

是的。

Yeah.

这简直都算不上一辆车。

It's like barely a car.

对啊。

Yeah.

售价都快四万美元了。

It's like $40,000.

这东西设计得真有意思，这么实用，完全由你来决定怎么用。

And you're like, this is a fascinating design object because it is so utilitarian, and it is so designed for you to do whatever you want with.

说实话，这产品更吸引人，因为它压根不带任何主观设计倾向。

And you're like, oh, this is a much more interesting product because the the way it is designed is to not have any opinions at all, really.

把这两样东西放一块儿，看看大家的反应，尤其是对Slate卡车价格的反应，挺有意思的。

And that it's funny just, like, put the two things next to each other and see how people react, maybe probably to the price of the slate truck.

没错。

Sure.

而且你还能把车顶拆了，换上别的车顶，想改成什么车就改什么车。

But also the idea that you can, like, take off the roof and put on a different roof and make it whatever car you want.

这和那个蓝色 blobs 是法拉利未来的样子，感觉完全不一样。

And, like, that's a lot different vibe than this blue blob blob is the future of Ferrari.

没错。

Right.

代表了汽车设计的巅峰之作。

And represents the absolute pinnacle of car design.

对吧？

Right?

没错。

Like Yep.

法拉利自带一种自我满足感，我觉得这种态度也让不少人感到不舒服。

There is a self satisfaction that comes with Ferrari that I think also kinda rubbed people the wrong way on this one.

意思是，既然它是法拉利，那我们肯定是对的。

That is like, by definition of the fact that it's Ferrari, we are correct.

对吧？

Right?

有这个可能

There's

你得明白，法拉利内部所有人都心知肚明。

You have to think everyone at Ferrari knew.

我倒想知道。

I wonder.

想象一下那些会议，他们说：咱们现在都在搞社交推广。

Like, imagine the meetings where they're like, we're we're all at the social plan.

咱们是不是该请刘易斯和查尔斯来拍个社交视频？

Like, are we gonna get Lewis and Charles in, and we're gonna do a social video?

结果社交视频的制作人说：各位，我觉得这事儿不太合适。

And the social video producer is like, guys, I don't I don't know we should do that.

但在意大利人听来，天哪，这下糟了。

But in Italian, like, oh, no.

到某个时候，他们才想起来，应该让汽车设计师和乔尼提前碰面，可那时候已经晚了，结果就……

At some point, they they forgot to put the car designer in a room with Jony until it was too late, and then they're

天哪。

like, oh god.

他还真以为自己做了一辆车呢。

He thinks he made a car.

确实。

Yeah.

说实话，这完全有可能。

It's honestly it's very possible.

我的天哪。

Mama Mia.

顺便问一下，我很好奇你们大家觉得这辆车怎么样。

I'm curious to know, by the way, what all of you think of this car.

我觉得，尤其是现在新闻热度稍微降了一些，大家都有时间看看照片，消化一下这些信息。

I think, especially now that the the news has died down a little bit, everybody's had time to, like, look at the pictures, digest it.

如果你特别喜欢Luce，一定要告诉我你的想法。

If you love the Luce in particular, I wanna hear from you.

欢迎收听《The Verge》播客，电话号码是866-11。

Vergecastoftheverge.com, 86611.

告诉尼莱和我，我们哪里错了，为什么乔尼·艾维才是对的。

Tell tell Nilay and I why we're wrong and why Jony Ive is right.

我是真的很好奇。

I I'm I'm genuinely curious.

我特别期待乔尼·艾维打来电话。

I'm excited for Jony Ive to call.

就像底特律的史蒂夫会说，你看吧。

Like, Steve from Detroit is gonna be like, look.

这早就注定了。

It was inevitable.

好吧。

Alright.

我们休息一下，之后再聊一堆人工智能的新闻。

We should take a break, and then we have a bunch of AI news to talk about.

我们马上回来。

We'll be right back.

本节目由Odoo赞助。

Support for this show comes from Odoo.

经营企业已经够难了，干嘛还要用十几个互不相通的软件让自己更麻烦？

Running a business is hard enough, so why make it harder with a dozen different apps that don't talk to each other?

这就是Odoo。

Introducing Odoo.

你这辈子只需要这一套商业软件。

It's the only business software you'll ever need.

它是一个全方位整合的平台，能让您的工作更轻松。

It's an all in one fully integrated platform that makes your work easier.

客户关系管理、财务、库存、电子商务，还有更多功能。

CRM, accounting, inventory, ecommerce, and more.

最棒的是？

And the best part?

Odoo 以不到几分之一的价格，取代了多个昂贵的平台。

Odoo replaces multiple expensive platforms for a fraction of the cost.

所以才有成千上万家企业选择了切换到 Odoo。

That's why over thousands of businesses have made the switch.

那你为什么还不试试呢？

So why not you?

去 odoo.com 免费试用 Odoo 吧。

Try Odoo for free at odoo.com.

就是 odoo.com。

That's odoo.com.

好的。

Alright.

我们回来了。

We're back.

尼莱，现在该聊聊我们一直说的事了——大多数人好像特别讨厌人工智能。

Nilay, it's it's now time for us to talk about a thing we keep talking about, which is that most people seem to hate AI a whole lot.

每次一有AI相关的消息出来，大家的反应几乎都是：别来了。

And and that every time AI things happen, the overwhelming response is, no.

谢谢。

Thank you.

没错。

Yeah.

求求你，快停下吧。

Please make it stop.

有个挺有意思的数据，我们上周聊了很多关于Google I/O的事，就是那些全新的AI搜索功能。

One interesting bit of data, we talked a lot about Google IO last week, the the all this new AI search stuff.

谷歌正在真正地把搜索框改造成一个AI产品。

Google is changing the search bar to make it an AI product in a very real way.

DuckDuckGo是个搜索引擎竞争对手，它一直以不被AI淹没为傲。

DuckDuckGo, a a search engine competitor that is sort of proudly not being overrun by AI.

现在这已经变成一种营销手段了。

This is now a a marketing thing.

有个叫Vivaldi的浏览器，最近在我所有动态里刷屏，它主打的卖点就是‘我们是不搞AI的浏览器’，结果还真管用。

Like, there's this browser, the the Vivaldi browser, that is all over all of my feeds now, and its entire pitch is we are the non AI browser, and it's working.

这简直太离谱了。

Like, it's bananas.

不过话说回来，DuckDuckGo说，自从这事出来后，他们在美区的iOS应用安装量周环比涨了33%，要说原因，恐怕就是用户想逃离谷歌了。

But, anyway, DuckDuckGo said that its own iOS app installs went up by 33% week over week in The US after this stuff came out, and there's just nothing to attribute that to except people are trying to run away from Google.

但这个增幅其实意义不大，毕竟33%是从一个很小的基数涨上去的，谷歌早就统治了这个市场，而且至今仍牢牢掌控着。

And that is going to be a very small number, up 33% to another very small number for DuckDuckGo because Google won this market a long time ago and continues to own it.

但确实，用户流失是真实存在的，大家开始寻找替代品也是事实。

But, like, the the attrition there is real, and people starting to look for something else is real.

我觉得，离开谷歌是互联网上最难改变的用户习惯之一，但现在，有人已经开始这么做了。

I think leaving Google is one of the hardest user behaviors on the Internet, and people are starting to do it.

把iOS安装量和搜索行为拆开来看，其实挺有意思的。

So I it's interesting to pull apart iOS installs and search behavior.

对吧？

Right?

你打开手机上的谷歌应用，里面全是让你用Gemini产品的提示，比如让你把照片调成亮蓝色，因为AI现在能帮你做到这一点，你不觉得烦吗？

So if you open the Google app on your phone and it's just loaded with prompts to use Gemini products and don't you wanna read tint your photos, like, bright blue because AI can let you do that now?

也许你会去下载DuckDuckGo，但谷歌搜索的使用可能根本不受影响。

Like, maybe you go and install DuckDuckGo, but Google Search might not be affected.

而且，正如你所说，DuckDuckGo的下载量本来就很低。

And, also, the amount of DuckDuckGo installs was small, as you're pointing out.

对。

Right.

我明白你的意思。

So I I I see it.

我特别小心，不会因为一个数字就过度解读，毕竟谷歌搜索的体量摆在那里。

I'm very careful not to overread, like, what one number means in the context of how big Google Search is.

是的。

Yeah.

这并不意味着我觉得谷歌搜索没变得越来越奇怪、越来越荒唐，也越来越难预测，我觉得这真的很糟糕。

That's not to say I don't think Google Search is getting weirder and goofier, and it's less predictable, which I think is really bad.

是的。

Yeah.

新的智能搜索框会让你不知不觉陷入无尽的深入探索中。

Like, the in the new intelligent search box can just lead you down infinite rabbit holes.

你有没有这种经历：搜个东西，点开一个结果，又跳到另一个，然后你脑子里就记住了这条路径？

Like, you know, the do you have this behavior where, like, you do a search and you kinda click one result and that leads you to another result and you kinda remember the pattern in your brain?

对。

Mhmm.

于是你得拼命复现那个路径，才能回到原来想找的地方。

And so you're like, I need to recreate that pattern to get back to where I was going.

新版的谷歌搜索根本不让你这么做。

The new Google Search does not let you do that.

没错。

No.

每次都是全新的界面，完全不一样。

It is much more whole cloth every time.

每次都不一样。

It's random every time.

现在它甚至可能给你推一个应用。

Now it can, like, might make you an app.

这种改变背后有种意思，他们希望我们变得特别聪明，用多模态和你对话，结果你没法再靠之前的关键词路径找到原来的地方。

Like, there's something about that change where, you know, they want us to be so intelligent and talk to you in multimodal that you can't pattern match your way towards, like, I've used these keywords to get here in this path.

我觉得他们该改一改。

I think they ought to change that.

我觉得这已经不只是AI的问题了。

I I do think that's, like, beyond just the AI of it all.

打破这种使用习惯对谷歌来说很危险，因为用户随时可以去别的地方找同样的服务。

Breaking that behavior is is dangerous for Google because then you can go get that anywhere else.

是的。

Yeah.

我搜索某个东西的时候，每次结果都应该保持一致才对。

The idea of when I search something, the same thing should probably happen every time

对。

Yeah.

这是个很有趣的问题。

Is such an interesting question.

因为还有一个情况，就是谷歌的一大收入来源，其实是你搜索效果差的时候。

Because there's also a version of this, like, a big thing that Google does, right, is make a lot of money when you do a bad job of searching.

因为直到现在，你都得回到顶部重新开始。

Because until now, you've had to go back to the top and start over.

这些年，这已经成了谷歌的一个大问题，因为它原本有很强的动机去让搜索做得不够好——因为你搜索次数越多，谷歌赚得越多，而谷歌一直最看重的就是搜索量增长。

And and, actually, this has become a huge problem for Google over the years because it had a lot of incentives to not be a very good search engine because you would do more queries, and everything at Google has always been about query growth.

而让你多搜几次的一个办法，就是第一次别给你正确答案。

And one way to get you to do more queries is to not give you the correct answer the first time.

但这次这种像抽盲盒一样的感觉，反而让体验变得更糟了。

But in this case, this idea of it feeling like a grab bag does just make the experience worse.

我现在去谷歌搜索，已经不像以前那样有信心了。

Like, I no longer go to Google sort of confident in in what's going to happen.

对吧？

Right?

就算答案是，每次搜索都会显示一个AI概览，右边再堆一堆链接，如果你想看更多内容。

Even if the answer was, we're gonna have an AI overview for every single thing that you search and and then a bunch of links on the right side, and then if you wanna make more stuff.

如果这就是谷歌最普遍的体验，那倒也说得通。

If that was just the sort of most common denominator Google experience, that would make sense.

但事实不是这样。

It's not.

确实不是。

No.

你一搜索，完全猜不到接下来会看到什么。

You just you search, and you just have no idea what's about happen.

你开启谷歌的上下文功能了吗？

Context turned on in your Google?

我开了。

I do.

我发现，由于它总在记住你搜过的东西，还读你的邮件，搞一堆个人化的情境分析，结果不仅答案越来越难预测，还特别自信地误解我。

So the thing I've discovered with that is because it's constantly remembering other things you've searched and, like, reading your Gmail and doing all the personal context stuff, not only are the answers getting less predictable, they're getting more, like, confidently wrong about me.

是的。

Yeah.

你有没有注意到，谷歌总假装它跟你住在同一个地方？

And so oh, have you noticed Google also pretends it lives where you live?

哦，这挺有意思的。

Oh, that's interesting.

没有。

No.

这可能是个小例子，但我们把空调换成了热泵。

So maybe this is just a silly example, but we we swapped out our air conditioners for heat pumps.

所以呢，你就搜搜看，这型号是怎么回事，工作原理是什么。

And so, like, you know, just googling at what model, like, how does it work?

我该不该搭配我的Ecobee一起用？

Should I use my Ecobee's with it?

就是这些乱七八糟的东西。

Like, all this stuff.

到最后，它居然说我们住在威斯特彻斯特县，这个型号特别棒。

And by the end, it's like, you know, where we live in in Westchester County, like, this is a great model.

结果第三次搜索却说：天啊，不对。

And then, like, the third search is like Oh, no.

你明明知道我已经买了那些热泵，我根本没买过任何东西。

You already bought those heat pumps, like, I didn't buy shit.

你到底在说什么？

Like, what are you talking about?

你根本不像个朋友。

You know, it's like, you're not my friend.

我只是想顺利找到答案而已。

Like, I just need to get down the path.

不过里面也有一些部分确实很有用，特别管用。

And there were some parts of it that were really useful, like, spectacularly useful.

我们现在家里装了太阳能板，打算换成这种热泵。

Like, I we have solar panels in the house now, and we're gonna switch to these heat pumps.

天然气价格实在太贵了。

The, you gas prices are really expensive.

但我们装太阳能板的时候，还没定下这个方案呢。

Like but we sized the solar panels before we had the thing.

你能帮我算一笔账吗？

Can you run the math for me?

空调更省电了，我真能省钱吗？

Like, am I gonna save money because the air conditioning is more efficient?

就是些琐碎事。

Blah blah.

典型的爸爸式考虑。

Like, real dad stuff.

是的。

Yeah.

家里没人愿意跟我聊这个，只有Gemini愿意。

Like, no one in this house wants to talk to me about this at all, but Gemini does.

你知道，这正是任何

You know, this is the promise of any

没错，确实如此。

sure that's true.

但我真不想去期待。

But I can't want to.

不。

No.

Gemini还说呢，你和我，哥们儿，住在韦斯特切斯特。

Gemini was like, you know, you and me, buddy, living in Westchester.

它还真做到了。

And it, like, did it.

它帮我算出了账，我用好几种方法都核对过，结果真没错。

It, like, figured out the math for me, and I checked it a bunch a bunch of different ways, and it actually gotten it right.

这些其实都非常有用。

And all this is really useful.

但下次我搜索的时候，它却开始聊起我之前另一次的互动来了。

But then the next time I ran a search, it was start it was just talking about this other interaction that I'd had.

没错。

Right.

这就是个人化的上下文。

And, like, that's the personal context.

所以就连那些在我看来完全不相关的搜索，也被我的过往经历影响了。

So it's like, even the other search, which in my mind is totally unrelated, is, like, infected with my own history.

你根本看不出这些怎么就能派上用场。

And it like, none of that you can see how it will get useful.

对吧？

Right?

桑达尔刚在《Decoder》上说过，这些东西迟早会融合在一起。

Sundar was just on Decoder, and he's like, these things are gonna come together.

比如说，你有个平台可以创建应用，搜索里的AI模式能直接给出答案，还有Gemini Spark这个智能代理平台。

Like, the idea that you have the canvas where you make the apps and the AI mode in search, we get the answers, and then Gemini Spark, the agent platform.

最终，你只要说‘我想找这个’，它就会直接帮你买下来，而不是给你一堆搜索结果。

And so eventually, you're gonna say, I'm looking for this, and it goes off and just buys it for you instead of showing you a search result.

或者它替你预约服务，或者帮你规划假期，你只要选好想去的地方和时间，它就自动帮你订好一切。

Or it books you the service appointment on your behalf, or it makes you the vacation planning app, and then you pick the places you wanna go on what schedule, and then it goes and books them for you.

这些功能迟早都会整合到一起。

This is all obviously gonna come together.

但问题在于，你描述的这种做法，其实会打乱原有的使用体验。

But the thing you're gonna do is what you're describing is you're gonna destabilize the actual experience.

每次使用都会变得更不可预测。

It will be more random every time.

是的。

Yeah.

而且我觉得，你刚才说的，其实是一种误解——以为现状就是最糟糕的状态。

Well and I think, again, you're you're describing what I think of as the this is the worst it will ever be fallacy Mhmm.

这其实就是每个人为一切辩解的方式。

Which is how everybody defends everything.

对吧？

Right?

你说这居然会是史上最差的版本，难道不觉得离谱吗？

Like, isn't it wild that this is the worst this is ever gonna be?

好像科技进步就是线性的、必然的，而且永远完美无缺似的。

Like, as if the march of progress was linear and inevitable and and perfect forever and ever.

但 Sundar Pichai 实际上说的是，我们正在让全网所有谷歌用户参与这场持续的测试，直到真正把这产品做好。

But what what Sundar Vichai told you, it cast another way, is we are running beta tests on every Google user on the Internet until we figure out how to make this product actually good.

我们会先让你用糟糕的版本，就是为了找出怎么把它变成好的版本。

We are going to subject you to the bad version of this in order to figure out how to make the good version of this.

现在发生的，就是这么回事。

And that's what's happening.

我觉得很多人对这些 AI 产品的不满，正是因为你在明明白白地告诉他们：这产品现在根本不好用。

And I think that's what a lot of people are reacting to with these AI products is you're you're actually telling me this product isn't any good.

但你的意思是，你得用它，我们才能把它做好。

But you're saying you have to use it so that we can make it good.

对。

Right.

因为如果不让你用，我们就没法改进。

Like, because we can't unless we make you use it.

所以我们干脆把你按住，把眼睛撑开，逼你用这个产品，好让我们收集数据，把它改好。

So we are going to basically, you know, sit you down and tape your eyes open so that you have to use this product so that we get what we need in order to make it good.

而在这整个过程中，谷歌现在提供的根本就不是什么好产品。

And nowhere in that trade is Google today is a good product.

你有没有看过我之前放的那个片段，关于搜索结果的？

Did you see the clip where I showed sooner or the search results?

他们根本不喜欢那样。

They don't they don't like it.

我每次做这个的时候，都特别享受。

When I do this, I I love doing it.

我每次都这么做

And I always do

最好的Chromebook。

best Chromebook.

最棒的Chromebook。

Best Chromebook.

对吧？

Right?

是的。

Yeah.

没错。

Yeah.

就是个很实用的搜索。

It's it's it's just a useful search.

对吧？

Right?

谷歌应该心里有数才对。

Like, Google should know.

你们要是做Chromebook的公司，理应知道哪款才是最好的。

Like, if you are the company that makes Chromebooks, you should know which one is the best one.

没错。

Yes.

其实很多情况下，你根本不需要上网去查哪款是最好的。

Like, you know, in a lot of ways, you should not need the web to tell you which one is the best one.

这本质上就是一个很商业化的搜索。

And it's just a really commercial query.

对吧？

Right?

The Wire Cutter这家公司，完全是靠‘最佳Chromebook’这个关键词做起来的。

Like, the wire cutter built an entire business on the back of ranking for best Chromebook.

没错。

Yep.

每个人都想拿到这个排名。

Everybody wants to win that rank.

我们也是这么做的。

We do it.

大家都这么做。

Like, everybody does it.

你还可以在页面里放购物链接。

And there's you can put shopping links in there.

所以，一个关于‘最佳Chromebook’的搜索结果页该怎么排版，对我来说，完全体现了这个产品的核心优先级。

So just how you organize a a page, a search results page for best Chromebook, like, to to me, in my mind, just demonstrates a lot about the priorities of the product.

我给桑达尔看了，他说得特别有意思。

And, you know, so I I showed Sundar, and he he said this really interesting.

他说，这个页面的观点太鲜明了，不该这样。

He said that page is more opinionated than it should be.

这个AI给出的结果更有主见，因为它直接给了我一个答案，但这个答案和The Wirecutter的不一样，也和Tom's Guide的不一样。

Like, this AI result is more it because it just told me an answer, which is not the same answer as wire cutter, which is not the same answer as Tom's guide.

我觉得这特别有意思。

Like, I found that really interesting.

对吧？

Right?

谷歌给出的答案和其他的都不一样。

The the Google answer is different than the other ones.

它推荐的是一台联想笔记本。

It told you a Lenovo laptop.

对吧？

Right?

我确实看到了。

I did.

而且还是台挺贵的。

It was, like, an expensive one too.

是的。

Yeah.

它现在告诉我的这个也不是。

It's also not the one it just tells me right now.

对。

Yeah.

真奇怪。

Weird.

对吧？

Right?

有时候它会说是宏碁。

It it says the acer some some of the time.

是啊。

Yeah.

真奇怪。

Weird.

谷歌连哪个语法最好都判断得这么不稳定，这不挺奇怪的吗？

It's it's weird, right, that, like, Google is not even consistent knowing which was the best grammar.

这并不是关键所在。

This is not the important thing.

我知道这没有标准答案，这个页面可以有上百万种不同的设计方式。

I I recognize that it's a it's not a subjective answer, and you can have you can make that page a billion different ways.

但如果你看我采访 Sundar 的视频，就会看到我把手机递给他，他早就知道我要这么做，因为我每次跟他聊天都这样。

But if you go watch the video of me interviewing Sundar, you see me hand him the phone, and he he knows it's coming because I do it all the time every time I talk to him.

他就会说，挺好。

And he's like, great.

在上万亿条搜索里，我们偏偏要挑这一条来分析。

Out of trillions of queries, we're gonna look at this one.

然后我问他，这个结果怎么样？

And then I ask him, like, is this good?

他会说，我们其实有一套很科学的方法来评估结果是否合格。

And he's like, you know, we actually have a very scientific way of measuring this to make sure it's good.

而且，我们是用统计学的方式来做的。

And, like, we do it statistically.

我一直在琢磨这个回答，想了好久。

And I have just been thinking about that answer for a long time.

对吧？

Right?

谷歌把质量简化成了一套指标，这些指标其实反映的是用户遇到内容后的实际行为。

That Google has reduced quality to some set of metrics that are in response to how people behave when they encounter the thing.

没错。

Right.

对吧？

Right?

比如你搜出结果来，

Like, you show up in the search result.

不会有哪个框写着‘我觉得这玩意儿真烂’，

There's not a box that says, I think this sucks.

有的只是你接下来会做什么。

There's just whatever you do next.

对。

Right.

对吧？

Right?

你没法对着麦克风大喊大叫。

You don't get to shout into the microphone.

奇怪的是，每次的答案都不一样。

Like, it's weird that all the answers are different.

你接下来做什么，才是关键。

There's just whatever you do next.

而这一切，居然被用来推测你对这个页面设计、页面构造，以及AI概览根据你身份、地理位置或搜索历史给出不同答案时的真实感受。

And somehow that's supposed to reveal, like, your emotional reaction to the design of this page, to how the page was constructed, to the idea that the AI overview confidently tells you a different answer depending on who you are or what part of the world you're in or your own search history.

你根本没法把这些反馈告诉谷歌。

There's no way for you to communicate that to Google.

所以谷歌在测量这些数据，他们说，单个搜索结果并不能代表整个产品，毕竟有上万亿条搜索查询，我们有一套科学的方法来衡量质量。

And so Google's measuring all this stuff, and they're saying no individual search result is, like, indicative of the whole product because there's trillions of search queries, and we have this, like, scientific process for measuring quality.

但当你看任何一个搜索结果时。

But then you look at any individual search result.

是的。

Yeah.

我其实说不清这个结果是好是坏。

I can't actually tell you if this is good or bad.

对。

Yeah.

只有从整体来看，只有根据我接下来点哪一个，才能判断它是好是坏。

Only in aggregate and only based on my next click do you have a sense of this is good or bad.

而这个AI世界里，有点特别。

And there's something about this AI world.

这正是你在说的。

It's what you're talking about.

在你告诉我们该改好之前，它就是有问题的。

Like, it's bad until you tell us to make it's good.

然后我就在想，你真能用这种方式衡量吗？

And then there's me being like, I don't know that you can measure this this way Yeah.

根本不行。

If at all.

这到底算好还是不好？

Like, is this good?

我们用这些工具的体验真的好吗？

Are we having good experiences with these tools?

靠几十亿人下次点击的行为，根本测不出来。

It's not measurable based on the next click of billions of people.

对。

No.

所以你们才会转向查询量增长这种说法，觉得人们搜得越多，就越喜欢谷歌。

Which is why you land on things like query growth based on the idea that if people do more Google searches, they must love Google.

但这里面其实有个巨大的逻辑跳跃——如果当初你做得对，我第一次搜索就得到答案了，根本不需要再搜第二次。

Despite the fact that there's actually, like, a complete logical leap in there that suggests that maybe if you'd done it correctly, I would've gotten the answer the first time, and I wouldn't have had to do more queries.

所以，也许真正的成功是我做的搜索更少了。

That, like, actually, maybe success looks like me doing fewer queries.

对。

Yeah.

成功看起来像是搜索量增长和广告收入。

It looks like growth and ad revenue.

对吧？

And right?

所以你们选了这些指标来衡量，而不是去想怎么确保页面质量，因为那需要做出一些实际的决定。

So you pick all of these things to manage against rather than how do we make sure the pages are good because that that involves some kind of, like, actual practical decision.

谷歌对页面上不该出现的内容非常严格。

And Google is very aggressive about what it doesn't want on its pages.

对吧？

Right?

而且我觉得，它越来越明确地表达出哪些内容不希望出现在页面上。

Like and I think it's got increasingly vocal about the stuff it doesn't want on its pages.

他们一直在跟SEO农场较劲。

They're they're they're picking fights with SEO farms.

多年来，他们一直大声地打这场仗，成效时好时坏。

Like, they they've been fighting that fight loudly for years to varying degrees of success.

但问题是，这到底该是什么样子？

But the idea of, like, what is this supposed to be?

从来没人会告诉你答案。

No one will ever tell you.

我也不确定他们自己有没有答案。

And I I don't know that they know the answer.

是的。

Yeah.

但你想想，再往远了说一点。

But if you think about it, like, again, cast this out even further.

你和我都搜过‘最好的Chromebook’。

You and I both Google best Chromebook.

对吧？

Right?

接下来，根据谷歌的设想，它应该能直接帮我买下那台Chromebook，或者至少具备这个能力。

The the next step of this according to Google is that it will buy me that Chromebook or that it should be able to.

对吧？

Right?

我只要说一句‘帮我买台最好的Chromebook’，它两天后就该送到我门口。

That I just say, buy me the best Chromebook, and it should appear at my door in two days.

这才是大家心目中未来的模样。

That is that is everybody's idea of the future.

对。

Mhmm.

如果它给你的结果和给我的不一样，这算成功还是失败？

If it's telling you and me different things, is that a is that a success or a failure?

为什么它给我们的结果会不一样？

And why is it telling us different things?

而且，它是根据什么标准来判断的呢？这一切简直太乱了。

And and based on what criteria is it like, the the all of this stuff is such a mess.

我来从好的方面和坏的方面分别回答你。

Let me answer that in the the good way or the in the bad way.

好的一面是，Google 会根据你的需求，推荐最适合你的 Chromebook，然后直接下单，把那台设备送到你手上。

The good way, the positive vision for you, Google, best Chromebook, and it determines a different answer for you and me, and then it goes off and buys you the the one that told you that it was the best.

这背后其实有一个积极的设想：Google 对你有足够的了解。

There's a positive vision for that, which is it Google has enough context about you.

它知道你是怎么上网的。

It knows how you use the web.

它知道你常用哪些应用和服务。

It knows what apps and services you use.

它甚至能访问你的 Gmail。

It is in your Gmail.

它能大致判断出你平时常浏览哪些类型的网页。

Like, it it has a sense of, like, what kinds of web pages you go to.

你整天都在用Canva，还是只是看文档？

Are you in Canva all day, or are you just reading documents?

就是你平时在做的任何事。

Like, whatever thing that you're doing.

它知道哪款Chromebook最适合你。

It knows which Chromebook is best for you.

没错。

Right.

而这对我的情况来说答案就不同了，因为我整天在玩基于浏览器的3D游戏。

And that is actually a different answer than what Chromebook is best for me because I'm, I don't know, playing Open WebGL games.

我都忘了那技术叫什么了。

I don't even know what that technology is anymore.

不是Open WebGL。

It's not Open WebGL.

因为我一整天都在玩网页游戏，没错。

Because I'm playing browser based games all day Sure.

还有在浏览器里用CapCut剪视频，或者那些乱七八糟的在线工具。

And editing video in in, like, browser cap cut or whatever nonsense exists.

对吧？

Right?

所以我需要的配置跟别人不一样，得要更长的续航和更快的处理器，诸如此类。

Like and so I have a different answer, and I need longer battery life and a faster processor, whatever it is.

所以我一搜‘最好的Chromebook’，搜索结果就全是针对我的需求推荐的。

And so when I Google best Chromebook, it's tailoring the answer to me.

我觉得这完全打乱了人们在买东西之前做研究时的基本预期。

I think that breaks people's fundamental expectations of what should happen when you begin the research process to buy anything.

对吧？

Right?

我问你‘哪款Chromebook最好’，你要是了解我，可能会直接推荐某一款给所有人。

Like, I ask you what the best Chromebook is, and maybe you know me, and you're like, just told everyone to get this.

多花两百美元，你就能买台MacBook Air。

For $200 more, you can buy a MacBook Air.

就这样，没别的了。

Like, end of story.

但如果你更了解我一点，答案可能就不一样了，那对话也会完全不同。

But, actually, if you know me a little bit more, maybe that's a different answer, and that's a different conversation you have with a person.

我觉得大多数人搜‘最好的Chromebook’时，都期待看到四五个不同的人或者网站，各自做测评，讲清楚原因。

I think most people Google best Chromebook, and they expect to find, like, four different people or four different websites, like, doing the testing and saying why.

然后他们自己读、自己判断，这种自主感让他们觉得安心，事情也就完了。

And they get to read it and make their own determination, and that fills them with a sense of agency, and then, like, you're done.

我觉得这事儿不光是Chromebook的问题。

And I don't think this is just about Chromebooks.

现在营销界有个说法，叫‘终极产品’，意思是所有东西都在宣传：我帮你做完了研究，你不用操心了，我直接给你出了款咖啡机。

There's sort of this concept in marketing right now about, like, terminal products where everything is marketed as I did the research so you don't have to, and I so I made the coffee maker.

我做的这款咖啡机，是你这辈子最后一台要买的，因为选择太多让人头疼，其实所有微波炉都差不多，都是中国产的。

I made the last coffee maker you'll ever have to buy, and it's because, like, choice is overwhelming and everything is actually the same microwave from China.

这说的就是这么回事。

That's what this represents.

这就是所谓的终极营销。

Like, terminal marketing.

对吧？

Right?

我们已经走到尽头了，谷歌直接告诉你买哪个，反正也就这么一个选择。

Like, at the we're at the end of the line, and Google's just gonna tell you which one to buy because there's only one choice anyway.

我觉得，事物总是在两个极端之间来回摆动。

And I I I think, you know, pendulums swing back and forth.

也许我们现在就处在这样一个阶段。

Like, maybe that's where we are now.

但这不会是未来的常态。

It's not where we're gonna be in the future.

AI模式无法应对另一种情况——那种你确实想要很多选择的世界。

And AI mode doesn't have an answer to that other world where actually you do want a lot of choices.

你可以试穿五十件大衣，看看哪一件才是最适合你的Chromebook。

So you you get to try on 50 different coats and see if that's the Chromebook that's best for you.

没错。

Right.

而且还得说清楚，就算你描述的那种未来真是你想要的，现在也离得十万八千里。

It's also important to say that even the first thing you described, even if that's exactly the future you want, not remotely close to being the case.

根本不可能。

Like No.

那种认为AI能凭我的数据精准判断我该买什么的想法，连我到底喜欢哪种卫生纸都猜不准。

This idea of you have perfect data about me so you can make correct purchasing decisions about me is it it can't tell you with any confidence what kind of toilet paper you like.

更别说其他那些了。

Forget any of the rest of this.

这些根本不可能实现。

Like, none of that is close to coming to pass.

我说的就是这个意思。

And this is what I mean.

再说一遍，我们现在被推送的，都是大家公认的产品糟糕版本，因为厂商根本没法做出好的版本，也没人知道怎么把真正好的东西单独拿出来交付。

Again, we're we're being served what everyone understands to be the bad version of this product because there's no way for them to build the good one otherwise, and no one has figured out how to ship just the good thing in the middle.