本集简介
双语字幕
仅展示文本字幕，不包含中文音频；想边听边看，请使用 Bayt 播客 App。
埃里克·塞弗特，你最近怎么样？
Eric Seufert, how are
你最近还好吗？
you doing?
很高兴见到你们，大家最近都怎么样？
Good to see How's going, guys?
谢谢邀请我来。
Thanks for having me.
是的。
Yeah.
感谢你们的到来。
Thanks for being here.
我重新连上线了。
Hopping back on.
我们一直在关注ChatGPT广告上线这件事。
We've been following the ChatGPT ads rollout, the story.
你之前做了很多精彩的分析。
You've had a ton of great analysis.
我们之前简单聊过一点，但这次特别想听你亲自讲，深入探讨一下。
We've talked about it a little bit on the but we wanted to hear it from you and dive way deeper.
所以能不能请你先给我们梳理一下，目前这个广告上线的情况怎么样？产品形态是怎样的？你当初的预期是什么？就从头说起，聊聊 ChatGPT 广告这一路的发展吧。
So maybe you could set the table for us on what the launch is looking like right now, what the product is looking like, what your expectations were, just maybe begin at the beginning of this saga for the ads in ChatGPT.
这个故事的起点，其实就是 ChatGPT 的起点。
Well, the saga begins where ChatGPT begins.
我们早就知道会走到这一步。
We all knew we were gonna end up here.
大家心里都清楚，广告迟早会来。
We all knew ads were coming.
我早在2025年5月就写过一篇相关文章。
I mean, I wrote a piece May 2025.
我当时就说，ChatGPT 肯定会通过广告来变现。
I said, obviously, ChatGPT will monetize with ads.
我当初为什么这么有把握？
Why did I feel so confident in that?
因为他们刚聘用了斐吉·西莫。
Because they had just hired Fiji Simo.
对。
Yep.
她以主导Facebook移动端信息流广告产品而闻名。
She's kind of known for having led the the mobile news feed ads product to Facebook.
那无疑是人类历史上最成功的广告业务之一，所以既然他们已经积累了相关领域的专业能力，推出ChatGPT广告也就顺理成章了，事实也确实如此。
That's obviously been one of the the most successful ads businesses in in the history of mankind, and so it seemed natural that that they would they would bring given the domain expertise they had acquired, that they would bring ads to bear for ChatGPT, and they have.
现在的上线方式，跟Netflix的广告上线很像。
Now the launch looks like the launch of Netflix ads.
对吧？
Right?
如果一直这样下去，对业务恐怕不太有利，但我相信他们会逐步优化的。
And if it remains that way, it's probably doesn't bode well for the for the business, but I think they're gonna evolve it over time.
他们现在公布的这些信息，媒体已经做了不少深入报道。
And what they've said now or what's come out, the information has had a lot of good reporting on this.
他们按千次展示收费，这在直接响应类广告里并不常见。
They're charging on a CPM basis, which is, you know, not standard for kind of direct response ads.
更像是一种品牌广告的定价模式。
It's more like a brand advertising formulation.
每千次展示收费60美元，正好和Netflix刚推出时的定价一样。
But so it's $60 CPM, which is exactly what Netflix came out of the gate charging.
是的。
Yep.
他们要求的投放承诺金额不到一百万美元。
They're asking for commitments of less than a million dollars.
这明显是在试水阶段。
So this is clearly like a testing phase.
他们提供的投放测量和定向功能会非常有限。
They're gonna offer very little in the way of of measurement or or targeting.
所以这感觉就像一个非常原始的最小可行产品。
So this feels like a very sort of primitive MVP.
但问题是，他们能多快把它升级成看起来更像Facebook广告平台的东西呢？
But, you know, the question is, like, how quickly can they evolve it into something that looks more like the the Facebook ads platform?
你觉得是什么原因让我们走到这一步，居然出现了Fiji Simo这种情况？
So what do you think like, what do you think got us to this place where you get Fiji Simo?
她以前做过类似的事。
She's done this before.
她人脉广、能力也强，但即便如此，要推出一个不够成熟的产品，还是得花上八个月。
She has massive connections and skill set, but it still takes, you know, eight months to roll something out that feels not as mature as something else.
是人才的问题吗？
Is it a talent issue?
有没有其他监管方面的障碍？
Are there other regulatory concerns?
还是说他们得应对一些别人根本想不到的特殊情况？
Or do they need to worry about certain edge cases that people might not be aware of?
还是说他们这么做是出于某种战略考虑？背后是不是有什么不为人知的原因？
Or is it specifically strategic to actually is there some secret reason why they're they're taking this approach?
这是我的推测。
So here's my hypothesis.
对吧？
Right?
这事儿可不只是关于 Fiji Simo 本人。
And so it's not just about, you know, Fiji Simo.
她是 Apps 公司的首席执行官。
She's the the the CEO of Apps.
是的。
Yeah.
但他们收购了 Statsig，这家公司有不少前 Meta 员工。
But they they acquired Statsig, which is, you know, staffed by a lot of X Meta people.
那是个实验平台。
That was an experimentation platform.
如果你要推出一个需要大力测试的广告平台，这种人才正是你最想引进的。
That's the exact kind of talent that you'd wanna ingest into a company if you're gonna roll out an ass platform, which needs to be aggressively experimented on.
我觉得事情大概是这样的。
Here's here's what I think happened.
我觉得，ChatGPT 这个项目，是 OpenAI 一步步发展起来的。
I think, you know, ChatGPN is so OpenAI kind of evolved over time.
它最初是个研究实验室，后来嘛，显然在转型成营利性公司时经历了不少风波。
It was a research lab, then it you know, obviously, they've had the the drama around trying to pivot into a for profit corporation.
我猜投资人希望他们请个更懂商业敏感度的人来掌舵。
My guess is the investors wanted them to bring in somebody that was a little bit more sensitive to, you know, the commercial sensitivities.
对吧？
Right?
所以，Fiji 当首席营销官，确实很符合这个定位。
And so, you know, Fiji CMO would be a very good fit for that profile.
他们请了她进来，但我感觉公司内部对变成一个以广告驱动的业务，多少有点抵触。
They brought her in, and my sense was there probably was a little bit of animosity internally towards the idea of this being an ads an ads driven business.
对吧？
Right?
主要是个广告驱动的业务，我觉得最终肯定会是这样。
Primarily an ads driven business, which which I think it ultimately will be.
是的。
Yeah.
就在两年前，萨姆还说，他认为广告是一种失败的表现。
Only only, like, two years ago, Sam said, I look at ads as like a failure state Yeah, for right.
没错。
Yeah.
后来他开始转变说法，说其实他挺喜欢一些Instagram广告的。
And then he started pivoting the rhetoric and started saying, well, I like some of the Instagram ads.
如果投放精准，广告其实挺让人愉快的。
When they're targeted well, they're actually delightful.
在合适的场景下，广告也可以是好的。
And in the right context, they can be good.
他逐渐接受了我们的观点，认为广告其实很不错。
And he was coming around to the three of ours worldview, which is that advertising is great.
但他对外传递这个信息时，内部也得花大量精力去沟通和调整。
But but he did have to message that externally and then obviously had a whole bunch of work to do messaging it internally.
也许正因如此，内部推动这件事的劲头才没那么足。
And may maybe that's why there was a little bit less energy internally to to to really push towards that.
这说得通。
That makes a lot of sense.
也许他只是被传达信息的那个人。
Well, he may have been the one that was messaged to.
他未必是掌控信息方向的人。
He may not have been the one controlling the message there.
他可能只是被传达信息的那个人。
He may have been the one that was messaged to.
我们需要广告，因为广告是不可避免的。
We need ads because ads are inevitable.
不是说广告有多好。
It's not that ads are great.
广告是不可避免的。
Ads are inevitable.
如果你想触达我所说的规模级用户，广告是不可避免的。
If you wanna reach what I call humanity scale, ads are inevitable.
如果你想在最大潜力规模上实现变现，那就必须靠广告。
If you wanna monetize that at the at the at the at the scale it can be at its total potential, you need ads.
这就是关键所在。
That's how you do it.
对吧？
Right?
所以他可能是那个被说服的人。
And so he may have been the one that was convinced.
对。
Mhmm.
而且，让一个有相关经验的外部人士加入，可能是投资者策略的一部分。
And and, you know, bringing someone external in who had been there, done that may have been part of an investor strategy.
但抛开我的推测不说，我觉得这个项目有潜力成为一项极其成功的举措。
But but putting aside my hypothesis, look, I think this has the potential to be an incredibly successful initiative.
关键在于，他们能多快转向类似Meta的模式。
The question is how quickly can they pivot into something that looks like the Meta model.
对吧？
Right?
如果你现在来看，我觉得你说的每一点，都跟Netflix当年的情况一模一样。
So if you look at it right now, it's, you know and and and I think, look, we to every everything you just said mirrors what would happen at Netflix.
多年来，Netflix一直说我们绝不会做广告。
For years, Netflix said we would never do ads.
广告太让人反感了。
Ad ads are offensive.
这简直是一种冒犯。
They're an affront.
这完全违背了我作为消费者的感受。
They're an affront to to to to my sensibilities as a consumer.
当时管理层就是这么说的。
That's that's what the leadership said.
可有一天，他们突然就不这么说了。
And then one day, they weren't.
对吧？
Right?
OpenAI 也经历了完全相同的变化。
And the exact same thing happened with OpenAI.
问题是，六个月后会怎样？
The question six months?
全押上去。
All in.
没错。
Yeah.
别忘了，他们都全押上了。
Don't forget all in.
他们都全押上了。
They're all in.
当初我们可是求着他们全押上的。
All in was and and we we begged them.
我们求着他们。
We begged them.
我们求着他们。
We begged them.
而现在
And now
我觉得吧，他们当然是自己做决定的，但现在他们已经开始打广告了。
And I you know, I think they made the decision independently, of course, but they're they're running ads now.
他们
They
一直都在。
always come Yeah.
我早就这样了。
I I've been Yeah.
从2025年5月起，我就一直在大力推动这件事。
I've been banging the drum for for since May 2025.
对。
Yeah.
但我觉得关键问题是，需要六个月吗？
But I I think the question is, is it six months?
是一年吗？
Is it twelve months?
要十八个月才能做出一个像Meta广告那样的东西吗？
Is it eighteen months to get something that looks like Meta Ads?
对吧？
Right?
这意味着要根据行为画像来做精准投放，对吧。
And that means targeting, right, based on behavioral profiles.
这意味着要进行效果衡量。
That means measurement.
这意味着要向广告主解释，你为他们带来的实际影响到底有多大。
That means allowing that that means explaining to advertisers the the the sort of impact of what that you've delivered on their behalf.
没错。
Yeah.
这意味着需要开发大量技术支持。
And that means, like, a whole lot of tech that needs to be built out.
对吧。
Right?
这还意味着创意优化，以及各种相关的工作。
And that means, you know, creative optimization, all these things.
单纯的转化优化广告，和他们目前提供的东西完全不同。
Just conversion optimized advertising is very different from what they're offering.
是六个月吗？
And is it six months?
是一年吗？
Is it twelve months?
得花十八个月才能达到吗？
Is it eighteen months to get there?
这是个大问题。
That's a big question.
对吧？
Right?
但这将决定成败。
But that'll determine the success.
如果你说，这跟奈飞的做法一样，那可真不是个值得参考的范本。
And if if you say, look, this mirrors what Netflix did, well, that's not really a playbook that you wanna draw inspiration from.
对吧？
Right?
我的意思是，他们不得不转型。
I mean, they had to pivot.
他们得来个彻底的转型。
They had do a hard pivot.
你看，广告套餐已经上线三年了。
I mean, the the ads tier has been live for three years.
他们在2025年赚了15亿美元。
They did 1,500,000,000 in 2025.
对吧？
Right?
所以，目前广告收入在总收入里还占不了多少。
So, you know, it's it's not a significant chunk of the revenue yet.
我觉得这块收入潜力大得多，但奈飞还有别的考量。
I think it could be a much larger chunk of the revenue, but Netflix has other concerns.
对吧？
Right?
他们担心内容质量。
They've got quality concerns.
他们还顾虑用户感受，没法完全放开广告。
They've got, you know, consumer sentiment concerns that where they can't just, like, fully embrace ads.
我觉得ChatGPT在这方面没这种限制。
I don't think ChatGPT is limited in that way.
是的。
Yeah.
对。
Yeah.
你回忆一下Netflix刚推出的时候。
Take me back to the that initial launch of Netflix.
你说当时的千次展示定价大概是60美元。
You said similar CPM, dollars 60 per thousand views.
但当时的用户反响怎么样？
But what was the response like?
比如，有没有广告主觉得这很棒，觉得物有所值？
Like, were there any advertisers that were like, yes, this is great.
觉得这笔钱花得值。
I got my money's worth.
效果还不错。
It worked out.
还是说结果很模糊，根本没法追踪？
Was it just vague and unattributable?
问题出在哪？
What's problem?
我觉得那些优秀的效果营销人员，都很愿意尝试新渠道。
My perception is like great performance marketers Yeah.
尤其是那种能带来高转化意向流量的渠道。
Are always excited to try new channels, especially a channel that's driving high intent traffic.
对吧？
Right?
如果你告诉一个效果营销人员，嘿，我这里有十亿用户，你可以去给他们打广告，他们肯定会很兴奋，不管整体架构怎么样，因为他们就是想先拿一笔测试预算试一试。你去跟营销人员聊，到了一定规模，他们就会说，行，咱们先投个十万美金试试。
If you tell a performance marketer, hey, I've got a billion users and I'm gonna let you advertise to them, They're gonna be excited no matter what the general setup is because they just wanna get in there with a test budget and like, you'll talk to marketers and at a certain scale, they'll they'll they're like, yeah, let's try a 100 k here.
再投个十万美金试试。
Let's try a 100 k there.
其实就是先撒钱，看看哪些渠道真正有效，再重点加码。
It's just like you're just kind of throwing around money, and then you're gonna double down on on what really works.
没错。
Right.
初期上线时，这就是推广广告平台的正确做法。
For the initial rollout, this is how you launch an ads platform.
我不是说这不对。
Don't get me wrong.
这就是推广广告平台的正确做法。
This is how you launch an ads platform.
所以我才说六个月和十二个月的区别。
That's why I'm talking about six months versus twelve months.
就得这么干。
This is how you do it.
得这么积累数据。
This is how you bootstrap the data.
得这么把用户拉进来。
This is how you get people onboarded.
得这么收集反馈。
This is how you get feedback.
你必须这么上线。
You have to launch it like this.
没别的路可走。
There's no other way.
你不可能一上来就做转化优化的广告平台，因为根本还没任何转化数据。
You can't launch a conversion optimized ads platform because by definition, you don't have any conversions data yet.
对吧？
Right?
所以这就是为什么我觉得他们推出了即时结账功能的另一个原因。
Now that's that's another reason why I think they they launched instant checkout.
我觉得即时结账其实是个幌子。
I think instant checkout is a stalking horse.
这本质上是一种收集转化数据的方法，之后他们就能用这些数据来做精准广告投放。
I think that's a method of of of gathering conversion data that they can use to then target against with ads.
我觉得这正是他们这么做的全部目的。
I think that was the whole purpose of that.
对吧？
Right?
我不觉得他们把这当成一个长期的收入来源。
I don't think that was seen as a long term revenue opportunity.
我觉得他们只是想用它来积累初始数据。
I think that was seen as a way to bootstrap the data.
但这就是推出广告平台的正确方式。
But this is how you launch in an ads platform.
我的意思是，这个新的转化优化广告平台确实得不断进化，但一开始不可能直接做到那样。
I mean, it's gotta evolve the new conversion optimized ads platform, but you can't be that at the start.
你没法凭直觉就设置好所有参数。
You can't intuit the settings.
你也不可能一开始就知道该调哪些地方才能让系统跑起来。
You can't intuit the tuning that you needed to implement to make this work.
但关键问题是，他们到底要花多久才能达到那个水平？
But but the question is how long does it take them to get there?
对吧？
Right?
所以这些事，咱们就等着看吧。
And and so that that, you know, that we'll just see.
但我觉得他们已经具备了核心能力。
But my sense is they've got the DNA.
你知道的，他们的数据报告里显示，有七百多名来自Meta的员工。
They you know, the information report, they have 700 x Meta employees.
说白了，Meta 的每个人都在做广告相关的工作，不管他们是做什么的。
Like, essentially, everyone at Meta is working on ads, whether they're working Yeah.
不管是做客户代表，还是做机器学习工程师，
You know, as an account rep or they're working as an ML engineer.
大家都是朝着同一个方向努力。
They're all sort of rowing in that direction.
没错。
Yeah.
你对即时结账功能有什么看法？
What was what's been your reaction to instant checkout?
关于收费结构，大家争论挺多的。
There's been a lot of debate on on the fee structure.
有人说，没有零售商能负担得起。
Some people saying, no retailer can afford this.
我们在很多情况下都反驳了这种说法。
We were pushing back on that in in a number of different cases.
很多品牌都会想，你愿意为获取一个客户花掉订单平均价值的4%吗？
There's plenty of brands where you say, like, would you spend, like, 4% of AOV to to acquire a customer?
我也能理解另一方的观点。
I can also see the other side of it.
我担心的是，你在Meta平台上投广告来寻找客户。
A concern that I have is like, you know, you're running an ad on a Meta platform to find a customer.
但客户却跑去ChatGPT搜索产品。
The customer then goes to ChatGPT if that's where they're searching for products.
这样一来，品牌实际上要付两次钱：一次是投广告吸引客户、激发兴趣，另一次是在转化环节再付4%的费用。
And then the the brand is effectively paying twice to, like, actually find the customer, drive them, get them interested, and then and then at the bottom of the funnel, they're paying, like, another 4% fee.
这感觉不太划算，但你怎么看？
That feels, like, not great, but what's been your reaction?
是的。
Yeah.
我以前属于那一派。
So I was in the former camp.
我觉得这个费率根本行不通。
I said that that's not a that's not a workable fee.
不过我得回头说说你提到的奈飞推出广告时的情况。
Let me let me come back to that, though, because you asked a question about what happened when Netflix offered ads.
是的。
Yeah.
当时报名的广告主太多了。
They had too many advertisers sign up.
真的太多了。
They had too many.
他们不是需求不足，而是广告位供不应求。
They were demand they were they were supply constrained, not demand constrained.
最后不得不退钱。
They had to return money.
我觉得这正是OpenAI说‘不行’的原因。
I think that's exactly why OpenAI is saying, no.
你听好了。
Look.
我们给这个设了个上限。
We're setting an upper limit on this.
一百万就是最高限额了。
A million is the max.
对吧？
Right?
因为他们没那么多用户可以投放广告。
Because they won't have enough people to advertise to.
这纯粹是为了测试一下，慢慢扩大规模。
This is just a way to test it out and to start scaling.
奈飞犯的错误是，他们从来不做决断。
And the problem Netflix made was they never act.
从一开始，奈飞的问题就在于他们和微软、Xanders达成了合作。
So he the problem that Netflix had from day one was they had the partnership with with Microsoft with Xanders.
他们当时用的是外部技术来投放广告。
They were using external tech to run the ads.
对吧？
Right?
Xanders 是一个广告平台。
Xanders is an ad an ad platform.
所以问题就出在这儿。
And so that was the problem.
他们没法实现自定义的广告位。
They couldn't implement, you know, custom placements.
很多数据统计功能也做不了。
They couldn't do a lot of measurement stuff.
由于没掌控广告投放的底层系统，很多定向投放也搞不了。
They couldn't do a lot of targeting because they weren't operating the actual pipes, the actual plumbing that was delivering the ads.
数据方面问题一大堆，主要是拿不到数据。
There's lot of, like, data data issues, just access to data issues.
OpenAI 这儿的做法根本不是这样。
That that's not what OpenAI is doing here.
OpenAI 是从头开始自建技术的。
OpenAI is starting out by building their own tech.
顺便说一句，Netflix 后来也转向了这条路。
And Netflix, by the way, pivoted to that.
但他们转向得太晚了。
But they pivoted to that very late.
所以我觉得 Netflix 的广告业务没做起来，也没带来多少收入。
That's why I think the ads initiative in Netflix was not successful and and and, you know, hasn't really delivered that much revenue.
再回到那 4% 的抽成，首先，我觉得接下来会发生什么。
But going back to the 4% fee, first of all, here's what I think is gonna happen.
我觉得他们会把费率降到零，或者几乎为零。
I think they're gonna drop that to zero or very low.
因为这整个目的就是为了促进转化。
It'll be tiny because I think the whole point of this is to just drive conversions.
关键是，当你作为零售商看到 ChatGPT 的转化效果，本质上你是在为这些转化付费，那你自然就想对广告投放做点控制。
And the whole thing is, like, when you see performance in ChatGPT as a retailer who's just basically paying for the conversions essentially with the fee, you're gonna wanna actually advertise against that.
你会想掌控局面，因为如果只是靠 ChatGPT 判断什么内容相关就展示，那你根本没任何主动权。
You're gonna wanna control it because you have no control if it's just surfaced based on what ChatGPT believes is relevant.
总有一天，想投放的商家会多到一个位置放不下，那时候就得靠竞价来协调了。
And at some point, you're gonna have too many retailers to place into a single placement, so you're gonna have to mediate that by bids.
对吧？
Right?
所以归根结底，我觉得这最终会演变成广告模式。
So, ultimately, I think this gives way.
总的来说，这终究会转向广告业务。
In in in in the main, this gives way to advertising.
但那 4% 的费率能不能持续，关键在于你怎么看待这些购买行为——这根本不是获客。
But whether the 4% is sustainable or not actually comes down to how you view what those purchases are because it's not user acquisition.
你不是在获取新用户。
You're not acquiring a user.
如果是那样，你就该拿它和用户终身价值做对比。
If you were, then you'd compare that to LTV.
但既然这不是获取用户，你也不是在考虑和这个用户建立长期关系。
But since it's not user acquisition, you're not acquiring that user as a sort of, like, with with a long term relationship in mind.
你其实只是用结账金额的一定比例来换取结账机会，但这个用户并不会直接跟你产生互动。
You're you're essentially trading a percentage of the checkout for the checkout, but that user doesn't interface with you.
对吧？
Right?
在很多情况下，
And in a lot of in a lot
你意思是，这些成本不会转嫁给底层供应商吗？
of terms There's not gonna be passed through to the underlying vendor, you're saying?
比如，
Like, a
不，不是这样的。
Well, no.
我的意思是，你想想看。
It I mean, look.
你是实际的商家。
You're the merchant of record.
对吧？
Right?
你得获取他们的信息，因为你需要完成订单。
And you get their information because you have to fulfill the order.
但ChatGPT或者OpenAI在条款里明确说了，你不能用这些邮箱做再营销。
But ChatGPT's or OpenAI specifically says in the terms that you can't use your email address for, like, remarketing.
对吧？
Right?
亚马逊也是这么规定的。
And, you know, Amazon says this too.
所以我觉得，你其实并没有真正获得客户。
So so the I think is you're not capturing a customer.
你只是为这一笔交易支付了4%的费用，但你完全无法掌控是否还能再次接触到这位客户，是否能把他纳入你的客户体系。
You're just getting you're giving 4% for that one off transaction, but you have no control over whether you reach that customer again, whether that customer is brought into your orbit.
这跟向人做广告、获取客户是完全不同的概念。
There's very that's a very different proposition from advertising to someone and acquiring a customer.
一旦你真正获取了客户，后续就有无数机会可以再次营销他。
You acquire that customer, you get a whole stream of opportunities to remarket to them.
如果你只是用4%的代价换一笔交易，那你就只能得到这些。
If you're just, you know, trading 4% for the transaction, that's all you get.
所以这根本算不上客户获取。
So it's not actually customer acquisition.
如果你现在想想这4%的交易费，再对比一下，当初这个数字出来的时候，人们常说有人在Facebook广告上花50%、60%的预算。
If you think about 4% per transaction now consider that if if you thought about the the people talk about you know, when that when that number came out, they talked about, well, some people spend, you know, 50%, 60% on Facebook ads.
那这两者有什么区别呢？
Why is that different?
因为那笔广告费是覆盖客户整个生命周期的，而不是只针对单次交易。
Because that's over the life cycle of that customer, not for the
正在发生的事。
one that's happening.
一次性的交易。
One Or time.
他们为这个客户付一次钱，之后客户还会回来十五次，或者选择订阅，这样他们才真正掌握了客户关系。
They pay for the customer one time, and then they come back, you know, 15 more times or they end up subscribing, and they've actually actually owned the customer relationship.
对。
Mhmm.
是的。
Yeah.
没错。
Exactly.
没错。
Exactly.
再换个角度想想。
And and and think, you know, think about it also this way.
你知道吗，这4%的利润，可能就是你作为电商运营者能拿到的全部 margins 了。
You know, that 4%, that might be all the margin you're getting as an e commerce effort, you know, as an e commerce, you know, operator.
对吧？
Right?
但这并不意味着你不能做广告。
And the thing is that's not preventing you from advertising.
也不是说，好吧，
That's not saying, okay.
你得把所有广告预算都省下来，但你还是得在各个地方投广告。
You have to save your entire advertising, but you're gonna still have to advertise everywhere.
这只是你订单中一小部分你无法掌控的部分。
This is gonna be a tiny sliver of your orders that you're not in control of.
你根本控制不了。
You can't control it.
这部分的订单量可能会忽高忽低。
This could be lumpy.
这个月可能上升，下个月就下降。
This could be up one month, down the next.
对吧？
Right?
这事你控制不了。
You can't control that.
你还是得继续做广告。
You're gonna have to continue to advertise.
就像你提到的，你根本不知道，这会不会抢走你在其他平台做广告本来能带来的订单。
Now you have no idea, to your point, whether this is cannibalizing purchases that would have happened from the people that you advertise on other platforms.
对吧？
Right?
所以这本质上只是增加了成本，可能还拖累了你在其他地方的广告投入。
So this ends up just being a cost on a drag probably on your advertising expenditure elsewhere.
你能解释一下归因问题，以及奈飞广告产品的演变吗？
Can you explain attribution and the evolution of Netflix's ad product?
因为我坐在沙发上看电影，看到广告后还得去手机上操作，感觉特别割裂。
Because if I'm sitting on my couch watching a movie and I see an ad and I go to my phone, it feels really disconnected.
即使在ATT政策出台之前，这也非常困难。
Even in a pre ATT world, it feels extremely hard.
他们是怎么做归因的？
How do they do attribution?
这种归因能做到多好？
How how good can this get?
再说到ChatGPT，在ATT之后的环境下，OpenAI的归因能力上限到底在哪？
And then and then with ChatGPT, what what's the upper bound on attribution for OpenAI given that we are in a post ATT world?
上限大概就跟Facebook看到的一样，因为是直接响应式的。
Upper bound is basically what you see with Facebook because it's gonna be direct response.
完全是基于点击的。
It's gonna be click based.
对吧？
Right?
Netflix 的情况有点不一样。
With Netflix, it's a little different.
在 CTV 领域，公司通常会搭建一个叫‘安全屋’的环境来做效果衡量和归因。
So with CTV, the way that companies do measurement and attribution in CTV is is traditionally they set up what's called a clean room.
对吧？
Right?
我会搭建这样一个集中式的环境。
So I set up this environment, it's a centralized environment.
然后把所有我掌握的广告互动数据都传进去。
I push all the ad interaction data that I have.
简单说，只要我知道给谁投了广告，就把相关数据全部上传到这个安全屋。
So basically, when I know I showed an ad to someone, I push all the data into this clean room.
很多时候，数据包括 IP 地址和设备信息。
A lot of times it's, you know, it's IP address and device information.
广告主也会把他们的数据传进来，然后我这边做匹配。
And then the advertiser pushes that information into the clean room and I match it up.
这传统上就是针对CTV的归因方式。
That's his attribution for CTV traditionally.
现在有其他方法可以做了。
Now there's other ways to do it.
还有其他完全基于概率的衡量方式。
There's other ways to do measurement that's just totally probabilistic.
比如媒体组合建模之类的。
This is like media mix modeling and stuff like that.
对。
Mhmm.
但说到归因，如果你真想用这个词，传统上就是像这样通过安全屋来做的。
But with with attribution, when you when you wanna use that term, traditionally, it's done with a clean room like that.
所以我只是把数据集根据某个关键信息关联起来。
So I'm just, like, linking the datasets together based on some key.
很多时候，就是用IP地址。
A lot of times, it's the IP address.
但如果你想想点击，比如用户点出去的情况。
But if you think about clicks, like, clicks out.
对吧？
Right?
所以这就像是我正在观看内容。
So that's like I'm viewing.
我正在用第二块屏幕。
I'm on second screen.
我可以查看时间信息。
I can look at time.
这里面其实涉及时间因素。
Like, there there's time elements.
做归因的时候，得考虑时间维度。
You you look at temporal elements due to do attribution.
但只要我点了出去，这个点击记录就有了。
But if I'm clicking out, I've got the click.
展开剩余字幕（还有 437 条）
对吧？
Right?
我已经有点击数据了。
I've got the click.
这背后其实包含了很多信息。
And so that carries a lot of information.
我还能分析时间维度，但通过IP地址也能做不少事情。
I can still look at temporal aspects of that, but I can I can do a lot with the IP address too?
所以你看CAPI，其实就是这么回事。
And so that's you know, if you look at a CAPI, that's what CAPI is.
CAPI就是一种把数据直接传到后端的方式，绕过应用、绕过移动设备。
A CAPI is just a way to get this data on the back end so it bypasses the apps you know, it bypasses the mobile device.
绕过浏览器。
It bypasses the browser.
这样一来，苹果就没办法通过ITP、ATP或者ATT来干预了。
So there's no way for, you know, Apple to interfere with that with ITP or ATP or ATT.
所以你还是能获取到数据，虽然从时间推移来看，IP地址的有效性会衰减，这种测量方式本质上还是概率性的。
And so you're able to get you know, it's still probabilistic in that way if if you consider, you know, the the decay of the validity of the IP address over time to induce probabilistic measurement.
但传统上，这之外还会用到像素追踪。
But that's traditionally outside and then a pixel as well.
不过，这正是OpenAI所拥有的资源。
But, I mean, that's what OpenAI has at its disposal.
它既有CAPPN，也有像素追踪，因为整个系统是基于点击来运作的。
It's got the CAPPN and the and the pixel because it's gonna be click based.
因此，在测量和归因上，它的精准度会非常高。
So so there's there's a there's a a high degree of fidelity that'll come with the measurement there, the attribution.
这比通过CTV能做到的要高得多。
That's that's way higher than what you can do with CTV.
对。
Mhmm.
三。
Three.
你觉得有多少公司会尝试搭建大语言模型的广告网络？
What are what do you think the prospects are for the number of companies that are trying to build LLM ad networks?
我至少被几家公司在一年多前拉过融资，说大模型以后会免费。
I got pitched at least a few companies that were maybe a year or so ago saying like, hey, LMs are going to be free.
他们会插入广告。
They're going to insert ads.
然后搭建一个网络，在多个应用里统一投放广告。
We're going to build this kind of network to serve ads across a bunch of different apps.
我当时对这种模式特别不看好，主要是觉得OpenAI肯定想掌控整个闭环。
I was super bearish on those, specifically just because I thought that OpenAI would want to own the full stack.
谷歌当然也会想掌控整个闭环。
Obviously Google would own the full stack.
我觉得消费级AI领域最终会形成一种寡头格局。
And I think we're heading towards an oligopoly of sorts in consumer AI.
真正的规模效应到底从哪儿来？
And where's the scale really going to come from?
不过，你对这方面有什么看法吗？
But any kind of takes on that front?
不过，会有很多智能代理希望靠广告盈利。
Well, there'll be a long tail of agents that are going to want to monetize with advertising.
所以我觉得，专门为它们服务的广告网络还是有空间的。
So I think there might be space for an ad network to exist that services them.
我不明白为什么不能直接用现有的广告网络。
I don't know why that wouldn't be an existing ad network.
对。
Mhmm.
我不明白为什么不能是，或者干脆就是程序化广告平台。
I don't know why that wouldn't be or, you know, just sort of DSP.
比如，为什么不能是The Trade Desk这样的公司呢？
Like, I don't want I don't I don't know why that wouldn't be The Trade Desk, say.
但确实如此。
But but yeah.
我的意思是，这些长尾型代理其实很有机会通过广告实现盈利。
I mean, I think there's probably an opportunity to to monetize some of those long tails of agents with with ads.
但如果你想想像Cheti这样的公司，他们肯定会自己搞一套。
But if you think about, like, a Cheti, of course, gonna build their own.
据我所知，Perplexity好像又要重启他们的广告业务了。
I I would you know, Perplexity is apparently restarting its ads initiative.
我觉得这些公司要是不自己建广告系统，那才奇怪呢。
I I think these companies, it's it would be silly to not build their own.
毕竟，拥有完整的广告技术栈，才能把所有东西都定制得刚刚好。
I mean, you wanna have that ad stack so you can make everything bespoke.
对吧？
Right?
而且还能让广告表现达到最优。
And and and and maximally perform it.
对吧？
Right?
看看Meta的财报，那正是他们大幅超预期的关键所在。
If we look at Meta earnings, I mean, that's exactly what delivered that big beat.
很多人忽略了AI在数字广告中的应用潜力，这些效果是会随着时间累积的，尤其是在以直接响应为导向的广告平台上。
And and I think what people forget about the the prospects of of of this sort of, like, AI enablement with advertising in in digital advertising is that these effects compound over time, especially when you're talking about direct response based ad platforms.
对吧？
Right?
从2025年第一季度开始，我就一直在强调这一点，尤其是对Meta来说。
So, like, with with Meta, I've been banging this drum since q one twenty twenty five.
这些效应是会不断叠加的。
These effects compound.
人们总爱拿3.5%、5%之类的点击率或转化率提升说事，比如Andromeda、Gem或者Lattice带来的变化。
Like, people point to, like, oh, 3.5% or 5% or whatever in increasing click through rates or increasing conversion rates from from Andromeda or from Gem or from Lattice.
谁在乎这些数字呢？
Who cares?
这些提升微乎其微。
These are tiny.
不是。
No.
但首先，每个季度他们都在记录这些性能提升，而且这些提升还会随着时间累积。
But they first of all, that's every quarter they're they're noting these performance improvements, but also they compound over time.
如果我的目标是90天内收回90%的广告投入，120天实现110%的广告回报率，那120天后我就会把这110%再投回你的广告平台。
If I'm targeting 90% a ninety day recoup on my ad investment and and, you know, a hundred and twenty day a 110% ROAS, what am I doing after a 120 I'm reinvesting the 110% in advertising on your channel.
所以如果你每个季度都给我带来3.5%、5%的提升，这些增长最终都会被重新投入广告。
And so if you're giving me 3.5%, 5% bumps every quarter, that's just gonna end up getting reinvested.
这就是为什么我们看到增长再次加速。
That's why we're seeing the growth reaccelerate.
进入2026年第一季度，增长正在重新提速。
Growth is reaccelerating going into q one twenty twenty six.
这太惊人了。
That's amazing.
所以昨晚Facebook上涨了10%，就是因为增长正在重新加速。
That's why Facebook was up 10% last night because growth is reaccelerating.
他们将看到自ATT政策低迷以来从未有过的增长速度。
They're gonna see growth rates they haven't seen since the post ATT doldrums.
这太惊人了。
That's incredible.
这种增长简直不可思议。
That's incredible growth.
任何曾经质疑历史资本支出的人，现在都不得不承认这个现实。
And anybody who doubted the CapEx going historically has to accept that reality now.
他们已经承认了。
And they did.
他们昨天就承认了。
They did yesterday.
是的。
Yeah.
对。
Yeah.
没错。
Yeah.
人们批评的点是，扎克伯格在资本支出上太敢下了。
What what people like, the the criticism is, like, Zuck is being so ballsy with the CapEx.
特别是生成式AI领域。
Specifically on Gen AI.
就是生成式AI。
On Gen AI.
具体说是生成式AI产品。
Specifically Gen AI product.
批评的声音说，他还没推出过一款爆款的生成式AI产品。
And the criticism is like, hasn't launched a hit Gen AI product yet.
我们之前在节目里提到过，这家伙实际推送的生成式AI内容，比地球上几乎任何人都多。
Our take earlier on the show was like, look, this guy is serving more Gen AI content than almost anyone on earth.
对。
Mhmm.
即使他没推出一个爆款的新产品，他的产品也因为这些因素获得了巨大的推动力。
Even if he doesn't have like a breakout net new product, his product has massive tailwinds because of all this.
他现在完全有底气决定：我要成为这个领域里最大手笔的投资者之一，提前好几年就做好规划。
Like, he's in a perfect position to decide, yeah, I'm gonna I'm gonna be, you know, one of the biggest spenders in this category and and plan, you know, however many years out in advance.
他们在人工智能上的盈利，仅次于谷歌，比谁都多。
They're making more money on AI than anybody except for Google.
你看看。
Look.
说他们没用生成式人工智能？你见过脸书上的广告吗？
The the the idea that they're not they're not utilizing generative AI have you seen the ads on Facebook?
那些全是生成式AI做的。
Those are all generative.
曾经爆发过一次抗议。
There was a revolt.
那是广告主们的集体抗议。
There was an advertiser revolt.
最近出了个丑闻，因为Meta在广告生成上太激进了。
There there was a scandal recently because Meta was being too aggressive with the ads generation.
说他们没用生成式AI，这简直荒谬。
They're the the idea that they're not implementing generative AI is absurd.
首先，生成式AI不只是你看到的最终输出结果。
Mean, I first of all, generative AI is not just what you see on on in the output.
他们提到把大语言模型和广告排序结合起来。
I mean, they talked about pairing LLMs with ad ranking.
这在广告领域可是非常前沿的研究。
That's really fascinating cutting edge research in in advertising.
用大语言模型来评估广告效果。
Using LLM to sort of give you feedback on the ad.
这个广告能引起共鸣吗？
Does this resonate?
有人会点这个广告吗？
Is this something that someone would click on?
想象一下你就是这个人群。
Like, pretend you're this demographic.
已经有研究证明，这样做对广告效果有显著提升。
There's already research that shows that that has a a really beneficial impact on advertising.
不只是你动态消息里看到的内容，连这些内容本身都是生成的。
It's not just like what you see in your feed, but also the stuff that you see in your feed is generated.
说他们没在广告里用生成式AI，这简直荒谬。
The idea that they're not utilizing generative AI in advertising, that's absurd.
你仔细看就能发现。
You can see it.
现在就去刷一下Facebook。
Just go on Facebook right now.
对。
Yeah.
很有道理。
Makes a lot of sense.
是的
Yeah.
我想聊点别的公司，但还想再问一个关于OpenAI的问题。
I I wanna move on to other companies, but I wanna have one more question about OpenAI.
你觉得，广告太精准了，会不会引发反感？
Do you think that there's any any sort of risk to push back of the creepy ads, like being too well targeted?
Facebook以前就遇到过，比如T恤上印着你的名字，还有你的整个职业轨迹。
Facebook went through that with the t shirts that said your name on them basically and your whole career path.
我猜，最终你看到的每一条广告，都是专门为你生成的。
Assumption is that eventually, like, you're only getting ads that were specifically generated for you.
以前呢，品牌会做一个广告，比如在超级碗上播一次。
Like this idea of like historically, you'd have, you know, one ad that a brand would make and they'd run it at the Super Bowl.
对。
They're like Yeah.
我们要把这个广告推给全美国。
We're gonna send this to all of America.
到了像Meta这样的平台，我猜每个广告都是单独生成的，系统完全知道该怎么定位产品，用什么颜色、什么场景来呈现。
And then eventually with, you know, on platforms like Meta, I would assume that every ad is like a one off generated and it knows it knows exactly how to position a product, what color, what environment to put it in.
对。
Mhmm.
想想以前，我们做广告，就拍一个创意素材，然后投给上千万人，真是不可思议。
And it'll be insane that we used to have ads and we'd make one piece of creative and then run it to like 10,000,000 people Yep.
因为那时候觉得这样已经不错了。
Because it was like good.
但现在不一样了，广告会越来越细分，越来越精准，越来越个性化。
And it's like, no, this will all just niche, niche, niche, niche, niche down Mhmm.
越来越细。
More and more.
是的。
Yeah.
你怎么看？
What do you think?
说实话，现在还觉得每个人打开Facebook看到的动态都一样，这简直荒谬。
Look, I mean, it's like the idea that everyone would see the same feed when they opened up Facebook is absurd now.
对吧？
Right?
你想想看。
I mean, like, look.
如果真有人那么讨厌那种让人不舒服的广告，Facebook的用户数也不会每个季度都持续增长了。
If if people truly hated creepy ads, Facebook user numbers wouldn't, you know, monotonically increase every quarter.
对吧？
Right?
而且，欧洲人早就该排队买订阅服务了。
And, you know, Europeans would be lining up to to buy subscription in Europe.
但事实上，欧洲只有1%的用户选择了这项服务。
And when it in fact, only 1% of users in Europe has taken that offer.
对吧？
Right?
是的。
Yeah.
那这个选项是
And is that an offer
取消广告，也就是Meta提供的那个服务吗？
to eliminate is that, like, an ad free, like, Meta offer?
你指的是这个吗？
Is that what you're talking about?
因为我知道他们正在计划推出
Because I know they're talking about rolling
订阅服务。
out subscription.
所以，由于遵守DMA规定，他们在欧洲推出了非个性化广告的选项。
So they've got the the less personalized ads option in Europe as a result of the DMA compliance.
对吧？
Right?
所以他们提供了三种选择。
So they have three options.
一种是完整的订阅服务，完全没广告。
They have full on subscription, no ads.
另一种是广告个性化程度较低的版本，还有就是普通的体验。
They get the less personalized ads and they get the normal experience.
只有1%的人选择了订阅服务。
1% of people opted into the subscription.
大家其实都不喜欢广告。
Like, people hate ads.
但比起其他盈利方式，他们还是更愿意接受广告。
They just hate them less than every other monetization model.
对吧？
Right?
其实人们并不讨厌广告。
People love ads.
如果你看看实际的行为表现，其实人们并不讨厌广告。
And if you look at, you know, demonstrated behavior, it people love ads.
人们喜欢免费使用各种工具。
People love access to free tools.
人们喜欢那种几乎无限使用的能力。
People love access to to to to sort of, like, endless endless capacity.
对吧？
Right?
一旦你限制了使用容量，或者设置了付费墙，
You start capacity constraining things or you start putting up paywalls.
很明显，这些东西的使用率就会下降。
You know, obviously, by definition, those are gonna get used less.
没错。
Yep.
对吧？
Right?
或者更有可能，按定义来说就是这样。
Or, you know, most likely by definition.
所以关键是，我觉得ChatGPT会遇到只有聊天机器人才有的这个问题，这我完全不认同。
And so the the thing is, like, the idea that, like, ChatGPT is gonna run into this problem that's unique to chatbots, I just don't buy that.
首先，也许他们真会这样，因为他们可能设计得不对，但我愿意给他们信任。
And first of all, maybe they will because maybe they'll design this the wrong way, but I give them the benefit of the doubt.
我认为他们完全可以做得对。
I think they can do it in the right way.
其中一个办法就是完全不把广告和聊天内容挂钩。
And one way to do it is to just not tether the ad at all to the chatbot context.
你完全可以这么说：
You could just say, look.
这是根据Capi和Pixel数据判断你可能感兴趣的商品展示广告，我们保证它和你跟聊天机器人聊的内容毫无关系。
This is a display ad for what we know you're in market for based on Capi, based on the Pixel, and and we're gonna make sure it has nothing to do with what you're talking to the chatbot about
对。
Yep.
所以你根本不需要怀疑聊天机器人是不是真心为你着想。
So you never have to even question whether the chatbot has your best interest at heart.
这是避开这个问题的简单办法，但我认为他们团队里都是很专业的人。
That's a very easy way to sidestep that problem, but I think they've got very skilled people there.
他们很清楚这个陷阱在哪，一定会避开的
They know exactly what that trap is, and they're gonna avoid
没错。
it.
是的。
Yeah.
我只想问一句，Siri在未来两年内会插入广告吗？
Let me just we're gonna get ads in Siri in the next twenty four months?
哦。
Oh.
不会。
No.
就是因为苹果做这种事的公众形象问题。
Because of just because of Apple's because of the optics of Apple doing that.
我觉得苹果其实挺想这么做的。
I think Apple would probably love to do that.
苹果内部肯定有人在推这个想法，但他们不会做，因为这跟他们对广告的定位不一致。
I think there's probably people inside Apple that are pitching that, but they they won't because it's not in alignment with their optics around advertising.
但即便如此，苹果并不会扩大广告业务。
But but nonetheless, Apple is not increasing.
我觉得，人们迟早会习惯苹果产品里出现广告。
We are talking like, I I think people will be normalized to ads in Apple products.
现在App Store里不就已经有了吗？
They obviously are in the App Store today.
德国那边提到，苹果地图很快就要加广告了。
German was talking about ads in Apple Maps Mhmm.
很快就会推出。
Coming in the near future.
从那以后，如果人们开始把Siri当搜索引擎用，广告迟早会出现。
And then I think from that point, if people are starting to use Siri as a search engine, there will be ads eventually.
毕竟，要在iPhone的界面里原生插入广告，对用户来说太自然了，他们迟早会接受的。
It will just be too to to to be able to serve ads like basically natively in the UI to iPhone consumers they
闹钟App里也得放广告。
gotta do ads in the alarm clock app.
我倒希望每天醒来都能看到不同的广告。
I wanna wake up to a different ad every day.
一睁眼就起来买东西。
Get out of bed and start buying.
赶紧上网。
Get online.
该结账了。
It's time to check out.
是的。
Yeah.
从早到晚，你都应该一直在消费。
Morning to night, you should be you should be consuming.
不对。
No.
我我
I I
我不确定。
don't know.
也许吧。
May maybe.
我觉得地图类播客肯定会有。
I think maps podcast for sure.
但关键是，我写这篇文章的时候，正是苹果推出ATT的时候。
But the thing is, like, I wrote this piece when when Apple introduced ATT.
我说苹果这是抢了黑帮的银行。
I said Apple robbed the mob's bank.
我说了，你看看。
I said, look.
这不过是他们把原本给Meta的预算，转到自己广告生态里的一次机会。
This is just an opportunity for them to to to shift some of the budget that's going to Meta into their own ads ecosystem.
他们之所以这么做，是因为扩大了苹果搜索广告的投放。
They did that because they expanded their Apple search ads.
他们把每次搜索的广告展示次数翻了一倍。
They just doubled the number of impressions per search.
现在苹果搜索广告里有两个广告位，而且他们还故意把广告和自然结果混在一起。
They've got two placements now in Apple search ads, and they're also blurring the lines between ads and organic results there.
对吧？
Right?
所以觉得苹果讨厌广告，这想法其实是错的。
So the idea that Apple hates ads, I think, is mistaken.
我觉得苹果真正要做的，是走钢丝——表面上装作讨厌广告，实际上却靠广告赚钱，毕竟这玩意儿能大幅提升利润率。
I think what Apple needs to do is they need to sort of walk this tightrope of appearing to hate ads while also benefiting from ads because they are such a, you know, an accretive sort of margin expansive way to make money.
对吧？
Right?
地图肯定算一个，我觉得播客也是个特别适合插广告的地方。
So maps, for sure, I think podcast is another, like, prime piece of real estate for ads to be inserted into.
至于Siri，我不太确定。
I don't know about Siri.
Siri要是放广告，感觉有点太直白了，不过谁知道呢？
Siri, that that feels like maybe a little bit too on the nose, but who knows?
那Gemini呢？
What about Gemini?
你对德米斯提到Gemini里放广告这事怎么看？
Did you what was your reaction to Demis talking about ads in Gemini?
这完全说得通，因为他们已经在用广告变现Gemini了。
That makes total sense because they're already monetizing Gemini with ads.
他们已经在AI概览和AI模式里这么做了。
They're doing it with AI overviews and AI mode.
他们对广告和AI概览的变现方式，和搜索是一样的。
They're they're monetizing ads and AI overviews at parity with search.
既然如此，为什么还要急着在Gemini聊天机器人里插广告呢？
Why would they rush to put ads in gem in Gemini, the chatbot?
别忘了，Gemini其实包含两个部分。
Keep in mind, Gemini is two things.
Gemini既是一系列模型，也是那个聊天机器人。
Gemini is the family of models and Gemini is the chatbot.
既然这样，为什么非要在Gemini聊天机器人里靠广告赚钱？
Why monetize with with ads in in in Gemini, the chatbot?
根本没必要这么做。
You never need to.
你完全可以靠推广来提升它的使用率。
You could just drive adoption of that.
他们已经在通过AI概览和AI模式，对Gemini模型系列进行变现了。
They're already monetizing Gemini, the family of models through AI overviews and AI mode.
AI概览每月触达二十亿人。
AI Overviews reaches 2,000,000,000 people a month.
这是目前最大的单一大模型广告展示平台。
That's the biggest single LLM output ad surface that exists.
他们已经在通过这个渠道实现盈利了。
They're already monetizing Gemini through there.
对。
Yeah.
这个产品到底是什么样子的？
And is that product just what's the shape of that product?
它和谷歌上的品牌关键词搜索完全一样吗？
Is that just the exact same as just branded keyword search on Google?
AI概览或者AI驱动的搜索广告，跟传统搜索广告有什么不同？有没有什么新发现？
Like, what is there are there anything different or any learnings from the the ads in AI search overviews or AI powered search?
AI模式其实就是那个聊天机器人。
So AI mode is essentially the chatbot.
AI概览就是现在占据谷歌搜索页面上方整个首屏的那些结果。
AI overviews is the is the results that's basically takes up now the the whole above the fold on Google search.
但它更好用。
But it's better.
比传统搜索更出色。
It's better than search.
你知道为什么吗？
You know why?
因为搜索的核心就是一次搞定。
Because the whole point with search is I want this to be one shot.
我想输入关键词，然后立刻看到最理想的结果。
I wanna type in my keyword, and I wanna get the best possible result Yeah.
就在第一组结果里。
In that first list of results.
如果看不到，那我就觉得这产品不行。
And if I don't, then I think your product is bad.
是的
Yeah.
对吧
Right?
所以谷歌有动力确保用户一次点击就搞定一切。
And so Google is incentivized to make sure that there was one click, and that's it.
你只有一次机会展示广告。
So you get one chance to show ads.
但AI概览的设计初衷，恰恰是引导用户继续提问。
With AI overviews, the whole point is to then prompt further queries.
对吧
Right?
这样一来，你就有好多次机会展示广告了。
So you get a bunch of opportunities to show ads.
对吧
Right?
所以如果这种模式的变现和以前持平，那很可能也推高了广告曝光量。
So if that's monetizing at at parity, it's probably also driving impressions up.
我觉得这比传统搜索的广告展示位要好得多。
I think that's a much better ad surface area than traditional search.
是的。
Yeah.
Lucky 应该怎么做呢？
How should Lucky.
创业者和商家该怎么看待新老板旗下的 TikTok？
Startups, business owners be thinking about TikTok under new ownership?
TikTok 很有意思，他们好像又全力投入电商了。
TikTok is funny because they're going all in on commerce again, it seems.
比如他们搞了个规模很大的黑色星期五活动，收入明显上涨了。
Like, they had a really big kind of Black Friday, you know, whatever event, you know, revenue pop revenue spike.
但话说回来，他们之前似乎有意在收缩 TikTok 的电商业务。
But, you know, they seemingly had had were walking back the TikTok commerce opportunity.
现在他们好像又开始全力推进了。
Now they seem to be leading into it.
我不太确定那边会有什么进展。
Not I'm quite sure what to expect there.
我觉得，TikTok 之前有一段时间有点停滞，主要是因为分拆计划的不确定性。
I think, you know, TikTok, I think they were in a little bit of the stasis for a while just because of the uncertainty around what was gonna happen with the with the spinout.
现在既然情况明朗了，他们大概会往自己最有优势的方向发力，而眼下电商可能就是个不错的选择，毕竟 Meta 已经完全退出了。
My sense is now that they have the certainty, they'll probably lean into where they feel like they're best positioned, and that might be commerce now that Meta has completely abandoned it.
你懂的。
You know?
但我总觉得，这种做法未必对路。
And I but I think, like, I just don't think that that's the right approach.
我觉得西方用户对这种模式没那么买账，他们更习惯纯粹的广告驱动模式。
I don't think Western audiences really appreciate that as much as just the ads driven model and so, you know, the the pure play ads driven model.
所以，我们还是看看这招最终能不能奏效吧。
So we'll see if that ends up working for them.
但说实话，我觉得如果Meta对TikTok也像对Snap那样下手，TikTok可能会陷入非常被动的境地，到时候他们可能不得不采取一些Meta不会做的调整，比如更大力度地投入电商。
But, yeah, I mean, I think TikTok you know, it it if Meta can do what they did to Snap with TikTok, they're gonna be in a pretty difficult position, and they may have to just adapt in ways that that Meta won't, and that would be leaning into commerce more heavily.
对。
Yeah.
我一直以为，也许有点天真吧，他们会主动退出电商，因为看起来那简直就是个无底洞。
The the I always assumed maybe naively that they would lean out from commerce because it seemed like it was just a money pit.
钱烧得飞快，新东家一接手就会说：好吧，我们刚买下它。
They're just lighting money on fire and the new owners would say like, okay, we just bought this.
我们还得把一半收入分给字节跳动。
We have to pass back 50 percent of our revenue to ByteDance.
整个商业模式看起来完全颠倒了。
Like, the the whole, like, economic model of the business seems like super upside down.
不如专心做广告，控制成本，先把用户基础稳住。
Let's just serve ads and try to run the business lean and maintain the user base.
但反过来大力推电商，总让我觉得不太对劲。
But leaning into commerce just feels like I don't know.
过去一年里，我听说了太多品牌，收入都涨到了数亿级别，而几乎全靠抖音小店支撑。
I I there's so many brands that that I've heard about over the last twelve months that are scaling into the hundreds of millions of revenue and and it's almost all on TikTok shop.
这背后，其实很长一段时间都是字节在补贴。
And it just feels like that was, you know, basically subsidized by ByteDance for a long time.
新东家还会愿意继续补贴这些带来大量销量的品牌吗？
And will the new owners wanna subsidize these brands that are driving real volume?
但说到底，我不确定这些品牌能不能自己站稳脚跟。
But ultimately, I don't know if they they would be able to stand on their own two feet.
是的。
Yeah.
他们确实已经退回来了。
I mean, they did walk it back.
他们在电商部门连续进行了两轮裁员。
I mean, they did two rounds of layoffs in the in the in the commerce division.
对。
Yeah.
你要知道，TikTok和字节跳动，还有这笔交易的整个背景，关键在于他们仍然在授权那个算法。
I mean, the thing about so one thing you have to remember about TikTok and the and the the ByteDance and the the whole, like, the parameters of this deal is that they're they're still licensing the algorithm.
对吧？
Right?
我觉得这影响了估值，也可能会影响其他一些经济模式。
So, like, I think that impacted the kind of the the the valuation, and that and that may impact some of these other economics.
但说实话，我觉得他们可能察觉到，Meta之前在这个领域退缩了，所以这反而是他们可以大展身手的地方。
But, yeah, I mean, I I just I just feel like they maybe they they maybe sense that that's an area where Meta retreated, and so that's where they could potentially flourish.
不过有点奇怪的是，他们之前确实撤回了，但现在又好像重新接纳了。
But it was odd because they did walk that back, and now they seem to be, you know, embracing it again.
是的。
Yeah.
你对奈飞和华纳兄弟的这笔交易现在怎么看？
What's your current thinking about the Netflix Warner Brothers deal?
感觉现在正处于一个平静期。
It's it it it feels like we're in a lull.
事情确实在推进，但我很想知道
It's moving forward, but would be interested to
看看，是的。
see Yeah.
我也很好奇，你觉得派拉蒙没了这些版权内容，还能做多好的广告平台？
I'm also interested just how how do you think about, you know, Paramount's prospects as an app ad platform without all of that IP?
是的。
Yeah.
对。
Yeah.
说起来，这挺有意思的。
Well, so it's it's interesting.
关键就在于，版权内容实在太重要了。
So, like, the the thing is, like, the IP is so so important.
对吧？
Right?
所以，我发现Netflix那笔交易里特别有意思的一点是，他们基本上把华纳兄弟游戏的估值定为零。
So, like, the one of the aspects that I I found really interesting about the the the Netflix deal was that they essentially valued WB Games at zero.
对吧？
Right?
要知道，华纳兄弟游戏可是有爆款作品的。
So WB Games makes remember, they're big hit games.
他们出了《权力的游戏》手游，还有格斗类游戏，各种主机游戏，比如《蝙蝠侠：阿卡姆》系列。
It makes their Game of Thrones game on mobile and they've got World Combat games, lot of different console games, Batman Arkham.
对。
Yeah.
而且有一段时间，它的售价在数千亿美元的范围内。
And at one point, it was for sale in the multi billions price range.
对吧？
Right?
所以Netflix基本上把它估值为零。
And so Netflix essentially valued it at zero.
为什么呢？
Why?
因为归根结底，他们必须有个IP授权协议，才能继续靠这些游戏赚钱。
Because ultimately, there have to be, like, a IP licensing agreement that allowed them to continue to monetize those games.
对吧？
Right?
我觉得WB这边的IP情况也差不多，总体上都是这样。
So I think the same is probably true on the IP side just generally with with WB.
我的感觉是，这其实说明了一件事——这一点在财报电话会里他们自己也明说了。
My sense is, like, what this indicates to me is, like and and it this was just I mean, this was said explicitly in the earnings call.
意思是，我们迟早得为内容买单，而且我们特别需要这种高端IP。
It's like, you know, we're gonna pay for content one way or the other, and, like, we need this high end IP.
对吧？
Right?
你想想，大家都看过乔·罗根那期节目，马特·达蒙和本·阿弗莱克聊到，Netflix老是让他们在对话里反复重述剧情，因为观众要是没听清楚，马上就没兴趣了——毕竟他们都在刷手机。
And, like, you know, if you look at I mean, everyone saw the the Joe Rogan clip with Matt Damon and and and Ben Affleck talking about, like, Netflix makes us restate the plot of the movie, like, every couple minutes in dialogue because if people don't hear it explicitly, they just lose interest because you guys are on their phones.
是的。
Yeah.
你不禁会想，嗯。
Like, you you wonder, okay.
这会不会是因为人们的注意力越来越短了？
Well, is that a result of, like, people's attention spans, you know, declining?
还是因为你们产出的内容本身不够吸引人，大家根本没心思看，觉得无聊？
Or is that just because the stuff you're pumping out isn't that interesting and people are not really paying attention because they're kind of bored.
对吧？
Right?
我觉得可能是后者。
And so I think maybe it's the latter.
如果你引入一批顶级IP，或许能让人更投入一些。
And so if you bring in a bunch of top notch IP, maybe that keep keeps people more engaged.
所以呢。
And then, hey.
也许这能提高广告的千次展示成本，甚至最终带动订阅人数的增长。
Maybe that boosts the the ad CPMs or may you know, and ultimately boost subscription numbers.
但我觉得，不管怎样，人们终究愿意为内容付费，也许一次性大量采购是最有效的方式。
But my sense is, like, they're gonna pay for content one way the other, and, like, you know, maybe just buying a lot of it in bulk at once is the best way to do that.
对。
Mhmm.
这周你还在关注哪些公司的财报？
What are you what else are you tracking this week across earnings?
比如微软和特斯拉。
You know, Microsoft and and Tesla.
不知道你有没有留意这些消息。
I don't know if you follow those stories.
这周还有什么其他让你感兴趣的事吗？
But what what else is interesting you this this week?
我们这边刚出了苹果的财报，是的。
Well, we got Apple Yeah.
现在就是如此。
Right now.
对。
Yeah.
没错。
Yeah.
我是说，你想想，昨天你请了那位德国人，他总是能爆出大量内部消息。
I mean, think, like, you know, you you had you had a German on yesterday, and, you know, he's just always an amazing wealth of of leaks.
他能搞到这些信息，真挺厉害的。
It's pretty impressive that he's able to get those.
苹果公司一向管得很严。
Apple's pretty buttoned down.
这确实令人印象深刻。
It's remarkable.
不过，确实如此。
But but yeah.
我的意思是，AI 这块儿现在真有点儿丢脸。
I mean, I think just the AI story there is is kind of embarrassing.
对啊。
Yeah.
我之前写过一篇稿子，说的就是这个。
And, you know, like, I wrote this piece a while back.
苹果现在是四面受敌。
Apple's besieged on all fronts.
我当时在讲，你看啊。
And I was talking about, like, look.
你要是看 AI，再看看 ATT 和欧洲那边的那些威胁，他们面临的问题可真不少。
If you look at AI, if you look at some of the threats to ATT and Europe, you know, there there there's a lot of issues that that they face.
他们宣布和谷歌合作，用 Gemini 做 Apple Intelligence，包括 Siri，我当时就在想，他们到底怎么处理隐私问题？毕竟他们以前对隐私可是寸步不让。
I think, you know, one of the big questions for me when they announced the deal with Google to use Gemini for for, you know, Apple intelligence, you know, which among other use cases include Siri, was how they would just manage the privacy aspects of that because they were so adamant.
对吧？
Right?
他们提到了私有云计算。
They talked about the private cloud compute.
他们一直强调，做这件事一定要按苹果的方式来。
They were so adamant that when we do this, we're doing it the Apple way.
要以隐私为中心来做。
We're doing it the privacy centric way.
可结果呢，他们却搞起了隐私洗白，转头就和谷歌合作，让谷歌去干那些收集数据的脏活。
And then they do the privacy washing thing, which is just then to partner with Google and allow them to do all the dirty work of sourcing the data.
对吧？
Right?
收集数据来训练模型，或者通过广告变现，他们就在后台抽成。
Sourcing the data to train the model, right, or of just monetizing with ads, like and they just get a cut on the back end.
这次是苹果在付钱，但迟早有一天，谷歌会为了打入iOS系统、获得曝光而付钱给苹果。
And in this case, Apple's paying, but at some but at some point, Google will be paying for penetration into iOS and the exposure.
没错。
Yep.
我觉得，这其实是个问题——这真的是主要原因吗？
And I think, you know, that's just a question of, like, is that really the reason?
这会不会是他们从不在这上面投资的原因之一？
Is that part of the reason why they never invested here?
因为他们没法以苹果那种注重隐私的方式来做。
Because they couldn't really do it in the Apple privacy centric way.
对。
Yeah.
如果你真的需要有人去干那些收集数据的脏活，比如手段不够光明正大，或者靠广告赚钱的话——
And if you want someone you know, if if if you understand that the dirty work of sourcing this data in ways that maybe you're not completely above board or just monetizing with ads.
苹果其实靠广告赚了非常多钱。
I mean, Apple makes so much money from advertising.
只是他们不亲自出面做罢了。
They just don't do it directly.
没错。
Mhmm.
对吧？
Right?
他们通过和谷歌的合作来实现。
They do it through the partnership with Google.
接下来他们要靠AI交互赚钱，但真正训练模型、采集数据这些可能有争议的事，他们自己并不动手。
Now they're gonna be making money through AI interactions, but they're not actually the ones training the models and sourcing the data in ways that could be questioned.
对吧？
Right?
我觉得这可能也是原因之一。
So I think that maybe plays a part in it.
而且我觉得，等到所有这些都能在设备本地完成的时候——当设备足够强大，能处理大部分需求，或许就会用更小的模型。
And I think maybe when you get to a point where all of this can be done on device, right, when the device is actually powerful enough to service most of these use cases, maybe you have smaller models.
不需要那种庞大的基础模型。
You don't have, like, foundational models.
我们可以用一系列能在设备上运行的小模型，那样的话，苹果可能会更感兴趣。
We have smaller models that can all run on device, and maybe Apple will be more interested in that.
但说实话，你得把AI这件事看成是一次重大失误，这似乎就是德国那边的看法，对吧。
But, mean, like, you you have to look at the AI thing as either as a massive whiff, seems to be what German's take was Yeah.
或者，就是觉得，你看。
Or it was just a sense of, like, look.
这真的很难跟我们的品牌形象和多年来一直依赖的宣传口径协调起来——比如隐私，这可是苹果的核心。也许我们会像外包搜索和广告给谷歌那样，先把AI也外包出去，等哪天所有东西都能在设备上本地处理了，再重新掌控这个叙事。
This is actually going to be really difficult for us to reconcile with our brand imagery and the sort of messaging that we've, you know, we we've relied on for years years, like, you know, privacy that's Apple, maybe we'll just outsource this in the same way that we outsource search, right, and advertising to Google until we're able to all process everything on device and and ensure that we can maintain that that messaging.
我的看法是，我同意‘失误’这个说法，但同时，他们这次的新收购也隐约透露出重建的迹象。
Mean, my take is, like, I agree with the whiff narrative, but, also, it does feel like there's the seeds of a rebuild between this new acquisition that they're doing.
他们开始愿意花钱了。
They're starting to open the pocketbook a little bit.
硬件确实是一流的。
The hardware really is best in class.
Mac mini的热销在一定程度上就证明了这一点。
We saw that with the Mac mini boom to some degree.
人们对苹果自研芯片的热情很高。
People are excited about the silicon, Apple silicon.
还有就是和Gemini的合作。
And then just the partnership with Gemini.
他们能接触到前沿技术，就能开始把它推广到各个领域。
Like, they're going to have access to something frontier so they can start to vend it into different places.
但一切都得按苹果自己的节奏来，而最近他们的节奏一直很慢。
But it's it's just on Apple's timelines, which have been slow lately.
所以嘛。
So yeah.
是的。
Yeah.
但iPhone Air确实失败了。
But the iPhone Air was the iPhone Air was a a flop.
对吧？
Right?
不过，折叠屏会怎么样，咱们还是拭目以待吧。
I mean, you know, we'll see what happens with the foldable.
但说实话，我觉得他们现在面临的问题是，手机的更换周期实在太长了。
But, like, I mean, I think they are facing you know, they're obviously, the the replacement cycle is really long.
我觉得有一件事被大家忽略了，我几年前写过一篇文章，当时就在问：十亿部iPhone意味着什么？
I think one thing that's, kind of under discussed is, like, I wrote this piece a while back, like, years ago, where I said, you know, what what are the implications of a billion iPhones?
那时候，全球iPhone总量还没达到十亿部呢。
And this is well before there were a billion iPhones circulation.
我的观点是，这些手机大部分都在发展中国家。
And and my point was the implications are that a lot of those are in the developing world.
对吧？
Right?
都是二手、甚至三手的iPhone在流通。
They're they're secondhand, thirdhand iPhone devices.
这就给苹果带来了一个挺微妙的困境。
And that's kind of like there's this strange dilemma that Apple faces.
手机性能越强，自然就越没必要频繁换新。
Like, the more powerful the devices get, obviously, the the the longer you can go without replacing them.
是的
Yeah.
对吧
Right?
因为 Spotify 的需求并没有变。
Because the demands of Spotify aren't changing.
从硬件角度来看，Facebook 和 Instagram 的需求也没变。
The demands of Facebook and Instagram aren't changing from, like, a hardware perspective.
对吧
Right?
所以问题就在于，你能把人工智能多大程度地融入核心使用场景？毕竟这需要多得多的计算资源。
And so the question is, like, how how how central can you make AI to the core use case, which is a lot more compute intensive Mhmm.
如果所有事情都靠设备本地处理，
If you do everything on device Yep.
那你怎么让这一点成为推动用户更频繁更换设备的关键因素？
How central can you make that to instigate a need to replace the device more frequently?
对吧？
Right?
不只是因为设备的性能到了某个阶段根本不再重要。
Not just like the because the power of the device at some point doesn't even matter.
是的。
Yeah.
好吧。
Like, okay.
你告诉我，下一代iPhone的算力是我现在手机的十倍。
You tell me, like, the compute capacity of the next iPhone is 10 x what I have now.
明白了。
Okay.
那Spotify的需求也涨十倍了吗？
What are the demands of Spotify 10 x?
因为那是我唯一用的App。
Because that's the that's the the only app I use Yeah.
每天好几个小时。
For hours and hours every day.
是的。
Yeah.
对吧？
Right?
所以这其实是个问题。
And and so that's kind of the question.
如果所有这些功能都能慢慢下放到 iPhone 7 这样的设备上呢？
And so it's like, if if all of that functionality can sort of, like, percolate down to, like, the iPhone seven.
对吧？
Right?
当人们只用 Instagram 和 Spotify 的时候，根本看不出新款 iPhone 和 iPhone 7 在使用体验上有啥区别。
And no one really sees a a difference in in sort of, like, the the in in sort of, like, lifestyle between, like, the the brand new iPhone and the iPhone seven with their only using Instagram and Spotify.
那你该怎么让人每年都换新手机呢？
How do you actually convince people to to upgrade on every cycle?
对吧？
Right?
你要知道，随着iOS用户基数的增长，大部分新增用户都会通过二手设备进入。
Because keep in mind, like, that that as that sort of installed base, like, you know, iOS installed base grows, most of that's gonna be through, like, secondhand.
是的。
Mhmm.
没错。
Yeah.
没错。
Yeah.
Jordi，你昨天说的真有意思。
Jordi, you're so It was interesting yesterday.
你提到，苹果在AI硬件方面的优势，其实源于他们当初做自动驾驶项目时积累的研发成果。
You shared that a lot of Apple's advantage on the hardware side in AI is because of the self driving push Our projects.
虽然那个汽车项目最终没成功，但那段研发经验，现在正被应用到其他产品上。
Of the car that ultimately failed, allowed them to do the r and d that that is now being rolled
是啊。
out Yeah.
苹果上个季度表现太出色了，对吧。
That's Apple had a had a huge quarter Mhmm.
营收超出了预期40亿美元，尤其是在中国市场。
Beat beat estimates on the revenue side by 4,000,000,000 with China Up specifically.
这真不错。
That's great.
盘后股价还上涨了。
And up after hours.
还有另一条挺有意思的消息，Eric，你走之前我想问问你。
The other news that's kind of interesting, I'd be curious before you leave, Eric.
ChatGPT从今天起要停用四o了。
ChatGPT is retiring four o as of today.
是的。
Yeah.
这挺有意思的。
That's interesting.
这是要彻底收尾了。
Putting putting it to bed.
我猜这会让很多人不高兴。
I'm sure that'll make a lot of people mad.
但对他们来说，流程更简单了。
But it's simpler process for them.
是的。
Yeah.
嗯，确实。
Well, yeah.
但他们同时也推出了 Go。
But they also launched Go
是的。
Yeah.
对，就是那种价格较低的订阅档次。
Right, which is the sort of like lower price point subscription tier.
所以我觉得，如果你真想推动大家使用这个，我觉得他们确实想这么做，那可能就得取消一些以前免费版依赖的老功能。
So my sense is like, you know, if you're trying to push people to adopt that, which I think they are, you maybe have to get rid of some of like the older models that the free tier relied on.
对。
Mhmm.
非常感谢你抽时间参加。
Well, thank you so much for taking the time to join.
整个流程都很快。
This is all fast.
是的。
Yeah.
真的很享受这次对话，Eric。
Really enjoyed it, Eric.
希望你尽快再回来。
Come back on soon.
是的
Yeah.
我们很快再联系你。
We'll talk to you soon.
保重。
Have a good one.
下次见。
Next time.
再见。
Goodbye.
关于 Bayt 播客
Bayt 提供中文+原文双语音频和字幕，帮助你打破语言障碍，轻松听懂全球优质播客。