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。
这是iHeart的播客节目。
This is an iHeart podcast.
百分百真人制作。
Guaranteed human.
做生意却不考虑做播客？
Run a business and not thinking about podcasting?
你再想想。
Think again.
听播客的美国人比听Spotify和Pandora广告版流媒体音乐的人还多。
More Americans listen to podcasts than ad supported streaming music from Spotify and Pandora.
作为排名第一的播客平台，iHeart的规模是第二名和第三名加起来的两倍。
And as the number one podcaster, iHeart's twice as large as the next two combined.
了解一下播客怎么帮你的生意增长。
Learn how podcasting can help your business.
拨打844844。
Call 844844
嘿，各位《波特兰迪亚》的粉丝们。
Hey, Portlandia fans.
我是凯莉·布朗斯坦和弗雷德·阿米森。
Carrie Brownstein and Fred Armisen here.
九十年代的梦想，现在以播客的形式复活了。
The dream of the nineties is alive in podcast form.
我们推出了全新播客《波德兰迪亚：重看版》，和大家一起重温《波特兰迪亚》每一集，深入剖析搞笑桥段，聊聊我们那些经典角色，还有幕后制作的那些事儿。
We're launching Podlandia AO Rewatch, our brand new podcast where we revisit every episode of Portlandia together, breaking down sketches, going deep on our iconic characters, and pulling back the curtain on how it all got made.
我们还会邀请参与这部剧的嘉宾、合作者和朋友们一起加入，比如导演乔纳森·克莱塞尔、市长本人凯尔·麦克拉克伦、传奇音乐人艾米·曼，还有更多人。
And we'll also be joined by the people who helped bring it all to life, Guest stars, collaborators, and friends, including director Jonathan Kreisel, the mayor himself, Kyle McLaughlin, legendary musician Amy Mann, and many more.
凯尔这次真是全情投入啊。
Kyle is going for it here.
你不仅即兴说了台词，连动作都是即兴的。
You fully improvised not just words, but
但我还以为你们都会写首歌呢。
but a song about Well, I thought you're all gonna write a song.
我记得你，是的。
I remember you Yeah.
我当时就是这么想的。
Thinking that.
在 iHeartRadio 应用、Apple 播客，或者你常用的任何播客平台收听《Podlandia AO 重看》吧。
Listen to Podlandia, AEO rewatch on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
我是尼克·塔图罗。
I'm Nick Taturo.
你可能在《纽约重案组》《最长的旅程》或者斯派克·李的《黑色追缉令》里见过我。
You probably know me from NYPD Blue, The Longest Yard, or Spike Lee's Black Klansman.
在我的新播客《尼克·塔图罗的送披萨幸福之旅》里，我会给一位新嘉宾送披萨。
And on my new podcast, Delivering Happiness with Nick Taturo, I deliver pizza to a new guest.
我已经和从赛斯·罗林斯——你拿我的腰带干嘛？——到比尔·伯尔的每个人都一起分享过披萨。
I've shared a slice with everyone from Seth Rollins What are you doing with my belt?
到比尔·伯尔。
To Bill Burr.
我从来没遇过这么跟我完全不一样的人。
I don't think I've ever met somebody so exactly out of
就像我这样。
their mind as I am.
接下来我们还会邀请更多精彩的嘉宾，比如著名的约翰·多图罗。
And now, we even have more great guests coming up, including the great John Dotturro.
这叫幸福。
It's called happiness.
送幸福。
Delivering happiness.
还有更多更多。
And many, many more.
打开免费的iHeartRadio应用，搜索‘Delivering Happiness with Nick Taturo’，立即收听。
Open your free iHeartRadio app, search delivering happiness with Nick Dotturro, and listen now.
这里是《亲爱的切尔西》的切尔西·汉德勒。
This is Chelsea Handler from Dear Chelsea.
每周新闻都越来越糟，世界也越来越疯狂，而亚曼eka就要告诉那些该负责的人：问题出在你们身上。
Every week, the news gets worse, the world gets crazier, and Yamaneka is here to tell whoever's responsible, you're the problem.
你知道吗，我刚发现
Do you know I just found
西德尼·斯梅尼到底是谁。
out who Sydney Sweeney was?
如果他能拥有一堆女生，那我也该有一堆男生。
If he got a bunch of women, then I should have a bunch of men.
你要是做不到更好，就少干点，别逼我干这么多。
Do better or do less so I don't have to do so much.
我是亚曼eka，我先走了。
I'm Yamaneka, and I'm out.
收听亚曼eka的节目《你才是问题所在》，在
Listen to You're the Problem with Yamaneka on
iHeartRadio应用、Apple播客，或者你常用的任何播客平台都能听到。
the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
Callzone Media。
Callzone Media.
您好。
Greetings.
你好。
Hello.
恭候大驾。
And salutations.
我是埃德·齐特隆，欢迎收听《更好离线》。
I am Ed Zitron, and this is Better Offline.
更好离线。
Better Offline.
我们回来了。
We are back.
你得订阅我们的通讯。
You need to subscribe to the newsletter.
你已经在听这个播客了，那就去下载一下吧。
You need to download the podcast you're already listening to.
拿上朋友的手机，也给他们下载上。
Take your friends' phones, download it on there.
如果你不这么做，那就让我失望了。
You will have failed me if you don't do this.
其实也不是。
Not really.
我真的很爱你们所有人。
I love you all dearly.
但你听好了。
But look.
今天，我邀请到了杰出的作家兼顾问尼克·苏雷什，他两周前发表了一篇关于人工智能的精彩文章，是我读过最好的之一——当然，不只是因为它印证了我的观点。
Today, I am joined by the incredible writer and consultant Nik Suresh, who two weeks ago put out one of the best pieces on AI I've ever read and not just because it confirms my biases.
这篇文章的名字叫《人工智能狂热正在摧毁全球决策》。
And it's called AI mania is eviscerating global decision making.
Nik，谢谢你再次来参加。
Nik, thank you for joining me.
你好。
Hey.
非常感谢你再次邀请我。
Thank you so much for having me on again.
这篇文章真让人害怕。
This piece was terrifying.
我得说清楚，我读到的这些内容，很多我都以为早就发生了。
I just wanna be really clear that, like, this the things that I read in here, a lot of them, I was like, I assumed this was happening.
我以为情况已经这么糟了。
I assumed that it was this bad.
简单来说，这篇文章指出，现在在商业圈里，AI简直像一种信仰，不管它到底有没有用，大家都一股脑儿跟风。
But to summarize the piece in short, it appears that just AI is this bordering on religious thing within business now that people get into regardless of whether it even does anything?
是的。
Yeah.
是的。
Yeah.
而且你知道，真正让这篇文章广泛传播的原因是，就连那些在工作中使用AI并声称从中受益的人，很多比较理性的人也在看着同事，心想：你们到底在搞什么？
And, you know, I I think the thing that got us so widely circulated is even the people who use AI in their work and say they're seeing benefits, a lot of the more sensible people are looking at their colleagues and going, what are you guys doing?
我觉得我的工作效率确实提高了一点，但我们的CEO却说：干脆把其他所有项目都取消吧。
I think I'm a little bit faster at my job, and here we have my CEO saying, let's just scrap every other project.
把所有员工都裁掉。
We'll remove all the employees.
我们就能实现一百倍的生产力。
We're gonna have 100x productivity.
就连那些本身喜欢用AI的理性人士，也开始对这种狂热程度感到担忧了。
And even sensible people that enjoy it are getting a little bit worried by the level the mania has gotten to.
对。
Yeah.
那你给我举个具体的例子吧。
So give me give me an example here.
第一部分叫‘AI投资基本全军覆没’，说实话，这标题已经把话说得明明白白了。
The the first section is called AI investments are generally total failures, and I realize the title really does say it all there.
但你能给我具体解释一下吗？
But can you break that down for me?
是的。
Yeah.
有意思的是，大多数公司里的项目本来就会失败，尤其是大公司。
So it's interesting because, you know, most projects at companies do fail, especially big companies.
他们本来就不太会搞技术项目。
They're just bad at running tech projects.
早在AI还没火起来的时候，我就写过这事儿。
I've written about this well before the AI thing was big.
现在更惊人的是，AI项目的失败率反而更高。
What is fascinating is that they have an even higher failure rate for the AI projects.
而且，咱们也不用跟人争论AI到底有没有用。
And, again, like, just to we we don't need to get into an argument with people about whether it works or not.
对吧？
Right?
我们可以先从基础说起——普通IT项目本来就会失败很多。
We can just start with the basis of normal IT projects fail a bunch.
没错。
Right.
AI项目是新兴技术，所以失败得更厉害。
AI projects are novel technology, so these are failing even more.
公司通常会设法控制成本，避免在IT上花太多钱。
And the protective factor to companies normally that they don't want to spend all their money on IT.
他们其实只想在IT上花最少的钱。
They actually wanna spend the smallest amount on IT possible.
对吧？
Right?
这本质上就是企业的成本支出。
It's mostly a thing you're it's a cost for the business.
对。
Right.
但AI的情况完全反过来了，公司们都在想：我得多争取点预算，能拿多少就拿多少。
With AI, this is totally inverted, and companies are going, I'd like to take as much budget as I can feasibly get.
我要把钱全砸进这个我根本不擅长的领域，还要挑最难的模式来玩。
I'm gonna put it all into this thing that I know I'm bad at, and I wanna do a harder version of that.
但我没法跟任何人说，我已经白白烧掉几百万美元了。
But I can't admit to anyone I've just wasted millions and millions of dollars.
所以全球几乎所有组织都在做同一件事：全力投入自己最不擅长的领域，选最难的难度，事后还假装一切顺利。
So it's sort of like all the organizations in the planet are going all in on the thing they know they're bad at, playing it on hard mode, and then sort of just lying about it after the fact.
你说到‘撒谎’，具体是指什么？
And when you say lying, what do you mean?
他们是说：‘没错，AI让我效率提升了上亿个百分点’？
Are they just being like, yeah, this is making me one bazillion more productive percents.
我玩AI可太在行了。
I'm so good at AI.
这太有意思了。
I love it.
我的意思是，他们到底在搞什么名堂？
Like, what what is the like, what are they doing?
他们到底拿这些玩意儿干了些什么？
What are they actually doing with this crap?
嗯，情况是这样的。
It's well, okay.
这其实有好几种表现方式。
So it it happens in a couple of ways.
对吧？
Right?
我说的‘撒谎’，听起来好像特别有预谋。
When I say lying, that sounds, like, very intentional.
确实有人会撒谎，毕竟哪儿都有这样的人。
And there are people that lie because there are people that lie everywhere.
但通常来说，无非是两种情况。
But usually it's two things.
从内部来看，往往充满自我欺骗。
Internally, it can be a lot of delusion.
很多高管投资了人工智能项目，却没意识到整个组织都在拼命向他们展示最乐观、最美好的假象。
There are a lot of executives out there who have funded AI projects, and they don't realize that the whole organization has shifted into presenting them with the rosiest view of what is happening that is possible.
比如，你是个CEO，听说了人工智能的好处，而你手下有两个家伙正争着想升职。
Like, if you're a CEO who has heard good things about AI, and let's say you've got two guys underneath you gunning for a promotion Right.
这两个人现在展开了竞赛，比谁的AI项目更多、生产力更高、应用范围更广。
Those two guys are now at an arms race to talk about their AI productivity and how much AI they're rolling out across the business.
这成了他们的目标。
Like, that that is the goal.
根本不是为了赚钱。
It's not to make money.
而是要显得自己最‘AI’，管它到底什么意思呢。
It's to be the most AI, whatever the fuck that means.
对吧？
Right?
所以这简直荒唐至极。
So it it's it's pretty nonsensical.
但这正是自我制造出来的局面。
That but that's that's self creation.
是的。
Yeah.
这现象特别值得观察。
It's it's very up to watch.
而且再说一遍，这只是我们平时看到的现象被放大到了极端。
And and again, it's it's an extreme version of what we already see.
对吧？
Right?
因为人们早就为了升职而欺骗CEO了。
Because people already lie to CEOs for promotions.
这简直就是加了猛药的版本。
This is just that on steroids.
你在这里提到，说你在一年半里看到过零成功率。
So you mention in here, you say you've seen 0% success in a year and a half.
真的吗？
Really?
真的是零吗？你一点都没看到？
Is it really zero you've seen nothing?
我知道我的听众会特别喜欢这个说法，但我还是想确认一下。
I just wanna be my audience is gonna eat this up, I realize, but, like, I just wanna make sure.
是的。
Yes.
我们得加个注释，因为有些人误解了这句话。
We we should, you know, add one asterisk there because some people misunderstood that.
我看到的是零，不代表全世界都是零，毕竟我没看过全球所有的项目。
Me seeing 0% doesn't mean, like, it's 0% across the whole world because I haven't seen every project in the world.
可惜有些人阅读理解能力不够，你得说得明明白白，不然就会冒出一堆AI吹捧者说：‘又不是真的一点都没有’。
Sadly, some people don't have enough reading comprehension for that, so you have to lay it out or you get, like, an AI booster who's like, it's not literally zero.
肯定有人在某些地方看到了一点成效。
Some must be getting some results somewhere.
我真的一点都没见过，过去一年半，我一直在澳大利亚和美国各地的公司做销售、跑项目。
I have seen nothing, and I've been doing nothing but sales and on sites at companies, like, all across Australia and The US for a year and a half now.
我确实没见到。
I do.
他们根本不会找我去做这些事。
I don't get brought in to do them.
我偶尔看到一些零散的案例，心里就明白：这根本行不通。
They're, like, things I see on the side, and I go, well, that's that is not working.
我听到有人说‘你也没看过所有的软件项目’的时候，就想到了一点。
Well, the thing I think about when I hear that people are like, actually, it's you haven't seen every software thing.
这就是AI吹捧者惯用的套路。
It's just this is how AI boosters work.
他们希望你作为现实的捍卫者，单方面去证明一切。
They want you as the defender of reality to be like, prove unilaterally.
他们说人工智能太差了。
This is the AI is terrible.
既然这么差，那你证明给我看啊。
If it's so bad, prove it.
但当你让他们拿出证据时，他们就看着你说：哎呀，数字嘛，今年我搞的数字可大了。
But when you ask them to prove anything, they look at you and go, well, you know, the the numbers is just the bigger the number number I made so big this year.
这是我见过最大的数字，可就是不肯给你看任何真实的东西。
It's the biggest I've ever seen, and they refuse to show you anything real.
我觉得现在他们已经给我发了15次那个Erdos问题了。
I think I've been sent the fucking Erdos problem thing 15 times now.
下次再有人发给我，我就画个大饼，纯属搞笑。
And the next time someone sends it to me, I'm gonna draw a gut that is a joke.
这纯粹就是个笑话。
That's a joke.
我不会拿那把枪做任何事。
I'm not gonna not gonna do anything with the gun.
但我可能会想象一下《电锯惊魂》里的拼图人。
But I am just gonna possibly imagine that Saw from Jigsaw.
抱歉。
Sorry.
拼图人，就是《电锯惊魂》里的那个，估计评论区又要骂我了。
Jigsaw from Saw or even they're gonna get me in the comments for that one.
我只是想把他们拉回来，问清楚：你觉得厄多斯问题到底是什么意思？
Just kinda tying them to radiate and say, show me tell me what you think the Erdos problem means.
告诉我，它到底意味着什么。
Tell me what it means.
说说看，你认为厄多斯指的是什么？
Tell me what you think Erdos means.
因为这才是关键。
Because that's the thing.
这些人根本没什么实质内容可以分享。
These people don't have anything tangible to share.
我是在机场读你的文章的，当时就坐在那儿，心里直骂‘天啊’，确实引来几道目光，因为那篇文章让我太震惊了——但它简直把我这两年一直想说却说不清的东西，一下子说透了。
And your piece kind of I read it in an airport, and I was just sitting there going up fucking hell for and peep I did get a few looks because I was so surprised by it, but it feels like your piece crystallizes everything I've been thinking for two years.
现在这种怪现象就是风格彻底死了，人人都在说：‘天哪，我太爱AI了。’
That just this is this weird death of style in style situation where people are just saying, goddamn, I love AI.
因为他们要是不说这话，怕让老板难过、客户难过、同事也难过，可实际上，谁也没从AI那儿得到真好处？
Because if they don't say it, they'll make their boss feel bad, their customers feel bad, and their colleagues feel bad despite none of them getting anything out of AI?
是的。
Yeah.
简单来说，每个企业通常都卡在一个关键问题上，这个问题直接影响了你赚钱的效率。
You know, the the simple way to think about it is that any given business, you know, you have one problem generally that is bottlenecking how you're making most of your money.
不管你觉得AI能做什么，只有当它能解决这个核心问题时，才值得投入。
And whatever you think AI can do, it's really only the correct thing to deploy if it solves that one problem.
写更多代码本身没用，如果写代码根本不是你赚钱的关键路径。
It doesn't matter if you write more code if writing code isn't the way to make more money.
没错。
And Yes.
我的意思是，事情其实就是这么简单。
I mean, it's it's basically just like as simple as that.
对吧？
Right?
人们都在全力投入一些根本不符合他们实际需求的东西。
People are going all in on a thing that doesn't really make sense for whatever context they're inhabiting.
你在文章里提到，你不得不停止提供AI服务，或者说是某一项AI服务。
So you mentioned within the piece that you had to basically stop offering AI services or a specific one.
如果我理解错了，请纠正我，但你是说在做网站项目时，根本不再提AI服务，因为客户一听到就反应过度。
Forgive me if I forgive me if I misread that, but you just won't mention AI services within surf software projects because people go outrageous over them.
是的。
Yeah.
我们发现，作为一个规模不大的公司，我们特别注重与客户的协作。
What what we found was, you know, we're a relatively small business, and we optimize a lot for working with people.
我们很喜欢。
We enjoy.
你是做软件的，
And you're software
澄清一下，你们是咨询公司。
consultancy, just to be clear.
对，我们是一家专注于数据的软件咨询公司。
Yes, we're a software consultancy in data specifically.
我的研究生学位是数据科学硕士。
So my post grad degree was a master's in data science.
在大语言模型热潮兴起之前，我就已经涉足人工智能了。
I was into AI nominally before the whole LLM thing took off.
我们发现，只要一提到人工智能，就会发生两件事。
And yeah, what what we found was whenever we mentioned AI, two things would happen.
我在文章里提到过，一说AI，客户就会变得特别热衷于找我们下单。
And and the one I mentioned in the piece was people would become rabidly interested in buying from us.
但他们也是特别难缠的客户。
But they were also really bad customers.
简直做了一堆糟糕的决定。
Like, they made bad decisions.
现金流总是出问题。
They had cash flow issues.
做事根本想不清楚。
They didn't think clearly.
说实话，这种客户我们根本不想接，生活方式上就不合拍。
Like, these were not people we wanted to work with as a matter of lifestyle.
对。
Mhmm.
而且你也知道，很多初创公司都这样。
But also, you know, it's like a lot of start ups and stuff.
你得把钱全预收了，不然真怕他们骗完就跑路。
The the sort of customer where you have to take all the money upfront because you're worried they're just gonna scam you.
这些根本不是你能跟他们搅和在一起的人。
Like, these are not people you wanna be in bed with.
而我们真正喜欢的客户，我们只是顺口提一句，他们立马就没了兴趣，因为总有人跟他们说现在到处都是骗局。
And the people we really liked, we would mention it just as an aside, and they would immediately become so disinterested because they just go, well, you know, a lot of people are trying to scam me with this right now.
我真的没兴趣。
I'm just not interested.
所以我们干脆在任何场合都不提这事儿。
So we just don't talk about it at all in any context.
我们也不喜欢谈这个。
We don't like doing it.
这活儿本身就很无聊。
The work's really boring anyway.
不过呢，我再跟你说个事儿。
But, yeah, like, it's just here's another thing I can tell you.
那些完全不提人工智能的客户，赚钱最多，因为他们只专注眼下最棘手的问题。
Our clients that don't talk about AI at all are, like, far and away the most profitable because they're focused on whatever their biggest problem is at any given moment.
这其实非常简单。
It's very, very simple.
这根本不是什么高深的技术。
Like, this is not rocket science.
对。
Yeah.
那些表现好的公司，都是在专心解决实际的业务问题。
It's just like it's like the companies that are doing well are focusing on solving business issues.
表现差的公司，反而在花大价钱买些花里胡哨的数字小玩意儿。
The companies that are doing badly are the ones getting very expensive digital fidget spinners.
没错。
Yeah.
事情就是这么简单。
Like, it's it's as simple as that.
我真没见过几家公司的核心问题，是靠大语言模型就能解决的。
I don't know many companies where LLMs unblock, like, whatever their core issue is.
我能想到的可能是客服，但我对客服这块其实不太了解。
There's stuff I can think like, maybe customer support, but I don't know much about customer support.
只要是我懂的领域，我就觉得，现在最不需要的，就是什么大语言模型。
Anything where I have some subject matter expertise, I'm like, an LLM is not the most important thing you need right now.
我真想不出还有什么例子。
I can't think of it.
我白白浪费了两三年，陷在这些破事儿里。
I went to two, three, four boobless years into this bullshit.
而且我现在觉得，
And I am
我是个离不开表格的人，但这东西根本做不好表格。
a spreadsheet hog, and it doesn't do spreadsheets very well.
我得说，要是别人听到我这么说，肯定得气炸了。
I will say people could be met people would be very mad at me for this.
但我还真找到一个大语言模型有用的场景。
I found one use of LLMs.
一个。
One.
我觉得我跟你说过这事，彭博终端，因为你在彭博终端上请求数据的方式是通过Python。
And I think I've told you about this, the Bloomberg terminal, because it the way that you act the way that you request stuff on a Bloomberg terminal is through Python.
我不写Python代码。
I don't write Python.
所以你只要说，‘彭博AI，告诉我怎么获取这个数据’，它就会给你代码，然后直接运行。
So you'd be like, Bloomberg LLM, tell me how to get this, and it gives you the code and it runs it.
搞定。
Boom.
简单得很。
Simple.
说白了，就是个高大上的搜索功能。
Probably the least you like, literally just sexy search.
干得漂亮。
Well done.
这11000亿美元的资本支出，我敢说肯定都值了。
That $1,100,000,000,000 of CapEx, I'm sure that was all worth it.
但有趣的是，你文章里举的那个Snowflake的AI聊天机器人例子，恰恰是LLM应该擅长的——数据库查询。
But what's funny is the example you give in the piece of the AI chatbot thing for Snowflake feels like the things that LLM should be good at, which is a database request.
对吧？
Right?
就是说，它本来
Like, the it
那为什么就不行呢？
why doesn't it work?
这明明才是LLM最该拿手的啊。
Surely, that is the thing that this should be good at.
问题到底出在哪？为什么LLM解决不了？
What is the issue there, and why don't LLM solve it?
是啊。
Yeah.
那里的复杂性在于，说实话，我对彭博终端了解不多，不过你刚才说的那些事儿还真挺有意思的。
The complexity there is, you know, I don't know much about Bloomberg terminals, although it's very funny everything you've told me about them.
这简直就是个自闭症患者专用的入口。
It's an autism portal.
功能太强大了。
It's so good.
对啊。
Yeah.
所以我觉得，它们能做的事情其实很有限。
So, you know, I imagine they're fairly constrained in what they can do.
没错。
Yes.
比如你提到写Python代码，这意味着有人已经开发了一个Python库，用来连接彭博终端，限制了你能问它些什么。
For example, you know, you mentioned writing Python, which means someone has written a a Python library that interfaces with the Bloomberg terminal and constrains what sorts of things you can ask it.
如果这个系统设计得足够好，把所有信息都整合进去，能让人顺利查询的那些方式，同样也能让大模型来操作。
So it's if it's very well designed and you put all the text in there, the same thing that lets me, human, interrogate it would let an LLM do it.
在商业环境中，最难的是很多东西根本没写下来，只存在于某个人的脑子里，现实世界实在太复杂了。
What's very hard at a business is a lot of that stuff is just not written down or it exists in someone's head or the real world is very, very complicated.
我们不妨用这个例子来说。
And the very people who are let let's use this example.
对吧？
Right?
那为什么Cortex会失效呢？
Why does Cortex not work?
你作为CFO，进去输入：我想知道去年我们赚了多少钱。
So you go in and you're the CFO and you type in, I would like to know how much revenue we made last year.
但它并不了解你业务里的那些细枝末节。
It doesn't know all the little nuances of your business.
对于这种关键问题，结果必须百分之百准确。
And for something really important, you need it to be correct.
但如果你是CFO，又在用大模型，你根本没法去排查它给出的结果哪里出了问题。
But if you're making use of an LLM and you're the CFO, you can't debug whatever it put out.
当数字完全错的时候，这倒也没关系，因为你肯定会发现：等等，
And that's okay when the number is totally wrong because you'll go, hey.
这不对劲。
This isn't correct.
如果偏差只有5%到10%，你很可能根本察觉不到。
If it's off by, like, 5% or 10%, you probably don't know.
就像你用彭博终端时那样。
And exactly what like with your Bloomberg terminal.
对吧？
Right?
如果你查到一个特别重要的数据，你可能会打电话问朋友：这个Python结果对吗？
If you looked up something really important, you might reach out to a friend and say, is is this Python correct?
因为我马上就要根据这个做一笔大额支出。
Because I'm about to spend a lot of money on this.
对。
Yeah.
澄清一下，当你在终端上发出请求时，它其实是在运行Python代码。
And to be clear, when you make a request on the terminal, it does the Python.
这代码只是用来向彭博获取真实数据的。
It's just the Python code to request Bloomberg's actual data.
我猜你不会自己写Python，但我想背后有个库，这个界面只是在上面加了一层，让像我这样的人能用，不用手动敲三四个字母、等弹出结果，再花几个小时折腾——当然，我以前没用LLM的时候就是这么干的。
Like, it hasn't I imagine don't write Python, but I imagine there is a library in this that this is just a layer on top of it so that people like me can actually use it without just punching the typing in, like, three letter things and making things pop up and then spending hours fucking with it, which, of course, I never did before I used the LLM side.
当然了，
Was it's, of course
是的。
Yeah.
没错。
Yeah.
但用了Cortex之后，
But with Cortex
但用Cortex的搜索功能时，问题好像也是它在编造内容？
but with Cortex, the search side, it seems that the problem is also that is it making stuff up?
我不是在装无知，但我觉得，面对这么庞大的数据库和信息，从中提取内容正是大语言模型擅长的，还是说，有人觉得数据库层是大语言模型解决不了的问题？
Because I and not being an ignoramus here, surely looking over a big database and information and pulling stuff out is what an LLM is good at, or is the suggestion that the database layer is not something LLMs can solve?
这事儿真的特别复杂，你知道的。
It's just very complicated, you know.
昨天我们去客户现场，他们内部有个账本，记录着所有进出的交易。
So, we were at a client site yesterday, and, you know, they have this ledger internally which tracks, like, all the transactions coming in and out.
光是账户类型就有四十多种，分散在不同的表里。
And there's, like, 40 different account types stored in all these different tables.
在不同情况下，你要用不同的交易类型来计算服务使用率之类的指标。
And in different situations, you wanna use different types of transactions to calculate certain things around, like, service utilization.
大语言模型一看这个，就说：我在这儿没干过八年，根本不知道怎么算这个。
And an LLM looks at that, and it just goes, well, I haven't worked here for eight years, so I don't know how to run this calculation.
真的，太复杂了。
Like, it's very, very complicated.
所以问题在于，数据本身没标清楚，还有其他各种问题。
So the data is so it's that the data isn't labeled properly and all these other.
所以问题在于数据是怎么呈现和查看的？
So it's how the data is presented and looked over?
对。
Yeah.
没错。
Yeah.
就算你把数据标注得再准确，人类专家的存在还是有道理的。
You can even label it properly, but there's a reason we have human specialists.
这跟我不让大语言模型帮我起草所有合同是一个道理。
It's the same reason I don't get an LLM to write all my contracts.
我当然可以这么做，但我会去找我的律师伊恩，问他：这件事我到底该怎么处理？
I could, but instead I go to my lawyer, Ian, and I go, what should I actually do here?
你知道的，就是这样。
You know, like Yeah.
这是经验跟数据之间的区别。
It's experience versus data.
没错。
Right.
你知道吗，我们刚创业的时候，他给我起草了一份合同，就是那种标准模板。
You know, like, when we started the business, he wrote a contract for me that, you know, took a standard one.
里面有些条款，我根本想不到会需要。
And it has things in there, like, I wouldn't have thought of.
比如我们和客户约定的邮件回复时间间隔，还有这些细节。
Like, for example, the agreement we have with clients on how long we take between email responses and Yep.
我们是怎么收集的。
How we capture
我以前做公关的时候，里面也全是这种东西。
When I was still doing PR stuff, that's exactly that's exactly what I had in there.
就是一些你没经历过根本不会考虑到的 basics。
Like, basic things that you wouldn't think of unless you'd experienced the thing.
但有意思的是，就连现在那些反对大模型的人，也常说：‘不过大模型用来查大数据库还挺管用的。’
But it's funny because even now, even people who are anti LLM, I will hear them say, but it's good for searching large databases.
听起来好像并不是这样。
And it sounds like it isn't.
也就是说，AI 在这方面并不擅长。
Like, AI is just not good at that.
我的专长是特定领域，就是大规模的信息检索。
I specialize in the specific area, which is just like big information retrieval.
它在处理数据库方面确实不错，不过在科技圈里，'数据库'这个词有特定含义。
It's pretty good at you know, the the a database specifically in tech means something.
它指的是像 PostgreSQL 这样的系统，但普通人根本听不懂这些术语。
It means post systems like Postgres or whichever, which which mean nothing to normal people.
我觉得它们特别擅长快速浏览海量文本，然后把人引到正确的方向。
I would say they're pretty good at reading gigantic amounts of texts and then pointing a human at the correct place.
比如，当我们刚到一个新客户那里，他们有一百万行乱七八糟的代码，我就会说：我得找找这段逻辑可能出现在哪儿。
So a good use is if we're, like, at a new client site and they have a million lines of some fucking thing, I can go, I need to look for the places where this logic might happen.
然后它们会给我指出来，虽然有时候还是不准，但至少我能有个方向了。
And they'll point me at them and it it's still wrong sometimes, but at least I can look.
这也不是没用，但问题是，我愿意为这个付多少钱呢？
That's not useless, but it's like, how much would I pay for that?
我不确定。
I don't know.
也许一个月二十美元吧。
Maybe, like, $20 a month.
你意思是说？
You know mean?
但这已经很划算了。
But that's a deal.
但关键就在这儿。
But that's the thing.
就连用命令行也是同样的情况。
It's like, even with the terminal, it's the same thing.
它还是会出错。
It will get things wrong.
它从来不会算错数字，因为计算的时候，数据是直接从特定来源提取的。
It never gets the numbers wrong because when it runs the numbers, it's pulling them from a specific place.
所以我每次用大模型的时候，都会再跑一遍，再跑一遍，然后自己去谷歌查证，因为我不信任何跟这些玩意儿沾边的东西。
And I still tell it if I ever touch the LLM stuff, run it again, run it again, just like and then I'll Google it, and I'll look up the stuff because I don't trust anything that touches them.
但归根结底，这其实就是一场盲目探索。
But still, it's just it is a fishing expedition.
你不能指望它有逻辑，而我觉得，大多数人用搜索引擎时，其实期待的就是逻辑——可人类自己提问时，往往根本说不清自己到底想要什么。
You can't rely on it to have any logic, which I think is what most people want from search when the, human beings are pretty terrible about asking for exactly what they want.
结果就是，它根本靠不住。
And I feel like it just falls down.
我老提这个，是因为我听说的大多数大模型应用，核心就是搜索大型数据库和做决策。
And the reason I keep going on about this stuff is because searching big databases and making decisions feels like 90% of the LLM integrations I've heard about.
人们说它有用，无非就是‘我该怎么用’‘这到底意味着什么’这类问题，这几乎成了他们描述业务价值的标配。
Like, it's just how do I do this and what does this mean is how people are kind of seem to at least be describing how it's useful in their business.
但它根本做不到这些，这问题就大了。
But it but it can't do that, which is bad.
是的。
Yeah.
归根结底，只要我们还得手动核对结果是否正确，哪怕假设它们大部分时候是对的——但事实往往并非如此。
It at the end of the day, as long as we're constrained by having to go in and check whether the results are correct, even if we grant that they're correct most of the time, which is not always true.
这事儿特别依赖具体领域。
It's very domain specific.
这就意味着你不可能快上一百倍，而正是这种提速的承诺，才让那么多钱砸进去。
That just means you can't get a 100 times faster, and that's what's justifying the bajillion million infinity dollars of investment.
对吧？
Right?
如果我们说，
If we went, hey.
这能让你快一半，
This makes you 50% faster.
那还行。
Like, alright.
这值得了一万亿美元吗？
Is that worth a trillion dollars?
我觉得不值。
I I don't think so.
我觉得它不值一万亿美元。
I don't think it's worth a trillion dollars.
这个市场逻辑完全建立在那种疯狂的宣称上。
Like, the the market thesis depends on the psychosis claim.
而且，说到底，这东西本来就是叫人工智能。
Also, I mean, it's meant to be artificial intelligence.
明确一点，‘智能’这个词可是写在名字里的。
Just to be clear, that's the intelligence part is in the name.
但我们说的这些，听起来简直蠢透了。
But everything we're saying makes it sound really fucking stupid.
根本就不是什么聪明的做法。
Just like it's just it's not smart to do.
展开剩余字幕（还有 480 条）
它根本什么都不知道，干的都是些傻事。
It doesn't know anything, and it does dumb guy stuff.
你没按我需要的方式问我。
Oh, you didn't ask me in exactly the way I needed to be asked.
我不太明白，老兄。
I don't understand, mate.
抱歉。
Sorry.
我帮不上什么忙。
I I'm not gonna be helpful at all.
而且，我们说的这些其实都很简单。
And, also, what we are describing is very simple stuff.
这值不了上万亿。
It's not a trillion dollar.
连五十亿的火腿产业都算不上。
It's not even a 50,000,000,000 ham industry.
就算它完全理解了我们说的，也还是没那么让人激动，对吧？
It's just even if it got everything we're saying right, it still wouldn't be that exciting?
是的。
Yeah.
说实话，刚推出的时候我觉得特别有意思。
You know, I I think it was it was very interesting when it came out.
那时候确实挺新颖的。
It was it was, like, pretty novel.
GPT-3刚出来的时候，我简直惊呆了。
When GPT three came out, I went, what the hell?
我上大学那会儿可没有这种东西。
Like, this was not available when I was at university.
这简直太疯狂了。
This is fucking wild.
我觉得它现在早就该过气了。
I think this really sort of overstayed its welcome.
这其实特别有意思，因为我们用的这项技术，如果你暂时抛开那些关于剽窃艺术家、环境破坏之类的伦理问题——我不是说这些不重要，但为了讨论，我先搁置它们。
And it it's so interesting because we've taken a technology that I mean, if you're willing to set aside many, many ethical claims around stealing from artists and environmental things, like, if we just hand wave all of that, which I'm not happy to do, but in this context for an argument, I will.
对。
Yeah.
是啊。
Yeah.
要是把这些都忽略掉，那确实，我们搞出个挺厉害的东西。
If you hand wave all of that, it's like, alright.
可结果呢？我们还是把一切都搞砸了。
We've made something pretty neat, and then we've still somehow completely managed to fuck everything up.
我们本来只要在怎么使用这项技术上保持理性十秒钟，再好好投入一下，就能做好，可我们偏偏全搞砸了。
Like, all we had to do was be reasonable for ten seconds about how we use this and invest in it, and we have completely failed.
所有人都疯了。
Everyone has gone insane.
做生意却不考虑做播客？
Run a business and not thinking about podcasting?
再好好想想。
Think again.
听播客的美国人比听Spotify和Pandora广告版音乐的人还多。
More Americans listen to podcasts than ad supported streaming music from Spotify and Pandora.
作为排名第一的播客平台，iHeart的规模是第二名和第三名加起来的两倍。
And as the number one podcaster, iHeart's twice as large as the next two combined.
不管你的客户听什么，你的信息都能传到他们耳边。
So whatever your customers listen to, they'll hear your message.
而且，只有iHeart能把你的信息扩展到广播电台的听众群体。
Plus, only iHeart can extend your message to audiences across broadcast radio.
你觉得播客能帮到你的生意吗？
Think podcasting can help your business?
想想iHeart、流媒体、广播和播客。
Think iHeart, streaming, radio, and podcasting.
去iHeart广告官网看看吧，我们来为你展示。
Let us show you at iHeart advertising dot com.
访问 iHeartAdvertising.com 了解更多。
That's iheartadvertising.com.
我是南希·格拉斯，播客《罪疚之重》第三季的主持人。
I'm Nancy Glass, host of the burden of guilt season three podcast.
这是一个关于一场噩梦般事件的故事，它摧毁了两个家庭。
This is a story about a horrendous situation that destroyed two families.
事情要追溯到七年前，辛迪和肯尼刚订婚，两人都刚满四十岁。
It started more than seven years ago when Cindy and Kenny were two 40 who had just gotten engaged.
几天后，他们卷入了一场完全意想不到的风暴。
Days later, they walked into a scorched earth war they never saw coming.
转过身，把双手放到背后。
Turn around and place your hands behind your back.
警察闯进我家，直接把我抓进了监狱。
The cops came into my house and threw me into jail.
令辛迪震惊的是，她的未婚夫肯尼被指控虐待她的一个孩子。
And to Cindy's surprise, her fiance Kenny was charged with abusing one of her children.
我从未怀疑过肯尼，我一直觉得这个指控有问题。
I never once doubted Kenny and I always doubted the accusation.
在《罪恶的负担》第三季中，你会听到一个关于人为了得到想要的东西会做到什么地步的故事。
On season three of burden of guilt, you'll hear a story about the lengths someone will go to in order to get what they want.
这个案子绝对是其中最卑劣的之一。
This case has got to be one of the most despicable.
请在 iHeartRadio 应用、Apple 播客或你常用的播客平台收听《罪恶的负担》第三季。
Listen to season three of burden of guilt on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
最近怎么样？
What's up, fam?
我是体育记者阿里·钱伯斯。
I'm sports journalist Ari Chambers.
嗨。
Hey.
大家最近都好吗？
What's up, y'all?
我是萨姆·杰ay，你的朋友。
It's your girl, Sam Jay.
我们是《人人都在看女子体育》这档新播客的主持人，由Together和iHeart女子体育联合推出。
And we're the host of everyone watches women's sports, a new podcast from Together and iHeart women's sports.
咱们得说实话。
Because let's be real.
最近女子体育精彩得让人聊都聊不完。
Women's sports is giving us way too much to talk about these days.
凯尔西·芬勒成了第一位独自划船从加州横跨到夏威夷的女性。
So Kelsey Finler, she became the first female solo rower to go from California to Hawaii.
我第一反应就是：她带了啥零食啊？
My first thought is just, like, what's up with the snacks?
到底吃的是些什么东西？
Like, what are we what are we eating?
那些高光时刻、激烈对抗、新星崛起，还有刷爆你动态的瞬间。
The highlights, the rivalries, the breakout stars, the moments that take over your entire timeline.
还有那些比赛期间开始聊、结果一整个星期都停不下来的讨论。
And the conversations that start during the game and somehow keep going all week.
每周我们都会深入剖析女子体育界最热门的新闻。
Every week, we're breaking down the biggest stories across women's sports.
娜奥米·大坂表现太棒了。
Naomi Osaka showing out.
她赢了萨巴伦卡。
She beat Sabalinka.
娜奥米，给你点赞！
Shout out to you, Naomi.
给你鼓掌，娜奥米。
You get the palm, Naomi.
给你鼓掌。
You get the palm
就为这个。
for that.
因为我们关心的不只是发生了什么。
Because we're not just interested in what happened.
我们更想知道，为什么大家都在讨论这件事。
We're interested in why everyone's talking about it.
因为真的有很多人看女子体育赛事。
Because everyone watches women's sports.
你听好了，很多人都在看女子体育赛事。
Listen to everyone watches women's sports.
在 iHeartRadio 应用里就能听。
On the iHeartRadio app.
Apple Podcasts 或者你平时听播客的平台都可以。
Apple Podcasts or wherever you get your podcast.
嘿，各位《波特兰迪亚》的粉丝们。
Hey, Portlandia fans.
我是凯莉·布朗斯坦和弗雷德·阿米森。
Carrie Brownstein and Fred Armisen here.
你们认识我们，或者说更熟悉他们——托尼和坎迪斯、妮娜和兰斯、斯派克，还有那只鸡。
You know us or rather you know them, Tony and Candice, Nina and Lance, Spike, and, yes, the chicken.
这么多年来我们演过很多经典角色，但今天，我们以真实的自己来跟你们聊聊全新的播客《Podlandia AO 重看版》。
We've played a lot of iconic characters over the years, but today, we're showing up as ourselves to tell you about Podlandia AO Rewatch, our brand new podcast.
每周我们会回看一集《波特兰迪亚》的首播剧集，拆解小品内容，挖掘我们那些标志性角色背后的故事，重访你们熟悉并喜爱的波特兰场景，也坦诚分享我们的创作过程。
Each week, we'll revisit an episode of Portlandia from the very beginning, breaking down the sketches, exploring the backstories of our most iconic characters, revisiting the Portland locations you know and love, and opening up about our creative process.
这些玩意儿到底是怎么拍出来的？
How did any of this get made?
我们当时怎么就觉得这主意不错？
Why do we think that was a good idea?
我们现在准备好聊聊这些了。
We're ready to talk about it.
我们还会邀请那些帮助把这一切变成现实的人一起参与。
And we'll also be joined by the people who helped bring it all to life.
包括客串演员、合作者和朋友们，比如导演乔纳森·克莱塞尔、市长本人凯尔·麦克劳克林、传奇音乐人艾米·曼，还有更多人。
Guest stars, collaborators, and friends, including director Jonathan Kreisel, the mayor himself, Kyle McLaughlin, legendary musician Amy Mann, and many more.
凯尔现在要
Kyle is going for
来这一手。
it here.
你当时完全即兴发挥，不只是台词，还编了一首关于
You fully improvised not of just words, but a song about
我以为你会写歌呢，我以为你们都会一起写首歌。
Well, I thought he was gonna write I thought you were all gonna write a song.
我记得你听着Podlandia时的样子。
I remember you Listen thinking to Podlandia.
哎哟。
Ayo.
在iHeartRadio应用、苹果播客，或者你常用的任何播客平台收听重播。
Rewatch on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
对。
Yeah.
这我也不太清楚。
It's I don't know.
我在想，要是大家都把平台当成服务来用，是不是就能接近真相了。
I'm try it's like if everyone did platform as a service and take it to a truth.
真是让人惊叹。
It's like, wow.
你只要写一行代码，整个系统就能完美运行起来。
You could just put a single line of code, and now your entire ecosystem is working perfectly on this.
对啊。
Yeah.
这听起来就像当初承诺过却没实现的那个愿景。
It sounds like the promise that was made with pass and then just didn't happen.
我之前接触过几个这样的客户。
I worked on a few of those clients.
简直糟透了，还得向记者解释什么是平台即服务。
Fucking terrible, explained to journalists what platform was a serviceman.
但再说一遍，不是这样的。
But, again, no.
没人突然就说大家都不再用自己的电脑了。
Everyone didn't suddenly say that no one was running their own computers anymore.
也没人再自己维护数据库了。
No one was running their own databases anymore.
这感觉完全脱离了现实，而你却深陷其中。
And it's just it feels like it's completely escaped reality, and you are very much in the trenches with this.
我还想问一下，我最喜欢的部分是‘异端者格杀勿论’。
And one thing I did wanna ask about as well was this the my favorite section of the piece, heretics will be shot.
你描述的是，背后有一种真正的宗教式狂热，一种咄咄逼人的科技传教心态：你不跟上AI，你就被排除在外。
And the way you describe it is that there is this genuine religiosity behind this, this aggressive kind of tech evangelism that says if you are not in with the AI stuff, you are out of the crowd.
是的。
Yeah.
看着这种情况，其实挺吓人的。
And it's it's actually kinda scary to watch.
一般来说，我写东西的时候，都是因为生气。
Like, normally, when I write something, I'm mostly angry about it.
但那篇文章里其实没多少愤怒，因为我更多是觉得你太离谱了。
And you also wasn't a lot of anger in that piece because I was more You were horrible.
后来我才意识到，确实如此。
Realizing yeah.
我当时就想，哎。
I was like, hey.
我们的医院里，居然没有真正靠谱的人在管事。
There are not adults running our hospitals.
你知道吗，结果发现是这样。
You know, they it turns out
哎。
that Hey.
哎。
Hey.
我觉得最糟糕的是，我意识到他们一直都很幼稚、孩子气，既不负责，也没有勇气。
I think the worst thing was realizing that they were always sort of immature and childlike and not responsible or brave.
他们一直都是这样。
Like, they've always been that way.
现在只是让他们能协同起来，一起做这种事而已。
This has just allowed them to coordinate around a thing that allows them to behave this way at the same time.
是的。
Yeah.
但凡有人不配合，天啊，我见过太多离谱的事了。
And anyone that does not get on board, like, god, I've seen some pretty wild stuff.
我们现在差不多到了这种地步：如果你跟一家大公司的CEO，尤其是市值超过十亿美元的公司高管说，你不相信人工智能是未来——我是认真的，我说的就是这几个字。
We're sort of at the point where if you tell a CEO at maybe most companies, especially big ones, right, if they're worth over $1,000,000,000, if you tell certain people, certain executives that you do not believe that AI is the future and I I mean those words.
对吧？
Right?
我不是说你要有个靠谱的商业计划，或者你相信某个具体项目能成。
I'm not talking like have a sound business plan or you believe a specific initiative will work.
如果你不肯认同‘AI正在改变一切’、‘AI将重塑这家公司’这类说法，他们就会直接把你开除。
If you will not sign off on statements like AI is changing everything, it will transform this business, they will just fire you.
我从来没见过这样的情况。
I've never seen anything like it.
以前不是也有过云计算吗？
Even you know, we had the cloud a while ago.
可人们还是老样子。
People still behave the same way.
简直有点像邪教一样。
It's like a little bit culty.
他们就说，得把所有东西都迁到云上。
They go, we gotta move everything to the cloud.
必须把所有东西都迁到云上。
We gotta move everything to the cloud.
我不记得有谁因为对云计算表示怀疑就被炒了鱿鱼。
I don't think I've seen someone get fired for expressing doubt about that.
但如果你说，我觉得AI不适合我工作，我见过好几个CEO都说过，那你就不该待在这个公司。
But if you say something like, I don't think AI is the way for me to do my job, CEOs I've seen multiple CEOs say things like, well, then you don't belong at this company.
收拾东西走人吧。
You know, pack your bags.
赶紧滚吧。
Get out of here.
这叫什么事儿？
What the fuck?
这有什么道理？
What's the justification?
因为你反对AI，那就说明你工作能力不行，因为以后公司根本不需要不用AI的人。
Well, you're against AI, which must mean you're bad at your job because we won't need people who don't use AI anymore.
说实话，我都不知道该怎么说了。
Know, it's just I don't know what to say, man.
因为当你问‘为什么’的时候，就暗示着……
Like, there because the question of why implies there's, like,
一个说得通的理由。
a coherent reason.
他们简直疯了。
They've gone insane.
你刚才也提到，怀疑AI还会让客户感到不安，这简直是一连串的荒唐事。
That's just because you mentioned in here as well that doubting AI also scares their customers, and that it's kind of a chain of idiocy.
你能再详细说说这个吗？
If you could talk about that a little bit.
是的。
Yeah.
有意思的是，我之前跟一位《财富》五百强企业的高管聊过。
So what was really interesting was, I was chatting with a Fortune five hundred executive.
对吧？
Right?
我还觉得特别有意思，因为有些在Hacker News上只会说些肤浅话的人就说，这不可能是真的。
And I thought it was really interesting because some of the, you know, two brain cell dudes on Hacker News were like, oh, this this can't be real.
我就说，你看看嘛。
And I'm like I mean, look.
这可是《财富》五百强的高管。
It's it's a Fortune 500 executive.
你说呢？
Like, what
你到底想怎么样？
do you what do you want
让我怎么说？
me to say?
这本来就是事实。
It's just it's the truth.
这人其实挺聪明的，但他公司说过一些特别离谱的话。
And this guy was pretty smart, but his company had said some pretty crazy things.
我就问他，咱们现在喝着咖啡，私下说句实话，你觉得呢？
So I went, well, you know, do you like, between you and me, we're sitting here having coffee.
这是事实，还是你只是这么认为？
Is it true, or do you believe that?
或者说，到底是什么让你这么说？
Or, you know, what is making you say this?
难道只是为了完成销售目标吗？
Is it just because you need to hit certain sales numbers?
然后呢，确实是这样。
And it's like, yes.
有些人太容易轻信了，真就信了这些说法。
Some people are very gullible, they buy into this stuff.
我们自己内部其实并不真信。
We don't really believe internally.
有些人是信的，但管他呢。
Some people believe it internally, but whatever.
到处都是些疯子。
It's crazy people everywhere.
当时这家公司规模很大，签了很多企业合同，有些首席执行官还和他们有特殊协议。
It was that at this company, because they're huge, they have tons of enterprise contracts, and some of those CEOs have deals with them.
如果那位CEO说，我们的生产效率提升了百倍，而我之前聊过的那个人上电视说‘百倍根本不可能’，那就等于让客户当众丢脸。
And if that CEO says something like, we have a 100 times speed up in productivity, and the original guy, the guy I was talking to, goes on the news and says, well, a 100 times is not possible, he has embarrassed his customer.
这些人特别小心眼，一不高兴就取消合同，结果第一个说话的人就被开除了，就因为冒犯了客户。
And these people can be very, very petty, so they'll cancel contract, and the first guy gets fired for insulting a customer.
我们陷入了一种奇怪的博弈困境，谁都不敢反驳其他高管，因为你根本分不清他们是真信这套，还是纯粹在演戏。
And we're in this weird game theoretic prisoner's dilemma, you know, standoff where you can't even contradict another executive because you can't tell if they're cynically playing the game.
要是你指出来，他们压根不在乎。
Because then you can say it and they don't care.
要是他们真信了，那你就是在公开场合让他们下不来台。
Or if they really believe it because then you've embarrassed them in public.
所以就连原本挺理性的人，也都选择闭嘴，不再跟他们唱反调。
So all of them have just sort of gone the eve even pretty responsible people have gone, I'm not gonna disagree with them.
我干脆就不提这事儿了，因为一说就惹祸。
I'm just not gonna talk about this because it's too dangerous to talk about.
这简直太疯狂了。
That is fucking insane.
我知道我今天得说好多遍，但这真的太疯狂了。
I realize I'm gonna say this a lot today, but it it it's fucking insane.
如果你上电视说不喜欢某样东西，就会丢掉合同，这简直是最荒唐的事之一，这也解释了那家公关公司为什么会出事。
That is one of the craziest thing that if you go on television and say you don't like something, you will lose a contract over it, which kind of explains what happened with the PR firm, I guess.
我不清楚他们原本打算怎么处理。
I mean, I don't know how they would have worked out.
我不喜欢人工智能，但也许 somewhere 早就有些迹象了。
I don't like AI, but maybe there were some clues somewhere.
这整个行业都疯了。
But it this is an industry gone mad.
这是我听过的最离谱的事，这已经不只是荒唐，简直是信息危险了。
This is this is the craziest thing I've heard in any this is like a an actual information hazard situation.
我真有点被吓到了。
I'm I am actually a little bit chilled by it.
是的。
Yeah.
我觉得这很有趣，因为一直以来，得罪客户都是不对的。
I think it's fascinating because it's always been bad to insult your customers.
没错。
Right.
显而易见。
Obviously.
但通常情况下，不会所有客户同时以这么具破坏性的方式说同一件事。
But usually, all your customers aren't saying the same thing at the same time in such a destructive way.
他们都在说同一件事，这个话题就是绝对不能碰的。
They're all saying one thing, that topic is just off limits.
因为如果你同意他们，其实你也没做成什么。
Because if you agree with them I mean, you haven't really accomplished anything.
他们根本不在乎你是不是同意。
They don't care if you agree with you.
你知道吗，比如你要是Xero这样的公司的CEO，肯定要跟一大堆大公司签大合同。
You know, they're just like you know, for for example, if you were the CEO at Xero or something, right, you're gonna have huge contracts with a bunch of big companies.
你根本没法说出自己真实的想法，怕惹恼了谁。
You can't actually say what you think on this topic because you might upset someone.
我觉得特别有意思的是，董事会施加的压力也越来越大。
And I I think what's been really interesting as well is seeing the pressure from boards.
我写完那篇文章后，好多高管都来信说，没错，董事会逼着我用人工智能。
I've had so many executives write in since I wrote that and go, yeah, my board has ordered me to use AI.
我自己并不相信这玩意儿。
I don't believe in it.
但他们说，我要是不用AI，他们就……
But they said if I don't use AI, they yep.
不行。
No.
董事会现在就是这么要求的。
Boards are saying this.
但我也有董事会成员告诉我，他们觉得如果不对所有人强制使用AI，自己就会被撤职。
But but I've also had board members tell me that they feel like if they don't tell everyone to use AI, they'll be removed from their positions.
我真的不知道该怎么跟你说才好。
It's just this I, like, I don't know what to tell you.
这太荒谬了。
It's insane.
完全说不通。
It doesn't make any sense.
这简直像一场大规模的集体幻觉，我觉得未来二十年我们都会在讨论这件事。
It's like this huge mass delusion event that I feel like we might be writing about the next twenty years.
真是让人着迷。
Like, it's just fascinating.
不过我认为，与此相反的是，这种情况也可能突然且剧烈地反转。
Well, I think what the the contra to all of this is it also feels like it could reverse very aggressively and suddenly.
因为现在很多人都只是装模作样地说‘是啊’，其实心里根本讨厌AI，或者有一群人，数量难以估量，就是纯粹在应付了事。
Because if everyone's just kinda sitting around being like, yeah, I actually fucking hate, or, like, there's a contingent of people that's impossible to measure of of people who are gonna just kinda fucking sucks.
我其实特别反感这事儿，但还是跟着做了。
I kinda fucking hate this, but I'm going along with it.
只要足够多的人开始说他们不喜欢这玩意儿，
And enough people start just saying they don't like it.
某一天，风向可能就突然变了。
It just feels like one day that the mood could shift.
我一直都在预测，迟早有一天，大家会停止用AI，或者某个他们敬重的人会说：算了，不玩了。
Because this is what I've been predicting the whole time that just at some point, everyone will stop doing AI or someone they like and respect will say, fuck this.
我不干了。
I'm not doing it anymore.
然后所有人都会跟着退出。
And then everyone will walk away.
他们只是在等一个借口。
They're just looking for permission.
没错。
Correct.
我们其实谈的是协调问题，对吧？就是一群人心里都想着要说同一件事。
We're basically talking about coordination problems, right, which is a thing where a bunch of people wanna say a certain thing.
但如果一个人先站出来，就会被围攻，所以谁都不愿意第一个开口。
If one person puts their hand up, that person will be killed, so no one wants to put their hand up first.
你需要的只是一个契机，让大家能同时说出同样的话。
What you need is just some event for everyone to say the same thing at the same time.
有意思的是，我们今天之所以在这里，是因为另一个问题被解决了——很多高管、董事会成员，还有其他人特别容易激动。
And what's interesting is we're here because a different problem was solved, which is a lot of executives and board members and people are very easily excitable.
他们一直都有种潜在的狂热情绪。
They they always had this sort of latent level of mania.
而LLM形式的AI出现，正好给了他们一个发泄口。
AI coming out in the form of LLMs solved that problem for them.
因为如果你站出来说，我要全力押注元宇宙，那简直是个笑话。
Because if you put your hand up and you said, I'm going all in on the metaverse, like, that's that's kind of a joke.
那时候这想法确实挺荒唐的。
It was silly back then.
所以我们先是给了他们一个可以一起疯狂聚焦的东西，解决了这个问题。
So we first solved the problem by giving them something to coordinate and be frantic around.
现在我们只等着第一个超大规模云服务商崩盘，来触发相反的局面。
And now we're just waiting for the first, like, hyperscaler to collapse for the reverse event.
或者可能是萨提亚·纳德拉，但不太可能是桑达尔·皮查伊，毕竟谷歌最近的财报表现太差了。
Or just Satya Nadella or probably not Sundar Peshai with the way that Google's earnings just went.
但总有人会最终忍不住说：算了，不玩了。
But someone is going to eventually just go fuck this.
我觉得可能是元宇宙，但谁说得准呢，马克·扎克伯格那个疯子下一步又会干出什么事儿。
I truly I think it could be meta, but also who knows with that wacky son of a bitch Mark Zuckerberg what he'll get up to next.
听说他正在搞一个电子宠物，真·AI驱动的那种，这人就是个怪才。
I hear he's working on a Tamagotchi, like an actual AI powered you know, he's wacky.
你真是个不按常理出牌的主儿。
You wacky son of a gun.
不过说真的，这事儿还挺酷的。
It's it it is really cool, though.
这让我想起你另一篇我特别喜欢的文章，《洗脑》和《高管今日》——但那是领英狂热最极端的版本，你做的每一件事，都成了你消费了多少领英的证明，只不过现在每个人都在当领英。
It kind of reminds me of your other piece I love, which is Brainwash and Executive Today, But the most extreme version of LinkedIn psychosis, where it's just everything you do is a is an artifact of how much LinkedIn you consume, except everyone is LinkedIn now.
每个人。
Everyone.
我能聊聊那篇《洗脑》吗？
Can I talk about that brainwash piece for a second?
当然，请说。
Yes, please, please.
请讲。
Please do.
好。
Okay.
那篇文章里，我以前还在职的时候，有个高管跑来，问公司里所有人关于一个叫Monte Carlo的产品，那玩意儿特别小众，简直是极客才懂的东西。
So the, you know, the the whole thing in the piece is one of my, executives when I when I still, like it didn't work for myself, came in and just asked everyone at the company about a product called Monte Carlo, which is, like, super niche, like, fucking nerd thing.
我们就纳闷了，你为啥在一对一谈话里跟谁都聊这个？
And we were like, why are you talking to all of us about this in our one on ones?
你一年才跟我们聊一次。
You only talk to us once a year.
这就像是我爷爷突然跑来，非得问我要不要了解量子加密。
This would be, like, my granddad coming in and going, like, tell me about, like, quantum cryptography.
这种话题根本不可能是你该关心的。
Like, there's no way that this is a top a topic you should be interested in.
对吧？
Right?
这完全超出了你的专业范围。
Like, it's just so far out of your wheelhouse.
是啊。
Yeah.
你是在哪儿听说这个的？
Where'd you hear about this?
就像我孩子突然蹦出个新词一样。
Like like, when my kid comes in and says a new word.
我当时就说，对啊。
I'm like Yes.
你从哪儿听说的？
Where'd you hear that?
没有。
No.
我确实听说过这事，是在领英上看到的。
I I did hear about this, and it was it was LinkedIn.
他们还告诉我消息来源，结果真是让人失望。
So they they did tell me where they heard about it from, which is dire.
不过，我不知道你有没有注意到这个。
But, I don't know if you saw this.
蒙特卡洛的首席执行官，不知道我之前跟你一起写过关于他们公司的文章——我倒不算批评，只是有些疑问。
The CEO of Monte Carlo, not knowing I'd written about the company with you know, I I wouldn't say I criticized them, but I had some question marks.
我就想，你们到底在向谁做广告？你们的网站到底哪儿出了问题？
And I was like, who are you advertising to, and what the hell is wrong with your website?
于是她拿了那篇文章，大力推广，说它正在被全球各地董事会的高管们传阅。
So she, took that piece and boosted it and said it's, you know, being shared with executives in boardrooms around the world.
这可能是目前董事会里流传最广的内容了。
It it might be the most circulated thing in boardrooms right now.
我没法去核实这个说法。
I have no way of measuring that.
如果你去查蒙特卡洛的官网，就会发现，一个刚看完那篇文章的人，比如我，其实挺喜欢这篇内容的。
And then if you look up the Monte Carlo website, you know, from someone who has just gone you know, I I like this piece.
我觉得它确实提出了一些不错的观点。
I think it raises some good points.
蒙特卡洛以前是不是这样我不确定，但现在他们的网址改成了 montecarlo.ai，整个宣传口径都变了。
Monte Carlo I don't know if it's always at this, but it's now montecarlo.ai, and they've completely shifted their messaging.
因为他们不再自称是数据可观测性平台了。
Because they are no longer a data observability platform.
现在他们自称是代理可观测性平台，管它到底是什么意思呢。
They're now an agent observability platform, whatever the fuck that means.
就像
Like
天啊，你那篇文章引发的反响也太棒了，我一直在关注。
Jesus fucking what's great as well, though, is the reaction to your piece, and I've been looking it up.
我偶尔也会看到有人转我的，但我表达对AI圈的人的看法更直接，所以他们不太愿意转发。
It is I I occasionally see it with mine, but I'm more direct with how I feel about AI people, so they're less willing to share it.
但我看到好多明显就是文章里讽刺的对象在转发，这简直太离谱了。
But the amount of people I've seen sharing it who are very clearly the subject of the piece is insane to
真的。
me.
大家的反应都特别有意思。
It's just like people very funny.
就是，大家都会说，对啊。
It's just like people like, yeah.
哇。
Wow.
真不敢相信有人居然毫无理由地用AI。
Can't believe people are miss people are all using AI for no reason.
当然，我跟这事儿一点关系都没有。
I, of course, have nothing to do with this.
我所有的AI方案都是完美的。
All my ace AI solutions are perfect.
我超爱AI的。
I love AI so much.
但那些用错AI的傻瓜，真是这样。
But these other imbeciles who are using it wrong, it's Yes.
真的快把我逼疯了，
It's genuinely driving me insane,
说实话，
if I'm gonna be honest.
我有个说法，我觉得挺新颖的。
I, I have a phrase for this, which I think is novel.
我可能真是第一个提出这个说法的人。
I might I might have invented this.
我叫它‘n减二傻瓜解决方案’——就是你环顾房间，觉得那个人是这里最聪明的。
It's the, n minus two idiot solution, which is if you look around a room and you go, that guy is the smartest person in here.
在一群足够荒唐或不堪的人里，大家都认这个，因为他们觉得自己是第二个。
Everyone will agree, and in a room full of sufficiently, you know, ridiculous or or compromised people, they're all okay with it because they assume they're number two.
对。
Right.
你只要给他们留点余地就行。
So you just need to leave them enough wiggle room.
你绝对不能说，‘这屋里只有一个人聪明’。
What what you can't do is go, only one person in this room is smart.
你可以说，‘你们俩是聪明的’，这样大家都能接受。
You can go, two of you are smart, and everyone's okay with that.
他们都会同意第一个那个人。
They'll all agree on the first person.
他们都觉得自己是第二聪明的那个人。
They all assume they're the second person.
而且，确实，很多报道都在谈这个荒唐的现象。
And, yeah, a a lot of the coverage has been, you know, this this insane phenomenon.
也许该给蒙特卡洛的首席执行官一些认可。
And and, you know, credit to possibly the Monte Carlo CEO.
她可能没这么直白地说过，但她的本意或许是：我被这个局面困住了。
Maybe she didn't say this in so many words, but maybe what she was actually saying was, I feel trapped by the situation.
我的整个公司都转向了AI代理，尽管这根本说不通。
My whole company has pivoted to AI agents even though it doesn't make sense.
也许这正是她以一种委婉的方式向虚空呐喊。
Maybe that was her gentle way of screaming into the void.
我其实也不确定。
I don't actually know.
但这真的挺搞笑的。
But it is very funny.
我打算哪天写篇博客，就说：其实你根本不用非得用AI。
I kinda wanna do a blog at some point that's just you can stop using AI.
真的，随时都可以停下来，完全不用继续。
It's just that literally you can stop this at any time.
你根本没必要这么做。
You don't have to do this.
如果你觉得被困住了，直接走人就行。
If you feel trapped, you can just leave.
你最后那部分其实就做了类似的事。
You kind of did that at the end of the piece.
你结尾写得特别有同理心，给出了应对公司里AI乱象的建议。
You did a very empathetic thing at the end saying, here's how to deal with the organizational AI stuff.
要是有人跟你说必须用AI，那你就干脆辞职。
If you're told this, well, just quit.
在别的场合，你也可以选择别的做法。
And in other situations, you could also do something else.
这真的让人作呕。
It's just it's genuinely sickening.
这种事简直让人气得不行。
It is actually sickening in a way that it pisses me off.
我以前也说过，虽然没这么直白，但这个行业本该是凭本事说话的。
And I've said this before in not so many words, but this is meant to be an industry of meritocracy.
本该是
This is meant to be
一个真正靠能力说话的行业。
the industry where people are.
我们做事情是为了提高效率，也是因为这些做法能带来好的商业成果。
We do things for efficiency reasons, and we do things because we do things because they they get good business results.
现在我才明白，你以前其实也暗示过我，大家其实都在骗人。
And now I'm realizing, you've kinda told me this in the past as well, everyone was lying about that.
他们只是在做让自己感觉良好的事。
They were just doing stuff that made them feel good.
是的。
Yeah.
我倒希望我们真能做到那样。
I I kinda wish we had that.
你偶尔会听到有人说，资本主义的问题就在于它把一切都优化成为了钱。
You know, every so often you'll hear someone say something like, you know, the the issue with capitalism is it optimizes everything for money.
对吧？
Right?
我们可以为了钱、效率和生产力，把其他所有价值都扔掉。
We can throw away every other value for money and efficiency and productivity.
天啊，我真希望这是真的，因为我们现在反而陷入了最糟糕的境地——既抛弃了所有其他价值，却也没能达到最高的效率。
And, god, I wish that was true because we somehow have, like, the the worst of both worlds where we're throwing away other every other value and somehow still not achieving peak efficiency.
对吧？
Right?
我们干了这么多可怕、恶劣的事，可你回头一问：我们到底赚了多少钱？
Like, we're we're doing all these horrible, hideous things, and then you're like, well, how much money are we making?
根本没赚多少钱。
It's like not even that much money.
我们浪费了这么多数据中心。
We're wasting all the data centers.
光是这些数据中心，就简直让人难以置信。
I I mean, the data centers alone are just fucking insane to me.
说到精英制度和数字方面，我最近一直在试着搞清楚一些最基本的问题，比如：这些东西到底赚钱吗？
I've been what's really good as well is on the subject of meritocracy in numbers, trying to just pull together very basic things like, is any of this profitable?
问了之后，人家就会对你发火。
And asking people and them getting angry at you.
特别生气，特别烦躁。
Just very angry, just agitated.
你干嘛老问这种问题？
Like, why'd you keep asking these questions?
你为什么总这么较真？
Why do you keep doing this?
我从小在个挺乱的街区长大。
And I grew up in a rough neighborhood.
我见过不少骗子。
I've met my fair share of scam artists.
只要有人因为你问这种问题就生气，那百分之百是心虚。
And 100% of the time, if someone gets angry at you for asking for things like that, that's because they are full of shit.
对。
Yeah.
没错。
Yeah.
我以前学心理学，后来才转行做IT，这种事在多层次营销骗局里太常见了。
You know, I was in psychology before I was in IT, and you see that a lot in things like multi level marketing scams and stuff.
你要是说，这不就是金字塔骗局吗？
You know, like, they're if you go, hey, isn't this a pyramid scheme?
他们早有标准话术，反问你：你什么意思，这叫金字塔骗局？
They've got the whole script, which is like, what do you mean it's a pyramid scheme?
你对金字塔骗局是怎么理解的？
What's what's your understanding of a pyramid scheme?
回答我，你这个混蛋。
Answer me, you fucker.
就是这么回事。
And it's Yeah.
这里的情况也差不多。
It's sort of the same thing here.
我们之前是不是聊过这个？你有没有看过那个对萨姆·奥特曼的非常温和的采访片段？
Didn't we maybe we talked about this previously, but have you seen, like, that clip of of, you know, a very, very gentle interview with Sam Altman?
最后，那个明显很欣赏他的采访者，语气特别轻柔地问：‘那你的盈利计划是什么？’
And at the end, this interviewer who clearly loves him very gently goes like, so what's the plan for profitability?
他就只是说：‘哦，对。’
And he's just like Oh, yeah.
他早就听腻了这种问题。
He had enough of these questions.
没有。
No.
他就说，我们今年赚的钱远超一千三百亿美元。
He's just like, we're we're we're making way more than $13,000,000,000 this year.
所以你是知道的吧？
So you have you know?
而且如果你需要人买你的股票，有趣的是，萨钦·阿代拉也在那通电话里。
And if you need someone to buy your stock, and and what's funny is Sachin Adela was on that call.
就像你的机器人，没错。
Like, just your bot your yeah.
萨钦·阿代拉就坐在那儿，一副神游天外的样子。
Just Sachin Adela just sitting there with, like, I don't know, the magic magic roundabout thing.
脑子里空空如也，等着下一条指令。
Just nothing going on up there waiting for the next input.
他脑子里就像那只转圈的小猴子，只会做后空翻。
The world's Just like most the little monkey with the symbol that does back flips just sitting in Sachin Adela's head.
他就在那儿等着，等我
He's going, waiting for I
我正想说，我脑子里浮现的是那个小小的 macOS 语音转圈动画。
was gonna say, I'm imagining, like, you know, the little Mac OS speech ball.
就只是头顶上一个小小的转圈小图标。
It's just a little spinning widget up there.
问题是，萨姆·阿尔特曼根本不会理我。
The thing is Sam Altman will never talk to me.
他们已经告诉我了，他不会见我。
They've already told me he won't.
但那是因为我知道，五分钟内他就会对我发火。
But that's because I know within five minutes, he will get pissed off at me.
因为只要我一开口，就很容易直接问：
Because the the it would just be really easy to just be like, hey.
你为啥赚不到钱？
Why don't you make any money?
萨姆，你怎么亏了这么多钱？
Why'd you lose so much money, Sam?
他可能会说，我们其实赚钱的啊。
It'd be like, well, we make money.
那你为什么就不行呢？
Why why don't you?
然后你就会越来越不讲理，最后他简直像只狗对着镜子里的自己狂吠一样，隔着屏幕冲你发火。
And then just you'd get more and more immature until he eventually tries to get at you through the screen like a dog that sees his reflection.
简单得很。
Easy peasy.
他一天之内就能想方设法弄死我。
He'd try and kill me within a day.
其实根本没必要这样。
I mean, there's no reason to.
对吧？
Right?
不可能。
Like No.
你知道的，那会是
It you know, it'd be
作为公关人员，我宁愿自杀，也不愿让萨姆·阿尔特曼和我通电话。
I as a PR person, I would I would rather kill myself than put Sam Altman on a call with me.
我的意思是，那才是真正该做的选择。
Like, I've just that that is the that would actually be the appropriate choice.
不管怎样，千万别让埃德·齐特隆和萨姆·阿尔特曼通电话。
Like, whatever you do, do not put Ed Zitron on a call with Sam Altman.
没错。
Yeah.
这对他来说纯粹是赔本买卖。
It's also it's like only a losing proposition for him.
没错。
Yeah.
假设他上场了，发挥出了有史以来最好的状态。
Because let let's say he comes on and he has the greatest showing of his whole life.
你知道的，他完美回应了每一个问题。
You know, like, he just he feels every question perfectly.
但结果和让一个记者夸他一小时没什么两样。
That is the same outcome as just talking to another journalist that's gonna jerk him off for an hour.
对吧？
Right?
既然有那么多根本不问他问题的人愿意给他正面报道，他干嘛要冒这个险呢？
There's no reason for him to take that risk when he can get so much positive coverage from people that aren't gonna ask him any questions.
我只会问他最烦人的问题——你开心吗？
And I will only ask him the most annoying like, I'll just ask him, are you happy?
他一听，脑袋估计都要冒烟了。
And he'll his head will start, like, erupting in smoke.
不过说来想去，我们好像跑题了。
Anyway anyway, so we've kinda got off topic, I realized.
我其实想跟你聊聊一件事，就是我听好几个人都提到过，关于AI做演示的问题。
I wanted to actually talk to you and break down something that I've heard with various people, which is this whole issue of demos with AI.
AI现在特别擅长做出一个看起来还不错的演示来打动客户。
The AI is really good at making a good enough demo to sell to someone.
在AI还没这么发达的时候，你就已经讲过类似的事，比如忽悠高管。
And you kind of covered this before AI was as far down the hole as possible with Brainwash an executive.
但我感觉，现在的AI就是特别会装样子，骗那些不懂行的人相信它。
But I get the sense that it's just very good at doing enough to sell an imbecile on something.
是的。
Yeah.
不过我愿意说得宽容一点，因为不一定非得是傻子才会上当。
And, you know, I I'm I'm willing to be a little bit more charitable here because it doesn't have to be an imbecile necessarily.
只要是对技术不太懂的人，就很容易被蒙住。
It just has to be someone who is not proficient with tech.
因为你一旦接触过足够复杂的东西，就会明白，部署软件、设计系统、实现商业价值，这些细节里头有多少麻烦。
Because when you work with something sufficiently complicated, you understand there's a a lot of difficulty around the fine details in deploying a piece of software and the design and achieving business value, whatever you wanna do.
但作为普通用户，我其实也不知道烤箱是怎么工作的。
But, you know, as a end user, like, I don't really know how my oven works.
我就是转几个旋钮，它就能正常运行了。
I just have some knobs that I twiddle and then the oven works.
对吧？
Right?
所以如果你跑来跟我说，我有纳米机器人能完美组装烤箱，然后它真能把派烤出来，
So if you turned up and you said, you know, I've got nanobots that assemble ovens perfectly, and then it does it, and it bakes a pie.
只要我不吃那个派，或者不管怎样，我可能会说，哇，
If I don't eat the pie or, you know, whatever, I might go, hey.
这真是太厉害了。
That was really, really cool.
而且这确实足够让人印象深刻，至少一开始没人会怪我投资了它。
And and it is genuinely impressive enough that possibly no one would blame me for investing in it initially.
是的。
Yeah.
因为你看到的东西看着挺靠谱的。
Because you got shown something that looked plausible.
对。
Right.
不只是靠谱，简直是2020年之前根本不可能做到的事。
And not just plausible, like, literally impossible prior to 2020.
以前你根本玩不了这种把戏。
Like, you you couldn't do this sort of trick.
可第二天你就开始发现，嘿。
And then the next day, you start to find out, like, hey.
你知道吗？
Well, you know what?
这饼其实难吃得要命。
Like, this this pie is really bad.
结果发现这烤箱只有一个档位，还时不时着火。
And it turns out the oven only has one setting, and sometimes it catches fire.
而且说实话，光看我一开始见到的那些东西，我还不至于是个傻子。
And, you know, I still might not be an imbecile because what I saw up front was so impressive.
我倒是觉得
I would
要是四年之后还继续买，那才真是傻了。
be an imbecile if I kept buying it four years later.
对吧？
Right?
确实有一段时间，过度投资还能说得通。
Like, there was a a period where the, overinvestment was understandable.
这其实又回到了洗脑高管的问题上——买软件的人，往往不是用软件的人。
Well, this actually kinda get gets back to Brainwash an executive because there's the point in there that the people that buy software are often the not the people that use it.
所以你只要让CIO、CTO或者CEO觉得厉害就够了。
So you just have to impress the CIO or the CTO or the CEO enough.
而这个人通常缺乏实际操作经验来做判断，但偏偏又是他做决定。
And that person does not have the practical on the ground experience to make that decision, but usually does.
是的。
Yes.
这正是导致这类骗局，以及大量企业级劣质软件得以存在的根本原因。
And that is basically what makes not just this con, but, like, a lot of enterprise bad software possible.
你就跟他们说，嘿。
You just go, hey.
你们的客服团队该换新平台了。
You know, your your customer support monkeys, they need a new platform.
要不要给他们用Zendesk？
Do you wanna put them on Zendesk?
然后就有人把Zendesk买下来了。
And then someone buys Zendesk.
作为CEO，你根本不会去登录Zendesk。
And as the CEO, like, you're never gonna log in to Zendesk.
你本该去用的。
You you should.
其实我觉得，我见过的大多数高管并没有那么忙。
I actually think, you know, most executives I meet aren't that busy.
他们确实挺忙的。
They're they're pretty busy.
是的。
Yeah.
他们根本没什么正事干。
They got nothing on.
我是认真的。
I'm serious.
对。
Yeah.
但他们完全可以下去，陪某个人坐上一个小时，看看他们平时都在做什么。
Like but they they absolutely could just go down and sit next to someone for, like, an hour and check out what their work looks like.
这种做法对公司的长期战略带来的帮助，可能远超过他们去参加那些大型会议。
That might do a lot more for sort of a long term company strategy than whatever they're doing, like going to big meetings.
我去看过了。
I look.
当然，如果你个人身家达到几十亿，那情况可能不一样。
There's probably some level of wealth where that's not true if you're worth personally worth enough billions.
但很多身家几亿的人，依然会花时间做这种事。
But, yeah, tons of people worth hundreds of millions of dollars still take the time to do that.
而且说实话，他们经营的公司利润很高，也不会在这些事情上浪费时间，除非非做不可。
And, again, they they run very profitable companies, and they don't waste their time on this stuff beyond what's necessary.
经营生意，却不考虑做播客？
Run a business and not thinking about podcasting?
你再想想。
Think again.
听播客的美国人，比听Spotify和Pandora广告版音乐的人还多。
More Americans listen to podcasts than ad supported streaming music from Spotify and Pandora.
作为最大的播客平台，iHeart的规模是第二名和第三名加起来的两倍。
And as the number one podcaster, iHeart's twice as large as the next two combined.
所以，无论你的客户听什么，都能听到你的信息。
So whatever your customers listen to, they'll hear your message.
而且，只有iHeart能把你的信息扩展到广播电台的听众群体。
Plus, only iHeart can extend your message to audiences across broadcast radio.
你觉得播客能帮到你的生意吗？
Think podcasting can help your business?
那就选iHeart。
Think iHeart.
流媒体、广播和播客，全都有。
Streaming, radio, and podcasting.
欢迎访问iheartadvertising.com，我们为你演示。
Let us show you at iheartadvertising.com.
网址是iheartadvertising.com。
That's iheartadvertising.com.
我是南希·格拉斯，《愧疚的负担》第三季播客的主持人。
I'm Nancy Glass, host of the burden of guilt season three podcast.
这是一个关于可怕境遇的故事，它毁掉了两个家庭。
This is a story about a horrendous situation that destroyed two families.
事情要从七年前说起，那时辛迪和肯尼刚订婚，都刚满四十岁。
It started more than seven years ago when Cindy and Kenny were two 40 who had just gotten engaged.
几天后，他们一头撞进了一场从未预料到的风暴。
Days later, they walked into a scorched earth war they never saw coming.
转过身，把双手放到背后。
Turn around and place your hands behind your back.
警察闯进我家，直接把我抓进了监狱。
The cops came into my house and threw me into jail.
让辛迪没想到的是，她的未婚夫肯尼被指控虐待她的一个孩子。
And to Cindy's surprise, her fiance Kenny was charged with abusing one of her children.
我从未怀疑过肯尼，我始终觉得这个指控有问题。
I never once doubted Kenny and I always doubted the accusation.
在《罪恶的负担》第三季中，你会听到一个关于人为了达到目的能走多远的故事。
On season three of burden of guilt, you'll hear a story about the lengths someone will go to in order to get what they want.
这个案子绝对是其中最令人发指的之一。
This case has got to be one of the most despicable.
在 iHeartRadio 应用、Apple 播客，或者你常用的任何播客平台收听《罪责之重》第三季。
Listen to season three of burden of guilt on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
最近怎么样？
What's up, fam?
我是体育记者阿里·钱伯斯。
I'm sports journalist Ari Chambers.
嗨。
Hey.
大家最近都好吗？
What's up, y'all?
我是萨姆·杰。
It's your girl, Sam Jay.
我们是新播客《人人都看女子体育》的主持人，这档节目由 Together 和 iHeart 女子体育联合推出。
And we're the host of everyone watches women's sports, a new podcast from Together and iHeart women's sports.
咱们得说实话。
Because let's be real.
最近女子体育赛事实在有太多值得聊的了。
Women's sports is giving us way too much to talk about these days.
凯尔西·芬勒成了首位独自划船从加州横跨到夏威夷的女性。
So Kelsey Finler, she became the first female solo rower to go from California to Hawaii.
我第一反应就是：她带了啥零食？
My first thought is just, what's up with the snacks?
到底吃的是些什么东西？
Like, what are we what are we eating?
精彩瞬间、激烈对抗、新星崛起，还有那些刷爆你动态的高光时刻。
The highlights, the rivalries, the breakout stars, the moments that take over your entire timeline.
还有那些比赛开始后就聊个不停、一聊就是一整周的讨论，我们每周都会深入剖析女子体育界最重要的新闻。
And the conversations that start during the game and somehow keep going all week, every week, we're breaking down the biggest stories across women's sports.
大坂直美表现亮眼。
Naomi Osaka showing out.
她赢了萨巴伦卡。
She beat Sabalinka.
给娜奥米点个赞。
Shout out to you, Naomi.
这掌声属于你，娜奥米。
You get the palm, Naomi.
这掌声属于你。
You get the palm
就因为这个。
for that.
因为我们关心的不只是发生了什么。
Because we're not just interested in what happened.
我们更想知道，为什么大家都在讨论这件事。
We're interested in why everyone's talking about it.
因为人人都在看女子体育。
Because everyone watches women's sports.
听听看，大家都关注女子体育。
Listen to everyone watches women's sports.
在 iHeartRadio 应用里就能听到。
On the iHeartRadio app.
Apple 播客或者你平时听播客的平台都行。
Apple Podcasts or wherever you get your podcast.
嗨，《波特兰迪亚》的粉丝们。
Hey, Portlandia fans.
我是凯莉·布朗斯坦和弗雷德·阿米森。
Carrie Brownstein and Fred Armisen here.
你们认识我们，或者说，你们认识我们演的角色。
You know us, or rather, you know them.
托尼和坎迪斯，妮娜和兰斯，斯派克，还有那只鸡。
Tony and Candice, Nina and Lance, Spike, and, yes, the chicken.
这些年我们演过不少经典角色，但今天，我们以真身登场，来跟你们聊聊我们的新播客《波德兰迪亚重播》。
We've played a lot of iconic characters over the years, but today, we're showing up as ourselves to tell you about Podlandia AO Rewatch, our brand new podcast.
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。