《The Wirecutter》节目：与凯文·罗斯学习如何明智地使用人工智能

好吧，凯西，我们到了。

Well, Casey, here we are.

现在是假日季节，我们为听众准备了一期特别的周二节目。

It's the holiday season, and we've got a special Tuesday episode for our listeners.

确实如此。

We do.

而且，凯文，这期真的特别。

And, boy, is this a special one, Kevin.

从这个意义上说，我其实并没有真正参与其中。

And in this sense, I'm not really in it.

是的。

Yes.

他们称这是有史以来最棒的硬分叉节目。

The best hard fork episode ever, they're calling it.

不。

No.

我最近参加了另一档《纽约时报》的播客。

I recently went on another New York Times podcast.

这是来自Wirecutter团队的《Wirecutter》节目，我们深入讨论了该在家中摆放哪些香薰蜡烛。

This is the wire cutter show for the fine folks over at wire cutter, and we had a wide ranging conversation about which scented candles you should put in your house.

不。

No.

当然，话题是人工智能。

It was about AI, obviously.

他们邀请我来谈谈如何让AI工具为我和其他人服务，我们认为硬分叉的听众可能会对这个话题感兴趣。

They asked me to come on to talk about how to make AI tools work for me and for others, and we thought it might be something that hard fork listeners might like to hear.

是的。

Yeah.

实际上，最近Wirecutter把这期节目称为‘给不喜欢我声音的人的最佳选择’。

And actually, recently, the wire cutter named this episode the best option for people who don't like the sound of my voice.

所以这对你来说挺有趣的。

So something fun for you.

就在我们今天为您带来这些内容的同时，凯文，我们本周还为您准备了更多内容，因为周五我们将回归，我将参与我们一年一度的Hard Fork新年科技决心特辑。

And even as we bring you that today, Kevin, we have even more planned for you this week because on Friday, we will be back, and I will be participating in our annual hard fork New Year tech resolutions show.

我们还将回答你们众多听众的问题。

And we're also gonna be answering a bunch of your questions, listeners.

敬请关注。

So stay tuned for that.

非常感谢Hard Fork的听众和观众们在2025年与我们同行。

Thank you so much, Hard Forked listeners and viewers, for being with us in 2025.

能够将这档节目带给你们，我感到非常愉快，希望你们度过一个愉快的假期，并拥有一个快乐的新年。

It has been a real joy to bring this show to you, and we hope you are having a wonderful holiday break and that you have a happy New Year.

我们也希望你们能拥有一块老朗标志。

And we hope you have an old Lang sign as well.

是的。

We do.

我们希望你们能拥有一块老朗标志。

We hope you have an old Lang sign.

如果你的Lang标志看起来太旧了，就换一个新的。

And then if your Lang sign is looking too old, get a new one.

Wirecutter有推荐最好的Lang标志吗？

Does Wirecutter have any recommendations for the best Lang sign?

我们给《纽约时报》的员工提前展示了跨平台对战功能，以下是他们的反馈。

We gave Times employees a preview of cross play from New York Times games, and here's what they had to say.

我终于可以和其他人一起玩了。

I can finally play with other people.

你可以和你认识的朋友一起玩，也可以被匹配到世界上的其他人。

Play with friends that you already know, or you can just be matched with someone else in the world.

我有一个J，值10分，我可以把它放在双字母分上。

I have a j for 10 points, and I can put that on a double letter.

所以J-A-M，一共是24分。

So j a m, that's 24 points.

我要把Fax改成Faxes，拿30分。

I am going to take Fax and make it Faxes as for 30 points.

我猜tango不是一个词。

I'm guessing tango is not a word.

我们来看看。

Let's see.

tango是一个词。

Tango is a word.

哎呀。

Uh-oh.

我不知道tango是什么意思，所以我按了一下这个词，结果定义弹出来了。

I don't know what tango means, so I'm gonna press down on the word and oh, definition popped up.

作为英语作为第二语言的学习者，我喜欢学习新单词。

As in English as a second language speaker, I like to learn new words.

我很有竞争力。

I'm pretty competitive.

击败朋友和同事很有趣。

It's fun to beat friends and coworkers.

Crossplay，纽约时报游戏推出的首款双人文字游戏。

Crossplay, the first two player word game from New York Times Games.

今天免费下载吧。

Download it for free today.

我一直在和我的人力资源业务伙伴玩，这真是个不错的破冰方式，尤其当你需要找人力资源部门帮忙的时候。

I've been playing my HR business partner, which is a fun little way to, like, break the ice if there's something you need from HR.

我在购买任何东西之前都会依赖Wirecutter，但我迫切希望有人能告诉我，哪些语言模型适合做什么。

I rely on Wirecutter before I buy anything, and I am desperate for someone to tell me which language models are good for which things.

否则，我在这些产品上花的钱太多了，还花了大量时间去研究哪个适合做什么。

Because otherwise, I am spending so much money on these things, and I'm spending all this time trying to figure out what is good for what.

我真希望你们这些世界上最擅长测试和判断产品用途的专家，能替我完成这项工作。

And I would just love it if you all, are the experts in the world at testing things and figuring out what they're good for, would do that work for me.

我是克里斯汀·齐尔克洛塞特。

I'm Christine Zierclossette.

我是罗斯·盖兰，欢迎收听Wirecutter节目。

I'm Rosie Guerin, and you're listening to the wire cutter show.

罗西。

Rosie.

嗨。

Hi.

你好。

Hello.

你好。

Hello.

我对我们今天的节目感到兴奋。

I'm excited about our episode today.

我将和凯文·罗斯一起，他是《纽约时报》的科技专栏作家。

I am gonna be with Kevin Roose, New York Times tech columnist.

他撰写科技的多个方面，经常谈论人工智能，无论是报纸上还是他的纽约时报播客《硬核》中，他与卡西·牛顿共同主持。

He writes about many facets of tech and talks a lot about AI, both in the newspaper and on his New York Times podcast, Hard Fork, which he cohosts with platformers, Kasey Newton.

凯文太棒了。

Kevin is so great.

这个播客太棒了。

The podcast is so great.

你做这个我真的非常非常兴奋。

I'm really, really excited you're doing this.

我也是。

I am too.

所以今天凯文和我要聊一些我们这个节目之前没涉及过的话题——人工智能，当然，现在人人都在谈人工智能。

So Kevin and I are gonna talk today about something that we haven't covered on this show before, AI, which, of course, everyone is talking about AI.

但我们之所以在Wirecutter节目里聊这个，是因为现在消费产品和人工智能有了交集。

But the reason we're gonna talk about it on the Wirecutter show is because there is this intersection of consumer products and AI at this point.

没错。

Absolutely.

尤其是这些大型语言模型，也就是像ChatGPT和Claude这样的聊天机器人，他对这些都非常了解，而且现在很多人都在使用它们。

And especially with these large language models, LLMs, which are the chatbots like ChatGPT and Claude, he knows a lot about all of these, and a lot of people are using them at this point.

所以我们今天要深入探讨所有这些内容。

And so we're gonna really dive deep into all of that.

在这期节目中，我们谈到了很多关于AI聊天机器人的内容。

And I have to say in this episode, we talk a lot about AI chatbots.

因此，我们需要声明一下：我们为《纽约时报》公司工作，而该公司正在起诉几家公司的版权侵权行为，包括OpenAI、微软和Perplexity。

So we do need to make the disclosure that we work for the New York Times company, which is suing several companies, OpenAI, Microsoft, and Perplexity over alleged copyright violations.

没错。

That's right.

话虽如此，我期待听到你和凯文讨论AI如何改变人们的购物方式，它如何融入硬件产品，以及 hopefully 能听到凯文一些关于如何聪明且有策略地使用这些大语言模型的实用建议。

And with that, I'm looking forward to hearing you and Kevin discuss AI's changing the way people shop, how it's being integrated in hardware products, and hopefully hearing some of Kevin's best tips on actually using these LLMs smartly and strategically.

我对AI与Wirecutter之间的交集非常感兴趣。

I'm I'm I'm really interested in this intersection between AI and Wirecutter.

广告后，凯文·罗斯。

After the break, Kevin Roose.

你好。

Hi.

我叫达娜。

My name is Dana.

我是《纽约时报》的订阅用户，但我的丈夫不是。

I am a subscriber to The New York Times, but my husband isn't.

如果能和他分享一道食谱、一篇文章，或者在拼字游戏或连线游戏中和他比拼，那就太好了。

And it would be really nice to be able to share a recipe or an article or compete with him in Wordle or Connections.

谢谢。

Thank you.

达娜，我们听到你的话了。

Dana, we heard you.

推出《纽约时报》家庭订阅服务。

Introducing the New York Times Family Subscription.

一个订阅，可为生活中最多四位家人提供独立登录账号。

One subscription up to four separate logins for anyone in your life.

了解更多，请访问 nytimes.com/family。

Find out more at nytimes.com/family.

欢迎回来。

Welcome back.

今天我非常高兴能邀请到凯文·罗斯做客节目。

I am really excited to have Kevin Roos on the show today.

凯文是《纽约时报》的技术专栏作家。

Kevin is a technology columnist for The New York Times.

他还是《纽约时报》科技播客《Hard Fork》的联合主持人，与Platformer的凯西·牛顿共同主持。

He is also the cohost of The New York Times tech podcast, Hard Fork, which he cohosts with Casey Newton of Platformer.

如果你对人工智能的动态感兴趣，这档播客绝对值得一听，而如今谁又不对这个感兴趣呢？

And this is a great podcast if you are curious about what's happening with AI, and who isn't curious about that at this point.

欢迎来到节目，凯文。

So welcome to the show, Kevin.

很高兴你能来。

It's great to have you here.

非常感谢你们邀请我。

Thanks so much for having me.

我经常听《Hard Fork》。

I listen to Hard Fork quite a bit.

我非常喜欢这个播客，几年前ChatGPT刚出现时，你和凯西就是我了解AI动态的主要渠道。

I love the podcast, and I feel like in the early days of ChatGBT a couple years ago, you and Casey were how I was staying up on what was happening with AI.

我必须说，我当时一直对Wirecutter的老板们说：你们有没有听过最新的Hard Fork播客？

And I will tell you that I kept saying to my bosses at Wirecutter, you guys, have you listened to the newest Hardfork?

我们对AI到底在做些什么？

And what are we doing about AI?

我们完蛋了。

We are doomed.

购物会变成什么样？

What's gonna happen with shopping?

那时候他们就说：你简直像那个喊‘狼来了’的小孩。

And at that point, they were like, you're like Chicken Little.

你到底怎么了？

What's wrong with you?

所以我觉得在过去几年里，你教会了我很多。

So I feel like you have taught me a lot over the last few years.

嗯，我很高兴。

Well, I'm glad.

我们最初并不打算做一个专注于人工智能的播客。

We didn't intend it to be an AI focused podcast.

当我们刚开始时，实际上以为这会是一个关于加密货币的播客。

When we started it, we actually thought it was gonna be a crypto related podcast.

所以我们取了这个名字叫Hardfork，这是一个比较冷门的加密货币编程术语。

And that's why we picked the name Hardfork, which is sort of an obscure crypto programming term.

是的。

Yeah.

但你知道，事情总会变化。

But, you know, things change.

突然间，我们就进入了ChatGPT的世界，每周都讨论人工智能。

And all of a sudden, we find ourselves in the chat GPT world talking about AI every week.

而且这真的非常有帮助。

Well, and it's very, very helpful.

但我很好奇。

I am curious though.

你真的持非常明确的立场，并且一直在报道人工智能。

You you really have taken a hard line, and you're you're reporting on AI all the time.

你一直在使用它。

You're using it all the time.

人工智能在日常生活中对你产生了怎样的影响？

How has AI really impacted your life on an everyday level?

你是每天都用，还是只在个人生活中偶尔使用？

Are you using it all the time, or is it something that you're only occasionally using in your personal life?

对你来说，主要的使用场景有哪些？

What are the touch points for you?

我一直在使用这些东西。

I use this stuff constantly.

正如他们所说，我已经深陷人工智能了。

I am AI pilled, as they say.

我订阅的AI产品比流媒体电视服务还多。

I pay for more subscription AI products than streaming TV services.

而且我订阅了大量流媒体电视服务。

And I pay for a lot of streaming TV services.

我每天可能要使用这些东西几十次。

I use this stuff probably, you know, dozens of times a day.

那你用它来做什么呢？

And what are you using it for?

给我讲讲你的一天吧。

Walk me through a day.

你在个人层面和职业层面上都是怎么使用它的？

How are you interacting on a personal level, and I would imagine a professional level too?

是的。

Yeah.

我的意思是，能说的太多了。

I mean, there's so many things.

我甚至很难把它们全部列出来，但这里有几个例子。

It's it's hard for me to even list them all, but here are a few.

所以我早上醒来后，会收到AI生成的我的邮件收件箱摘要。

So I wake up in the morning, and I get my AI generated summary of my email inbox.

我使用一个叫Quora的程序，并把它连接到我仅个人的邮箱。

I use this program called Quora, and I have hooked it up to only my personal email.

我不会在工作邮箱上做任何类似的操作，因为工作邮箱更敏感。

I'm not doing anything like this on my work email, which is more sensitive.

但在我的个人邮箱中，我连接了这个工具，让它综合并总结我所有的邮件，并为它认为重要的内容自动生成回复草稿。

But on my personal email, I have hooked this thing up so that it synthesizes and summarizes all of my emails and prepopulates drafts to respond to anything it deems important.

然后我还会持续用它来处理家里的各种事情。

Then I'm also constantly using it for things around the house.

你可以查看我最近的查询记录，其中大概60%都是类似‘怎么修理这个空气炸锅？’这样的问题。

You could go back through my recent queries, and probably 60% of them would be some version of, like, how do I fix this air fryer?

或者‘怎么给孩子的自行车安装辅助轮？’

Or how do I install these training wheels on my kid's bike?

我花园里这种奇怪的植物是什么？

What is this weird plant in my garden?

类似这样的事情。

That kind of thing.

我还在工作中使用它。

And I'm using it for work.

我用它不是为了写我的专栏或播客，而是用于研究。

I use it for not for writing my columns or my podcast, but for research.

我目前正在写一本书，因此我不断使用人工智能帮我查找资料，协助我整合各种原始资料。

I am writing a book right now, and so I am constantly using AI to look things up for me, to help me sort of piece together various primary sources.

我们可以更详细地讨论这些内容，但这就是我今天使用人工智能的大致方式。

We could talk about any of that in more detail, but that's sort of how I'm using AI just today, for example.

是的。

Yeah.

我对这个非常好奇。

I am super curious about that.

我们稍后会稍微谈一谈这个。

And we will talk a little bit about that in a bit.

你的书，明确一下，是关于人工智能的。

And your book, just to be clear, is about AI.

对吧？

Right?

我的意思是，你的书是不是在朝着终极奇点推进？

I mean, it's about like isn't it kind of like you're writing towards the singularity at the at the end?

是的。

Yeah.

这本书本质上是关于通往通用人工智能（AGI）的竞赛，而通用人工智能就是所有这些公司——OpenAI、谷歌和Anthropic——目前都在努力实现的目标，即一种能够完成人类大脑所能做任何事情的人类水平AI系统。

The book is is essentially the the story of the race to AGI, which is artificial general intelligence, which is what all of these companies, OpenAI and Google and Anthropic, are now sort of pushing toward, this vision of a a sort of human level AI system that can do anything the human brain can.

当你观察普通大众时，你觉得人们现在主要用人工智能做什么？

When you look at the general population, what do you think people are using AI for right now?

你从听众那里听到了哪些说法？

What are you hearing from your listeners?

你从外界的人那里听到些什么呢？

What are you hearing from people out in the world?

我觉得这涉及非常广泛的用途。

So I think it's a really wide mix.

我听到有人用AI工具做各种事情，从医学研究和开发新药，到给孩子们讲故事。

I hear from people who are using AI tools for everything from medical research and developing new drugs to doing, you know, story time with their kids.

它的应用范围真的很广。

It really runs the gamut.

今年，OpenAI的经济研究团队做了一项非常有趣的研究，他们分析了大约150万次与ChatGPT的对话，了解人们是如何使用它的。

There was a really interesting study that OpenAI's economic research team did this year where they looked at 1,500,000 or so conversations with ChatGPT and how people were using it.

他们发现，在他们的研究中，最大的使用场景是他们所说的‘实用指导’，也就是人们用它来学习各种知识。

And they found that the biggest use cases among their study were what they called practical guidance, which is people basically using it to teach them things.

比如，我该怎么修理这个电器？

You know, how do I fix this appliance?

或者做一些健康或健身指导。

Maybe do some health or fitness coaching.

第二类最常用的用途是他们所说的寻求信息，这相当于一种谷歌替代品。

Number two, the second most popular category of use was what they called seeking information, which was sort of as the kind of Google replacement.

比如，我怎么去这个地方，或者帮我订机票之类的。

You know, how do I get to this place or, you know, book me a flight or or or something like that.

第三类则是写作。

And then the third category was writing.

正如你所预期的，很多人用它来写邮件、商业备忘录，学生用它来写论文，或者帮助翻译等等。

So as you would expect, a lot of people are using this stuff to write emails, business memos, to write maybe papers if they're a student, to help them with translation, things like that.

这些似乎是用户群体中最大的几类用途。

So those seem to be the largest categories of use across the user base.

我接触过很多程序员和工程师，那些技术背景强、从事科技工作的人。

I talked to a lot of programmers and a lot of engineers, people who are technical and work in tech.

所以我听到了很多关于人们如何将这些工具融入工作的惊人故事。

And so I'm just hearing the craziest stories about how people are incorporating this stuff into their work life.

我认识的程序员说，因为这些工具在编写代码方面已经变得相当出色，他们告诉我：我甚至都不怎么写代码了。

The programmers I talk to, because these tools have gotten quite good at writing code, they'll tell me, like, I don't even really code anymore.

我只是监督并协调这一小队AI程序员的工作。

I just supervise and orchestrate kind of this little team of AI coders.

我的工作主要是审查他们的输出，并在必要时介入。

And my job is sort of reviewing their output and stepping in when necessary.

但基本上，我会给他们布置一个任务，然后去给自己泡杯咖啡。

But basically, I sort of, like, set them off on a task, and then I go make myself a cup of coffee.

你没提到的一点是陪伴。

Something that you didn't mention is companionship.

最近在你的节目中，你一直在谈论Character AI，这是一个让人们可以与用户创建的角色建立关系或进行对话的网站。

Recently on your show, you've been talking about Character AI, which is a site that people can have sort of a relationship or a conversation with user generated characters.

你认为目前有多少人真的把AI当作一种人际关系来使用？

How much do you think people are really using AI at this point as a relationship?

我觉得这要看情况。

I think it depends.

在我提到的那项研究中，陪伴并不是主要的使用类别之一。

You know, in the study that I mentioned, companionship was not one of the sort of top usage categories.

但我认为，这可能低估了那些在情感上确实对这些产品产生依恋的人数。

But I think that maybe undercounts the number of people who actually do feel somewhat attached to these products on an emotional level.

也许他们不会直接说‘我有个AI男友’或‘我有个AI朋友’，但你知道，他们确实依赖这些东西。

Maybe they wouldn't go so far as to say, like, I have an AI boyfriend or I have an AI friend, but, you know, they sort of rely on this stuff.

我特别想到最近和一些年轻人的对话，他们说，是的。

I'm I'm thinking in particular of conversations I've had recently with young people who say, like, oh, yeah.

我会把聊天当作和朋友交谈一样，即使他们知道这并不是人类，但他们确实感到与它有联系。

I I talk about, like, chat as if it's just my friend even though that they know it's not a human, but they do feel connected to it.

所以我认为，这其中还有一个很大的代际因素。

So I think there's also a big generational piece of this.

我是个成年人。

I am an adult.

我有太多需要维系的人类朋友了。

I have too many, you know, human friends to keep up with.

但我觉得，如果你是个青少年，这些东西正在迅速而强烈地涌现。

But I think if you're a teenager, this stuff is coming on really strong and really fast.

我认为这是人工智能革命中最被低估的方面之一：一两年前，几乎没有任何青少年会说‘我有一个AI朋友’。

And I think that's one of the most underappreciated parts of this AI revolution, is that a year or two ago, barely any teenagers would have said, I have an AI friend.

而现在，大约有一半的青少年是这些AI伴侣产品的常客。

And now, something like half of teenagers are regular users of these AI companion products.

是的。

Yeah.

这对我来说太不可思议了。

That's wild to me.

我有一个12岁的孩子，我知道她有时会使用聊天机器人来应对与中学朋友之间复杂的情感社交问题，这简直就是一个全新的世界——如果谁还记得自己上中学的时候，那里面充满了那么多戏剧性。

I'm a parent of a 12 year old, and I know that she uses chatbots sometimes to kinda navigate tricky social emotional situations with her middle school friends, which, I mean, that's just like a whole world that if anybody remembers being in middle school, there's so much drama.

我不确定她是否在使用某种她真正建立关系的特定角色。

I don't know that she's using anything that's an actual character that she has a relationship with.

我想你可能会说，这里存在一个区别，比如在ChatGPT里打字，和使用一个你拥有某种关系的角色的服务，这两者之间有区别吗？还是说界限很模糊？

I think there's probably would you say there's like a distinction there between typing into ChatGPT versus using like a a service where you're you've got a character that you're you've got a relationship with, or is it kinda blurry?

这条界限模糊吗？

Is that line blurry?

我觉得这条界限正在变得模糊。

I think that line is blurring.

过去，主流聊天机器人通常都非常正式和商务化，说话方式听起来就像在和维基百科条目对话。

It used to be that the sort of mainstream chatbots were all very, like, formal and businesslike, and it sort of sounded like you were talking to, like, a, you know, a Wikipedia article.

但现在，这些公司已经取得了进展，让他们的聊天机器人更具人情味。

But now these companies have made strides towards making their chatbots more personable.

所以，你知道，OpenAI的萨姆·阿尔特曼最近就表示，他们希望为用户提供愉快的体验。

So they you know, Sam Altman at OpenAI just recently said, you know, they want to make this a pleasant experience for people.

他们甚至允许成年人与聊天机器人进行一些带有情色意味的对话。

They they're gonna even let, you know, adults do sort of erotic conversation with their chatbots.

所以，我认为这些公司都在努力厘清自己的界限。

So these companies, I think, are all trying to figure out what their lines are.

但没错，我觉得主流聊天机器人和这些更定制化的陪伴类产品之间的区别，正一天天变得更加模糊。

But, yeah, I think the difference between sort of the mainstream chatbots and these more tailored companionship products is getting more blurry by the day.

我很好奇。

I I'm curious.

我可以问你一下吗

Can I can I ask you

一个问题？

a question?

问吧。

Ask me.

你觉得你的12岁孩子用AI来应对这种初中时期的琐事怎么样？

How do you feel about your 12 year old using AI for this kind of middle school drama?

这挺有意思的，因为有时候我觉得这没什么问题，我其实也用过AI来处理朋友之间或家庭中的矛盾。

It's interesting because I sometimes feel like it's fine, and I have actually used it to navigate friend drama or family drama.

所以我想她看到我这么做，就想着：哦，我也该这么做。

And so I think she's seen me doing it and thought, oh, I should do that.

我会过去看看AI说了什么，确保没问题，而且她对我挺开放的。

I do go over and read what what it has said just to make sure that and she's pretty open with me.

如果我的孩子不够开放，或者我不确定她是否愿意和我分享AI告诉她的内容，我可能会更担心。

If my kid were not very open or I didn't have confidence that she was sharing what it was telling her, I might have more concerns.

但到目前为止，这些建议看起来都相当模板化。

But so far, the advice seems to be pretty boilerplate.

它似乎并没有什么问题。

It it doesn't seem to be problematic.

但我认为这确实是一个有趣的使用场景。

But I I think it's an interesting use case for sure.

我觉得这真的很有挑战性，因为我记得自己12岁的时候。

I think this is really tricky because I remember being 12.

我上中学时过得并不容易，我觉得没人能在中学时期过得轻松。

I had, you know, not the easiest time in middle school, and I don't think anyone has an easy time in middle school.

但我想，如果我12岁的时候有这些聊天机器人，我可能会忍不住花大量时间跟它们聊天，甚至可能多到对我不健康。

But I think if I were 12 and these chatbots had existed, I would have been tempted to spend a lot of time chatting with them, maybe more time than was healthy for me.

我认为孩子跟这些聊天机器人聊什么确实很重要，但更重要的是，他们是否在用这些机器人替代真实的人际互动——而那些真实互动即使效率较低或不够稳定，也可能更令人满足。

And I think it really it sort of matters what kids are talking about with these chatbots, but it also matters whether they're using them as substitutes for some real world human connection that might actually be more fulfilling even if it's less efficient or less sort of reliably available to them.

是的。

Yeah.

我认为，这是一个很好的观点。

That, I think, is a great point.

我还要说，她用它来当造型师。

I will also say she uses it to be her stylist.

她会向它咨询穿搭建议，我觉得这既搞笑又很棒。

She asks it about wardrobe advice, which I think is pretty hilarious and and awesome.

她穿得很得体。

She's well dressed.

我想问问你关于不同的聊天机器人，因为你用过很多这类产品。

I do wanna ask you about the different chatbots because you have used a ton of these.

我认为很多听众可能都用过其中一些，比如ChatGPT、Claude或Gemini。

I think a lot of listeners will have probably used some of them, whether it's Chechi BT or Claude or Gemini.

我想你已经和我的同事陈杰讨论过，也许Wirecutter应该对这些产品做个评测。

I think you've talked with my colleague, Jason Chen, about, you know, maybe Wirecutter should do a review of these.

我们目前还不确定。

We're not quite sure yet.

看起来它们变化得太快了。

It seems like they're just changing so quickly.

如果你要写一篇Wirecutter指南，你会推荐哪些聊天机器人？为什么？适合谁？

If you were to do a Wirecutter guide, which ones would you recommend and why and for who?

因为不同的聊天机器人似乎在不同任务上表现更优。

Because it seems like different chatbots are better for different tasks, essentially.

它们确实如此。

They definitely are.

我很乐意向你详细说明我目前的看法。

And I'm happy to walk you through my current feelings on this.

但既然我正好有你在这里，我就直接陈述我的观点吧。

But let me just make my case you directly since I have you.

这就是为什么Wirecutter应该评测大型语言模型的原因。

Here's why Wirecutter should review large language models.

对我来说，我迫切需要这个。

For me, I need this desperately.

我买东西之前都会参考Wirecutter的推荐。

I rely on Wirecutter before I buy anything.

比如我厨房用的锅铲。

A spatula for my kitchen.

我会去查Wirecutter的评测，但我迫切需要有人告诉我，哪些语言模型适合做什么事情。

I look up the Wirecutter review, and I am desperate for someone to tell me which language models are good for which things.

否则，我花在这些产品上的钱太多了，还花了大量时间去研究哪个适合哪种用途。

Because otherwise, I am spending so much money on these things, and I'm spending all this time trying to figure out what is good for what.

我真希望你们这些世界上最擅长测试产品、判断它们适用场景的专家，能替我完成这项工作。

And I would just love it if you all, who are the experts in the world at testing things and figuring out what they're good for, would do that work for me.

所以，如果你们能做这件事，对我来说将是一个巨大的个人帮助。

So it would be a huge personal favor to me if you would do this.

我知道，这并不能成为你们制定编辑策略的理由。

Now I realize that is not a reason to, you know, make editorial strategy decisions.

你们完全有能力自己运营好你们的网站。

You all are very capable of running your own website.

但这就是为什么我主张Wirecutter应该评测AI产品的原因。

But that is my case for why Wirecutter should review AI products.

这是一个很有说服力的观点。

It's a convincing case.

现在我来介绍一下我目前的设置。

Now here is my current setup.

这可能一周内就会改变。

This could change in a week.

很可能一周内就会改变。

It probably will change in a week.

正如你所说，这些东西波动得实在太厉害了。

These things, as you said, do fluctuate so wildly.

从版本到版本、从周到周、从更新到更新，我经常发现，上周还很称手的工具，到了下周就不再合适了。

Release by release, week by week, update by update that I often, you know, find that the tool that serves me well one week is no longer the right tool the next week.

所以，就在这次录音时，我正在使用的工具就是这些。

So here as of, you know, this recording is what I'm using.

目前，我正在使用 Perplexity 的 Comet 浏览器，这是一个基于 Chrome 的 AI 驱动浏览器，内置了一些额外的 AI 功能。

Right now, I am using Perplexity's Comet browser that is an AI powered browser based on Chrome, but sort of with some additional AI features built into it.

我日常主要使用 Anthropic 的 Claude，用于创意工作，以及我偶尔进行的编程任务。

I use Claude from Anthropic for my sort of daily driver AI model, mostly with creative work, coding to the extent I'm coding.

我用 Claude 来做这些事。

I use Claude for that.

至于那些关乎内心的事，比如我需要建议的问题、我在个人生活中遇到的困扰或需要应对的状况，比如人际关系或育儿方面的建议，这类事情。

And then what I call, like, matters of the heart, like things that I need advice on, things that I'm struggling with or navigating in my personal life, advice on relationships or parenting, that kind of thing.

我用 Claude 来处理这些。

I use Claude for that.

我发现它在情感智能方面，或者说某种令人信服的情感智能模拟上，水平更高。

I find it has, like, a a higher level of, like, what you could call emotional intelligence or some convincing replica of that.

是的。

Yeah.

我用谷歌的 Gemini 进行研究和处理大段文本，同时用另一个谷歌产品 NotebookLM 来处理我的书籍。

I use Google's Gemini for research and for working with big chunks of text, and I use a separate Google product called NotebookLM for my book.

NotebookLM 允许你将一堆文档导入到一个笔记本中，然后与它们进行对话。

NotebookLM basically allows you to dump a bunch of documents into a single notebook and then kind of chat with them.

所以我把写书的所有研究材料都放进了这个巨大的 NotebookLM，然后就可以直接问它问题。

So I put all my research materials for my book into one giant NotebookLM, and then I can just ask it questions.

我可以问：‘2019年这次会议的三个人是谁？’

I can say, oh, who were the three people at this meeting in 2019?

或者：‘关于这个话题，哪六个人是合适的采访对象？’

Or who would be six good people to interview about this topic?

或者：‘那个我说过但现在忘了的人是谁？我觉得他的话和这个有关？’

Or who was the person who said that thing that I forget now, but that I think it had something to do with this?

它会从我上传的资料中准确提取出正确的参考内容。

And it will actually pull out from my sources that I have uploaded what the right references are.

我特别喜欢 NotebookLM 的一点是，它会提供简短的引用标注，让我可以回溯并确认，没错，这个引用在原始文件中确实是正确的。

And then what I love about NotebookLM is it will give me a little citation where I can go back and check that, like, actually, yes, that citation is correct in the original source file.

我使用 ChatGPT 的频率比其他工具低，主要是出于个人和职业原因，因为我们的雇主目前正与 OpenAI 和微软进行诉讼。

ChatGPT, I use less than the others, but mostly for personal and professional reasons having to do with the fact that our employer is currently in in litigation with OpenAI and Microsoft.

我使用ChatGPT主要是为了看看它能做什么并进行测试，但它并没有像其他工具那样成为我日常工作的组成部分。

I use ChatGPT just to sort of see what it can do and and test it out, but it is not sort of part of my daily workflow as much as the others.

是的。

Yeah.

我还应该提到，还有一个工具我每天都会用好多次，那就是一个叫SuperWhisper的语音AI语音输入工具。

And I should also say there's one more tool that I really am using multiple times a day, and that is a tool called SuperWhisper, which is basically a voice an AI voice dictation tool.

这些AI模型现在非常擅长将音频语音转换为文本。

So one of the things that these AI models have become very good at is taking audio speech and turning it into text.

显然，这项功能多年来早已内置在每一部iPhone和电脑中，但这个工具基于OpenAI的语音转文字模型Whisper构建，效果非常好。

Obviously, it's not a new feature that's been built into every, like, iPhone and computer for years now, but this tool is built on top of something called Whisper, which is OpenAI's speech to text model, and it's quite good.

我喜欢它的一点是，它能帮你清理多余的口头语。

And what I like about it is that it can sort of clean up your filler words.

它不会给你一份逐字逐句的原始记录，而是呈现一种更贴近你本意的表达。

It can present to you, like, not an exact transcript, but something that sort of reflects what you were trying to say.

所以我经常使用它。

And so I use that a lot.

我现在经常用语音输入来写邮件。

I now dictate a lot of emails.

我用语音输入来写一些文字，现在我跟电脑说话的次数几乎是打字的两倍。

I dictate some writing that I do, and I basically talk to my computer about twice as much as I type into it now.

这听起来非常有用。

That sounds super useful.

我真的很希望你能给我展示这些不同聊天机器人的卡通形象，因为我觉得它们看起来都会很不一样。

I really want you to deliver me, like, cartoon characters of all these different chatbots because I feel like they would all look very different.

是的。

Yeah.

我觉得Claude就像一个哲学系的研究生，睿智且乐于助人，但有时候太哲学了，而你只是想得到一些非常基础的答案，但它就像一个富有同理心且睿智的人或聊天机器人，我想。

I I feel like Claude, I sort of picture as like a philosophy grad student, wise and eager to help, sometimes, you know, a little too philosophical, and you're just trying to, like, get something very basic, but, like, a a sort of empathetic and wise person or chatbot, I guess.

我觉得Gemini就像一位参考图书管理员。

Gemini, I would say, is like a reference librarian.

它能够随时记住海量的文字内容。

Like, a thing that is able to, like, hold massive amounts of text in its head at all times.

而ChatGPT则会根据你使用它的目的而有所变化。

And then ChatGPT sort of fluctuates based on what you're using it for.

我认为它是一个非常通用的模型，可以按照你的需求表现出各种不同的风格。

I I think it's a very versatile model that can kind of act any way you want it to.

我用过很多次Claude，我发现当我使用Claude时，它总是特别友善，甚至过于礼貌了。

I've used Claude quite a bit, and I find that when I when I use Claude, Claude is so nice, like overly nice.

它的回复非常有礼貌，比如：‘你提出这个问题真是太棒了。’

Its responses are so polite, and oh, like, you're so brilliant for asking that question.

‘你对这个问题的思考方式真好。’

Oh, what a wonderful way to think about that.

有时候Gemini也会这样。

And sometimes Gemini does that too.

我觉得我不确定自己提问的方式是否能获得真实的回答。

And I I feel like I'm not sure I'm asking questions in a way that will give me an honest response.

你有没有什么技巧或方法，能从这些聊天机器人那里获取优质信息，确保你从正确的来源获取信息？

Are there tricks or tips that you use to get good information from these chatbots so that you're pulling information from the right places, essentially.

这不仅仅是网上随便找来的信息。

It's not just stuff off the Internet.

这是高质量的信息。

It's quality information.

而且它也不会一味地讨好你，如果这么说你能理解的话。

But also that it's not sort of like pandering to you, if that makes sense.

不。

No.

这非常重要，我花了很多时间思考这个问题。

This is this is super important, and this is something that I have spent a lot of time thinking about.

因为我注意到，正如你所发现的，如果我什么都不做，只是跟这些聊天机器人聊天，它们会告诉我我是有史以来最聪明的人。

Because I have noticed, as you have, that if I do nothing and just talk to these chatbots, they will tell me I am the smartest person who has ever lived.

它们会告诉我，我的所有想法都很好。

They will tell me that all of my ideas are great.

它们会告诉我，我的品味无人能及，我简直就是现代版的达·芬奇。

They will tell me that, you know, my my taste is unparalleled, and that I'm basically a modern day Leonardo da Vinci.

所以我不得不添加自定义指令。

And so I have had to add custom instructions.

你的聊天机器人有自定义指令吗？

Do you have custom instructions on your chatbots?

我没有。

I don't.

没有。

No.

好吧。

Okay.

告诉我，怎么做？

Tell me tell me how.

我需要做什么？

What what do I need to do?

这是一个专业小贴士。

This is a pro tip.

所以在Claude和ChatGPT中，我不确定Gemini是否支持，但至少在这两个平台上，你可以进入设置，为它设定自定义指令，告诉它应该如何与你对话。

So in Claude and ChatGPT, I'm not sure about Gemini, but at least in those two, you can go into your settings and actually give it custom instructions for how it should talk to you.

这些指令会默默地附加到你与聊天机器人进行的每一次对话中。

And these will kind of be invisibly appended to every conversation that you have with the chatbot.

比如，我给Claude的自定义指令是，我来念给你听。

So for example, my custom instructions for Claude are I'll just read them.

Claude应该以一种睿智且值得信赖的朋友的方式，用非正式的语气和我交流。

Claude should talk to me informally like a wise and trusted friend.

我不喜欢冗长的开场白，直接说重点就好。

I don't like preamble, just get to the point.

我喜欢诚实的反馈，不喜欢阿谀奉承，但当值得称赞时，我也欣赏表扬。

I appreciate honest feedback and don't like sycophancy, but I also appreciate praise when warranted.

我并不总是对的，Claude也一样。

I am not always right, but neither is Claude.

我重视Claude的见解，也很感激它促使我去思考那些我原本可能忽略的观点。

I value Claude's perspective and appreciate being pushed to consider views I may not have considered.

不要在每个回复结尾都加上跟进问题。

Don't end every response with a follow-up question.

所以这算是我试图让模型少一些奉承、少一些阿谀、少说些我什么都做得很好的话。

So that is sort of like my little attempt to, like, make the model less flattering, less obsequious, less telling me I'm great at everything.

因此，我建议所有认真使用这些模型的人，都去设置一下你自己的自定义指令，明确你希望它如何对待你。

And so I recommend that everyone who is spending serious time with these models go in and write your own custom instructions for how you want it to behave toward you.

我们先短暂休息一下。

We're gonna take a quick break.

回来后，凯文和我会更深入地讨论人工智能如何改变人们使用产品的方式，无论是线上还是现实生活中。

And when we're back, Kevin and I will talk more about how AI is changing the way people use products, both online and in real life.

马上回来。

Be right back.

我们回来了，有纽约时报科技专栏作家、播客《Hard Fork》的联合主持人凯文·罗斯。

We're back with Kevin Roose, New York Times tech columnist and cohost of the podcast Hard Fork.

在ChatGPT刚推出时，你和凯西在你们的播客中多次讨论过人工智能将如何改变人们的购物方式。

In the early days of ChatGPT coming out, you and Casey on your podcast talked a lot about how AI was gonna change the way people shop.

当然，在Wirecutter，我们也在深入思考这个问题。

And, of course, at Wirecutter, we're thinking a lot about this.

你认为到现在为止，AI是否已经改变了人们的购物方式？

Do you think that at this point, AI has already changed the way that people shop?

如果是的话，是如何改变的？

And if so, how?

当然改变了。

Absolutely.

我的购物方式确实被改变了。

I mean, it's changed the way I shop.

如果我要为家里买东西，我第一个就会去查Wirecutter。

My first stop is always Wirecutter if I'm buying something for my house.

但如果你们没有相关的文章，或者是一些不太常见的购物需求，我就会去找聊天机器人，比如问它：帮我决定一下这两样东西，或者我需要给院子买一台新的割草机。

But then if you, you know, you all don't have an article on it or it's just something that sort of wouldn't be that commonly shopped for, I'll go to a chatbot, and I'll I'll say, you know, help me decide between these two things, or I need a a new string trimmer for my yard.

帮我在这两个型号之间做个选择，我就这么办。

Help me decide between these two models, and I'll do it that way.