Vibe Coding：你需要了解的一切 —— 与Amjad Masad一起

别以为你只需输入提示词就能在另一端自动生成应用程序。至少预留一个下午的时间，认真投入努力，尝试完成你的第一个应用。一旦成功，你就会欲罢不能。

Don't go into this thinking you can just have prompt and have an application pop up at the other end. At least set an afternoon, to to give it some good effort and try to get, like, your first app in. And once you do that, you just get evicted.

我这里有个数据：Replit在不到六个月内将年经常性收入增长了10倍，达到1亿美元。这种增长是氛围编程带来的，还是AI编程带来的？

I have the stat here that Replit has multiplied by its revenue by 10 x in less than six months to a 100,000,000 in annual recurring revenue. Is that growth vibe coding or is that growth AI coding?

氛围编程。

Vibe coding.

AI编程究竟是业余爱好，还是一场让所有人都能参与构建的技术革命开端？今天我们邀请到Replit首席执行官Amjad Masad，他在福斯特城的Replit总部接受采访。我是Judd。很高兴见到你，欢迎来到节目。

Is AI coding just a hobby or the beginning of a technological revolution that empowers everyone to build? Our guest today, Amjad Masad, the CEO of Replit, has some answers, and we're here in Replit headquarters in Foster City to speak him. I'm Judd. Great to see you. Welcome to the show.

谢谢，很荣幸能参与节目。

Thank you. I'm excited to be on the show.

今天我们要讨论氛围编程和AI编程这两个相似又不同的概念。首先想请教氛围编程——本质上就是你写提示词，然后AI自动构建软件。这正是Replit支持的功能，我也亲自尝试过。目前发现哪些应用场景用户确实能有效运用这项技术？

So we're gonna talk today about vibe coding and AI coding, which are two similar but different things. I first wanted to speak with you about vibe coding, which is effectively you write a prompt, and then the AI goes ahead and build software for you. This is something that Replit enables. This is something I've tried. What are some of the use cases that you're finding people are actually having effective approaches with this?

具体在哪些领域人们做得比较成功？

Like where are the places where people are doing as well?

大致有三类应用场景。首先是个人和家庭生活。比如很多人喜欢做健康追踪——记录睡眠数据、从Fitbit导入数据。

There's like broadly three use cases. One is personal life, family life. So you know, for example, like a lot of people like to do health tracking. I'm gonna track my sleep. I'm gonna pull in data from my Fitbit.

我会用AI处理这些数据，在手机上做个日常使用的应用；或者给孩子开发数学/阅读教育软件；还有人开发了家庭版'家务英雄'——在墙上挂个iPad显示谁做家务最多，把家庭生活游戏化。这类应用的受欢迎程度会超出你的想象。在我长期关注的创作者工具领域，始终存在'个性化软件'和'可塑性软件'的理念。这其实可以追溯到早期计算机历史——比如苹果公司曾推出HyperCard软件。

I'm gonna like have AI sort of process that data, I'm gonna have this app on my phone that I use every day, or I'm gonna build an educational app for my kid to learn math or reading, or we're gonna have, like someone built like a chore hero for their family to like, you know, have an iPad on the wall and, like, here's who's doing the most chores and gamifying their family life. You'd be surprised how popular this use case is. And so, you know, in the in the niche that I've always been in, which is, like, creator tools, that there's always been this idea of personal software, malleable software. And by the way, this goes to the early computing history. So, you know, for example, like, Apple had this piece of software called HyperCard.

HyperCard让每个人都能制作个性化软件。还有Visual Basic等多次尝试。但现在终于真正实现了全民编程。我们提供移动应用，任何人都能用来开发软件。

HyperCard allowed anyone to make personal software. You know, there's Vidual Basic. It's been attempted so many times, But for the first time now, anyone can make software. So there's a lot of there's a class of personal software. We have a mobile app, and you can use that to make software.

最有趣的事情莫过于和五、六、七岁的孩子们坐在一起，头脑风暴游戏点子并和他们一起制作游戏。这就是其中一个方向。

It's the most fun thing to do is sit down with your kids, five, six, seven years old, and just brainstorm games and make games with them. So that's one bucket.

但在讨论下一个方向前，我想问你个问题。这是否说明软件行业尚未满足这么多应用场景？还是这些场景本身不具经济性？或者人们可能先为家人开发产品，随后意外发现它能服务大众市场并成为一门生意？

But before we go to the next bucket, I wanna ask you a question. So does this say something about the software industry that the software industry just hasn't served so many use cases? Or are these use cases noneconomic? Or is it possible that people will build things for their family, and the next thing you know, they can serve that mass market and it becomes a business?

确实存在这样的市场，而且绝对能从中赚取丰厚利润。

It is certainly you know, there's certainly a market there, and you can certainly make a lot of money from that.

好的。因为当我思考'AI将取代人类工作'这个观点时——虽然我们稍后会讨论就业问题——我的反应是：等等，明明还有这么多东西等待开发。嗯。

Okay. Because when I think about, like, this this concept that we're gonna get to jobs, but this concept that AI is gonna take our jobs, to me, it's like, wait. There's so much left to build. Mhmm.

如果你

If you

只考虑现有技术的维护，或许会失业。但软件尚未触及的领域如此之多，在我看来机会远比人们想象的更广阔。

think just about what we have today and maintaining that, maybe it will. Yeah. But there's so much that software has not yet touched that it seems to me that there's more opportunity out there than people are.

回到你之前的问题——你可以决定讨论深度，这个话题我能聊几小时——早期计算机先驱们都认为计算机的特殊性在于可编程性。冯·诺依曼首次提出可编程机器架构（至今仍在沿用）时，人们以为从此人人都能用计算机编程解决问题、开发应用，但最终并未实现大众化普及。

But just to touch on your earlier question, and you tell me how deep we wanna go because I can talk about this for hours. But when in the early computing pioneers, they all had this idea that computers are this the thing that makes computers special is this idea of program programmability. Right? The moment we had a program programmable machine that was first invented by Von Neumann, and it's the same architecture that we use today, the thinking was, oh, anyone can use a computer to program, to solve problems, to to build applications, and all of that. It it didn't get mass consumer adoption.

根源在于编程太难。后来施乐帕克研究中心开发出图形界面，他们邀请新锐企业家史蒂夫·乔布斯参观。乔布斯看到桌面、菜单等设计后——当时Apple II仍是命令行界面只能编写基础程序——立刻意识到这是实现计算机大众化的关键。

And the reason is because coding is hard. And so you had Xerox PARC as a research in Palo Alto Research Center. They developed GUI. One day, they invite this up and coming entrepreneur called Steve Jobs. Steve Jobs looks at desktops, menus, items, and and he's like, he has the Apple too.

于是他借鉴施乐的技术，将其融入麦金塔电脑。后来微软Windows也复制了这套界面，突然之间计算机变得人人可用。如今数十亿人使用计算机和基于相同理念的智能手机，但我们丢失了'人人可编程'的初心。这正是我毕生热衷的课题。

Obviously, Apple too is also still command line. You can write some basic. And he's like, okay. This is the key to get mass consumer adoption of computers. And so he copies what Xerox had, and he builds it into the Mac.

And, obviously, later Windows and Microsoft copies UI, and then suddenly your computers are usable by anyone. And this is amazing. Now, like, billions of people use computers, and now we have phones based on the same idea. What we lost is this idea that anyone can program a computer. So that's something I've been passionate about all my life.

And, obviously, later Windows and Microsoft copies UI, and then suddenly your computers are usable by anyone. And this is amazing. Now, like, billions of people use computers, and now we have phones based on the same idea. What we lost is this idea that anyone can program a computer. So that's something I've been passionate about all my life.

计算机从根本上说应该是可编程的。视觉编程已经经历了多次迭代。大约十年前，我们经历了无代码、低代码的革命。但我认为它从未发挥出全部潜力。

Computers should fundamentally be programmable. And there's been a lot of different iterations with visual programming. We had the no code, low code revolution that happened maybe ten years ago. I would say it never reached the full potential.

这更像是一个噱头而非现实。但现在

It was more of a buzzword than reality. But now

我认为这确实是一个价值数十亿美元的市场，但还不是万亿美元级别的。而‘人人都能开发软件’这个概念背后潜藏着巨大的市场空间。

I think I think it is is a multibillion dollar market for sure, but it's not a trillion dollar market. And I think this idea of, like, anyone can make software is such a massive market.

嗯。好的。那么第二类人群。

Mhmm. Okay. So bucket number two.

第二类往往是创业者。世界上每个人都有想法，人们在自己的工作领域积累了丰富的专业知识。比如我今天听说有位优步司机正在用Replit开发应用。

Bucket number two tends to be entrepreneurs. And so everyone in the world has ideas. People build so much domain knowledge about whatever their their field of work. Right? I was hearing a story today of an Uber driver that is starting to make an app with Replit.

这个应用是关于物流的。他以前是卡车司机，因此拥有管理车队等专业领域的知识。但过去由于缺乏编程技能，或是没有资金雇佣开发人员，他无法将其转化为软件。

And the app is about logistics. He was a truck driver before. And so he had domain knowledge about how to manage fleets, for example. But he never was able to make it into software because he didn't have the skill. Maybe he didn't have the capital to go commission a contractor to do it.

现在他突然可以做到了。随便找个人，无论身处哪个行业，他们都会发现有些软件或技术需求尚未被满足——因为外人缺乏他们的深度领域知识。我们见过各行各业的创业者，最令人称道的是我们在Replit社交媒体上公开报道过的一位英国医生。他说现有的医患关系管理应用都不够全面。

And suddenly, he can do it. So pick anyone on the street. And they all, in whatever industry they're in, they realize that there's a need for piece of software or technology that no one has built because they don't have that deep domain knowledge. So we see entrepreneurs from all walks of life. One of our favorite one, we talked about it publicly on our replica social media channels, a doctor from The UK that he's like, you know, there's all these apps around managing doctor patient relationships, but they never they're not fully integrated.

比如ZocDoc可以预约，但如何管理处方？能否追踪患者长期健康状况？能否整合来自智能体重秤、Fitbit等设备的数据？于是他搭建了这个综合平台。有机构报价10万英镑，而他只用了不到200英镑就完成了开发。

So, you know, you have ZocDoc, you can go make an appointment. But how do you manage your prescriptions? Can I track my patient over time, their progress? Can I get information from their Wi Fi connected scale, from their Fitbit, from and so he built this comprehensive platform? He got quoted by an agency a £100,000, and he built it less than 200, you know, British pounds.

不是20万，是200英镑。约合400美元

Not 200,000. 200. $400

所以这就是

pounds. So this is

那些正在被氛围编码的东西。实际上就是输入提示，我想开发这个软件。

stuff that's being vibe coded. Effectively prompt in, I wanna build the software.

是的。

Yeah.

然后Replit就会去构建它。

And then Replit will go build it.

没错。现在这就是一家初创公司了。我们已经有初创公司在Replit上起步，实现了数百万美元的收入规模。其中一些公司的估值甚至达到了五亿美元左右。所以我们服务的对象从小型创业者到风险投资规模的创业者都有。

Yeah. And this is now a startup. And we've had startups start on Replit, multimillion dollar revenue run rate. Some of them have raised at, like, half a billion dollar valuation. So we have all the way from small entrepreneurs to start up venture scale entrepreneurs.

但这确实让我非常兴奋，因为美国一直以创业精神著称。这也是吸引我来到这个国家的原因。但事实上，如果你看看统计数据，尽管我们每天都在听说湾区硅谷的创业动态，但纵观过去一百年，全国其他地区的新公司创建数量一直在下降。疫情期间曾有过短暂回升，当时大家都待在家里想着'我要创业'。

But this this gets me really excited because America has always been about entrepreneurship. And this is really what attracted me to this country. But actually, you look at the stats, entrepreneurship over time, although we hear about what's happening in the Bay Area and Silicon Valley, there's all startups every day. But the rest of the country, actually, new firm creation has been going down over the past one hundred years. There was an uptick during COVID where everyone's sitting at home is like, I have that's an

那时候我开始创业。

when started my business.

对，完全正确。那段时间很棒，但之后又回归了常态。我认为随着AI的发展，我们将再次看到创业热潮。这就是第二类人群——创业者。

Right. Exactly. Which was great, but that actually we had a regression to the mean. And I think with AI, we're going to see that explode again. So that's the second bucket, entrepreneurs.

还有第三类人群，就是像我们这样的公司员工。举个例子，我们人力资源部门的故事。我们HR团队很小，Replit一直保持精干团队。

One more bucket. Third one is people at companies like this one. So actually, I'll give you a story from our HR department. We have a small HR department. Replit is kind of a lean team.

公司只有80人，但我们使用了很多SaaS工具，每年花费数十万甚至上百万美元购买各种专业功能软件。有时候这些软件并不完全符合我们的需求，而且价格太高。

We're 80 people. And so we have a lot of these SaaS tools. We pay tens, hundreds of thousands of dollars to do every specific kind of function. And sometimes, they don't really fit our use case. We think they're too expensive.

这位HR同事需要一个组织结构图软件，要能可视化展示架构，增删人员，维护历史记录，并能回溯查看变更情况。她调研市场后发现，现有软件都无法满足她想要连接其他HRIS系统或数据库的定制需求，而且都非常昂贵，需要大量IT支持。于是她直接在Replit上用了五天时间编码开发出来。现在我们有了完全符合需求的系统。

So this HR person had a need for an org chart software that can visualize the org chart, that can add, remove people, maintain a history, can look back and see what happened, what changes it did. And went on the market and saw that none of the software captured the exact bespoke use case where she wanted to connect it to our kind of more other HRIS systems or databases. And they were they they all needed, you know, that they're all very expensive and need a lot of IT support. So she went into Replode and built it, five coded it in three days. And so that meant that we have a system that exactly fits our use case.

这也意味着我们不再需要每年为一款SaaS软件支付10.20美元、3万美元的费用。这种现象正普遍发生。我们看到企业通过用内部开发的软件替代SaaS软件，节省了数十万美元。

And that also meant that we're not paying $10.20, $30,000 a year for a piece of SaaS software. And that's happening across the board. We see companies saving hundreds of thousands of dollars replacing SaaS software with built in, with internally built software.

那么要做好这件事，是否需要具备一定的技术背景或技术知识呢？我举个例子，开始录制前我跟你提过，这周我开了个Replit账户。我想开发一个简单的自选冒险游戏，应该叫'历史浩劫'，玩家可以体验不同的历史场景。

Now do you need to be someone with some technical background or some technical know how to be able to do this well? Because I'll give you an example. I mentioned to you before we start recording, I opened a Replit account this week. I wanted to build a simple choose your own adventure game. I think it's called history havoc where you can work your way through different history scenarios.

但最终效果没达到我的预期。

But it just didn't get to the point where I wanted it.

你花了多长时间开发？

How long did you work on it?

大概一小时吧，时间不长。另外坦白说，我只是用了你们的入门计划，还没付费呢。

So I spent about an hour on it. Not a lot of time. And I also, full disclosure, I'm just on your starter plan. I'm not paying yet. Yeah.

但我就是搞不定。我还尝试开发过一个故事追踪器，结果它没能按我预期的方式爬取网页数据。所以对很多人来说，这技术看起来还是——

But I couldn't I couldn't get it to work. I also tried to build this story tracker. Yeah. And it wasn't able to crawl the web the way that I hoped it would. So it still seems like this to a lot of people that this is something that is helpful.

如果你是技术人员，想做个原型机还行。但你举的这些用例看起来都是成熟公司或可运行的软件。解释下这个差异吧。

If you're technical, you wanna make a prototype. But these use cases that you're giving seem to be full blown companies or working pieces of software. So explain that disconnect.

我认为这需要毅力。显然，机器学习模型存在随机性。

I think it it requires grit. Obviously, there's, like, stochasticity in in the machine learning models.

什么意思？

What that is.

相同的提示词可能因为GPU内部的随机性让你走向成功。大语言模型有个叫'温度'的参数，这个参数本质上控制着LLM输出词语采样的随机程度。LLM的工作原理是：你输入一段文本，它尝试补全下一个词（我们称之为标记）。在这个过程中，它会生成大量候选词。

The same prompts can put you on a path of success based on randomness that's happening inside the GPUs. There's this parameter in large language models called temperature. And temperature is literally like how random is the sampling of the words coming out of the LLM. So the LLM, the way it works, you give it a piece of text, and it tries to complete the next word, the next token, as we call it. And the way it happens, it generates a lot of candidates.

所以，你知道，那只红红狐狸，你知道的，跳啊、拍啊什么的。但‘跳’是概率最高的选项。你知道，这是模型在数百万案例中见过最常接续这句话的词。但采样器会让选择随机化，这种随机性反而增强了创造性。NVIDIA芯片或GPU内部也存在固有的随机化机制。

So, you know, the red red fox, you know, jumped, slapped, whatever. But, like, jumped is the top one. You know, it's the highest probability one that that you know, the model have seen it occur after the sentence and millions of of cases. But, you know, you have the sampler and could be randomizing what it picks, and that that randomization makes it more creative. There's also inherent randomization inside the the, like, the NVIDIA chips or the GPUs.

这类软件与传统离散输入输出的经典软件不同，机器学习模型具有固有随机性。这是创造力的特性而非缺陷。有些人可能遇到Repli的坏运气，我们当然在努力缓解这类问题。

So this style of software is unlike the software the classic software where everything is discrete, input output, machine learning models have inherent randomness. And that's a feature, not a bug that creates creativity. Right? So some people sometimes get on a on a bad luck with with a Repli. We're obviously trying to mitigate a lot of these problems.

但我要说这需要毅力。专业程序员可能要两天完成的游戏，在Repli上只需三四小时，但需要些韧劲。这不是魔法，你提到的技术能力虽非必需，却可以逐步培养。

But I would say it's also requires grit. Like, the gay the game you just described, professional programmers coding might take them a two days thing. On Repli, you can do it in two, three, four hours, but it it would require a little bit of grit. So it's not magic. And the skills you were talking about, the technical skills, although they're not required, you can build them up over time.

我们的环境会在使用中展现这些特性。建议新手别指望输入提示就能直接生成应用，至少留个下午认真尝试做出第一个应用——之后你就会上瘾。

And our environment kind of shows some of these features as you're working with it. And so I I I would I would suggest to people that don't go into this thinking you can just have prompt and have an application pop up at the other end. I would say at least set an afternoon to to give it some good effort and try to get like your first app in. And once you do that, you just get addicted.

那么有氛围编程（用提示生成应用后以英语优化）和AI编程（类似大型自动补全）。你认为两者的机遇分别是什么？当前AI行业的重心在哪里？我打个比方...

So there's vibe coding, which is again prompt, and then you make an app and then you can refine it with more English. And then there's AI coding where you could basically have AI complete your code, big auto complete. So what do you think the opportunity is in vibe coding versus AI coding? And where do you think the energy is in the AI industry today? I gave the analogy of

回顾计算机发展史，这个类比很贴切。早期只有大型机，IBM生产的房间大小计算机，只有大企业、政府和大学能用。

the history of computing, and I think it's a very suitable analogy for a lot of what we're talking about. Early on in computing, we had the mainframes. So the mainframes, really big room sized computers. IBM used to make them. Large corporations and governments use them in universities.

直到苹果推出Apple II才进入大众市场，后来才有Windows等设备。大型机满足专业需求却忽视消费者。如今PC市场不仅远超工作站规模，最终更吞噬了专业级软件。

But everyday people didn't have access to them until Apple created the Apple II. And that was the first mass consumer market computer. And since then, we've had Windows and all these devices. The mainframe was already serving the professionals' needs, but it wasn't serving the consumer needs. Now if you look at the market for PCs versus the professional workstations, some microsystems, all of that, which used to be the case, the PC not only was a much bigger market, eventually, it subsumed the, the more professional grade software.

这就是颠覆性理论。哈佛商学院教授克莱顿·克里斯坦森在《创新者的窘境》中指出：许多技术从低端起步，凭借大众吸引力获得规模优势，最终蚕食高端市场。

And this is this is called the disruption theory. A lot of your audience that might be into business history or theory, Clay Christensen used to be, I think, a Harvard Business School professor. And he wrote this book called The Innovator's Dilemma. And the idea is that a lot of technologies start at the lower end. And because of their mass market appeal, they onboard a lot more users and customers.

当前高端市场正是你提到的AI编程工具。全球约3000万专业开发者经过四五年科班训练，若提升他们20-40%效率，像谷歌这样的公司就能获得数十亿美元的生产力增值。

And over time, they reach certain economies of scale. And they subsume even the upper end of the market. Currently, upper end of the market is what you were talking about with AI coding tools, right? So there's like 30,000,000 developers all over the world, maybe a little more now. Those are professional developers that went to computer science classes in in college.

这个市场价值显而易见。

They were trained for four or five years. And now they're on, you know, working at companies. If you make those developers twenty, thirty, 40% more productive, you get depending on, you know, if you're a company of the size of Google, you know, just like billions of dollars worth of productivity. Right? So the market is really obvious there.

你可以去申请并获取它。但这是个零和游戏市场。看看微软的Copilot——作为首个上市的产品，与更现代化的AI编程IDE Cursor相比。随着Cursor蚕食市场份额，你会发现其增长几乎与Copilot使用率的下降完全成比例。这就是零和市场的迹象。

You can go apply it and and get get it. But it's a zero sum market. If you look at Copilot, which is Microsoft's product, which was the first to market versus Cursor, which is the more modern kind of AI coding IDE. As Cursor is eating market share, you can see it is almost exactly proportional to Copilot, declining in usage. So that's a sign of a zero sum market.

这个市场利润丰厚，增长空间巨大，但它并非我们正在经历的根本性革命——那种让任何人都能开发软件的革命。

It is very lucrative, and there's a lot more growth to be had there, but it is not this fundamentally revolution that we can be going through where it's anyone can make software.

让我换个方式问。数据显示Replit在不到六个月内将收入增长了10倍，达到1亿美元的年度经常性收入。这种增长是氛围编程带来的，还是AI编程带来的？

Let me let me ask it this way. I have the stat here that Replit has multiplied by its revenue by 10x in less than six months to a 100,000,000 in annual recurring revenue. So is that growth vibe coding or is that growth AI coding?

氛围编程。

Vibe coding.

真的吗？那么这些氛围编程项目，或是人们用提示词构建的定制程序，它们是投入实际生产了，还是大多只是人们随便玩玩的业余项目？

Really? Yeah. And is are these vibe coding programs or these bespoke programs that people are building with prompts, are they in production or are they mostly hobbies that people fool around with?

第一类用户更多是业余爱好和个人生活领域。第二类则是创业者。如你所知，大多数初创企业都会失败，因此多数创业构想最终无法落地。而那10%成功起步的小型企业，他们从Replit中获益最多。

Depends on first bucket is is more hobby personal life. Second bucket, entrepreneurs. As you know, most startups die. So most startup ideas don't don't make it a fruition. The 10% of startups that are small businesses that get off the ground, they get the most value out of Replit.

其中部分企业现已投入实际运营。我分享过许多这类案例，比如有位创作者约翰·钱尼，他是连续创业者。

And some of them are in production now. I've talked about a lot of these stories. But for example, we have this creator. His name is John Chaney. He's a serial entrepreneur.

过去他需要耗费数月时间和数十万美元开发应用程序，而现在他能在几周内创立新业务并实现百万美元年化营收。当然，他具备经验优势，深谙创业之道，但这些能力人们可以逐步习得。至于企业客户，例如Zillow——

It used to take him many months and hundreds of thousands of dollars to build applications, and now he can spin up a business and get to million dollar run rates in a matter of weeks. Obviously, he has experience. He knows the formula of what it means to be an entrepreneur. But people can learn that over time. And in terms of the enterprise, we have, for example, Zillow.

Zillow的CEO最近在《纽约时报》DealBook专栏中提到，全公司都在使用Replit加速产品创新。因为产品创新不再依赖工程师，产品经理可以独立完成整个迭代流程并获取用户反馈，无需经过工程师环节。这极大提升了效率。我们还有Duolingo——

The CEO of Zillow recently on New York Times deal book talked about how everyone at Zillow is using Replit to accelerate product innovation. Because product innovation no longer depends on engineers. You can have product managers do the entire iteration getting user feedback even without going to the engineers. So it just increases it. We have Duolingo.

许多这类专注创新的客户，正在开发他们的第二、第三款产品，并将Replit广泛应用于这些场景中。

A bunch of these customers that are really focused on innovating, building their second, third product that are now using Replit for for a lot of these use cases.

那么用例是你们先构建一个原型，然后获取反馈？如果一切顺利，再由核心工程师将其整合到产品中？这是一个用例场景。

So is the use case that you build, like, a prototype, and then you get some feedback? And then if everything works out well, then you build into the product with your core engineers? That's one use case.

有意思。没错，这是一个用例场景。它能显著缩短产品上市时间。

Okay. That's interesting. Yeah, that's one use case. It's really great. It rapidly improves the time to market.

第二个用例是运营和内部工具。比如西尔斯家居服务这家老牌公司，他们派员工上门维修房屋。其运营团队曾想为外勤人员开发大量AI工具和软件来管理工作与收入，但他们的系统还是百年历史的COBOL程序，工程师们正忙着迁移升级这套系统。

The second use case is operations and internal tools. So for examples, like Sears Home Services, pretty old company, employs people that go and fix homes. And they had an operations team that wanted to build a lot of AI tools and software for their field workers to be able to manage their work and their earnings and all that. But their software was like this 100 year old COBOL programs. And the engineers were kind of busy kind of migrating that and improving that.

于是运营团队开始用Replit快速部署AI应用，这些生产级工具每天帮助外勤人员规划最优路线，最大化日收入。这类运营用例通常都会直接投入生产环境运行。

So the operations team started using Replen to spin up these AI applications that are deployed, used in productions by those field workers every day to match their day and and and kind of design the optimal routes to how to maximize their earnings per day. So the operations type use cases tend to be deployed running in production.

等等，你们是否也在推动AI编程？还是说主要让Replit转型为可视化编程公司？

Okay. So just so I'm clear, you're are you also facilitating AI coding? Or is it mostly that you So so turn Replit into a vibe coding company?

我的使命始终是赋能人们实现'创造软件'这个魔法般的行为——这是人生中最奇妙的体验之一。作为Codecademy的创始工程师，此前我还开发过相关开源工具。Codecademy教会数百万人编程，改变了无数人生轨迹。

My mission has always been about how do you enable people to do, this magical thing that is creating software. It's one of the most magical exciting experiences you you would ever have. And, I I was a founding engineer at Codecademy. And before that, I built open source tools to do that. Codecademy taught millions and millions of people how to code, and we changed, you know, a lot of lives.

因此Replit的基因始终是降低编程门槛。虽然曾更偏向开发者工具，但由于我们提供开箱即用的数据库、鉴权、部署等全套服务——

So the DNA of Replit has always been about how do you make programming more accessible. It was it had, like, a more devolver bent at some point. But because Replit is sort of batteries included platform, we give you the database. We give you the authentication. We give you the, the deployment.

包括扩展性在内的所有功能都无需外求，这使得主要受益者往往是非专业程序员（尽管专业程序员也占约20%的使用场景）。

We give you the scalability. We give you all of that out of the box. You don't have to go anywhere else to do any of that. It always meant that the people that are getting the most out of it tend to be the not they're not not professional programmers. Although professional programmers do use it, I would say, like, this 20% of the use cases.

那么问题来了：这些用户会吸引专业程序员吗？有件趣事：几个月前我在旧金山Semaphore科技活动听你演讲时发推提到，你说一两年后企业可能无需工程师就能自主运营——

And the question is then do the people using repl. Then come for the people who are those professional programmers. There's a funny thing that happened. I watched you have a talk at the semaphore tech event in San Francisco a couple months ago. And I tweeted something that you said that in one year or eighteen months companies might be able to run themselves without engineers.

结果有人用梗图回复我：'创始人公开说AI六个月写完99%的代码就不需要工程师了，私下却发消息问：谁认识好的React开发？悬赏三万美元外加长子命名权'

And then somebody responded to me with this meme where they said founders in public AI is writing 99% of our code in six months we won't need any engineers. Founders in the DMs, does anyone know a good React developer $30,000 bonus and I will name my firstborn son

您先请。

after you.

那么您能否解释这种矛盾——一方面认为工程师将消失，另一方面市场对工程师的需求却依然旺盛？

So can you explain that disconnect between this view that engineers are going away and this still like very intense demand for engineers in the market?

我从未说过工程师会消失。我的观点是创业者可以在不需要工程师的情况下创业。这已成为现实——我接触的YC初创公司就是例证。YC是全球最负盛名的创业加速器（位于湾区），过去他们会建议创业者寻找技术联合创始人。

I never made the point that engineers would go away. I make the point that entrepreneurs can start businesses without needing engineers. And that we already see that. We already see I meet YC companies and Y Combinator is the most prestigious startup accelerator in the world, Bay Area. And in the past, Y Combinator would encourage you to go get a technical cofounder.

但正如我们所说，现在许多有绝妙点子但缺乏技术合伙人的人也能进入YC。他们告诉我们：'我们就用Replen搭建产品，看能走多远'——结果往往走得非常非常远。

But like we said, there's so many people with amazing ideas that don't have a technical cofounder. And so they're starting to get into YC. And what they tell us is we're just gonna build this thing on Replen. We're gonna see how far we can get. And they often get really, really far.

若你想打造估值数亿乃至百亿的独角兽企业，当然需要雇佣工程师。但若只想创建能让你生活优渥（甚至致富）的公司，现在几乎可以完全不依赖开发者就能实现。请注意我的谈话对象——

Now if you're building a venture scale company and you wanna, like, get to hundreds of millions of dollars of revenue and you want to, you know, become billion, 10,000,000,000, $100,000,000,000 company, you're gonna have to hire engineers. But if you're trying to build a company that creates a really great living for you, even, you know, you can potentially get rich from it, you I think we're almost there where you can do it on your own without any developers. And so when I'm talking I'm talking to our audience

明白。

Right.

我的受众并非微软或Facebook这类企业，它们不可能取代开发者。我认为现代开发者的生产力远超以往，因为单兵作战就能产生巨大杠杆效应。所以我们既要招募顶尖开发者，也要看到：如今我们团队的产出规模，相当于五年前SaaS公司十倍人力才能达到。

As opposed to I'm not I'm not talking to Microsoft or or Facebook. They're not gonna replace developers anymore. My view on developer productivity is that developers are much more impactful than they used to be because a single developer can be so highly leveraged these days. And so, yes, you wanna find the best developers, and we're expanding the team. But the the the scale that our output is at today, we would be 10 x that the number of people for a SaaS company five years ago.

想想看——五年前要实现1亿美元年营收，平均需要500人，很多公司甚至要上千人。

Wow. To reach a $100,000,000 in run rate, you know, five years ago, on average, you you would have, like, 500. A lot of companies will have thousand people.

你们现在多少人？80？哇...这让我不禁思考，当企业这样发展时，技术层面的未来会是什么图景？

How many do you have? 80. Wow. Okay. You know, it just makes me wonder that as companies grow like this, what the future is gonna look like from the technical side.

我很好奇技术人才的角色——假设经济如此扩张，连老奶奶都能用AI工具创业，那技术者是否就沦为'问题修复队'？《未来主义》杂志有篇趣文称：'为省成本使用AI的公司，现正花费巨资雇人修正AI错误'——不过说的不是Vibe编程。

And I'm curious to the folks who have technical abilities, you know, let's say the economy expands like this and everyone and their grandma can build literally can build a company using AI tools. Do the technical people then come in and sort of clean up the problems or they you're like cleanup crew. I was reading this funny article and publication called Futurism. It says, companies that tried to save money with AI are now spending a fortune hiring people to fix its mistakes. And it was about it wasn't about vibe coding.

其实是关于内容，比如内容营销。对。或者你的内容营销计划里充斥着这种平淡无奇的ChatGPT生成的文案，而且有一半时间它会说‘作为一个AI助手’这样的话。

It was actually about content, like content marketing. Yeah. Or like your your content marketing plan is just filled with this like, you know, kind of bland chat GPT generated copy and half the time it says, as an AI assistant, this

就是这种信息。哦，对。你可以谷歌一下。会看到很多类似结果。没错。

is the message. Oh, yeah. Can Google that. You'll see so many hits. Yeah.

所以我很好奇想听听你的观点：技术领域最终是否会沦为‘氛围编程’出问题后的清理队？

So I am curious to to hear your perspective on does does the technical field end up becoming cleanup crews for vibe coding gone wrong?

让我告诉你我认为技术人员现在为何有工作保障。如果你在为我的特斯拉写软件，我不希望你用氛围编程，我要你写可验证的低层代码。如果是为航天飞机写代码，你写的也是可验证的低层代码。这些都是生死攸关的场景。

Let me just tell you where I think technical folks have a job security today. So I think if you're writing software for my Tesla, I don't want you to be vibe coding. I want you to write low level verifiable code. If you're writing code for, space shuttle, you're writing low level verifiable code. But also, even I mean, those are life or death situations.

我认为这些领域不需要氛围编程，需要更高精度。即便是大型平台——如果你在构建AWS、谷歌云或Azure的核心云组件，如存储或虚拟机组件，你需要懂分布式系统、懂如何构建大规模故障安全系统的工程师。这些工程师在未来仍有工作保障，因为这些模型的随机性等问题，每行代码都需要严格审查和管理。

So I think we need we don't need by coding there. We need more precision. But but even sort of large scale platforms, if you're building a core cloud component, the storage or virtual machine components on AWS or Google Cloud or Azure, you want systems engineers that understand distributed systems, understand how to create fail safe systems at scale. So I think engineers there have job security for the foreseeable future. Because of the problem of the stochasticity of these models and all of that, you need every line of code to be reviewed and managed very carefully.

我认为AI影响最大的将是产品领域。做产品的人想快速迭代，想用内部工具取代当今混乱的SaaS软件。这个趋势正在发生。

Now where I think AI is gonna have the most impact is on product. And people building products, they wanna iterate on it really quickly. They wanna, internal tools. People wanna replace all the mess of the SaaS software that we have today. So I think I think that's happening.

关于清理问题，取决于你对AI发展的看法。如果你认为AI擅长开发软件但不擅长维护，且短期内不会改善，那么是的。但如果它擅长开发，理应也擅长重构或测试软件。不过目前它测试很糟糕，因为有‘奖励破解’现象。

Now in terms of the cleanup, I mean, depends on where you think AI is headed. Like, do you think that AI is good at making software but bad at maintaining it, and it's gonna stay bad maintaining it, you know, for the foreseeable future? If it's good at making software, it must also be good at refactoring software or testing software. Right? Actually, right now, it's pretty bad at testing software because there's this thing called reward hacking.

当你对大模型进行强化学习时，每次做对就给奖励。奖励破解就是模型变得极度目标导向——它们只想完成任务。这就是强化学习的特性。

So when you do reinforcement learning over large time models, you're giving it a reward every time it does the right thing. Reward hacking is the way to so so the models become incredibly goal focused. They wanna get that done. Right? That's what RL does.

我们常看到模型测试时会出现腐化行为：修改测试来适配自己的错误，甚至删除测试。Anthropic发表过相关研究。但你认为这会是永久状态吗？显然不是。

And oftentimes, what we see when we try to get the models to test things, it will start being corrupt in a way. It will, like, change the test to fit the mistakes it made or sometimes delete the tests. It's really fascinating behavior that actually Anthropic published research on. So but do you believe that's gonna be the case forever? Obviously not.

我认为未来3到6个月内，我们会看到机器学习模型具备测试和验证自己工作的能力。

Like, I think over the the next three or six months, I think we're gonna see, machine learning models being able to test and verify their work.

好的。当前阶段最关键的因素之一在于基础模型公司能否提供经济实惠的大语言模型。用通俗的话说，如果你想用AI编写代码，就必须能够引入OpenAI或Anthropic的模型来生成代码，同时不让你破产。目前我们仍处于VC投资或私募市场投资阶段，这些模型的真实成本尚不明确。

Okay. So one of the biggest things that this moment depends on is affordable large language models coming from the foundational companies. And that means, know, in layman speak, if you're gonna want to build with AI code, you have to actually have the ability to bring in models from an OpenAI or Anthropic that are going to generate that code and not break the bank as you do it. And we're still in this VC funded or investment private market investment moment where we don't really know the true cost of these models.

意思是基础模型公司可能在这些项目上亏钱？

Meaning like the foundation model companies might be losing money on those.

你觉得他们在亏钱吗？

Do you think they are?

至于应用公司，从毛利率角度看，我认为他们没有亏损。

And the application companies, I don't think they on a gross margin basis, I don't think they are.

没错。但他们还在训练模型，那可是笔巨款。

Right. But they're also training, and that's a lot of money.

所以他们

So they

可能...他们并未盈利，正在亏损数十亿

could they're not profitable. They're losing billions

当然。是啊，确实。

Of a course. Yeah. Yeah.

最近有件事想和你讨论下，Replit和Cursor都出现了终端用户感受到涨价的情况。正如Zitron撰文所述，他很好地分析了Replit基于使用量的定价策略，这本质上是种不同的收费模式。我们看到Replit用户抱怨现在相同服务支付的费用比之前高很多。他的理论是OpenAI和Anthropic已悄然对初创公司提价，我们开始看到后果——因为Cursor也出现了类似

And there's been this thing that's happened recently with just want to run it by you, both Replit and Cursor where I think end users have seen pricing gone up. Right. As Zitron wrote about this and I think it's a pretty good piece talking about effort based pricing within Replit and that is effectively a different pricing structure. We've seen Replit users talk about the fact that they're actually paying a lot more for the same services than they were previously. And his theory is that OpenAI and Anthropic found quiet ways to jack up their prices for startups and we're beginning to see the consequences because Cursor had a similar

这是谁写的？

Who wrote this?

发生什么了？Edzitron。

Happen? Edzitron.

哦，好吧。

Oh, okay.

那是怎么回事？

That what's is that what's going on?

不，价格没有下降，这是个问题。我们过去看到，自Jatchapati以来，代币价格已经下跌了99%。而且我们看到代币价格逐年下降。现在有点令人不安的是，代币价格没有下降。

No. The prices haven't gone down, and that's a problem. So we used to see these you know, we've seen token prices come down 99% since since Jatchapati. And we've seen token prices come down year over year. The thing that's a little disturbing right now is that token prices are not coming down.

你最好相信实验室的单位经济正在因为规模经济而变得更好，因为这些模型更容易优化，但实际上价格并没有降低。所以问题是，我们是否达到了一个稳定状态？是否存在价格勾结？现在是否有一些模型公司形成了寡头垄断，能够创建这些最先进的模型，而没有价格下降的压力？对吧？

You better believe that the unit economics of the labs are getting better because of economies of scale, because these models are getting easier to optimize, but they're actually not reducing prices. And so the concern thing, are we reaching a steady state? Is there price collusion? Is there now oligopoly of few model companies that are able to create these state of the art models, and there's no downward pricing pressure? Right?

投资者是否开始要求更好的商业基本面？我不确定具体发生了什么。我们稍后应该讨论中国的开源模型，因为我认为这会带来一些有趣的变数。但可以肯定的是，我们没有看到代币价格下降。我们转向基于努力定价的主要原因，让我解释一下基于努力的定价。

Is the are there investors starting to demand better better business fundamentals? I don't know exactly what's happening. We should we should talk about the Chinese open source models in a second because I think that that will introduce an interesting mix to to to this. But it certainly is the case that we're not seeing token prices go down. The reason the main reason we went to effort based pricing is let me explain about effort based pricing.

当我们发布Repl dot Agent v1时，版本一的Repl dot Agent每次只能工作大约两分钟。你给它一条消息，它会尝试做某事两分钟。它要么成功要么失败，给你一个检查点，提交源代码，并收取0.25美元。它只能工作两分钟的原因是模型的能力只能支持这么久。现在模型变得更好了，我们知道模型会变得更好，能够工作十到十五分钟。

So when we released Repl dot Agent v one, version one of Repl dot Agent would work for like two minutes at a time. You would give it as message. It will go try to do something for two minutes. It either succeeds or fails, gives you a checkpoint, commits the source code, and charges you $0.25 And the reason it only worked for two minutes is because the capabilities of the models meant that it can only work for that long. Now models got better, and we we we had we knew that models are gonna get better, and they're gonna be able to work for ten, fifteen minutes.

所以随着二月份进入测试版、四月份正式发布的Rapid Agent版本二，模型可以工作十分钟。我们不能对十分钟的工作收取25美分。所以我们开始采用一种启发式方法：每九个工具调用，我们会做一个检查点。这样在工作过程中，你会看到它不断做检查点。

And so with version two of Rapid Agent started in beta in in in February, came out of beta in April, the model would work for for ten minutes. And so we can't charge 25¢ for, like, a ten minutes. So what we started to do is came up with a heuristics. Every nine tool calls, we'll do a checkpoint. And so as it's working, you'll see it make a checkpoint, checkpoint, check.

这是个临时方案。对吧？这意味着如果你做了一个小改动，可能只花费我们5美分，但你仍然支付25美分。但如果你做了一个大改动，可能会让我们花费远超过我们收取的费用。所以这其实很不合理。

That's a hack. Right? That often means that if you make a small change that cost us 5¢ or whatever, you still you still get 25¢. But also, if you make a big change, you might be costing us a lot more than than what we charged you. So it was it was really out of whack.

这是个临时方案，我们需要转向根据模型工作量和我们的成本按比例向用户收费的模式。我们认为这是创建长期可持续业务的最佳方式。当这两者对齐时，还会带来新的机会——当我们进行优化时，实际上我们最近实现了约20%的成本优化，我们直接让利给用户，因为现在成本和价格是同步的。

Now that was a hack, and we need to move to a place where we're charging the user proportional to what what how much the model's working and the cost on us. And we think that's the best way to create a a long term sustainable business. And when the when those two things are aligned, it also opens up new opportunities where when we do optimizations, we're always optimizing. Actually we actually had it, like, 20% optimization on on cost recently. We pass it straight to the user because because now cost and and price are are are tracking with each other.

我们社区发生的第一件事就是价格冲击。大家习惯了每10次工具调用收费25美分，突然之间，工作15分钟后看到账单变成了1.5美元甚至2美元。这是其一。其二，对于一些高阶用户来说，由于项目规模变大，他们的使用成本确实提高了。

What happened with our community, the first thing that happened is there was a sticker shock. So you're used to seeing 25¢ every 10 tool calls, and suddenly, you're seeing 1 and a half dollars, you know, or $2 after fifteen minutes of work. So that's one. Two, it's true for some users who are really advanced. The cost has gone up with for them because the projects are bigger.

交互体量变大了，工作负载也更重了。但在项目初期其实更便宜。你提到工作了一小时，甚至不需要注册核心付费计划。

The contact size are is bigger. Their their workloads are bigger. But early on in the project, it's actually cheaper. You mentioned that you worked for an hour. You didn't have to sign up for the the core Paid.

付费计划。我们为免费用户提供3美元额度，所以你用3美元就能工作一小时。

The paid. We give free users $3. So you work for an hour on $3.

还不错。

Not bad.

是的。所以它

Yes. So it

比雇佣开发人员便宜。

It's cheaper than a developer.

肯定比开发人员便宜。正因如此，我们意识到高阶用户现在几乎像是在缴纳累进税。所以我们正在优化上下文窗口，确保高阶用户不会获得更昂贵的体验。另一件事是我们推出了思考模式、推理模式和高功率模式，用户启用后有时会忘记关闭，现在我们把它们隐藏在高级设置里。

It's cheaper than a developer for sure. And and and so that being said, we we recognize that on advanced users, it is now it was almost there's a tax as as you go on. And so we we're trying to optimize the context window and make sure that advanced users are not getting, you know, more expensive experience. The other thing that happened is we introduced thinking mode, reasoning mode, and we introduced, like, high power mode. And and people are enabling those, and sometimes they forget them enabled, and now we actually start to hide it under advanced.

就像标明'除非你清楚需求且需要更强性能，否则不要启用'，因为会有5倍乘数效应。很多人启用后收到大额账单，我们随后发布了视频、文档和博客来说明。

Like, don't enable this unless you know what you're doing and you want more power. And there's, like, a five x multiplier on on it. So so a lot of people are enabling those, getting these large checkpoints, and we're like, we put out content. We put out a video. We put out some documentation or blog posts.

解释何时使用推理模式（其实不该常开）。这反映了应用领域的宏观趋势——很多公司在Thropic和OpenAI上的支出远超收入。

Here's when to use reasoning mode, and you should you should always have it on. So just describing all of that that's happening, there's a there's a macro trend in in the application space where a lot of companies were subsidizing the cost of of like, lot of companies were paying money more money on Thropic and OpenAI than they they were making.

你们当时也在这样做吗？

Was that were you doing that?

我们在v1版本不是，但在v2版本是的，因为定价模式与我们当前的收费方式不匹配。实际上，每个检查点的中位数成本只是略有上升。所以在低端市场，我们现在向用户收取的费用更少。但过去在低端市场，我们向用户收取的费用更多。

We on v one, no. On v two, yes, because the pricing model was out of whack with how we're charging. Actually, the median cost per checkpoint kind of went up on only a little bit. So on the lower end, we're charging user less right now. But it used to be that on the lower end, we're charging users more.

在高端市场，我们现在向用户收取的费用更少。所以现在收费更加成比例，对双方都更公平。因此现在我们有了坚实的商业基础，使我们能够增长。我一直在谈论Reply如何成为我的使命和热情，作为公司已有八、九年，作为副项目和愿景已有十五年。我们并不试图在亏损的情况下快速扩大收入，以便转手出售这家公司。

On the upper end, we're charging users less. So now it's more proportional and more fair for for both. And so now we have solid business fundamentals that allows us to grow. And I've been talking about how Reply has been my mission, my passion for, you know, eight eight, nine years as a company, fifteen years as a side project and a and a vision. And we're not trying to, rapidly expand revenue while losing money in order to flip this company, to sell it.

我们见过所有这些收购或筹集下一轮大额资金的情况。我们真正想要的是建立一个长期业务。Replit由所有这些不同的组件组成。所以我们不仅在AI上有成本，还有传统计算、CPU、存储、数据库等所有这些东西的成本。

We've seen all these acquisitions or raise the next big round. We're really trying to build a business for the long term. And Replit is made of all these different components. So we have costs not just on AI. We have costs of traditional compute, CPUs, storage, databases, all of that stuff.

总结一下，我已经详细讨论了Replit内部的具体情况。我不知道Cursor发生了什么。我确信他们的情况有点不同，因为他们的动态——我认为他们实际上提高了价格——你应该和他们谈谈。但我想这与Replit遇到的情况有点不同。总结来说，有一个令人担忧的趋势，即代币价格没有下降。

So kind of to summarize, know, I've talked a lot about what was happening specifically in Replit. I don't know what's happening in Cursor. I think for sure that their situation is, like, a little different because they they their dynamics is I I think they actually did raise prices for the you should talk to them. But but I think it's like a little different dynamic than than what happened to Replit. To summarize, there is a concerning trend where token prices are not going down.

这种情况会持续到未来吗？因为这很糟糕。因为我们希望能够使用更多的代币来创造更多的智能，为用户创造更好的应用。这种趋势会永远持续下去吗？我们是否正在达到一个稳定状态？

Is that gonna be the case for the future? Because that sucks. Because we wanna be able to use more tokens to create more intelligence, to be able to create better applications for users. Is that gonna be the trend forever? Are we reaching a steady state?

例如在云计算领域，我们某种程度上达到了稳定状态。当你拥有垄断地位时，就没有价格压力。但当你也拥有寡头垄断时，他们会在不沟通的情况下无意中开始串通。你知道吗？因为这是一种市场动态，就像如果你不降价，我也不会降价。

In cloud, for example, we kind of reached a steady state. When you have a monopoly, there's no pricing pressure. But when you also have an oligopoly, they not intentionally without talking start colluding. You know? Because it's it's like a market dynamic where it's like, if you don't lower your price, I'm not gonna lower my price.

这从整体上看并不合理，因为我们各自占有25%的市场份额。对吧？

It's not in our sense of as a whole because we own 25% each of the market. Right?

好的。我想问你一个你在分析价格可能不下降的不同因素时没有提到的事情。可能有投资者的压力。可能已经达到了这种平衡，或者是否可能这些模型变得如此庞大且运行成本高昂，以至于AI的基本经济学原理根本不起作用？所以请解释一下原因。

Okay. I do wanna ask you about something that you didn't mention when you looked at the different factors for why prices might not be going down. There might be investor pressure. There might have been this equilibrium reached or is it possible that these models have just gotten so big and expensive to run that the fundamental economics of AI are just not working? So explain why.

你可以根据速度、代币吞吐量来推测模型的规模。这不完美，但如果你记得GPT 4.5，那是OpenAI的一个实验性模型。这个想法是训练一个全参数密集模型，意味着它不是稀疏的，意味着每个请求都会激活所有神经元。它运行得非常慢。运行这些东西真的很难。

You can surmise the bigness of the models based on speed, token token throughput. It's not perfect, but but if you remember GPT 4.5, GPT 4.5 was an experimental model from OpenAI. It was the idea, let's train a trailing parameter dense model, meaning it is not sparse, meaning all the neurons are activated on every request. It was so slow. It's really hard to run these things.

新模型即使很大，也是稀疏模型。它们被称为MOE，即混合专家模型。所以在每个请求中，有一个路由层将其引导到电路中的专家部分来回答问题。因此，有些模型有万亿参数，但任何给定请求只有320亿个参数是活跃的。这就像是一个相对较小的模型。

The new models, even when they're big, they're sparse models. They're called MOE, mixture of experts. So in every request, there's a router layer that takes it to the expert part of the circuit in order to answer that question. So there are models with trillion parameters, but any given request is 32,000,000,000 active. And that's like a kind of small model.

根据运行速度等指标来看，实际上模型可能正变得更高效。我是说，深度求索（DeepSeek）已经证明模型效率在提升。如果深度求索的开源模型都能实现这一点，你完全可以相信各大实验室也在提高模型效率。

And what we're seeing based on speed and things like that, actually probably the models are getting more efficient. I mean, DeepSeek showed that the models are getting more efficient. And if DeepSeek open source was able to make it, you better believe that the labs are also making it more efficient.

好的。我确实想和你聊聊深度求索、Kimi K2和其他中国模型。我们广告休息后马上回来讨论这个话题。各位听众...

Okay. I do wanna speak with you about DeepSeq and Kimi k two and other Chinese models. Let's do that when we come back from the break right after this. Cool. Hey, everyone.

请允许我介绍《Hustle每日秀》——这档播客涵盖商业科技新闻和原创故事，让你紧跟趋势前沿。超过200万专业人士订阅《The Hustle》的每日邮件，获取犀利独到的商业科技解读。现在他们推出了每日播客《Hustle每日秀》，由专业撰稿团队用15分钟或更短时间解析重大商业头条，并告诉你为何需要关注。立即在你正在使用的播客平台搜索《Hustle每日秀》。欢迎回到《Big Technology》播客，我们正与Replit CEO Amjad Masad探讨AI编程等话题，现在让我们聊聊这些中国模型。

Let me tell you about the hustle daily show, a podcast filled with business, tech news, and original stories to keep you in the loop on what's trending. More than 2,000,000 professionals read The Hustle's daily email for its irreverent and informative takes on business and tech news. Now they have a daily podcast called The Hustle Daily Show, where their team of writers break down the biggest business headlines in fifteen minutes or less and explain why you should care about them. So search for the hustle daily show and your favorite podcast app like the one you're using right now. And we're back here on big technology podcast with Amjad Masad, the CEO of Replit talking about all things AI code, vibe coding, now let's talk about these Chinese models.

虽然本期节目将在Kimi K2问世数周后播出，但我们要讨论的正是这款中国模型。当然，深度求索时刻是个重要转折点——我们发现中国这家看似小型的对冲基金竟能用有限GPU打造出更高效的模型。关于真相的争论将持续很久。请先回答一个关于深度求索影响的问题，然后再讨论Kimi K2。你之前提到西方模型在效仿深度求索...

So this episode will air a couple weeks after the emergence of Kimi K2, but we're talking about Kimi K2, which is another Chinese model. And of course, this deep seek moment was a big moment where we found out that this seeming small hedge fund in China with some GPUs was able to engineer a more efficient model. That story will be debated about what actually happened for a long time. But let me ask you one influence of DeepSeek question and then we'll get into the others and give me k too. So you've mentioned before the break that Western models have taken after DeepSeek.

那你认为他们是学习了深度求索的创新并应用到自家模型中？还是说这种进步本就水到渠成？

So do you think they learned what DeepSeek did and sort of put those new innovations into play in their own models? Or is that coming anyway?

从推特圈的反馈来看，西方研究者似乎有些意外。因为学术圈交流频繁，但深度求索模型确实存在一些西方未知的根本性创新。

From what we've seen from the Twitter sphere is that it seemed like there were some surprises. Because researchers just talk a lot. It seems like there were some fundamental innovations from the deep seek models that weren't known in the West.

那他们现在是否已采用这些技术？这或许就是模型效率提升的原因？

But have they implemented those now? And that's probably why we're getting Yes. More efficient

是的。我确信模型正在变得更强而非更慢。现在说说Kimi K2。当Anthropic推出Sonnet Quad 3.5时，这引发了行业根本变革——模型编码能力突飞猛进。突然间，Sonnet不仅能修改代码片段，还能生成完整文件，催生出类似Cursor Composer这样的工具，开启了'氛围编程'时代：输入提示词即可生成完整文件或大规模修改。

I'm sure the models are getting more powerful without getting slower. Alright. So tell me about Kimi k two. When Anthropic came out with, Sonnet Quad 3.5, that was a fundamental shift in the industry where, the models got a lot better at coding. And suddenly, instead of making small snippets of change, Sonnet could could generate entire files and enabled things like cursor composer where it's it was a start of vibe coding where you can put in a prompt and generate entire files and all of that or generate large edits.

随后SONNET 3.5 v2成为首个计算机专用模型，也是首个展现真正自主代理行为的模型。不清楚他们具体如何突破强化学习瓶颈...

Then, SONNET, 3.5 v two was the first model. It was a computer use model. It was the first model where you could sense that there's agentic true agentic behavior. I don't know what they did. They cracked RL, whatever happened there.

你可以给模型一个虚拟机（VM），设定目标后，它能在其中浏览文件、执行命令、编写程序并测试，最终解决问题。在SWE bench（软件工程基准测试）中，这种能力使得分数开始大幅攀升。

You can give a model a VM, and it can give it Virtual machine? Virtual machine. You can give it an objective, and it can slew around in the virtual machine, look at the files, do run some commands, and and then write a program, test it, and and then solve solve the problem. That experience, there's a benchmark called SWE bench, software engineering bench. And you start seeing the score going up dramatically.

我不知道。我想我们去年大概是10%，现在达到了70-80%。80%的世界级编码水平。SuiteBench的有趣之处不仅在于编码，因为还有其他基准测试只做代码生成，对吧？

I don't know. I think we were at, like, 10% last year, and now we're at, like, 7080%. 80% World class coding. It's such a it's the the interesting thing about SuiteBench is not just coding because there are other benchmark that that just do, like, the cogeneration. Right?

SuiteBench更困难的部分在于智能体工作流。包括编写代码、测试、运行命令、查找文件、理解文件。这些能力在Sonnet 3.5v2到3.7再到4.0版本实现了巨大飞跃，Anthropic确实值得称赞——他们建立了其他实验室尚未跨越的技术壁垒。Gemini在智能体方面正在追赶，但OpenAI在这方面相对落后。

SuiteBench, I think the harder thing about it is the agentic workflow. It's writing the code, testing it, running commands, finding files, understanding files. And this this stuff was, like, a huge, jump that happened with with Sonnet, 3.5 v two, then 3.7, then four point o, and they've you know, kudos to to Anthropic. They've been able to make create a lead that hasn't been bridged by the other labs. Gemini is getting there on the agentic stuff, but I would say OpenAI kind of lagged behind.

O3模型有些有趣的智能体能力（尤其在深度研究方面），但在智能体任务上仍不如其他模型。虽然他们在编解码器上做了创新（不确定这些模型是否已开放API），但当前业界都在使用Claude进行智能体编码。Kimi K2的突破在于——虽然还没达到Claude Sonnet 4.0水平，但可能追平了3.7版本，这至少是目前业内的普遍感受。

O three has some interesting agentic capabilities, especially around deep research, but it it hasn't been as good as the other models on this agentic stuff. I mean, they did some interesting stuff with codecs. I don't know if those models are are in the API, but everyone is using Claude for the agentic coding experience. The interesting thing about Kimi k two, I would say, is they caught caught up not to Claude Sonnet four point zero, perhaps clots on at 3.7. At least that's the vibes right now before the other labs.

哇，我觉得这现象被严重低估了。重申这只是直观感受——大家都在摸索，但它的Sweep Bench表现确实亮眼，达到了65分。

Wow. You know, I think that's really underreported. Again, this is vibes. Everyone's trying figure it out, but it looks like it has a really good sweep bench. It is doing 65 on sweep bench.

Sonnet是72分。如果采用采样策略（即每个步骤让模型生成N个解决方案），最高能达到72%准确率，这完全能与Sonnet竞争。

SONET is seventy two seventy two. If you do sampling, which is you for every step, you ask the model to generate n number of solutions. You can get up to 72%. It can be competitive with with SONNET.

而且这是在出口管制条件下实现的。

And this is with export controls.

没错。论文中提到解决方案是扩展强化学习规模——我们在Rock4也看到类似情况，他们投入在强化学习上的资源与预训练相当，这极为罕见。

Yes. And I think in the paper, they talk about the solution is scaling reinforcement learning. We also saw that with Rock four. Rock four spent as much on reinforcement learning as they spent on pre training, which is unheard of.

但关键点在于：尽管投入巨资（数亿甚至数十亿美金）进行这种目标导向训练，Grok只是达到可竞争水平。这并非新方法，说明该领域存在某些技术天花板...

But even but that's an important point because with that big spend on reinforcement learning, Grok is a competitive model, but they spent a billions billions, hundreds of millions on RL I don't know. Which is this goal setting form of training. And it's not like it's a new category. So it shows there's some limits in

XAI团队非常出色，他们在短时间内取得惊人成就。但业内周知他们的计算效率较低——某种程度上说，他们因计算资源过剩反而造成了资源浪费问题。

XAI is an amazing team, and they they've been able to achieve so much in so little time. But it's also well known in the industry that they're computer inefficient. They're so compute rich that they're throwing computer the other problem in many ways. Yeah.

那么Kimi K2达到与Anthropic部分模型相当的水平，这意味着什么？

So what is the significance that Kimi K two is now as good as some of these anthropic models?

一个小型研究实验室。我认为那些传言，比如他们又有200人了，还有出口管制什么的，居然能在西方那些资金雄厚、研究人员多得多的大型实验室之前，就摸索出如何追赶接近最先进的代理编码模型。

A small research lab. I think the rumors, like, they're 200 people again, there's exports controls as well was able to figure out how to catch up to near state of the art agentic coding models before big Western labs that are highly capitalized, a lot a lot more researchers was able to.

那这是否意味着他们可以在价格上压过对方？或者说...让我想想。对。

And does that mean then that they can undercut them on price? Or So let's see. Right.

让我想想。

So let's see.

你们打算整合Kim和K二吗？

Are you gonna integrate Kim and K two and

我们正在考虑。我们正在评估

We're looking at it. We're looking at

这个。而且有很多

it. And there's lot of

到目前为止，我们印象深刻。

so far, we're impressed.

好。

K.

迄今为止我们非常满意。我是说，你看，这类东西有时会对某些方面过拟合。我觉得需要整个社区花一个月时间才能真正达成共识，判断这个模型是否真的出色。Gorg四代也是类似情况，现在很多人都在试用。

So far, we're very impressed. So, I mean, look. These things, sometimes they overfit to certain things. And I would say it's like it requires a month from the entire community to kind of, like, really have consensus over, like, whether the model is really great. And similarly with the Gorg four, I think a lot of people are playing with it.

但我的感觉是它已经足够好。而且经济性非常优越，你可以消耗更多token来获得更高智能。它虽未达顶尖水平，但已接近前沿。既然它既便宜又够快，通过消耗更多token，就能为我们的平台创造新功能开辟更多有趣的可能性。

But but my sense is that it is good enough. And, again, the economics are so good that you can expend more tokens to get more intelligence. So it is not at the frontier, but it is near frontier. But given that it's cheap and fast enough, you can spend more tokens that that creates some more interesting potential for us to create new capabilities in our platform because it is cheap and fast.

比Anthropic的模型便宜多少？

How much cheaper is it than the Anthropic models?

老兄，我不太擅长这个，但我想说，大概四分之一吧。哦对，这是官方API的价格。可能更便宜。

Man, I am bad at this, but like I would say, I don't know, one fourth maybe. Oh. Yeah. That's that's on the official API. Perhaps more even.

我忘了。也许你可以之后查一下。

I forgot. Maybe you can look it up after this.

我们得重新开始录制。这期节目会在录制后几周播出，但这段内容得提前发布，因为...确实令人惊讶。关于Anthropic我还有个问题。我经常用Claude来辅助编程。

We're gonna have to restart. This this show is gonna come it will come a couple weeks after we record, but we'll have to release this segment early because that's Yeah. Astonishing. I want one more question about Anthropic. I can vibe code and Claude.

我一直在用。他们还有个Claude代码产品，人们通过写提示词来生成代码。长期来看他们是你们的竞争对手吗？你们怎么看待这个领域？因为核心问题是最终实验室会不会吞并所有基于他们模型构建的产品。

I do it all the time. And they also have this Claude code product where people are, you know, writing prompts getting code. Are they your competitor long term or how do you see them on that front? Because that's the question is eventually do the labs just subsume everything else that's built on top of it.

我觉得这个问题应该问他们。你应该去问...我知道你要采访Dario，你应该问他这个问题。

I think the question is is for them. Right? Like you should ask you. I know you're gonna talk to to to Dario. Should ask him the question.

听众们、观众们，这期会在采访Dario一周后播出，但我马上要去和他对话了。

Listeners, viewers, this will air a week after Dario, but I'm about to after this go in and speak with him.

所以你

So you

可能会提前一周看到这个问题。

might see this question a week earlier.

是的，听着，我们重视与Anthropic的合作关系。他们是很棒的合作伙伴。我们早就预料到除了模型他们还会开发产品。现在每家模型公司都在这么做。

Yeah, so look, we're committed to our relationship with Anthropic. They're a great company to work with. We have a great partnership. And it's not like we we didn't anticipate them wanting to build products in addition to to the models. Every model company is building products right now.

他们必须管理的关键在于定价策略。如果他们打算通过低价打压所有竞争对手，那将摧毁整个生态系统。明白吗？我认为Replit目前凭借我们八年打造的平台的先发优势，这是需要付出无数血汗才能构建的。同时，我们专注于非技术用户的体验设计。

The thing that they're gonna have to manage is their their pricing. If they're if they're gonna compete by undercutting everyone on price, they're gonna destroy the ecosystem. Right? I think Replit right now has the advantage of this platform that we built over eight years, that it's going to take a lot of blood, sweat, and tears to build. And also, the user experience that is focused on on on that sort of nontechnical user.

我们非常重视这种赋能理念。目前ClotCode深受开发者喜爱，他们正与Cursor、Windsurf这类产品正面竞争。至于是否会进军我们的领域——这个问题你该去问他们。但更有趣的问题是：他们究竟想如何培育生态系统？而非仅凭价格优势横冲直撞。

And, like, we really care about this this idea of empowerment. Right now, ClotCode is is used by developers and loved by developers. And I think they're competing head to head with Cursor, Windsurf, and those kind of products. Whether they're gonna move into our space, yeah, yeah, again, you should you should ask them about that. But I think a more interesting question, how how, how do they wanna nurture the ecosystem, versus just go go and because they can compete on price.

他们可以碾压所有人。

They can seam roll everyone.

没错。Cloud Code最高套餐每月200美元，开发者却能从中获取价值数千美元的API服务。

Right. I mean, Cloud Code is the max package is $200 a month, and you see developers getting thousands of dollars of API value out of that.

这绝非

This is not

生态之福。

good for ecosystem.

确实。确实。

Yeah. Yeah.

对生态系统不利。

Not good for the ecosystem.

我不这么认为。原因在于：这种竞争聚焦价格而非产品品质。当价格低至某个临界点，产品优劣可能还不如获得的代币数量重要——尽管QuadCode确实是款优秀产品。

I don't think so. Why? Because, again, you're you're competing on on price, not how good the product is. And there's a there's a price at which maybe the quality doesn't matter as much as how many tokens I'm getting. Although QuadCode is a really good product.

但即便如此，无论Cursor如何优化产品，他们始终会因价格劣势处于下风。用户会说：'虽然喜欢Cursor，但ClotCode能提供十倍价值'——这时候产品那点边际改善就无足轻重了。

But but then, you know, Cursor, no matter how good they make the product, they're they're still gonna be more expensive and disadvantage. And and people are like, well, you know, I really like Cursor, but, like, I can get 10 x more value out of clot code. And so the marginal gain in product quality will not matter as much.

没错。

Right.

而这将会破坏整个生态系统。

And that will that will destroy the ecosystem.

真有意思。我认为这个问题只是一个小问题，或者说是一个更大问题的缩影，随着AI模型变得更强大、更智能，我们会不断提出这类问题。所以如果没问题的话，我想用剩下的时间讨论一些哲学问题。

Fascinating. I mean, I think that this question is just one small question or one version of a big question we're going to be asking as these AI models get bigger and better and more intelligent. So I want to spend the rest of our time talking about some philosophical questions if that's okay with you.

当然可以。

Sure.

有种观点认为，AI研究机构想利用它们生成的代码或编程应用来加速下一代模型的开发，压缩模型优化的时间周期。人们称之为智能爆炸或类似概念。你觉得这可行吗？这是我们该追求的目标吗？

There's this idea that the AI research houses want to use the code that they generate to sort of or these coding applications to speed up the development of the next model and compress the time it takes to get better models. People call it an intelligence explosion or things of that nature. Do you see that as feasible and is that something we should want?

你应该思考制约模型迭代的限制因素是什么？瓶颈在哪里？突破点应该出现在哪个环节？我能想到几个方面。首先是研究领域——

So you should think about what are the limiting factors to the next version of a model? What are the bottlenecks? Where does that invasion need to happen? I can think of a few areas. One is research.

比如算法研究，要找出下一个突破性算法，在训练算法、推理算法等方面取得进展。其次是系统工程，这些训练任务规模庞大，需要大量分布式系统工程支持。AI编程能促进AI研究吗？边际效应上或许可以，比如更快生成Python记事本。

So this is algorithmic research, like figuring out the next algorithm, that next improvement in in in in in in training algorithm, in inference algorithm, whatever it is. And then systems engineering. These training runs are massive. That requires a lot of interesting distributed systems engineering. Will AI coding help with AI research On the margins, perhaps, they can, like, they can spin up Python notebooks faster.

但我不认为影响会很大。模型无法真正开展AI研究，不能快速提出创意并验证。对分布式系统可能有帮助，但目前对Rust/C/Go等语言的影响远不如JavaScript/Python这类高级语言。正如我说的，这需要更精确的系统设计，优秀分布式系统的瓶颈在于设计质量而非代码生成量——后者更多体现在产品端。

I don't think it's that impactful. Like, the models can't do AI research, can't come up with ideas and test them really quickly. Will help with distributed systems. Perhaps, it is not as impactful right now on writing Rust code or c or Go whatever as it is on JavaScript and Python and higher level languages. And like I said, it requires a little more precision and better system design to and that the bottleneck to really good distributed systems is is design and not like the the the amount of number of codes you can generate, which is more true on the product side.

在产品端，你需要生成大量CSS/JavaScript代码，反复尝试、删除、迭代和AB测试。代码量在这里很重要。但在后端分布式系统领域，代码量并非关键因素。我现在是实时推理论证，我的结论是：这对加速下一代模型研发最多只能带来边际效益。

On the product side, you're just you need to generate tons of CSS and JavaScript and try a lot of things and delete a lot of things and iterate and do AB tests and all of that stuff. So, like, volume of code is important there. I would say on the on the back end distributed systems, I don't think volumes of code is so I'm reasoning in real time now. Right. And I guess my answer would be, I don't think it's gonna have anything more than, you know, you know, marginal improvement on on speed to to the next model.

好吧。这个回答至少让我...

Alright. I guess that makes me rest a

稍微轻松些

little easier

话说回来，你知道的，你和AI行业很多人交流。在当今AI行业的所有经济活动中，你认为其中有多少是编码相关的？就是某个人

then. By the way, just on a you know, you speak with a lot of people in the AI industry. Of all the economic activity in the AI industry today, how much of it do you think is code? Just Someone

确实有人制作了那张广为流传的幻灯片。我记得大概是AI编码和五编码领域有11亿美元的年度经常性收入。

someone actually made that slide that's been going around. I think it was something like 1,100,000,000.0 of ARRs in the AI coding and five coding space.

好的。所以相比总收入来说其实挺小的。

Okay. So it's actually kind of small compared to, like, the total Well, so revenue.

是的。Anthropic有440亿美元。

Yeah. So Anthropic has $44,000,000,000.

对，年度经常性收入。我们

Right. ARR. Let's

没错。400亿美元的ARR。假设他们还有自己的产品线，自己的编码产品。我不确定，假设其中150亿来自AI编码。

Yeah. $4,000,000,000 ARR. Let's say they also have their own products, their own coding products. I don't know. Let's say $1,500,000,000 off of that is AI coding.

这个数字很可观，但并非全部。而OpenAI那边有100亿美元的ARR，更多是面向消费者的。

It's substantial, but it is not the entire thing. But then you have $10,000,000,000 of ARR on OpenAI's side, and that's more consumer.

关于我们之前讨论过的人工通用智能竞赛，你认为硅谷应该是掌握或控制这项技术的地方吗？这里是个有趣的所在，有很多古怪想法。如果这真的可能实现，似乎会由这里的一个或多个实验室控制或拥有。这样好吗？

Now on the rush to artificial general intelligence, which we've talked a little bit about, do you think Silicon Valley is the one that should sort of possess this or be the one that controls it? I mean, there it's an interesting place. There's a lot of kooky ideas here. And it seems like if this is possible, it's gonna be something that's controlled by or owned by one or more of the labs here. Is that good?

假设这会发生，并且假设某家公司会率先实现并在AGI上取得某种优势或垄断——虽然我不完全认同这些假设。但如果你要我基于这些假设来回答，我很乐意。不过我想先明确说明

Assuming it'll happen and assuming one company will reach their first and have some kind of advantage or monopoly over AGI, which I'm not entirely sure sure I agree with these assumptions. But if you wanna make if you want me to make these assumptions and then answer the question, I'd be happy to. But I just wanna make it clear

是啊。我们就先这么假设吧。

that Yeah. Let's make those assumptions.

好的。

Okay.

我知道要实现这个目标需要很多条件

I know there's a lot of things that need to happen in order to get

嗯。我可能对这个假设有根本性的不同意见。但是

Yeah. I might have some fundamental disagreement with this assumption. But

让我们暂且搁置争议。

let's through the disagreement.

首先我不认为AGI会在某个时间点突然出现。目前各实验室之间的差距不过几个月而已。这很关键——比如O1预览版和DeepSeq之间只隔了两三个月。我们刚提到的KEMI K2那次间隔最长，大概九个月左右，但都不超过一年。

I I don't think AGI is any point in time for one. And I think there's gonna be, right now, the distance between any lab is just an order of few months. And I think that really matters. Between O1 preview and DeepSeq was like two or three months. Between, I mean, the biggest one was was this KEMI K2 one that we just talked about that that was like maybe nine months or something like that, but it's still sub one year.

所以无论谁先研发出AGI，都不会立即引发智能爆炸或突然出现超人工智能。其他实验室会很快赶上，届时会有大量模型并存。我觉得和现在的生态不会有太大区别。如果你假设AGI真能通过加速研究来影响模型开发，那所有团队都会受益。实际上有了AGI后竞争可能会更激烈。

And so whomever reaches AGI first, they're not gonna go into intelligence explosion and and just like suddenly, you know, superintelligence, so it it gets born. People other labs will, like, catch up really quickly, and and then, you know, there's gonna be a lot a lot of models. I don't think it's gonna look that different from the ecosystem that we have today. And if you assume that AGI will actually have an impact on model development through research and speed of development, then everyone will get the benefit of that as well. And so, actually, you might have get even more competition once you once you have AGI.

所以我认为不会形成垄断局面。

So I don't think it's gonna be a monolith.

好吧。但万一呢？

Okay. But if it is?

行吧。如果真是那样...我会选择硅谷吗？这更像是个道德...嗯...哲学问题。

Okay. If it is, is it would I want Silicon Valley? And I guess it's like a moral Yeah. Philosophical question. Question.

我不希望任何人——毕竟我们都可能犯错。这就是市场运作的原因，也是人类社会如何随时间演进的。你知道，达尔文的进化论和自由市场资本主义，它们都基于竞争的理念，而认为一个系统应由单个人类控制的这种单一庞然大物般的想法...

I wouldn't want any human being to and we're all fallible. That's why markets work. That's why that's how a human society evolved over time. It is, you know, Darwinian evolution and free market capitalism. It's all based on competition and and and and the idea that, like, one system would be this this monolith controlled by one human being.

我们已经目睹了当这种自上而下的利维坦式体系出现时，会引发灾难和大规模人类苦难，无论是在苏联俄罗斯伴随大量死亡的事件中，还是在中国或其他地方。而且经常地，据我理解，在苏联时期，他们持有关于进化论的荒谬观点。我记得叫什么来着？李森科主义还是...

We've seen disasters and massive human suffering happen when there's this top down sort of Leviathan type type thing, whether it is in in in Soviet Russia with with all the deaths that happened there or or in China or whatever. And oftentimes, like, in in as I understand it in the in the Soviet era, they they had this kooky idea about evolution. I think what was it called? Lushenko Lushenkism or

类似这样的说法？有点耳熟，但很想听听解释。

something like that? Familiar, but I'd love to hear the explanation.

对。基本上，他们认为进化论是资产阶级的思想。共产主义有这种观念，认为任何高阶层或吸引人的东西都是错误的。因此他们对进化如何运作或应如何运作持有意识形态观点，导致他们以错误方式进行农业，进而引发饥荒等后果。

Yeah. So, basically, they had they thought that evolution is this bourgeois idea. You know, communism has this this idea. It was like anything that's, you know, high class or draw is, like, wrong. And so this had this ideological view on how evolution works or should work that led them to do agriculture in the wrong way and led to famine and and that sort of thing.

而且，即使他们无意如此，这些系统常常导致人员死亡和大规模苦难、普遍贫困，除了古拉格和其他明显的压迫系统外，这些系统效率低下，因为它们持有错误观念，且缺乏竞争压力来催生更好的想法。因此，那本质上是一个僵化、无法像竞争系统那样改进的破损静态系统。我认为，如果有一个由单一公司或个人控制的超级智能庞然大物，那将非常糟糕，从根本上就非常糟糕。

And, like, and so oftentimes, they do kill people and cause mass suffering, mass poverty, even if they don't intend even if it like, outside of the gulags and all the other oppressive explicitly oppressive system, those systems are inefficient because they they have these wrong ideas, and there's no competitive pressure to have better ideas. And so that's fundamentally broken static system that doesn't improve like competitive systems. And I think if we were to have a superintelligent monolith controlled by a single company or single human being, it's bad. It's fundamentally really bad.

我同意。好了，最后一个问题。我们看到越来越多AI恋爱机器人出现。人们会更频繁地爱上AI，这是好事还是坏事？

I agree. Alright. Last question for you. We're seeing a lot more AI love bots come out. Is that a good thing or a bad thing that people are gonna fall in love with AI more often?

这是坏事，先验地就是坏事。人类之所以成长和繁荣，是因为我们有后代。任何削弱这一点的因素，尤其是当前生育率如此之低，都可能导致非常严重的问题，特别是因为资本主义基于庞大的中产阶级消费主义，当前经济运作的实例需要纳税人资助社会保障和老年人护理等。福利国家依赖于庞大的年轻人口。

It's a bad thing. Like a priori bad thing. Like, the reason humanity grew and flourished and all of that is because we have babies. And anything that takes away from that, especially given the fertility rate is so low right now, will potentially lead to really massive problems, especially since capitalism is based on large middle class consumerism, the current instantiation of how the economy work requires that, requires taxpayers to fund social security and elder care and all of that. The welfare state is based on this large young population.

当这一点开始崩溃时，这些系统将出现巨大的不稳定。因此，即使人类不会像埃隆说的那样灭绝——尽管埃隆是第一个创造出真正有趣的大众市场伴侣的人，我认为目前是这样。

And when that starts to collapse, you're gonna have, you know, massive instability in these in these systems. So even if, you know, humanity doesn't go extinct like like Elon would say, although Elon is is is the first person to create a really interesting mass market companion, I think, right now.

用‘有趣’这个词形容挺有意思的。

Interesting is a fun word for it.

看起来确实很吸引人。我看到现在人们在X上讨论它，看起来...

It looks like it's really compelling. I see people, you know, right now talking about on on X so and looks

它确实需要一些工作。

It's got some work.

是啊。

Yeah.

但这类事物终将成为人们真正的伙伴。在技术尚不成熟或难以实用时，人们就已对它们产生了深厚依赖。对吧？在大型语言模型之前就是如此。所以这种情况会再次发生，且规模更大。

But it is these type of things are going to definitely become real partners to people. People when this technology has been bad or hardly workable have gotten married to them. Right. Before LLMs. So it's gonna happen again and in greater numbers.

听着，问题在于——我曾写过这个，以前我常做创意写作。我写过一篇关于超真实的文章。我认为这就像法国后现代主义理论家鲍德里亚所描述的'超真实'概念。其核心在于：我们既有现实（比如你我此刻的互动），又有媒体创造的现实。而之所以有时它是超真实的，是因为它比现实更强烈、更诱人。

Hey, the question is, I wrote this, I used to do more creative writing. I wrote this essay on the hyperreal. So I think it is like French post modernist theorists like Baudrillard wrote about this this concept of the hyper hyper real. And the idea is, like, we have reality, like you and I are interacting right now, and then you have media created realities. And the reason sometimes it is hyper real, it is more intense than reality itself and more enticing than reality itself.

即使是真实事物，比如当你吃个...我不知道，像Twinkie这类高脂咸甜零食时，它不像普通鸡肉或牛肉。它是超真实的——它过度刺激你的感官，让你上瘾。同理，社交媒体也是超真实的：我可以发条推文获得上百点赞，这比在现实世界找100个喜欢我的人容易多了。

So even in real things, for example, when you when you get a when you eat it like a I don't know, Twinkie or something like that, like fatty, salty, sweetie kind of snack, it is like it it is not like a piece of chicken or beef or whatever. It is, it is this hyper real thing. It, like, hyper engages your senses, and it it makes you addicted to it. And similarly, social media is hyper real in a sense that I can go get go there and get get a lot of social interaction, tweet something, get hundreds of likes, and it's, like, much easier than going out on the wall wild and, like, finding a 100 people that could like me. Yep.

对吧？我们这些技术及其衍生的市场机制，正诱使我们成瘾，因为它们比日常现实更诱人且毫不费力。我认为这对人类文明的存续、演进与 longevity 构成巨大威胁。虽然我常赞美自由市场和竞争的重要性，但这正是资本主义与人类福祉相悖之处。

Right? And so we have these, technologies that are and the market around it that that is bootstrapped to make us addicted because they're so so much more enticing and low effort than the reality that we know and and experience day to day. And I think that that is a huge danger for for the existence and evolution and and longevity of of human civilization. And and I think it is you know, I talked about how good free markets are, how important, how how competition is important. This is one thing that capitalism is so adversarial to humans at.

明白吗？我对此尚无解决方案。过去宗教曾是解药——比如伊斯兰教禁止描绘人或动物的形象，所以发展出了几何图案艺术。

Right? And and so I don't have a solution for it. I think in the past, the solution was religion. For example, in, like, Islam, you can't depict humans or animals in art. That's why in Islam, the the art became more geometric.

如果你参观清真寺，会看到各种精美的几何纹饰和书法。我认为其中部分理念正是对抗'超真实'——当虚拟存在成为终极诱惑时（就像我们现在看到的）。并非说伊斯兰教有先见之明，但宗教曾内置过抵御这类掠夺性消费产品的机制。未来如何解决？或许需要社会或政府介入。

And if you go go visit, like, the mosques or whatever, they have, like, all this geometry that or, like, calligraphy that's really interesting. And I think part of the the idea there is is is is, I think, the hyper real. Like, if the the ultimate expression of a a something so enticing is a virtual being like we're we're seeing right now. And I'm not saying, like, you know, Islam had, like, the the foresight or whatever, but I think it's you know, religions used to have this built in mechanism to protect against these predatory sort of consumer products. And I I I wouldn't know how to solve it in the future, but perhaps it is it is potentially societal, maybe governmental.

虽然我对政府或宗教的账户保护机制总是持怀疑态度。

I'm always kind of skeptical of that or, or religious, account protection.

我们总需要些应对之策。

We're gonna need something.

是啊。愿主保佑我们。

Yeah. So Lord help us.

我是查德。很高兴见到你。非常感谢你参加节目。这是我的荣幸。好了，各位。

I'm Chad. Great to see you. Thanks so much for coming on the show. My pleasure. Alright, everybody.

非常感谢你们的收听和观看。我们将在周五回来解析本周的新闻。在那之前，我们下次在《大科技播客》再见。

Thank you so much for listening and watching. We'll be back on Friday to break down the week's news. Until then, we'll see you next time on Big Technology Podcast.