Modal CEO Erik：千亿美元推理市场机遇，开源大模型推理尚未见顶

我认为有机会打造一家千亿美元市值的公司，专注于大量专用模型的推理。会有回调，我不介意，因为我认为长期机会太大了，真的非常大。今天，我非常高兴欢迎 Modal 的联合创始人兼 CEO Erik。Modal 的年度经常性收入刚刚突破 10 亿美元，他们刚完成了由 Excel 领投的 7.5 亿美元融资，今天估值 157.5 亿美元。但今天我们深入探讨推理市场、未来走向、所有人都该看到却没人看到的东西，以及更多内容。

I think there is an opportunity to create a hundred billion dollar company with a lot of specialized models in inference. There's going to be a pullback. I'm okay with that because I think the long-term opportunity is so big. It's like so large. Today, I'm thrilled to welcome Erik, co-founder and CEO of Modal. Now, Modal just hit a billion in ARR. They just raised $750 million led by Excel and they're priced at $15.75 billion today. But today we dig into the inference market, what's to come, what no one sees that everyone should see, and so much more.

我们失去了去一些实验室的新毕业生，他们拿到了 2000 多万美元。我是说，这些是杰出的人，非常杰出。我不想卖掉这家公司，因为我会直接去创办 Modal 2。

We've lost new grads going to some of the labs for like 20 plus million. I mean, these are exceptional people. Exceptional. I don't want to sell this company because I would just go and start Modal 2.

准备好了。Erik，听着，兄弟。你能来节目真是太好了。你知道，我之前在楼下和你聊天，我当时就觉得，你能感觉到什么时候会是一场好节目，因为人们做自己。那是一次很棒的对话，我们会玩得很开心。所以，谢谢你加入我。

Ready to go. Erik, listen, dude. It is so great to have you on the show. You know, I was sitting downstairs with you chatting before and I was just like, you can feel when it's going to be a good show because people are themselves. And that was an amazing conversation and we're going to have a lot of fun. So, thank you for joining me.

能来这里太兴奋了。我是说，我听过很多期节目。现在身处其中感觉有点超现实。

It's so exciting to be here. I mean, I've listened to so many episodes. It's kind of surreal to be inside of it right now.

那真是太棒了。就像我跟你说过的，你创办公司的时候可以给我发个短信，也许我会稍微更富有一点。

That's really awesome. As I said to you, you could have texted me when you started the company and maybe I would have been a little bit richer.

是啊，但那时候你可能还不知道我。

Yeah, but you probably didn't know about me then.

不，我不知道。但我会欢迎投资另一个项目。嗯，我想从——我昨晚和 Zach Frankle 聊天，他谈到了你的童年，以及早期成功的迹象和非凡的能力。你是怎么爱上技术的？而对我作为投资者来说，看到你的童年时，什么会让我兴奋？

No, I didn't. But I would have welcomed investing in another suite. Um, I wanted to start with—I was talking to Zach Frankle last night and he said about your childhood and like early indicators of success and like exceptional ability. How did you fall in love with technology? And for me as an investor, what would have made me excited in seeing your childhood?

我大概 1992 年开始编程，那时我八岁。就是爱上了它。你知道，就像那种事情，我意识到你可以控制计算机。我有一台 Mac Plus，1 兆字节内存之类的。然后开始做游戏，我是说它们很糟糕，但你知道，回想起来，我某种程度上非常感恩。我回顾我的童年，我的朋友们在收集冰球卡之类的，那是他们的爱好。而我的爱好是编程，你知道，现在我从这上面赚了很多钱。这不是有点疯狂吗？你想想，我能靠我从八岁起的爱好赚很多钱，这有点不可思议。所以我非常感恩我爱上了技术和编程，然后我 16、17 岁开始参加编程竞赛，非常有趣。

I started coding 1992 or something like that. I was eight. Uh, and just loved it. You know, it's like one of those things like I realized you can control computers. I had a Mac Plus, you know, 1 megabyte of RAM or whatever. Um, and yeah, started making games and like I mean they were terrible, but you know, it's just I think back at like in a way I'm like so grateful. I look back at like my childhood. My friends, they were like collecting hockey cards or whatever. That was like their hobby. Like my hobby was programming, you know, and now I make a lot of money from it. Isn't that kind of crazy? Like you know when you think about it like the fact that I can make a lot of money like doing what was my hobby since I was eight it's kind of nuts. So I'm just so grateful about you know the fact that I fell in love with technology and coding and then you know I started doing programming competitions when I was 16, 17 and so much fun.

Modal 是你的第一家公司，还是你的第一家公司——为什么花了这么长时间？

Modal was your first company or is your first company—why did it take so long?

我认为我生活中一贯的模式是，我从来没有对自己有足够的信心。嗯，很多缺失的就是信心，也有一点钱的原因。所以直到我 36、37 岁，我才创办 Modal。我想我从来没有一定要创办公司的想法。就像那种事情，我只是看到了市场上的巨大机会。我几乎对自己说，为什么没有别人在做这个？这么明显的想法。然后这让我想，也许我应该创办一家公司。然后，你知道，五年后，我们就在这里了。进展相当不错。我想事后看来，也许我应该更早开始。

I think the consistent pattern in my life is like I never had quite enough confidence in myself. Um, and a lot of it was that was missing was just the confidence, a little bit the money too. But, uh, so it wasn't until I was like 36, 37, I started Modal. And I think I don't think I've ever wanted to necessarily start a company. It was like one of those things where like I just saw this big opportunity in the market. I almost thought to myself like why is no one else doing this? Such an obvious idea. And and that led me to think like maybe I should start a company. And then, you know, five years later, here we are. Uh has gone pretty well. I think in hindsight maybe I should have started much earlier.

你希望自己早点开始吗？

Do you wish you had?

我看着一些斯坦福的孩子，他们才 21 岁，我有点觉得他们可能太自信了。你懂我的意思吗？他们觉得自己要接管世界，才 21 岁。嗯，我觉得理想的年龄也许是，我不知道，32 岁左右。你多一点经验。你知道如何与人打交道。你知道如何招人。你知道如何解雇人。所有这些事情。嗯，我不知道我 20 多岁的时候会不会是一个好的创始人。

I look at some of these Stanford kids and they're like 21 and I kind of feel like they maybe have too much confidence. You know what I mean? Like they think they're going to take over the world and they're 21. Uh I think the ideal age is maybe like I don't know 32 or something like that. You have a little bit more experience. You know how to navigate people. You know how to hire people. You know how to fire people. Like all these things. Um I don't know if I would have been a good founder like in my 20s.

在我们进入市场和下面所有内容之前，我们之前在楼下聊天。嗯，我们经常听说创办公司就像吃玻璃。很难，很痛苦。你怎么看这些说法？

Before we get into the market and everything beneath that, we were chatting downstairs. Um we often hear that founding companies is like eating glass. It's hard. It's painful. How do you feel about those?

兄弟，我讨厌那些说法。我不理解。就像你每天都在吃玻璃，那你为什么还要做？我不知道。你期待某种神奇的奖励，最后有个大饼干。我从来不理解那种心态。我觉得很多人有这种心态：我要承受痛苦，然后最后会有巨大的回报。我不知道，我不那样想。我享受每天去上班。当然，有艰难的时刻，比如你必须进行艰难的对话，也许解雇人，也许做困难的事情，但总的来说，生活就是这样。我享受生活。我享受建立公司。我不认为有一天我们会把这家公司卖掉然后——我希望在这上面工作 20、30、40、50 年。我太开心了。我不想卖掉这家公司，因为我会直接去创办 Modal 2。

Dude, I hate that stuff. It's like I don't get it. Like you're eating glass like every day. Like why are you doing it then? Like I don't know. You expect some like magic reward like some big cookie in the end. Like I never understood that mindset. I think it's like this mindset that a lot of people have. I'm gonna suffer pain and then there's going to be this big reward in the end. I don't know. I don't think about it that way. I enjoy every day I go to work. Of course, there's like hard times. Like there's, you know, you have to have hard conversations, maybe fire people, maybe like do hard things, but like in the grand scheme of things, like I mean, such is life. And I enjoy life. I enjoy building a company. And I don't think of it as like in one day we're going to flip this company and be—I hope to work on this for 20, 30, 40, 50 years. I'm having so much fun. I don't want to sell this company because I would just go and start Modal 2.

你真的不会吗？

You really wouldn't?

不，我不会。

No, I don't.

这很有趣。我有机会卖掉 20 VC。这个节目一直很成功。我们之前说过，我会以大约 8000 万美元分享一切，我拥有 100% 的股份，那对我来说很多了。我母亲和我一起散步，她说：“你明天会做什么？”我说：“我会创办 19 分钟 VC。”她说：“是啊。”她说：“别卖。”

It's funny. I had the chance to sell 20 VC. Uh the show's always successful. We said earlier I'd share everything for like 80 million and I own 100% of it and that for me is a lot. And my mother went for a walk with me and she said, "What would you do tomorrow?" I said, "I'd start the 19minute VC." She was like, "Yeah." She was like, "Don't sell."

是啊。你会做什么？我不认为钱一定会让你快乐。我是说，有一点。有钱是好事，但你知道，超过某个点，你知道，我喜欢构建东西。

Yeah. What would you do? I don't think money necessarily makes you happy. I mean, a little bit. Like it's good to have some money but you know beyond a point like what you know I like building stuff.

你真是个可爱的——

You're such a sweet—

我爱技术。我爱编程。太有趣了。

I love technology. I love coding. So much fun.

你真是个可爱的——那是瑞典式的吗？那是瑞典人的心态吗？

You're such a sweet— Is that Swedish? Is that Swedish mentality?

也许是吧。

Maybe it is.

是啊。从来没有美国人上来之前说这不是关于钱。

Yeah. Never has an American come on before and said it's not about money.

我是说，我确实有点在乎钱，但也许你说得对，因为我确实认为瑞典人的心态是自我实现。做你喜欢的事在某种程度上是最高形式的声望。可以是音乐，可以是诗歌，什么都行。几乎不重要。

I mean I do care about money a little bit but maybe you're right because I do think in Swedish mentality it's like self-actualization. Like doing what you enjoy is like kind of the highest form of prestige in a certain way. And it could be music, it could be like poetry, whatever. Like it almost doesn't matter.

你做什么是你不喜欢的？

What do you do that you don't enjoy?

是啊，老实说，我有点讨厌融资。我是说，还行吧。我不知道。如果遇到好人，我挺享受的，但这挺有挑战性的。你必须对很多人说不。我想这是个好问题，但你知道，总的来说，这——

Yeah, I kind of hate fundraising, honestly. I mean, it's fine. I don't know. Like, I enjoy it when it's like good people, but it's like kind of challenging. You have to say no to a lot of people. I guess that's a good problem to have, but you know, in general, it's—

哦，你必须对很多人说不。

Oh, you have to say no to a lot of people.

Larry，你只能拿到 1000 万、5000 万。哦，你从来没做过路演 PPT 吧？

Larry, you can only have 10 million, 50 million. Oh, you've never done a deck, have you?

我从来没做过路演 PPT。到目前为止我们做的每一轮融资，我都喜欢在融资之前提前很久去了解投资人。所以我们做的每一轮，最近我们昨天刚宣布了和 Excel 的一笔交易。我和 Excel 的 Rich 和 Gonzalo 聊了我觉得至少一年。到某个时候，我们说，其实你知道吗，我们可以拿点钱。条款是什么？我们谈谈条款吧。然后很快就完成了，因为我们已经有那层关系了。

I've never made a pitch deck. Every round we've done so far, I like to get to know people well in advance of raising money. So every round we've done, most recently we announced a deal yesterday with Excel. I've been speaking to Rich and Gonzalo at Excel for I think at least a year. At some point, we're like, actually, you know what, we could take some money. What are the terms? Let's discuss terms. And then it happened very quickly because we already had that relationship.

你怎么看价格优化？这是一场争取最高价格的游戏吗？

How do you think about price optimization? Is it a game of getting the highest price?

不，我不这么认为。我觉得也许在后期阶段，它变得更像一种商品，就像钱本身一样。但我认为在早期，有几个原因。一是存在激励错位，风投总是想投入比公司需要多得多的钱。而你会承受大量稀释。作为股东这并不好。但另一件事是，你其实想克制一下，我认为通常更好的做法是以更低的估值少融很多钱，然后让风投对下一轮保持渴望。

No, I don't think so. I think maybe in later stages it becomes a little bit more of a commodity like the money itself. But I think in the early days, for several reasons. One is that there's a misaligned incentive like VCs always want to put in way more money than a company needs. And you just take on a lot of dilution. It's not good as a shareholder. But the other thing is like you actually want to hold back, I think, and it's often better to raise a lot less money at a lower valuation and then keep the VCs hungry for the next round.

我能问你吗，这难道不是资本为王的时代吗？你需要比你想象多得多的钱。你说风投想给公司比需要更多的钱。我以为 AI 终于成了风险投资资金所需的输出出口。

Can I ask you, is this not the generation though where actually capital is the mode and you need so much more money than you think? You said that VCs want to give you more money than the company needs. I thought AI was finally the output exhaust for venture dollars that was needed.

我认为市场上有时候你会受到资本约束。在公司的大部分时间里，我们并没有受到资本约束。我们一直极其资本高效。我们每月只烧几百万。我是说只，你知道，这是一大笔钱。我尊重这些钱。

I think there are times in the market where you are capital constrained. We have not been capital constrained for most of the time of the company. We've been extremely capital efficient. We burn only a few million a month. I mean only, like, you know, it's a lot of money. I have respect for the money.

你每月烧几百万。

You burn a few million a month.

不，我是说直到去年左右。现在我们不得不在容量上花很多钱。

No, I mean up until like last year. Now we have to spend a lot of money on capacity.

你现在每月烧多少钱？

How much do you burn a month now?

很多。我是说我们实际上在损益表上是盈利的，但现金流相当负，因为我们不得不在获取 GPU 和建设容量上花很多钱。所以这是公司的一个不同时期，以及我们如何看待资本。资本是——我不想称它为护城河，因为我觉得人们滥用这个词来指代那些——对我来说护城河有非常严格的定义——但在这个节点资本肯定有帮助。

A lot. I mean we actually are profitable on a P&L basis but cash flow is quite negative because we have to spend a lot of money on getting GPUs and building out capacity. And so that is a different period in the company and how we think about capital. Capital is—I don't want to call it a moat because I feel like people throw around that word to mean things that—to me moat has a very strict definition—but certainly capital helps at this point.

对你来说护城河的定义是什么？

What's the moat definition for you?

对我来说护城河就像你去读《七种力量》之类的书。那是本好书。但我认为人们读了之后就会产生错觉。他们说，对，我们正在创办这家公司，它有所有这些护城河。你在早期从来没有护城河。

I mean moat to me is like you go read like Seven Powers or like one of those things. It's a good book. But I think then people read that and they get delusional. They're like, yeah, we're starting this company, it has all these moats. You never have moats in the early days.

谷歌可以做任何他们想做的事。

Google could do anything they wanted to.

谷歌可以做。是的。但这也像，你知道，那七种是什么？我都不记得了。像是垄断资源、市场力量之类的。那些东西直到你成立 10 年左右才会发生。我总回想起 Ben Thompson 的一篇旧博客文章。它讲的是 Snap 和姜饼人策略。你知道这个吗？重点是 Snap 没有护城河。他们只是跑得更快。他们是姜饼人。我其实总是回到这一点。我觉得这对如今几乎每家初创公司都如此。你只需要比别人跑得更快。

Google could do. Yeah. But it's also like, you know, what are the seven? I don't even remember. It's like cornering resources, market power, whatever. None of those things happen until you're like 10 years in. There's an old blog post I always come back to by Ben Thompson. It talks about Snap and the gingerbread man strategy. Do you know this thing? And the whole point is Snap doesn't have a moat. They just run faster. They're the gingerbread man. And I actually always come back to that. I think that's so true for almost every startup these days. You just have to move faster than everyone else.

我其实认为如今我投资时重要的三件事是：一是团队。他们是前 1% 的团队吗？

I actually think like the three things that matter today when I invest is like one team. Are they top 1% team?

是的。

Yeah.

二是产品决策的速度，非常具体。你看看今天的 Instinct。功能发布的速度真的非常令人印象深刻，他们非常——三是品牌。

Two speed of product decision-making very specifically. You look at like Instinct today. The speed of feature shipping is really very impressive and they're very—and then three is brand.

是的。我认为对我来说这是可防御性护城河的三个决定因素，但它是一种组合。

Yeah. And I think those are the three determinants to me of defensibility moat but it's a kind of package.

是的。是的。我是说， again 我不知道我会不会，但我有点同意你的观点，我觉得品牌——我不想称它为护城河，但事实上我认为品牌的价值已经上升了很多。我能冒昧问一句吗？

Yeah. Yeah. I mean again I don't know if I would but I kind of agree your point like I think brand I don't want to call it a moat but it's actually I think the value of brand has gone up a lot. Can I be a dick?

请便。

Yeah please.

你觉得 Modal 在品牌方面做得好吗？

Do you think Modal have done well with brand?

我觉得相当好。我认为我们有——它与开发者体验、产品、网站、视觉识别以及所有这些事情相结合。我们只是在包装方式、活动和视觉外观与感觉上做了很多。

I think pretty well. I think we have—it's coupled with the developer experience and the product and the website and the visual identity and all these things. We just did a lot of how we package it, the event and the visual look and feel.

你觉得你的品牌比 Baseten 和 Fireworks 更大吗？

Do you think you have a bigger brand than Baseten and Fireworks?

我觉得我们有。我觉得我们有非常非常好的氛围。我永远在线。人们在推特上谈论 Modal 多得多。我觉得有更多的社区感。我觉得我们周围的氛围好得多。但我对那些公司很偏执。我对竞争很偏执。我当然会追踪。我当然看到所有关于我们竞争对手的推文。但这很重要——每个人都想成为社区的一部分。我认为人们误解的是，我喜欢 Jeb 的 TypeSafe。

I think we have that. I think we have really really good vibes. I'm terminally online. People tweet way more about Modal. I think there's a little bit more of a community. I think there's much better vibes around us. But I am paranoid about those companies. I am paranoid about competition. Of course I track it. I of course see everything that gets tweeted about our competitors. But this is important—everyone wants to be a part of the community. What I think people misunderstand is I love TypeSafe with Jeb.

是的。

Yeah.

因为每个人都想克隆 Jeb。没人想上台穿粉色毛皮大衣，然后以一种疯狂的方式谈论 Jevolution。对我来说品牌就是这样。那家伙在完全不同的层次上玩游戏。我觉得这对今天的公司来说太真实了。你必须考虑这一点。

Because everyone wants to clone Jeb. No one wants to get up on stage and wear a pink fur coat and just talk about the Jevolution in an unhinged way. That is what brand to me is. That guy is playing the game on a whole different level. And I think it's just so true about companies today. You have to think about that.

所以你能跳出来说说为什么吗？因为他真实且不同。我认为你只需要创造一种个性，你得弄清楚你公司的个性是什么，然后保持一致，创造一种几乎像运动的感觉、一种社区感，就像这是很酷的东西，人们想用你是因为你是外面很酷的产品之一，就像“你试过 Modal 吗？挺酷的。”那么 Modal 做了什么 Baseten 和 Fireworks 没做的来建立这些？

What so can you jump out why cuz he's authentic and different I think so I think you just have to create a personality you kind of have to like figure out like what is the personality of your company and just be consistent and create like almost like a sense of movement a sense of like community like this is something cool and and people want to use you because you're one of the cool products out there like have you tried modal It's pretty cool. What what has modal done then that base 10 and fireworks haven't to build that?

我认为我们从开发者体验开始。所以那是——我们有一种非常严肃或非常清晰的 PLG 产品市场契合、产品增长，意思是人们自下而上地来。他们只是在产品上注册。我认为来自 Spotify 的消费者背景对我如何思考构建现在这家有效销售技术的 B2B 公司非常重要。

I think we started with the developer experience. So that was—and we had a very serious or very clear sort of PLG product market fit product growth in the sense that people came bottoms up. They just sign up on a product. I think the consumer background of coming from Spotify is very important in how I thought about building what's now B2B company selling tech effectively.

某种程度上，这是一个非常无聊的类别，但我认为拥有个性和一点锋芒是创造那种运动感、兴奋感以及让你感觉与众不同的重要部分。我想从你访问我们网站、尝试产品的那一刻开始，你就能感觉到它不同，那种神奇的时刻。我可以问你吗，我们已经谈到了三个不同的提供商。推理作为一个市场有许多不同的投机者在评论。当你展望未来 5 年，推理市场会如何发展？

In a way, it's like a very boring category, but I think having personality and having a little bit of an edge is a big part of how you create that sort of movement and that sort of sense of excitement and the feeling that you're different. And I wanted to start the moment you go to our website and the moment you try the product and you just feel that it's different, the magic moment. Can I ask you, we've spoken about the three different providers there. Inference as a market has many different speculators commenting on it. When you look at the next 5 years, how does the inference market play out?

首先，我认为推理将成为一个巨大的市场。它基本上就像新的云。但我不认为推理本身可能是你唯一需要的东西。我非常强烈地感觉到的一件事是，它必须伴随一整套不同的产品。它必须伴随后训练、沙盒、智能体式的东西、各种其他类型的推理，如低延迟推理、批量推理。所以构建那个平台——顺便说一句，我们过去 5 年一直在做这个——我认为对我来说这不是护城河，但它是一个非常强大的技术，是一个非常难构建的基础设施产品。

First of all, I think inference is going to be an enormously large market. It's like the new cloud basically. But I don't think inference in itself is maybe the only thing you need. One of the things that I feel very strongly about is that it has to come with a whole set of different products. It has to come with post-training, sandboxes, agentic stuff, all kinds of other types of inference like low latency inference, batch inference. So building that platform—and by the way, we've been working on this for the last 5 years—I think that to me is not a moat, but it's a very strong tech, it's a very hard product to build infrastructure.

为什么你需要拥有那个产品集的广度，否则你只是在一个压缩层？

Why do you need to have the breadth of that product set otherwise you're just in a compression layer?

是的，因为否则你只是在卖商品化产品，我认为。如果你只是用美元卖 token，就没有转换成本。人们可以直接切换到不同的提供商，对吧？这只是一场逐底竞争。但如果你把它与周围所有这些其他东西结合起来——沙盒、后训练、强化学习以及所有这些——你实际上可以创建一整套工具。没有人想买一个只会一招的小马。人们想买一个不同东西的组合，一整套工具，基本上让他们在其上构建惊人的 AI 应用。对我来说，以及底层的基础设施，是很多差异化，很多市场持久力。构建所有这些产品所需的基础设施非常难以复制。

Yeah, because otherwise you're just selling commodity product, I think. And if you're just selling tokens for dollars, there's no switching cost. People can just switch to a different provider, right? It's just a race to the bottom. But if you couple it with all these other things around it—sandboxes and post-training and reinforcement learning and all these things—you can actually create a whole set of tools. No one wants to buy a single-trick pony. People want to buy a portfolio of different things, a whole suite of tools that basically lets them build amazing AI applications on top of it. And that to me, and the infrastructure underneath, is a lot of differentiation, a lot of staying power in the market. It's very hard to copy the infrastructure required to build all of these products.

关于推理市场，没有人看到什么，或者你很少看到评论但对你来说非常明显的是什么？

What does no one see about the inference market or you don't see commented on much that is very obvious to you?

人们没有看到推理市场的一件事是，我实际上认为长期来看，这个领域有如此多的毛利率机会，因为你可以围绕它构建所有这些事情。所有的沙盒、所有智能体式的东西、工作流、后训练、强化学习。如果你做所有这些事情，实际上有如此多的赚钱机会。

One thing that people don't see about the inference market is I actually think in the long term there is so much gross margin opportunity in the space because you can build all these things around it. All the sandboxes, all the agentic stuff, workflows, post-training, reinforcement learning. If you do all those things, there's actually so much opportunity to make money.

这是一个有软件利润率的市场吗？因为我们现在看到的利润率在 25% 到 40% 范围内。那是传统玩家。我不知道那是不是你的，但那是推理提供商普遍的情况。随着你拥有越来越多的东西，它们会达到 70% 吗？

Is it a market that has software margins? Because we're looking at margins now in the 25 to 40% range. That's kind of traditional players. I don't know if that's yours, but that's kind of what inference providers are generally. Do they go to 70 as you own more and more of the stuff?

我认为基本上今天几乎没有公司还有软件利润率了，就像你过去在硅谷 10 年前看到的 90% 以上的利润率。我认为那基本上消失了，原因实际上几乎是拓扑的。就像你有一堆供应商。今天你有应用、有模型、有基础设施、有云、有 Nvidia。所以当客户在堆栈顶部支付一美元时，它必须流经许多不同的层，每一层可能有 20% 到 30% 的毛利率。所以我认为有很多利润率机会，有很多赚大钱的机会。我的意思是，空间如此之大，但你也把它分散在堆栈的许多层上。所以 90% 软件利润率的企业已经消失了。我认为 Modal 长期有机会达到 40%、50%、60% 的利润率，但我们永远不会达到软件利润率。

I think basically almost no company today has software margins anymore like what you used to see in Silicon Valley 10 years ago with 90% margins plus. I think that's basically gone and the reason is actually almost like topological. It's like you have this stack of vendors. Today you have applications, you have models, you have infrastructure, you have cloud, you have Nvidia. So when a customer pays a dollar at the top of the stack, it has to flow through a lot of different layers and at every layer there's going maybe 20% 30% gross margins. So I think there's a lot of margin opportunity and there's a lot of opportunity to make a lot of money. I mean, the space is so big, but you're also splitting it across a lot of layers in the stack. So the 90% software margin businesses are gone. I think Modal has an opportunity long-term to get to 40 50 60% margins, but we're never going to get to software margins.

关于推理市场的未来，你有什么疯狂预测，你认为很多人会不同意？

What's a wild prediction you have on the future of the inference market that you think many people would disagree with?

我认为关于推理市场的一个预测是，它将分裂成许多不同的东西，因为市场在过去变得如此之大——我不认为任何人预料到，包括我自己，我不认为任何人预料到它会如此之大，它比我曾经想象的大 10 倍或 100 倍。对我来说，这意味着有如此多的专业化空间：低延迟、高延迟、各种类型的微调和后训练。我认为会有一些古怪的新技术让你用更少的资源做到这一点，新硬件，各种不同的东西。

I think one prediction about the inference market is it's going to split into a lot of different things because the market got so big in the last—I don't think anyone expected, me included, I don't think anyone expected it to be so big, it's like 10x or 100x bigger than I ever thought. And that to me means there's going to be so much room for specialization: low latency, high latency, various types of fine-tuning and post-training. I think there's going to be some wacky new technologies that let you do it with way less resources, new hardware, all kinds of different things.

非常抱歉，这是否与你刚才说的需要拥有全部并拥有产品广度相矛盾，如果它大得多以至于你需要真正的专业化？

I'm really sorry, does that not go against what you just said about the need to own it all and have that breadth of product if it's so much larger that you need real specialization?

按产品专业化。我认为它将分裂成许多不同的产品。现在我们可以构建很多这样的产品，并全部托管在 Modal 下，这对我来说非常令人兴奋，因为公司将需要很多不同的产品。对于一个产品，他们需要低延迟推理；对于一个产品，他们需要批量推理；对于一个产品，他们需要沙盒，我们有；对于一个产品，他们将预训练自己的模型，也许是某些专门的音频模型或其他什么。除了语言模型推理，我们做很多其他事情。所以对我来说，这创造了很多机会，基本上重新思考 AI 时代的云是什么样子，并解构云，为这个新的软件时代重建它。

Specialization by product. I think it's going to fragment into a lot of different products. Now we can build a lot of those products and host it all under Modal and that to me is very exciting because there's going to be a lot of different products that companies are going to need. For one product they're going to need low latency inference, for one product they're going to need batch inference, for one product they're going to need sandboxes which we have, for one product they're going to do pre-training their own models, maybe it's certain specialized audio models or whatever. We do a lot of stuff besides LM inference. And so that to me creates a lot of opportunity to basically rethink what a cloud looks like in the era of AI and deconstruct the cloud and rebuild it for this new era of software.

我可以问你吗，你之前提到用 GPU 构建容量？

Can I ask you earlier you said about building capacity with GPUs?

是的。

Yeah.

今天做那件事是什么感觉？像报道说的那么疯狂吗？

What's it like doing that today? Is it as crazy as reports suggest?

所以它相当疯狂。三月份我在圣何塞的 GTC，Nvidia 的大型会议上待了一段时间。我会把那个描述为露天毒品市场，但用 GPU 代替毒品。你到处跑，工作，与 500 个不同的云供应商交谈。你说，你有 GPU 吗？请弄一些——最后你出来，你做了很多交易，那是我们必须培养的能力。我认为 Modal 价值主张的一部分是我们的客户不必考虑那个。我们在幕后做所有 GPU 交易的脏活。

So it is pretty crazy. I spent some time in March at GTC, Nvidia's big conference in San Jose. I would describe that as like an open air drug market, but GPUs instead of drugs. You're running around, working, talking to 500 different cloud vendors. You're like, do you have GPUs? Please get—and in the end you come out, you do a lot of deals and that's something we've had to develop the muscle. I think part of the value prop of Modal is our customers don't have to think about that. We do the dirty work of doing all the GPU deals under the hood.

我们现在与 25 家不同的 Neo 云合作。我们控制着全球算力的 0.1% 左右——我甚至不知道具体数字。但我对长期目标的思考是：随着我们控制越来越多的全球算力，你可以把供给侧的容量和需求侧的需求聚合起来，并极其高效地管理这种容量和需求。所以这对我来说是我们真正认为可以做到的事情——我们可以填补一个巨大的空白。

We work with 25 different Neo clouds right now. We control like 0.1% or whatever—I don't even know the number—of the world's compute. But how I think about the long-term goal is: as we control more and more of the world's compute, you can aggregate that capacity and aggregate demand on the other side, and manage that capacity and demand extremely efficiently. So that to me is something we actually think we can do—we can fill a big gap.

作为那个聚合层，这让你能做什么？

What does that allow you to do, being that aggregation layer?

嗯，我们很多基础设施都是自己构建的。我们不用 Kubernetes 和 Docker 之类的东西。我们构建了自己的容器文件、自己的文件系统、自己的容器运行时、自己的调度器。我们构建了自己管理容量的基本方式。这让我们能够把全球所有算力当作一个大容量池来对待。所以如果你想想，我们已经有几千个付费客户，而且随着时间只会越来越多。每个客户都有非常可变的需求，或者他们的工作负载在一天中会变化，也许会有突发，单独来看非常不可预测。但如果你把所有客户放在一起，你最终会得到非常平滑的整体需求，而这是我们可以极其高效管理的。所以 Modal 的价值主张之一，除了‘我们可以和你一起扩展’之外，还有‘你甚至不需要做容量规划——直接来找我们，我们给你 100 块 GPU。’我觉得这太神奇了。回到 TypeSafe 的例子——我们在他们发布之前很久就聊过，但他们一夜之间在 Modal 上发布，我们甚至都不知道；它就自动扩展了，我们给了他们大量 GPU，一切自动发生。所以那种魔力我认为是人们在 Modal 上真正看重的：我们可以在非常短的通知内给他们大量 GPU，并且我们能处理那种可变性。如果你来找我们，你不必锁定一个矩形的 GPU 块三年，也不必知道确切数量——它是按需付费的，你可以扩展和缩减，按 GPU 秒付费。

Well, so we built a lot of our infrastructure ourselves. We don't use Kubernetes and Docker and things like that. We built our own container files. We built our own file system. We built our own container runtime. We built our own scheduler. We built our own—basically how we manage the capacity. And that enables us to treat all of the world's compute as one big pool of capacity. And so if you think about the fact that we already have many thousand paying customers and over time it just grows and grows. Each of those customers have very variable requirements, or their workloads vary over the day, and they have maybe bursts, and individually they're very unpredictable. But if you pull all of those customers together, you end up with a very smooth overall demand, and that is something we can manage extremely efficiently. And so one of the value props of Modal besides 'we can scale up with you' is also 'you don't even have to do capacity planning—just come to us, we give you 100 GPUs.' I think it's so amazing going back to TypeSafe for instance—we had talked to them a long time ago before they launched, but they launched overnight on Modal and we weren't even aware of it; it just scaled up and we gave them a ton of GPUs and it just happened automatically. So that magic I think is something that people really value with Modal: we can give them a lot of GPUs on very short notice and we can handle that variability. If you come to us, you don't have to lock up a rectangular block of GPUs over 3 years and know the exact number—it's pay as you go, you scale up and down, you pay per GPU second.

那么你们在预测方面如何进行 GPU 采购？因为你不想让大量过剩的供应闲置在账上。

How do you do GPU buying in terms of forecasting then? Because you don't want to have an insane excess of supply just sitting on your books.

完全正确。目前 GPU 的交货时间是三到四个月。所以我们基本上做直线或指数线外推，我们看所有客户的总体情况。我们有一组非常多样化的客户。你实际上可以很好地看总体需求，并根据月增长率来预测——过去 12 个月大约是每月 30-40%——你只需将其投射到未来 3 到 4 个月，最终会得到大约 2 到 3 倍。你出去购买那个容量并确保逐步增加。

Totally, yeah. The GPU lead times right now are three to four months. So we basically do a straight line or exponential line extrapolation and we look at an aggregate across all our customers. We have a very diversified set of customers. You can actually look at the demand in aggregate pretty well and predict based on monthly growth rates—which have been about 30-40% month-to-month for the last 12 months—and you just project that 3 to 4 months into the future, you end up with like 2 to 3x. You go out and buy that capacity and make sure to ramp up.

那你能直接买到那个容量吗？

And can you just buy that capacity?

可以的。是的。我的意思是，我认为有一点误解，认为我们没有容量了。英伟达每个月出货的 GPU 数量惊人。我不知道有多少 10 万块 Blackwell 正在安装。价格肯定上涨了。你必须支付更多才能获得同样数量的 GPU。我不想说这是一个有效市场，但如果你愿意比其他人支付更多，你可以获得很多 GPU。

You can. Yeah. I mean, I think there's a little bit of a misconception that we're out of capacity. Nvidia is shipping an insane amount of GPUs every month. I don't know how many 100,000 Blackwells are getting installed. Prices are up for sure. You have to pay more to get the same amount of GPUs. And I don't want to say it's an efficient market, but if you're willing to pay more than other people, you can get a lot of GPUs.

有没有一个价格对你们来说不合理？当然有。但比如在什么点？

Is there a price at which it doesn't make sense for you? Of course there is. But like at what point?

我不这么认为。事实上，我们做多 GPU 而不是做空。因为我们可以向客户收取更高的价格——实际上存在市场价格——而且动态不同，因为我们锁定长期预订并以稍短的期限出租。但如果基础价格上涨，人们也愿意支付更多。

I don't think so. I think in fact we're actually more long than we're short GPUs. Because we can charge a higher price to our customers—there's effectively a market price—and it's different dynamics because we lock up longer-term reservations and rent them out on slightly shorter duration. But if the underlying price goes up, people are also willing to pay more.

完全理解。价格随时间会怎样？H100 本应贬值并下降。

Totally get you. What happens to the price over time? H100s were meant to depreciate and go down.

我完全不知道。是的。我的意思是，实际上这很有趣。

I have no idea. Yeah. I mean, actually it's funny.

好吧老兄，如果你都不知道，那我该怎么办。

Well dude if you have no idea what am I meant to have.

你看到黄仁勋了吗？昨天他在伦敦，有人问了他这个问题，他说嗯显然需求会赶上供应——或者说供应最终会赶上需求，然后我们看看定价会发生什么。可能一年后，可能五年后。他说的话和我一直告诉人们的一模一样：硬件价格就像重力。长期来看硬件价格总是下降。现在我们处于供应短缺，这意味着价格上涨。它们可能继续上涨一年、两年、五年——我不知道——但最终会下降。

Did you see Jensen? Yesterday he was in London actually someone asked him this question and he's like well obviously demand is going to catch up with supply—or sorry supply is going to catch up with demand eventually and then we'll see what happens with pricing. It might happen in a year it might happen in 5 years. He said exactly the same thing I always tell people: hardware prices are like gravity. Hardware prices always go down in the very long run. Right now we're in a supply shortage and that means that prices have gone up. They may continue to go up for another year, two years, 5 years—I don't know—but eventually they will come down.

我们有多早？因为我用 Instinct 做所有事情。我在 Instinct 上有超过 100 个自动化，做的事情简直疯狂——从买演唱会门票到策划演出，再到帮我母亲处理药物。我使用的算力量是杰文斯悖论的 100 倍。如果这对社会上许多人来说也是等效的——因为它可以在 WhatsApp 内访问——那么我们到底有多早？

How early are we? Because I run Instinct for everything. I have over 100 automations on Instinct doing things that are just absolutely wild—from buying concert tickets to planning shows to helping with medication for my mother. The amount of compute I am using is Jevons paradox times 100. If that is equivalent for actually many people in society—because it's accessible within WhatsApp—and how early are we though?

我认为我们非常早期。我想你说的是同一件事。我的意思是，是的，杰文斯悖论大概怎样都行。我认为对这些东西有太多潜在需求。而且我认为最终我们会开始遇到新的资本约束——利率在上升，我们就是无法足够快地建造数据中心。也许那是好事。我不知道。在某种程度上，这是防止过度建设的保证，但我也不知道我们是否能过度建设这些东西，因为需求太大了。

I think we're very early. I think you're saying the same thing. I mean, yeah, Jevons probably whatever. I think there's just so much latent demand for these things. And I think eventually we're going to start to run up against new capital constraints—interest rates are going up, we just can't build data centers fast enough. Maybe that's good. I don't know. In a way it's a guarantee against overbuild, but I also don't know if we can overbuild these things because there's just so much demand for it.

是的。

Yeah.

我认为未来 5 年不会过度建设。

I don't think you can overbuild in the next 5 years.

可能不会。不。不。我认为可能发生的是人们意识到——Instinct 是个好例子。据称他们还不收费或者还没有收入。所以我认为这是个好例子，但显然 Instinct 未来能赚很多钱。但这是市场的一种动态。有很多公司还没有赚钱，却在使用大量 GPU。

Probably not. No. No. I think what could happen is like people realize—Instinct is a good example. Allegedly they don't charge or they don't make revenue yet. So I think it's a good example, but clearly Instinct can make a ton of money in the future. But that's a dynamic in the market. There's a lot of companies that are not yet making money yet and they're using a lot of GPUs.

他们最终总得开始赚钱，但这个滞后可能会持续一段时间，对吧？我们可能会撞上这个问题。你提到了推理市场内部出现的专业化，我们也聊到了 Nvidia。我们看到芯片市场越来越专业化。Nvidia 今天的业务是比以往任何时候都更强，还是说其实比以往任何时候都更弱——因为市场正在碎片化，分裂成非常专业的玩家，从 Cerebras 到 Etched 到 Groq 到 Fractile，这些以前都不存在？

They just have to eventually start to make money, but that lag could take a while, right? And we might run up against that. You mentioned the specialization that comes within the inference market, and we spoke about Nvidia there. We're seeing more and more specialization within the chip market. Is Nvidia's business today stronger than ever, or actually weaker than ever, as you see the fragmentation of the market into very specialized players, from Cerebras to Etched to Groq to Fractile, that weren't there before?

我觉得你是想把它说成非此即彼。我的看法是，这些东西都会增长。我认为 Nvidia 会继续增长，他们会继续出货很棒的芯片。我认为 Cerebras、Etched 以及其他这些公司也会出货专用芯片，它们也会增长。所以我认为这里面会有所有人的市场。

I think you're trying to phrase it as one versus the other. My view is that all of these things are going to grow. I think Nvidia is going to keep growing. They're going to keep shipping amazing chips. I think there's also going to be Cerebras and Etched and all these other ones shipping specialized chips, and they're also going to grow. So I think there's going to be markets for all of it.

那为什么 Cerebras 跌了 30%？

So why is Cerebras down 30%?

我不知道。我不怎么关注股市。

I don't know. I don't really follow the stock market.

你不关注股市？

You don't follow the stock market?

不关注。我是非常被动的那种，你知道，买指数，主要是因为时间。我觉得我没有精力去盯股市。

No. I'm in a very passive, you know, buy-the-index approach, mostly because of my own time. I don't feel like I have the energy to follow the stock market.

CUDA 在多大程度上是大家所说的那种护城河？

To what extent is CUDA the moat that everyone says it is?

我其实觉得这有点被夸大了。我不认为 CUDA 本身就是护城河。CUDA 用起来其实挺烦的，是个相当烦人的编程模型。但我确实认为整个 Nvidia 生态——所有的 kernel 优化，以及你知道，每个 torch 版本总是对 Nvidia 的支持比别的都好得多，还有像 vLLM 和 SGLang 这样的开源引擎，正因为如此，它们总是对 Nvidia 支持得好得多。所以今天一切都是 Nvidia 优先，而且这大概会持续下去。有意思的是，过去 15 年我一直在说，明年是非 Nvidia 加速器的元年。结果从来没发生。我总是错的。Nvidia 就是展现出了极强的持久力，直接统治了市场。我们拭目以待吧。我是说，有 Trainium，有 TPU，有 AMD，还有很多别的东西看起来相当不错，它们在追赶，但我仍然觉得它们落后不少。

I actually think it's a little overstated. I don't think CUDA itself is the moat. CUDA is actually kind of annoying to work with. It's a pretty annoying programming model. But I do think the whole Nvidia ecosystem — all the kernel optimizations and all the stuff that, you know, every torch always supports Nvidia much better than anything else, and the open source engines like vLLM and SGLang, because of that, they always support it much better. So everything is Nvidia first today, and that's probably going to continue. It's funny: for the last 15 years I've been saying next year is the year of non-Nvidia accelerators. It never keeps happening. I'm always wrong. Nvidia has just shown to be incredibly good at having that staying power and just dominating the market. We'll see. I mean, there's Trainium, there's TPUs, there's AMD, there's a lot of other stuff that does look quite good and they're catching up, but I still think they're quite behind.

这些新云（neocloud）在多大程度上是朋友？因为它们在很多情况下和你们是合作伙伴。我相信，就像你提到的——我记得你说 25——或者说是敌人。你知道，我请过 Nebius 的创始人来做节目，他说，我们要进军推理市场，我们需要拥有全栈。我不是针对他，我完全理解。但他们是朋友还是敌人？

To what extent are the neoclouds a friend? Because they're partnered with you in a lot of cases. I'm sure, as you mentioned — I think 25, you said — or a foe. You know, I had the founder of Nebius on and he was like, we're coming for the inference market, we need to own full stack. And I don't mean that against him, I totally understand it. But are they friends or are they foes?

我觉得创业圈里每个人同时都是朋友和敌人。他们和所有人都保持着这种奇怪的关系。我认为一些顶级新云，包括 Nebius，它们正试图往上层走。我个人有点怀疑。我觉得要既做好核心业务，同时又试图进入另一个市场，是非常难的。

I feel like everyone in startup land is a friend and a foe at the same time. They have this weird sort of relationship with everyone. I think some of the top neoclouds, Nebius included, they are trying to move up the stack. I'm personally somewhat skeptical. I think it's very hard to be good at your core business and also try to move into a different market at the same time.

为什么对他们来说会很难？就是说，他们能够高效地进入推理层，看空理由是什么？

Why would it be hard for them? Like what's the bear case on them being able to move into the inference layer efficiently?

我是说，我有偏见，但我们花了过去 5 年搭建了一套极其复杂的基础设施栈。我认为要复制我们建的东西非常非常难。

I mean, I'm biased, but we spent the last 5 years building an incredibly complex infrastructure stack. I think it's very, very hard to replicate what we built.

最难的是什么？

What's the hardest thing?

就是所有的代码，以及围绕文件系统、容器运行时、调度器的一切——构建一个系统，能每分钟放置数百万个沙箱，能扩缩数千块 GPU，能支撑数十万个并发运行的任务。这些都是非常难的工程问题。我是说，我是个基础设施极客，我喜欢这个，做起来特别有意思，但这些是难题。

Just all the code and everything around the file system, the container runtime, the scheduler — building the system that can place millions of sandboxes every minute, scale up and down many thousands of GPUs, power hundreds of thousands of tasks running concurrently. These are very hard engineering problems. I mean, I'm an infra geek, I love that, it's so much fun to work on, but these are hard problems.

Nebius 是比 CoreWeave 更好的生意吗？

Is Nebius a better business than CoreWeave?

我其实不——不，我不知道。我觉得它们两家都执行得极其出色。我对我们做成的事怀有极大的敬意。我觉得归根结底，我最敬佩的其实更多是在金融工程这一侧。他们非常擅长这类事情。

I actually don't — no, I don't know. I think they've both executed extremely well, both of them. I have tremendous respect for what we've done. I think at the end of the day, the thing that I have the most respect for is actually more on the financial engineering side. They're very good at that kind of stuff.

你为什么这么说？

Why do you say that?

因为我认为快速搭建一个新云，很多约束都在金融这一侧。你得非常擅长——先锁定一份长期客户合同，然后拿它去融一大笔钱，然后去买芯片，然后交付芯片，然后在接下来几年里还清债务。和财务数字跳这种舞，是挺有意思的一种游戏。而 CoreWeave 做得极其出色。这是资本的游戏，是金融的游戏。

Because I think a lot of the constraint for building a neocloud very quickly is on the financial side. You need to be very good at — you go lock in a long-term customer contract, then you go raise a bunch of money against it, then you go buy the chips, then you deliver the chips, and then you pay off the debt over the next few years. Doing that sort of dance with the financials, it's kind of an interesting game. And CoreWeave has done it exceptionally well. It's a capital game. It's a financial game.

当你往下走一层，我们聊到了 Jeb，他显然在 Modal 工作。你知道，我们看到了开源社区惊人的进展，这很令人兴奋。但你也看到 Haiku 和 Anthropic 最新的模型性能极强、价格极低。如果前沿模型几乎一样便宜、一样高效，开源模型的好处是什么？

When you move a layer beneath them, we spoke about Jeb who obviously works with Modal. You know, we've seen incredible movement from the open source community, which has been exciting. But then you also see Haiku and Anthropic's newest models being incredibly performant and incredibly cheap. What's the benefit of open models if frontier models are almost as cheap and as efficient?

我觉得开源永远会——永远会有很棒的模型出来。我们也要记住，开源模型都在转向授权模式。所以它们越来越擅长捕获这些模型的价值，我认为这意味着它们会有更多资金去训练更好的模型。我真心希望会有欧洲和美国的模型，因为我觉得有这么多中国模型不是个好局面。我是说，显然我对所有中国实验室怀有极大的敬意，它们在做的事上极其出色，但我认为更多力量分散开来对世界是好事。不过不管怎样，回到问题，我确实认为开源——永远会有人想要微调、想要对自己的模型有更多控制、想调参数、要低延迟高吞吐，不管是什么，开源会在其中扮演重要角色。

I think open source is always going to — there's always going to be great models coming out. Let's also remember the open source models are all moving into licensing models. So they're getting better at capturing the value of those models, and I think that means they're going to have more funding to train better models. And I really hope there's going to be European and US models, because I don't think it's a good situation that we have so many Chinese models. I mean, obviously I have tremendous respect for all the Chinese labs. They're extremely good at what they're doing, but I think it's good for the world if there's more power spread out. But anyway, going back to the question, I do think that open source — there's always going to be people who want to fine-tune, want to have more control over their models, and sort of tweak the parameters, go low latency, high throughput, whatever it is, and open source is going to play a big role in that.

Windows 对 Linux。恕我直言——抱歉——你觉得欧洲会产出一个有竞争力的开源模型吗？

Windows versus Linux. With the — I'm sorry — with the greatest of respects, do you think Europe will produce a competitive open source model?

会，我是说，我爱欧洲。我在欧洲长大。

Yeah, I mean, I love Europe. I grew up in Europe.

我在瑞典长大，但我确实认为欧洲在积累大量资本方面有困难。我认为这两件事都是欧洲的难题，而中国在这方面做得非常好。他们积累了庞大的人力资本，而美国显然非常擅长筹集巨额资金和训练非常大的模型。我不知道欧洲是否已经达到了其中任何一个。

I grew up in Sweden, but I do think that Europe does struggle with accumulating large amounts of capital. I think both of those things are things that Europe struggles with, and that is something that China has done really well. They've accumulated massive amounts of human capital, and the US obviously has been very good at raising enormous amounts of money and training very large models. I don't know if Europe has passed to either of those.

如果英伟达通过先收购 Poolside 又收购 Reflection 来拥有美国的开源模型层，这要紧吗？

Does it matter if Nvidia owns the open-source model layer in the US with first Poolside and now Reflection?

我认为英伟达非常聪明地专注于生态系统，构建生态系统。我认为这包括支持许多开源模型。他们可能看到了收购开源模型的机会，因为这些公司非常资源密集、资本密集。我认为英伟达收购其中一些并有效地赞助它们以创建美国制造的模型是明智的。顺便说一句，我想收回我说过的一件事：欧洲确实拥有惊人的人力资本。我认为欧洲非常不擅长的是聚合他们，把他们放在同一家公司。我认为这是美国非常擅长的。

I think Nvidia is very smart to focus on the ecosystem, building up an ecosystem. And I think that includes propping up a lot of open source models. I think they probably saw an opportunity to pick up open source models like these companies are very resource intensive, very capex intensive. And I think it's smart of Nvidia to buy some of those and sponsor them effectively in order to create American-built models. By the way, I want to take back one thing I said: Europe does have an insane amount of human capital. I do think what Europe is very bad at is aggregating them and putting all of them at the same company. That is something I think the US is very good at.

而且我认为激发雄心和巨大规模的心态。我想回到你之前说的那点，比如“哦，我对自己没有信心”。我尊重地认为，如果你在美国，你可能被周围的影响所包围，告诉你要早点创业。

And I think inspiring a mindset of ambition and huge scale. I think to the point that you said earlier, which is like, oh, I didn't have the belief in myself. I respectfully think if you're in the US, you might have been surrounded by influences that told you to start a company earlier.

是的。是的。我的意思是，就那点而言，几乎有点过头了。

Yeah. Yeah. I mean, it's almost like too much on the other side to that point.

哦，当你四岁时，你被鼓励离开学校，退学。幼儿园不值得。正是。

Oh, when you're four you're encouraged to leave school, drop out. Kindergarten wasn't worth it. Exactly.

我完全同意你。我确实想谈谈你提到的模型，特别是 Jav。我们是否看到模型按功能基础进行解绑，而 Jav 显然在某些任务上非常擅长？

I totally agree with you. I do want to talk about you we said about models and Jav in particular. Are we seeing like the unbundling of models on a like per function basis where Jav is obviously very good for its certain tasks?

我认为是这样，我认为这是我之前谈到的函数，即推理市场如此之大，你知道当一个市场是比如十亿美元时，你不会有很多子市场，但当市场开始达到 100 亿、1000 亿或 whatever 时，你会有很多不同的子市场，而 Jev 在分类和回归类型的事情上做得非常非常好，并且有足够大的市场。

I think so and I think that's a function of what I talked about earlier where like the market for inference is so large and that you know when a market is like you know billion dollars like you're not going to have too many submarkets but when a market is starting to get 10 billion 100 billion whatever right like you're going to have a lot of different submarkets and and Jev does classification and and things like that regression type things very very well and and there's a large enough market for that.

好的，欢迎，这是你作为 20VC 基金合伙人的第一天，我们在投资委员会，非常友好地我们有机会投资 Jav 的新一轮。

Okay welcome it's your first day as a partner at 20VC fund and we're in the investment committee and very kindly we have the opportunity to invest all in Jav's new round.

全押。全押。

All in. All in.

Jav 价值 1000 亿美元的看涨理由是什么？

What's the bull case for Jav being worth 100 billion?

我之前告诉过你，但我认为品牌、社区和围绕的运动，你知道，他们所做的一切都如此不可思议。对我来说，这在这个新时代是一个巨大的资产。

I told you before, but I think the brand and the community and the movement around, you know, this just the sort of everything they're doing is so incredible. And that to me is like a massive massive asset in this new age.

那是一家千亿美元的公司。

That's a hundred billion dollar company.

我认为有机会通过许多专门的推理模型创建一家千亿美元的公司。

I think there is an opportunity to create a hundred billion dollar company with a lot of specialized models in inference.

不过你会在 85 亿美元时投资。

You would have done it though at 8.5 billion.

我会在 85 亿美元时投资。我认为这里有一个相当大的结果。可能是退出，可能是，你知道，像一家上市公司。谁知道呢？但我确实认为这里有很多上行空间。

I would have done it at 8.5 billion. I think there's a pretty big outcome here. Could be an exit, could be, you know, like a public company. Who knows? But I do think there's a lot of upside here.

老兄，我很想看到他在纽约证券交易所敲钟。你明白我的意思吗？

Dude, I'd love to see him ringing the bell on the New York Stock Exchange. You know what I mean?

是的。很棒的人。

Yeah. Fantastic guy.

喜欢那个人。

Love that guy.

他很棒吗？我的意思是，你知道，他的魅力，他的，就像你必须围绕你的品牌创造一种个性。你必须创造这种视觉识别。这显然不是我会选择的视觉识别，

Is he great? I mean, you know, his charisma, his like you have to create like a personality like around your brand. You have to create this like visual identity. It's obviously not the visual identity I would have picked,

但我认为，我听到你了，但老实说，Eric，我认为你应该更响亮一些。

but I think I think then I hear you, but I think then you should be more loud if I'm just honest, Eric.

Modal。

Modal.

不。嗯，是的，但你

No. Well, yeah, but you

是的，我的意思是那是可能的，对吧？从你的角度看，我不认为 Modal 有个人崇拜。

Yeah, I mean that's possible, right? I don't think there's like a personality cult from modal from your perspective.

我认为这是一个合理的批评点。我认为我可以更活跃。我发推，我尝试发推。我尝试写博客。我最近有点不擅长。

I think that's a fair point of criticism. I think I could be out there even more. I am tweet I try to tweet. I try blogging. I've been kind of bad at that recently.

没人在乎博客。

No one gives a about blogging.

是的，也许。对。也许我只需要在 Twitter 上更放飞自我。这基本上就像有趣的营销，就像在 Modal 的早期，那是我增长的重要部分，就像我一直发推关于我们正在构建的东西，我认为这非常重要，我的意思是每家公司都必须找到自己的上市策略，但在早期，我们花了很多时间在 Twitter 上与人交谈，就像联系人们尝试产品，你觉得怎么样，一半的人最终成为了天使投资人，你知道他们不是 Modal 的支持者，嗯，你知道你必须倾向于你擅长的，在早期我做了很多发帖

Yeah, maybe. Right. Maybe I just need to like go more unhinged on Twitter. It's like fun related marketing basically like in the early days of modal like that was a big part of our growth right like me just like tweeting all the time about what we're building and I think that's really important if I mean every company has to find its own sort of you know go to market strategy but in the early days like we spent a lot of time just talking to people on Twitter just like hitting people up try the product what do you think half of them ended up angel investing you know they're not supporters of modal uh and and you know you have to lean into what you're good at and in the early days I did a lot of posting

你知道吗，对于在 5 万之前关注我的每个粉丝，我会给他们发一条个人的感谢信息。

Do you know for every follower that followed me before 50,000, I would send them a personal like thank you message.

太棒了。你自己打出来的。

That's amazing. You typed it out yourself.

每天晚上从 1 点到 2 点半。我会亲自说，比如，我看到你来自瑞典。我喜欢 Grand Hotel。你在伦敦时告诉我。我每晚发送 50 到 100 条。

Every night from 1 till 2:30 a.m. I person and I'd say like, I see that you're from Sweden. I love the Grand Hotel. Let me know when you're in London. And I would send like 50 to 100 a night.

那是奉献。太棒了。

That's dedication. That's amazing.

不，但那是客户获取飞轮。

No, but it was the customer acquisition flywheel.

是的。

Yeah.

就像他们第二天去办公室会说，“哇，你认识这个人，他给我发了消息，你应该听他的节目。相当不错。”很聪明。我喜欢你在创建这个媒体品牌方面所做的一切。太棒了。

Like they would go into the office next day and be like, "Wow, you know this guy, he'd send me a message and like you should listen to his show. Pretty good." It's smart. I love what you've been doing in create in terms of building up this like media brand. It's fantastic.

Neil Ma 有一个很棒的想法，就像我如何将客户体验提升 10 倍，实际上让我们来做吧，让我们玩个游戏。我如何将当前的 Modal 客户体验提升 10 倍？

Neil Ma has a great thing which is like how can I 10x the customer experience and actually like let's let's do it, let's play a game. How do I 10x the current modal customer experience?

我认为我们需要的最大的事情是我们需要在 Modal 中有更多的产品。就像我们想创建一个基本上全面的体验，你构建 AI 应用程序所需做的任何事情都应该是 Modal 内部的产品。我们可以做的另一件事是雇佣更多的 FD 和更多的销售代表。我会说我们在这方面落后了，但我们应该雇佣更多的人员亲自联系你并说，“嘿，我看到你开始在 Modal 上工作。我看到你在生物领域工作。我们如何帮助你？这里有一些生物示例，你知道，蛋白质折叠，whatever，对吧？就像我们与许多其他类似公司合作。你似乎怀疑。”

I think the biggest thing we need is we need way more products in modal. Like we want to create a basically comprehensive experience of anything you need to do in order to build an AI application should be a product inside modal. The other thing we can do is hire way more FDs and way more sales reps. I will say we've been behind on that, but we should hire like way more people that reach out to you personally and say, "Hey, I saw you get started in modal. I saw you working in the bio space. How can we help you? Here's some bio examples, you know, protein folding, whatever, right? Like we work with a bunch of other similar companies. You seem skeptical."

嗯，不，我一点也不怀疑。

Well, no, I don't at all.

我其实只是在想，我喜欢这种产品野心，但作为一个欧洲人，我会想，哇……你身处 B2B 领域，所以我觉得，你知道……但你会担心摊子铺得太大。

I'm actually just thinking about, like, I like the product ambition, but I go as a European who goes, wow... you're in a B2B space, so I think it's like, you know... but you worry about being stretched too thin.

是的。

Yeah.

就像你说的，你知道，为什么 Nebas 不能像你那样高效地做你所做的事？因为你太擅长了，而且你花了那么多时间。当你扩展到所有细分领域时，每个领域都有已经做了多年的玩家，而我会想……

To your point on, like, you know, why can't Nebas do what you do so efficiently? Because you're so good and you spent so much time as you expand into all the specialties, there are players in each of them who've been doing that for years, where I'm like...

我觉得那是真的。我不认为他们很擅长，很多人。但我的意思是，我有点同意你更高层次的前提。就像，如果有什么是超级关键且永远不会消失的，那就是优先级和资源分配，对吧？就像，你不能把水烧开的同时试图做所有事情。你需要极其无情。而拥有多产品战略，我认为是一个非常难的平衡。就像，你什么时候扩展到第二个产品或第三个产品？对吧？就像，我们花了四年才看到第二个产品的成功。

I think that's true. I don't think they're very good, a lot of them. But I mean, I sort of agree with your higher-level premise. Like, if there's anything that's super critical and never goes away, it's prioritization and resource allocation, right? Like, you can't boil the water and try to do everything at the same time. You need to be extremely ruthless. And having a multi-product strategy, I think, is a very hard balance. Like, when do you branch out into a second product or third product? Right? Like, it took us four years to see success with a second product.

那就是沙盒。

Which was sandboxes.

四年才看到第二个产品的成功？

Four years to see success with your second product?

我们推出五年就达到了十亿美元营收。

We launched five years to a billion in revenue.

是的。所以我们在 2023 年推出了沙盒。早期并没有看到很大的成功。我的意思是，有一些成功。你知道，不算糟糕，但直到 2025 年夏天，模型，我想，比如 4.5 那个版本，它们真的非常擅长生成代码和构建整个应用，然后它就爆发了。我的意思是，今天的沙盒是一个 3 亿美元年营收的业务。大约是我们营收的三分之一。所以那是一段疯狂的旅程。现在我们在思考，你知道，第三个和第四个是什么。我们刚刚推出了一款训练产品，我非常兴奋，正在做批处理产品，正在做强化学习，我们正在做各种其他东西。所以那种演变成平台的进化，它实际上就在那里。就像，如果你回到 Modal 早期的演示文稿，你知道，我整理了一些只是为了我自己的思考，它都在那里。就像我们有这个市场地图和所有这些，你知道，最终的产品和我们想象的很不一样。但对我来说，创建，你知道，专注于一个类别和一组解决方案，而不仅仅是成为一个单一组件的供应商，是非常重要的。

Yeah. So we launched sandboxes in 2023. Didn't see a lot of success in the early days. I mean, there was some success. You know, it wasn't horrible, but it was a relatively insignificant part of our revenue up until summer of 2025 when models, I think, like saw the 4.5 one of those, like they just got really good at generating code and building whole apps, and then it just exploded. I mean, sandboxes today is a $300 million run rate business. It's about a third of our revenue. And so that's been a kind of a crazy journey. And now we're thinking about, you know, what's the third and the fourth. We just launched a training product I'm super excited about, working on batch products, working on reinforcement learning, we're working on all kinds of other stuff. And so that evolution into becoming a platform, it's actually there. Like, if you go back to the very early decks of Modal, you know, I put together some just for my own thinking, it was all there. It's like we had this market map and all these, you know, it ended up being very different products than we thought. But it's so important for me to create, you know, to focus on a category and focus on a set of solutions, not just be a single widget vendor.

关于今天的沙盒市场，你希望你在开始并发布第一个产品时就知道的是什么？

What do you know about the sandboxes market today that you wish you had known when you started and released the first product?

它实际上就是这样的产品之一。我认为身处基础设施领域的好处之一，但奇怪的是也是缺点之一，就是你只能相信自己的直觉，判断什么是重要的原语，然后构建出色的构建块。我认为我们处于提供构建块的空间。我们试着观察工程师需要什么，而不是太仔细地倾听，因为他们会要求各种随机的东西，但要提炼出一些构建块。我们在 2021 年推出了无服务器函数产品。它在 2023 年随着生成式 AI 推理而取得成功。我们在 2023 年推出了沙盒产品，因为我们看到很多人使用大语言模型来生成代码，然后它在 2025 年取得了成功，就像我说的。所以我们对构建产品的很多思考一直都是，让我们推出好的构建块，然后它就爆发了。但回答你的问题，是强化学习，是氛围编程，后台智能体，也许未来还有其他东西。而那些是我们实际上没有预料到的。所以我认为基础设施往往就是这样。就像我们开始构建无服务器函数产品时没有预料到稳定扩散。它恰好成为了杀手级应用。

It actually is like one of those products. I think one of the benefits of being in infrastructure, but also one of the drawbacks in a weird way, is like you kind of just have to trust your gut on like what are important primitives and just build amazing building blocks. I think we're in the space of providing building blocks. We try to look at what do engineers need and not listen too carefully because they'll ask for all kinds of random stuff, but kind of distill it down into what are some of the building blocks. And we put out a serverless function product in 2021. It saw success in 2023 with Gen AI inference. We put out a sandbox product in 2023 because we saw a lot of people were using LLMs to generate code, and then it saw success in 2025 like I said. So a lot of how we think about building products has always been kind of like let's just put out good building blocks and then it exploded. But to answer your question, it was reinforcement learning, it was vibe coding, background agents, and maybe some other stuff too in the future. And those were things we actually didn't anticipate. So I think infrastructure is often like that. Like we didn't anticipate stable diffusion when we started building our serverless function product. It just happened to be the killer app.

你从中得到的启示是什么？唯一重要的是你必须在游戏中待足够长的时间，让你的时机正确。

What's your takeaway from that? The only thing that matters is you've got to stay in the game long enough for your timing to be right.

有时我认为那是对的。我的意思是，在某些方面，我们在时机上确实非常幸运。就像我们在 2021 年公司成立之前就推出了产品，然后，你知道，稳定扩散和 ChatGPT 在 2022 年出现。我们的时机真的很好，对吧？它创造了如此多的需求，而我们处于非常有利的位置。但其中一些也像基础设施一样，再次，我认为这是关于构建你预计未来会有用的产品，并且从你自己的经验来看，是你一直想看到的东西。我认为你必须相信自己的直觉。

Sometimes I think that's right. I mean, we've definitely been extremely lucky in certain ways with timing. Like we put out products like right before we started the company in 2021 and then, you know, stable diffusion and ChatGPT came out in 2022. We had really good timing with that, right? It just created so much demand and we were very well positioned. But some of it is also like with infrastructure, again, I think it's about building products that you anticipate will be useful in the future and from your own experience are things that you always wanted to see. I think you have to kind of trust your gut in that.

老兄，你在那里说到了资源分配，并且在你思考产品广度时，这一点非常非常重要。事后看来，你在哪些地方分配了资源，你希望你没有这样做，你如何反思？

Dude, you said there about resource allocation and that being very, very important as you think about this breadth of products. Where did you allocate resources where, with the benefit of hindsight, you wish you hadn't, and how do you reflect on that?

我认为我们应该更早地向语言模型推理分配更多资源。我认为我们有一段时间有点袖手旁观。我们没有意识到开源推理会变得多大。我们认为它会是一场逐底竞争，只是一个商品化空间。然后 DeepSeek 出现了，我们就像，我们必须快速行动。所以我们立即组建了一个团队，你知道，从一两个人做这个，到 12 个月后大约 30 人。现在我们是前三的提供商之一。

I think we should have allocated more resources towards LM inference like earlier. I think we're a little bit on the sidelines for a while. We didn't realize how big open source inference would be. And we thought it would be kind of a race to the bottom, just a commodity space. And then DeepSeek came out, we're like, we got to move very fast. And so we staffed up a team immediately, you know, went from like one or two people working on this to like 30 people 12 months later. And now we're one of the top three providers.

Jason Lanin，我的好朋友，我每周四和他一起做节目，他总是说，现在我们已经达到了开源大语言模型推理使用的顶峰。我们将看到前沿提供商直截了当地提供极其便宜的产品，就像他们现在这样。美国企业不想使用中国模型。这将继续。因此，实际上，在这两股核心盛行风的影响下，前沿将实际上研究，我们将看到开源大语言模型推理模型的减少和趋于平稳。

Jason Lanin, dear friend of mine who I do a show with every Thursday, he always says that now we've reached peak actually open source LLM inference usage. We're going to see frontier providers bluntly provide incredibly cheap products that they are now. US enterprises don't want to use Chinese models. That will continue. And so actually with those two core prevailing winds, frontier will actually research and we will see a reduction plateauing of the OS LLM inference model.

我认为那有点固定馅饼大小的心态。就像我认为，你知道，恕我直言，我确实认为两个空间在未来 12 个月及以后都会增长 5 倍、10 倍、20 倍。就像，所以很难说，你知道，哪个会比另一个更大，但我认为两者在未来几年都会以数量级增长。未来 12 个月增长 5 倍、10 倍、20 倍。

I think that's a little bit of a fixed pie size mentality. Like I think, you know, with all due respect, I do think both spaces will grow 5x, 10x, 20x over the next 12 months and beyond. Like, and so it's hard to argue like, you know, which one's going to be bigger than the other, but I think both will grow like in orders of magnitude over the next few years. 5x, 10x, 20x over the next 12 months.

是的，当然。

Yeah, for sure.

你刚刚达到了十亿美元营收。是的。干得好。

You just hit a billion in revenue. Yeah. Well done.

是的。

Yeah.

那感觉挺酷的吧？

Does that feel pretty cool?

感觉挺酷的。是的。我的意思是，你知道，就像我不知道。就像我是这个，你知道，瑞典人，你知道，在一个随机的地方长大……对我来说，这些数字是荒谬的，但就像是的。感觉很棒。

It feels pretty cool. Yeah. I mean, it's, you know, like I don't know. Like I'm this like, you know, Swedish guy that, you know, grew up in a random... I for me it's like these numbers are absurd, but like yeah. Like it feels great.

这钱也太多了。

It's so much money.

Erik，你真是——我没想到你这么讨人喜欢、这么酷。

Erik, you're so— I wasn't expecting you to be so likable and cool.

我不知道。就是很多个零。十亿里面有多少个零？

I don't know. It's like a lot of zeros. Like how many zeros are in a billion?

很多。你查看的时候，你的财富经理会告诉你的。12 个月后你会到什么位置？

It's a lot. Your wealth manager will tell you when you look at that. Where will you be in 12 months?

我不知道。我能说的是我们一直在增长。过去 12 个月我们增长了 50 倍。所以你可以外推一下。我是说，我不知道——市场到了一定程度就会变得非常大，然后开始稍微放缓。但以现在的增长率——增长率并没有放缓。我们每个月增长 30% 到 40%。

I don't know. What I can say is we've been growing. We grew 50x in the last 12 months. So you can extrapolate that. I mean I don't know— at some point markets just get very big and they start to slow down a little bit. But with the growth rate right now— the growth rates are not slowing down. We're growing 30-40% every month.

40 亿。

4 billion.

我是说，我绝对认为我们应该——我希望远远超过那个数字。

I mean, I definitely think we should— I hope to be way beyond that.

恕我直言，那 150 亿感觉就相当便宜了。就像如果我是大使，那现在这是我的地盘。就像如果你觉得你会达到 40 亿，比如说降到，我不知道，12 亿到底线。对。那他们实际上只付了，多少，明年年底底线的 14 倍。收入的 3 倍。

With the greatest of respects, 15 billion then feels like quite a cheap price. Like if I'm an ambassador, so this is now my ballpark. Like if you think you were going to be at 4 billion, say dropping, I don't know, a billion two to the bottom line. Yeah. Then they actually only paid what, 14x end of year next year bottom line. 3x top line.

对。对，是挺便宜的。

Yeah. Yeah, it's kind of cheap.

那感觉挺便宜的。

That feels pretty cheap.

我觉得是便宜。我是说，如果我不认为我们能增长并成为一家世代级公司，我就不会做这件事。我认为上升空间太大了。我认为这是有史以来最大的市场机会之一。我认为对算力的需求、对 AI 应用的需求，将会极其巨大，它们都需要一个运行的地方，我希望很多都会跑在 Modal 上。

I think it's cheap. I mean, I wouldn't do this if I didn't think we could grow and become a generational company. I think there's so much upside. I think this is one of the largest market opportunities that's ever existed. I think that the demand for compute, the demand for AI applications, it's just going to be enormous and they're all going to need a place to run and I hope a lot of them are going to run on Modal.

今天你还没有、或者很小的哪条收入线，你认为在三年后、五年后会非常重要？

What revenue line do you not have today or is tiny that you think will be very significant in three years time, five years time?

这是个好问题。我认为训练是我们刚推出的东西。我认为那可能会非常大。我希望它能在未来 6 到 12 个月内达到 1 亿美元的流水。后训练类似。在做批处理任务。我也对很多传统的数据仓库东西、数据库、工作流编排、特征存储比如向量数据库非常兴奋。这里面有太多部分，它们都能很好地契合在一起，这对我来说就是我如此兴奋的原因。一旦你开始在这个领域构建，就有太多天然的邻近领域。这是我的一个问题——我太容易兴奋——我不缺想法。在 Modal，我们有太多想做的方向——CI，很想进入持续集成，那是个非常有趣的空间。智能体框架，也许我们应该训练模型，也许——有 5000 个方向，我那颗注意力缺陷的大脑想往所有这些方向跑。你显然必须在优先级上无情，只在正确的时间做一两个事情。但我们可以走的方向太多了。

That is a good question. I think training is something we just launched. I think that could be very large. I hope that could get to 100 million run rate in the next 6 to 12 months. Post-training similarly. Working on batch jobs. I'm very excited also about a lot of traditional data warehousing stuff, databases, workflow orchestration, feature stores like vector database. There's just so many parts of this and they all kind of fit together very nicely and that to me is what I'm so excited about. Once you start building in this space there's so many natural adjacencies. This is a problem with me— I get excited about so— I have no lack of ideas. At Modal we have so many ideas of things we want to go— CI, would love to get into continuous integration, that's a very interesting space. Agent harnesses, maybe we should train models, maybe— there's just 5,000 directions and that's my ADD brain that wants to run in all of those directions. You have to be obviously ruthless with prioritization and only take on one or two things at the right time. But there's so many different directions we can go.

你妻子找你做决定真是太神奇了。她肯定得到的是非常不连贯的答案。

It's amazing that your wife calls you for decisions. She must get a very incoherent answer.

我知道。我知道。

I know. I know.

我只是想要一些窗帘。

I just wanted some curtains.

对。你知道，关于窗帘，我就喜欢——你选一个。我们看过《辛普森一家》，Rain Wolf Castle 成为总统，他们说：“你想选哪个？”你知道，一到七。他说：“我生来是领导的，不是读文件的。”然后选四号。

Yeah. You know, with the curtains, I just like— you pick one of them. We've seen Simpsons and like Rain Wolf Castle becomes the president and they're like, "Which one do you want to pick?" You know, one to seven. He's like, "I was born to lead, not to read." And number four.

对。我完全理解。

Yeah. I totally get that.

我能问一下——我们看到重心上确实发生了转变，资源分配也从预训练转向后训练。你对此怎么看？你觉得这是新常态吗？

Can I ask— we've seen this real shift, I think, in emphasis, but also in resource allocation from pre-training to post-training. How do you feel about that? And do you think that is the new normal?

我不知道。我认为后训练有效，是因为人们在现实生活中做的很多任务都有相当有限的反馈机制。所以，编程是个很好的例子，对吧？你可以用强化学习把编程做得很好，因为你可以在沙盒里运行代码，很多实验室正是为此在使用 Modal 沙盒。网络安全是另一个领域。数学显然有 Lean——我是说，我们见过一些很棒的产品——我觉得前几天那些 OpenAI 的证明挺不可思议的。法律也许在某种程度上可以做评估。我认为现在后训练——大家如此关注它，是因为有所有这些传统上只有人类能做的任务，但实际上智能体和 LLM 已经触手可及。但我不知道它是否会扩展到那之外。我不知道——外面有很多任务反馈循环长得多，我不确定后训练是否已经够得着。很多人问我软件工程师长期会怎样。我认为我们可能会把智能体能在 24 小时或一周内做的所有事情都自动化。但我感觉工程师的工作时间尺度更长，反馈循环也更长。所以我有种感觉，强化学习在这方面会开始遇到一些限制。但我已经错了很久了。所以也许我会错，它就继续下去，最终我们能在人类需要 6 个月的事情上做强化学习。我不知道。

I don't know. I think post-training works because there's a lot of tasks that people do in real life that have fairly finite feedback mechanisms. So, coding is a great example, right? You can do coding with reinforcement learning pretty well because you can run code in sandboxes and a lot of labs are running Modal sandboxes for that exact purpose. You start to look at cyber security is another area. Math obviously with Lean— I mean, we've seen some fantastic product— I think it's kind of incredible all these open AI proofs the other day. So law maybe to some extent you can do evals. I think post-training right now— there's such an incredible focus on it because you have all these tasks that traditionally only humans can do but actually well within reach for agents and for LLMs. But I don't know if it extends beyond that. I don't know— there's a lot of tasks out there that have way longer feedback loops that I don't know if post-training quite yet is within reach. A lot of people ask me like what's going to happen with software engineers in the long term. And I think we're probably going to automate everything that an agent can do within 24 hours or a week or something like that. But I kind of feel like engineers work on a longer time scale and they have feedback loops that are longer. So I have a feeling like reinforcement learning we're going to start to run into some limitations around that. But I've been wrong for so long. So maybe I'll be wrong and it just keeps going and eventually we can do reinforcement learning on things that take 6 months for a human. I don't know.

三年后工程师的一天是什么样的？

What does the day look like in 3 years for an engineer?

我对软件工程极其看好。我写代码 30 年了，见过编程语言、云计算以及各种框架的太多变化。每一次我们让程序员更高效，都只是增加了对软件的需求。世界上有太多地方迫切需要软件，所以我对软件以及软件工程的需求非常非常看好。我认为外界最大的误解之一是，开发者——就是把英语翻译成代码。那从来不是软件工程师的工作。首先是理解需求，并成为你为之构建的人类——归根结底是人类——之间的接口，试着弄清楚这在流程和工作流上到底意味着什么。而这会不断向栈的上层移动。60 年代，我们写汇编和打孔卡，然后转到 COBOL，再到 C++，再到 Python，现在我们用智能体。这就像是——这绝对是我过去 30 年见过的最大的变化。

I'm extremely bullish on software engineering. I've been coding for 30 years and I've seen so many different changes with programming languages and the cloud and so many different frameworks. Every single time we make programmers more efficient it just increases demand for software. There's so many parts of the world that desperately need software and so I'm very very bullish on software and the need for software engineering. I think one of the biggest misconceptions from the outside is that developer— it's about translating from English to code. That has never been the job of software engineers. It's to understand the requirements in the first place and be kind of an interface between the humans you're building for— humans at the end of the day— and trying to figure out what does that even mean in terms of processes and workflows. And that continuously moves up in the stack. Back in the 60s, we used to write assembly and punch cards and then we moved to COBOL and then to C++ and then to Python and now we're using agents. And that's just like one— and it's definitely the biggest thing that I've seen in the last 30 years.

但我确实认为，在软件和计算机的整个历史中，我们有很多很多的类比和很多平行之处。对我来说，这让我觉得，你知道，软件工程师在未来 10 年、20 年里会没事的。

But I do think we have many, many analogies and many parallels throughout the history of software and computers. To me, that makes me feel like, you know, software engineers are going to be fine for the next 10, 20 years.

Modal 每个月在编程智能体上花多少钱？

How much money do you spend on coding agents a month at Modal?

我想大概五十万左右吧。

I think a half a million or something like that.

你们有多少工程师？150 个。

How many engineers do you have? 150.

老兄，这低得让人担心。

Dude, that's so worryingly low.

你们应该更多地榨干 token。应该是这个数的 10 倍。

You should be token maxing more. It should be 10x that.

是啊，也许吧。我不知道。

Yeah, maybe. I don't know.

Brandon 说 McCord，呃，他们在推理上的花费是人类薪资的 3 倍。

Brandon said McCord, um, that they spend 3x their human salaries on inference.

是啊，是啊，是啊。我们还没到那个程度。但我要说，我们的软件工程师都极度痴迷智能体，包括我自己。我是说，我一直在用智能体写代码。

Yeah. Yeah. Yeah. We're not quite there yet. But I will say like our software engineers are extremely agent-pilled, including myself. I mean, I write code with agents like all the time.

但我觉得，如果你看 Anthropic 的 IPO，最大的问题是——别担心，我不会像彭博社那样把你拉进来问，你怎么看 Anthropic 的 IPO？我讨厌那种。

But I think like the big question if you're looking at like an Anthropic IPO and don't worry, I'm not doing a Bloomberg and like pulling you in to be like, what do you think of the Anthropic IPO? I hate those.

是啊，挺好的。AI 圈里没人不看好。

Yeah, it's great. It's no one in AI is like not a fan.

是啊，大家都在冲。呃，但我确实认为，一大块——可以说是主矿脉——是 Claude Code。它有 600 亿美元的营收。它在很大程度上决定了那 2 万亿美元的估值。呃，你看 Mark Benioff 说，他们今天把工程师薪资的 5%，也就是 3 亿美元，花在 Claude Code 上。现在，如果随着效率提升，这个数字大幅下降，诸如此类，那是个挑战。还是说它会继续上升，因为它吃掉越来越多工程师的工作流？

Yeah. Everyone is go. Um, but I do think a big, uh, the motherlode, so to speak, is Claude Code. It's $60 billion of revenue. It largely kind of, um, dictates the $2 trillion price. Um, and you look at Mark Benioff who says that they spend 5% of engineers' salaries, 300 million, on Claude Code today. Now if that comes way down as efficiencies improve, you name it, that's a challenge. Or will that go up as it just eats more and more of engineers' workflows?

是啊，它只会吃掉更多。我是说，我们又见过太多次了。你让某样东西更高效，你让某样东西更强大，它就解锁了五个新用例，需求最终只会上升。这就是过去 50 年计算机和软件的历史。但如果不是这么乐观，我就不会在这里。这就是我创业的原因，我就是这个世界上对软件和计算机最乐观的人，我就是热爱这个，我认为它只会继续增长。

Yeah, it's just going to eat more. I mean, like again, like we've seen this so many times. Like you make something more efficient, you make something more powerful, it just unlocks like five new use cases and the demand just goes up in the end. Like that's just the history of like the last 50 years of computers and software. But like I wouldn't be here if I wasn't such an optimist. That's why I started a company and like I'm just like the biggest like optimist that there is in the world about software and computers and like I just love this and I think it's just going to keep growing.

数据市场很有意思。我们有这么多玩家扩张得如此之快。

The data market is an interesting one. We have so many players that have scaled so fast.

是啊，太不可思议了。

Yeah, it's incredible.

如果我 20 岁，只想赚很多钱，我会出去创办一家环境公司，直接把数据卖给前沿实验室。那个领域能赚的钱太多了。有点不可思议。

If I was 20 and I just wanted to make a lot of money, I would go out and start an environment company and just sell data to frontier labs. There is so much money to be made in that space. It's kind of incredible.

我们再深入一层。你接下来会怎么做？你专攻某个特定市场。

Let's go a layer deeper. What would you do next? You specialize on a specific market.

我是说，编程领域是目前需求最大的地方。如果你能向实验室出售代码环境，你就能赚很多钱。那里仍然有很大的空间。

I mean, the coding space is where you see the most demand right now. Like if you can sell like code environments to the labs, like you can make a lot of money. Like there's a big space in that still.

仍然。是啊，是啊。我经常遇到初创公司。我天使投资了一堆。我是说，它们是很棒的生意。我不觉得——我要说的是，我的一条建议是，你今天能赚很多钱，但这是一只风投基金吗？你想踏上风投的跑步机吗？我不确定。我觉得这些公司今天会赚很多钱。我不知道它们在非常长期是否还能赚很多钱，但谁在乎呢？你知道，拿钱就是了。那是很棒的钱。它今天是一门极好的生意，这就够了。

Still. Yeah. Yeah. I meet startups all the time. I angel invested in a bunch of them. I mean, they're incredible businesses. I don't think—what I will say, like my one piece of advice is like, you make a lot of money today, but like, is this a venture fund? Like do you want to get on the venture treadmill? I'm not sure. Like I think these are companies that are going to make a lot of money today. I don't know if they're going to make a lot of money in the very long run, but like who cares? Like you know, take the money. It's like great money. Like it's a fantastic business today and that's all that matters.

我认为有数据市场，也有数据与服务。你看领先的供应商，他们一开始是数据采集方，然后就把数据卖给你，差不多就这样。而我认为他们现在真的走完了全生命周期，一直到部署、实施、评估、监控，具备真正的企业级手把手能力，类似于 FDS，他们会真正成为那个合作伙伴，确保完整落地。

I think there's data marketplaces and then there's data and services. And when you look at the leading providers, they started as the data acquirers and then just sell you the data and that's kind of it. And I think they've really gone full life cycle now to deployment, implementation, evaluation, monitoring with a real enterprise handholding ability similar to the FDS where they will really be that partner to ensure full implementation.

是啊，这就是你扩大利润率、在长期拥有某种生意的方式。

Yeah, that's how you expand margin and have some form of business in the long term.

当然。是啊。而且我认为从非常长期看，这会是一门艰难的生意，但今天你卖这些数据集可以有极好的现金流。

For sure. Yeah. And I think it's going to be a tough business in the very long run, but like today you can have extremely good cash flow selling these data sets.

我能不能——当我们想到那种与合作伙伴如此紧密合作的实施，你说过，Lynn 也从 Fireworks 说过，拥有自己的智能很重要。你能展开讲讲这看起来是什么样吗？因为以这种方式为公司做专用模型，陪审团还没定论。我们有 Harvey 说这太棒了、非常惊人，然后又有 Lora 说这很荒谬——底层模型变化太快，这是错误的策略。

Can I—when we think about that kind of implementation working with partners so closely, you said, and Lynn has said too from Fireworks, about the importance of owning your own intelligence. Can you unpack how this looks? Because specialized models for companies in this way, kind of the jury is still out. We have Harvey who says that it's fantastic and it's been amazing, and then you have Lora who say ridiculous—the underlying models move so fast, it's the wrong strategy.

是啊，对某些东西来说，拥有自己的智能是好事，对另一些则不是。我觉得这并不重要。我们有很多客户在 Modal 上运行专有模型，他们的客户，你知道，用的就是专有模型。我不觉得每个人都拥有自己的智能是合理的。

Yeah, for some stuff it's good to own your own intelligence, for other stuff it's not. Like I don't think it matters. Like we have a lot of customers that run proprietary models on Modal, like, and their customers, you know, they use a proprietary model. Like I don't—I don't think it makes sense for everyone to own their own intelligence.

为什么对你来说合理，而对别人不合理？

Why does it make sense for you to versus not?

我认为对某些用例，你可以拿一个开源模型，微调它，就能得到非常好的性能。好处是，你知道这个微调模型基于什么。它不会变。他们不会把模型从你手里拿走。拥有那种价值、那种一致性，以及持续调整它的能力，而不用管底层模型变化——这在一些前沿实验室那里确实发生过——这很好。我支持这个。但你知道，我认为还有很多其他东西对人很重要。拥有自己的智能只是众多因素之一。我认为延迟重要，成本重要，性能重要，可改造性重要，拥有一个有很多相邻服务的平台——这些也很重要。拥有自己的智能，是啊，我支持，但你知道，还有很多其他东西。

I think for certain use cases, you can take an open source model and you can fine-tune it and you can get really good performance. And the nice thing is you know what the fine-tuned model is based on. It's not going to change. Like they're not going to take away the model from you. It's nice to have that value, like that consistency, and the ability to like continuously tweak it without, you know, the underlying model changing, which has happened in some cases with the frontier labs. I—that's like, you know, that's nice, like, you know, I support that. But like, you know, I think there's many other things that matter to people. You know, owning your intelligence is like just one thing out of many. Like I think latency matters, I think cost matters, performance, I think, you know, hackability, I think, you know, having a platform of a lot of adjacent services—like these things matter too. Owning your own intelligence, like yeah, I'm supportive, but you know, there's many other things.

这难道不是有点违背你的生意吗？

Isn't that kind of going against your business?

不，我不这么认为。我——我们——

No, I don't think so. I—we—

我喜欢这份坦诚，但老兄，说“是的，每家公司都会需要它，而我们是帮你实现的合作伙伴”对你更有利。我都有点把球喂到你脚下了，而你——

I love the honesty, but like, dude, it would be in your favor to say yes, every company's going to need it and we're the partner to help you. I kind of teed you up there and you—

嗯，我是说，也不尽然。我是说，你想想 TypeSafe 和 JVM，对吧？你知道，那是一种高度专有的模式。我希望人们用 JVM。我是说，它运行在 Modal 上，所以我们有很多那样的客户，他们本身就是专有模型，托管在 Modal 上，服务他们的客户，我也非常非常支持那个。

Well, I mean, not really. I mean, if you think about TypeSafe with JVM, right? Like, you know, it's a highly proprietary model. Like, I want people to use JVM. I mean, it runs on Modal, like, so we have many customers like that, that themselves are proprietary models that are hosting on Modal, that are serving their customers, and I'm very, very in favor of that too.

我认为 Modal 的美妙之处始终在于它是一个通用平台，我们支持许多不同的东西。对我来说，美妙之处就在于此：如果你想用开源模型，那很好——我们会帮你。有很多客户在开源模型之上构建了令人惊叹的 AI 应用，对它们进行微调、强化学习，或者一堆相邻的沙盒用例或训练什么的。但也有很多公司建立了非常成功的业务——Suno、Runway、Typesafe 以及许多其他公司——在 Modal 上运行自己的专有模型。他们的客户并不在意。这让我们处于非常有利的位置。

I think the beauty of Modal is always that there's a general-purpose platform and we support so many different things. To me, that's the beauty: if you want to use open-source models, that's great—we'll help you. There are many customers that build incredible AI applications on top of open-source models, fine-tuning them, reinforcement learning, or a bunch of adjacent sandbox use cases or training whatever. But there are a lot of companies that have built highly successful businesses—Suno, Runway, Typesafe, and many others—running their own proprietary model on Modal. Their customers don't care. That puts us in a really good spot.

在快速提问之前，我还想问两个问题。

I want to ask two more questions before I do a quick fire.

好的。

Okay.

模型路由，正如亲爱的美国人所言。

Model routing, as the dear Americans say.

那是一门有护城河的生意吗？

Is that a defensible business?

你们在英国怎么说？路由。

What are you saying in the UK? Routing.

路由。

Routing.

我喜欢这个说法。

I love that.

嗯，我的意思是，我们其实是原版语言；美国只是拿走了我们的语言并把它糟蹋了。但没错。我的问题是：这是价值链中一个有护城河的环节，还是一种商品化的产品，每个人都可以自己构建？

Well, I mean we are actually the original language; America just took ours and bastardized it. But yes. My question to you is: is that a defensible segment of this value chain, or is that a commoditized product that everyone just builds themselves?

我认为在模型路由领域建立成功的企业是可能的。我认为今天作为独立产品你可以做到。我怀疑长期来看它是否必须成为更全面平台的一部分。归根结底，我认为这些公司想要的是评估、微调的方法、对性能（如延迟、吞吐量）的更多控制，所有这些。我认为这是当今合理的事情之一。这些是非常好的业务。我个人投资了 OpenRouter，他们取得了很好的成果。我实际上认为他们作为独立公司本可以做得更好。但非常长期来看，感觉它属于那种类别——这是我对 AI 中一切事物的看法——它真的应该成为更大、更全面工具集的一部分。

I think it's possible to build a successful business in model routing. I think you can do that today as a standalone product. I wonder if long-term it has to be part of a more comprehensive platform. At the end of the day, I think what these companies want is eval, ways to fine-tune, more control over performance like latency, throughput, all these things. I think it's one of those things that makes sense today. These are very good businesses. I invested in OpenRouter personally, and they had a great outcome. I actually think they could have done even better on a standalone basis. But in the very long term, it feels like one of those categories—and that's my view on everything in AI—it should really be a part of a larger, more comprehensive set of tools.

恭喜，你进入了 OpenRouter 董事会，收购要约来了。请论证为什么他们会这样做。是的。请论证。

Congratulations, you're on the OpenRouter board and the acquisition offer comes in. Please make the case for why they would have done that. Yeah. Please make the case.

我认为 OpenRouter——增长率和收入——坦率地说还有品牌，他们有一个非常强大的品牌。还有 500 个其他竞争对手，但这个领域增长如此之快，收入如此之多。我会告诉他们继续前进。

I think OpenRouter—the growth rate and the revenue—and frankly the brand, they have a really strong brand. There are 500 other competitors, but the space is growing so fast and there's so much revenue. I would have told them to keep going.

最后一个。我认为一个非常能说明问题的问题是：你相信什么不利于你的业务？例如，我认为今天的种子投资有点死了，这种“2 人 10 万，慢慢构建 18 个月”的想法。

Final one. I think a really telling question is: what do you believe that does not go in the favor of your business? And so for example, I think that seed investing today is kind of dead and this idea of like two on 10 and we'll slowly build for 18 months.

嗯，这有点糟糕。如果你想赚钱，这有点糟糕，对吧？但如果你想与工程师或创始人建立关系，那很棒。

Well, it's kind of bad. Like it's kind of bad if you want to make money, right? But it's great if you want to develop a relationship with engineers or, I mean, with founders.

当然。但这与我所做的相悖。我

Sure. But that goes against what I do. I'm

你反对与……的关系

You're against the relationship with

不，我是说我认为种子投资的死亡，2 人 10 万构建 18 个月已经死了。

No, I'm saying like I think the death of seed investing, the two on 10 and build for 18 months is dead.

是的。不，我明白你在说什么。

Yeah. No, I understand what you're saying.

那么，你的会是什么？

So, what would yours be?

我不知道。我的意思是，我有一个想法：显然在某个时候，AI 这东西增长如此之快，我们在这里，我们将去那里，然后会有一次回调，你知道，股市 20%、30% 的回调。每个人都会惊慌失措，人们会说这是世界末日。我几乎确定这会发生，但我如此看涨，因为之后我们会继续前进。所以这将是我认为的一件事：会有回调。我对此没问题，因为我认为长期机会如此之大。如此之大。

I don't know. I mean, I think one thought I have is: obviously at some point this AI stuff is growing so fast and we're here and we're going to go to here and then there's going to be a pullback, you know, 20%, 30% in the stock market. Everyone's going to freak out and people are going to say it's the end of the world. I'm almost certain it's going to happen, but I'm so bullish because then we're going to keep going. So that would be one thing that I'm like, you know, there's going to be a pullback. I'm okay with that because I think the long-term opportunity is so big. It's so large.

完全同意。你接受老股转让吗？

Totally agree. Do you take secondaries?

很少。

Very little.

为什么？

Why?

我个人不需要很多钱。我对生活很满意，我有足够的钱。所以个人而言，我实际上支持员工这样做。我认为作为创始人，让股权尽可能有价值非常重要。做到这一点并告诉未来候选人的一种方式是：不，这不仅仅是纸面财富，这是真钱，就是有进行员工老股转让的先例。然后你可以指出，你知道，上一轮我们让员工卖了这么多或什么的，最多 10%。我认为必须有护栏。你不希望员工卖太多。你希望他们仍然有大量切身利益，并非常有动力让公司成功。但我认为这是一种将股权变成流动性很强的硬通货的方式。

I personally don't need a lot of money. I'm pretty happy with life and I have enough. So personally I actually support it for my employees though. I think it's quite important actually as a founder to make the equity worth as much as possible. And one of the ways to do that and to tell future candidates that no, this isn't just paper money, this is real money, is to have a precedent of doing employee secondaries. Then you can point to, you know, last round we let employees sell this much or whatever, up to 10%. I think you have to have guardrails. You don't want employees to sell too much. You want them to still have a lot of skin in the game and be very incentivized to make the company successful. But I think it's a way to create a very liquid hard currency out of the equity.

我完全同意。当你考虑人才市场更具竞争力时，它非常

I totally agree. And as you think about a talent marketplace more competitive, it's very

那里竞争极其激烈。是的。

It's extremely competitive out there. Yeah.

这疯了吗？

Is it nuts?

疯了。疯了。我的意思是，我们看到的报价——我们失去了去一些实验室的新毕业生，比如，谈到 2000 多万。我的意思是，这些是杰出的人。杰出。

It's nuts. It's nuts. I mean, the offers we see—we've lost new grads going to some of the labs for like, talking, you know, 20 plus million. I mean, these are exceptional people. Exceptional.

什么造就了杰出的新毕业生？

What makes an exceptional new grad?

就是聪明的人，你知道？我一看就知道。他们只是极其渴望、勤奋、了解技术的一切。

Just sharp people, you know? I know it when I see it. They're just extremely hungry, hardworking, know everything about technology.

我的意思是，他们去一些大实验室。就这样？他们去 OpenAI。

I mean, they go to some of the big labs. Is that it? They go to OpenAI.

是的，大实验室。是的，他们付很多钱。

Yeah, big labs. Yeah, they pay a lot.

但就这样？人才来自那里吗？

But is that it? Is that where the talent comes?

还有几个其他的，但主要是那些人以支付非常多的钱而闻名。

There's like a few other ones, but mainly those people are known for paying very a lot of money.

而你不会付那么多。

And you won't pay that much.

我们付很多。我们可以非常激进。我们希望支付市场第 95 百分位的水平，我知道有几家公司可能会付得稍多一点。我认为我们的股权有更大的上升空间。回到我关于老股转让所说的，我认为出于这个原因，在股权货币中创造真正有形的价值也很重要，这使我们的报价看起来非常非常强。再加上我们有很大的上升空间。我的意思是你说得对，我们的股票看起来很便宜。我有点同意我们的股票有很大的上升空间。

We pay a lot. We can go very aggressive. We want to be out there paying 95th percentile for in the market and there's like a few companies that I know like they're probably going to pay a little bit more. I think we have way more upside in the equity. And going back to what I said about secondaries, I think it's important for that reason also to create a real tangible value in the currency in the equity that makes our offers seem very very strong. Plus the fact that we have a lot of upside. I mean you said it right like our stock seems cheap. I kind of agree there's a lot of upside in our stock.

所以这让我们不必非得和那些实验室对标，但我认为我们能打造一个有说服力的卖点，说明我们的薪酬方案基本上比实验室更好，而且我们愿意下很大功夫去招揽最优秀的人才。

So that enables us to not necessarily have to match with the labs, but I think you know create a compelling sales pitch for why our comp package is basically better than the labs, and we're willing to go very far to get the best talent.

好，我们来做一轮快问快答，不然我能跟你聊一整天。你有一个很棒的的天使投资组合——Open Router，你之前提到过 Lovable。你只能把所有钱投到一个项目里。

Okay, we're going to do a quick fire round, because otherwise I could talk to you all day. You have a great angel portfolio — Open Router, you mentioned Lovable before. You can put all of your money into only one.

我很喜欢 Turbuffer。我跟 Simon 很熟，他是个丹麦人。所以有这层斯堪的纳维亚的联系。他是个非常出色的创业者。我就是很喜欢这个生意。

I love Turbuffer. I know Simon really well, as a Danish guy. So you have the Scandinavian connection. Just a fantastic entrepreneur. I just love the business.

那个生意能做多大？

How big can that business be?

这是个好问题。我是说，它确实有些专门化。他们在做向量数据库。我认为这可能成为一个非常大的品类。如果我来经营那家公司，我会有 500 个不同的想法，去扩展到数据库、数据仓库、其他类型的查询系统等许多相邻领域。它有可能成为一个大品类。

That's a good question. I mean, it is somewhat specialized. They're doing vector databases. I think it could be a very large category. If I ran that company, I would have 500 different ideas of how to expand into many adjacent things around databases, data warehouses, other types of querying systems. It could be a large category.

过去 12 个月里，你改变想法最大的是什么？

What have you changed your mind on most in the last 12 months?

大概是沙盒这个领域的规模。我觉得这有点疯狂。我看着营收，它一直在涨，我就想，好吧，它现在显然该停下来了。我觉得我对营收在搞某种反向心理，因为它就是一直涨。我就想，这是怎么回事？它一直在增长、增长，而我已经错了太多次——还是往好的方向错。所以那个领域变得这么大，真是不可思议。

Probably the magnitude of the sandbox space. I think it's kind of insane. I look at the revenue and it just keeps going up, and I'm like, well, clearly it has to stop right now. And I think I'm doing some sort of reverse psychology on the revenue because it just keeps going up. I'm like, what is going on? It just keeps growing and growing, and I've been wrong so many times — in a good way. So it's incredible how large that space has gotten.

你认为今天 AI 领域最过热的品类是什么？

What do you think is the most overheated category within AI today?

天哪。我不知道。我不觉得有什么是过热的。我要说的是，让我有点担心的一件事，是那些依赖持续注入股权资本的公司。我想我不会投资它们。有很多这样的公司会让我担心——任何一旦风险投资收缩、就会难以再融资的公司。我认为重要的是，你把每一轮融资都当作最后一轮来融，并且有一条真正通往商业化和营收的路径。

Oh boy. I don't know. I don't think anything is overheated. I will say one of the things that I'm a little nervous about is companies that are dependent on continual infusion of equity capital. I think I would not invest in them. And there's a lot of them that I would be nervous about — anything where if there's a venture capital pullback, they will struggle to raise more money. I think it's important you raise every round as if it's your last round, and there's a real path towards commercialization and revenue.

你有两个孩子。

You got two kids.

两个孩子。

Two kids.

关于做个好爸爸和做个好 CEO，最大的心得是什么？

Biggest lesson on being a good dad and being a good CEO.

有时候我觉得这两者挺相似的。你是指如何平衡，还是指单纯做个好爸爸？

Sometimes I feel like it's kind of similar. Do you mean in terms of how to balance it, or do you mean in terms of just being a good dad?

做个好爸爸需要什么？

What does it take to be a good dad?

为他们的人生成功打好基础。我总是琢磨该给孩子多大程度的推动。我不是虎爸，但我认为你得稍微推孩子一把，让他们多接触一些东西。我不想在未来出现这种情况：孩子回来跟我说，“爸，我希望你当初逼我去做 X，或者希望你逼我去做 Y。”但我也不想出现这种情况：孩子 20 岁时回来跟我说，“爸，你当初为什么逼我们去做 X，或者为什么逼我们去做 Y？”所以，稍微推一推。我让他们学钢琴。我也不知道为什么，但我觉得这是学一样东西的好办法，我希望他们 20 岁、30 岁时会真正感激。他们没那么喜欢，但我觉得这类事情你得稍微推一把，但别太过。我父母非常宽松，我发现了数学和编程，他们就说，“Erik，你做那么多数学干嘛？你应该像别的孩子一样出去玩。”所以某种程度上，我其实希望他们当初逼我更狠一点，但反正我自己逼自己逼得很狠，所以也无所谓。所以，就是恰到好处的推动。

Setting them up for success in life. And I always think about the right amount to push the kids. I'm not a tiger dad, but I think you need to push the kids a little bit to get them exposure. I don't want to create a situation in the future where my kids come back to me and say, "Hey, Dad, I wish you'd made me do X or I wish you made me do Y." But I also don't want to create a situation where my kids come back to me when they're 20 and say, "Hey, Dad, why did you make us do X or why did you make us do Y?" So a little bit of pushing. I make them play piano. I don't know why, but I think it's a good way to learn something that I hope when they're 20, 30 they will actually appreciate. They don't like it that much, but I think with stuff like that you've got to push them a little bit, but not too much. My parents were very lax, and I discovered math and programming, and they were like, "Erik, why are you doing so much math? You should go out and play like the other kids." So in a way I actually wish they would have pushed me even harder, but I pushed myself so hard anyway, so it doesn't matter. So, the right amount of push.

你是怎么把自己逼得这么狠的？抱歉——你其实来自一个从小被告知“出风头不好”的国家。我忘了那个词了，但我交过一个瑞典女朋友。

How did you push yourself so hard? Sorry — you're actually from a nation where you're told that it's bad to stand out. I'm forgetting that word, but I had a Swedish girlfriend.

詹特法则。

Jantelagen.

对。就是那种出风头不好的观念。

Yeah. Where it's like it's bad to stand out.

对。别以为自己很特别。

Yeah. Don't think you're special.

对。为什么？

Yeah. Why?

但我不确定那是真的。我觉得关于斯堪的纳维亚文化有很多误解。我确实认为，人们很尊重那些追随梦想、去做事情的人——不管是音乐、写书、做家具、艺术，还是创业。我其实觉得那很受尊重。我认为这种自我实现、追随梦想是很被看重的，我觉得这就是非常宽松的文化的另一面——你某种程度上鼓励人们去追随梦想。而我的梦想就是做数学和编程。我热爱那个，所以我就去做了。

But I don't know if that's true. I think there's a lot of misconceptions about Scandinavian culture. I do think there's a lot of respect for people who follow their dream and do things — whether it's music or writing books or making furniture or art, or being an entrepreneur. I actually think there's a lot of respect for that. I think this self-actualization and following your dreams is highly regarded, and I think that's the flip side of having a very lax culture — you kind of encourage people to follow their dreams. And my dream was doing math and programming. I love that, so I did that.

除非你来自挪威，那里的税收结构基本上让你没法创业。

Unless you're from Norway, where a tax structure exists that basically makes it impossible to start a company.

是啊，我是说挪威确实有些问题。

Yeah, I mean Norway has some issues.

他们也不太工作。

They also barely work.

是啊，这非常对。但我们确实有一个主权财富基金。

Yeah, this is very true. But we do have a sovereign wealth fund.

是啊，他们很有钱。

Yeah, they are rich.

谢天谢地。

Thank god for that.

也许它像是第一个后经济时代的社会。我不知道。

Maybe it's like the first post-economic society. I don't know.

几乎字面上就是，从这个词的多重含义来说。

Literally almost, in multiple meanings of the word.

最后一个问题。你最期待什么？什么都可以。

Final one for you. What are you most excited for? It can be anything.

我看好慢性病的药物研发。我妈妈——

I look at drug discovery for chronic conditions. My mom—

其实，药物研发是我觉得酷得不得了的一个应用。我迫不及待想服用 AI 生成的药物。我老了，皮肤也变差了，每次喝啤酒第二天都难受得不行。把所有我能吃的 AI 生成药物都给我吧。我想活得久，我想活得健康。而且我觉得——

Actually, drug discovery is one of the applications I think is so incredibly cool. I can't wait to take AI-generated drugs. I'm old and my skin is getting bad, and every time I drink a beer I feel horrible the day after. Give me all the AI-generated drugs I can take. I want to live long, I want to live a healthy life. And I think—

你觉得那真的是一个现实的未来吗？

Do you think that is actually a realistic future?

我觉得是，但考虑到 FDA 和其他因素，这需要一段时间。不过我认为今天在 AI 辅助药物设计方面有太多惊人的进展。有太多很酷的东西，太多东西了。

I think so, but I think with the FDA and other things it's going to take a while. But I think there's so much incredible progress today in AI-aided drug design. There's so many cool things, so many things.

老兄，我真的很享受这期节目，说实话我没想到会这么享受。你这么有趣、开放、有魅力。你太棒了。非常感谢你。

Dude, I've so enjoyed this show in a way that I honestly wasn't expecting. You've been this fun, open, charismatic. You've been amazing. So thank you so much.

非常感谢你邀请我。这非常有意思。

Thank you so much for having me. This is a lot of fun.