The Untold Story of Higgsfield | $1B ARR in 18 Months | Why Moats in AI are BS | CEO, Alex Mashrabov

On average at Higgsfield, person on the team spends over $10,000 a month on various models. So internal usage of models a month is over 4 million.

在 Higgsfield，团队成员平均每人每月在各种模型上的花费超过 $10,000。因此，公司内部每月使用模型的费用超过 4 million。

Higgsfield. This is the story that no one has told in startups yet. The company has just hit a billion in revenue. It is the fastest growing company in consumer land to hit this milestone. It even surpassed Cursor.

Higgsfield 的故事，此前还没有任何一家创业公司讲述过。这家公司的营收刚刚达到 a billion，是消费领域达到这一里程碑速度最快的公司，甚至超过了 Cursor。

Alex, the founder, is an incredible genius. This is the story that you don't know that you need to know. My parents told me that I must get to the United States cuz this is the place where technology matters. By the age of 19, I was able to get to top three in the world in competitive programming. I just caught a guy who spent over 30k in a week on Astra model. Many people spend over 10,000 in a week. Ready to go.

创始人亚历克斯是一位不可思议的天才。这是一个你尚不了解、却必须了解的故事。我父母告诉我，我必须去 United States，因为那里才是技术真正重要的地方。到 19 岁时，我已经跻身全球竞技编程前三名。我刚刚发现，有个人一周内在 Astra 模型上的花费超过了 30k。很多人一周的花费都会超过 10,000。让我们开始吧。

Alex, I am so excited for this dude. We were talking downstairs and I said, I don't think the Higsfield journey has been told before and it's it's an amazing journey. So, thank you so much for joining me today.

亚历克斯，我非常期待这次对谈。刚才在楼下聊天时，我提到，Higgsfield 的这段历程似乎从未有人讲述过，而这确实是一段精彩的旅程。非常感谢你今天来参加我的节目。

Uh that's very special opportunity for us. Thank you for having me. Obviously, your story is inspiring as well, like how social media has become like an elevator for you, opportunity to create fun and so on. Dude, it's very kind of you to say. I do just want to go back though because you're not the Stamford, Silicon Valley, born and bred engineer.

对我而言，这是一次非常特别的机会。谢谢你的邀请。你的故事同样很励志，例如社交媒体如何成为你的上升通道，让你有机会创造乐趣等。你这么说太客气了。不过，我确实想先回到之前的话题，因为你并不是那种在斯坦福或硅谷土生土长的工程师。

You were a competitive programmer in Kazakhstan. Can you just take me back?

你曾在哈萨克斯坦参加竞技编程。能带我回顾一下那段经历吗？

How did you first find and fall in love with computers and become a programmer so early?

你最初是如何接触并爱上计算机，又是如何在这么小的年纪成为程序员的？

So, first you need to understand where I come from. So my father is from Usbekiststan. Usbakistan is a country in central Asia where like if a family of five people makes $1,000 a month, it's considered to be wealthy. So it's like not very high standards of living unfortunately. So um but both my parents are professors of mechanical engineering. Since I remember myself since I was eight, my parents told me

首先，你需要了解我的出身。我父亲来自乌兹别克斯坦。乌兹别克斯坦是一个位于中亚的国家，在那里，如果一个五口之家每月能赚 $1,000，就算得上富裕。遗憾的是，当地的生活水平并不高。不过，我的父母都是机械工程教授。从我大约八岁、刚开始记事时起，他们就一直告诉我：

that I must get to the United States because this is the place where technology matters.

我一定要去美国，因为那里才是技术真正重要的地方。

So um my mother had to work three jobs because basically my education was to compete in programming competitions all the time and to go to various educational camps where I could learn from the best like certain data structure data structures algorithms and so on. Can I ask you a question? Did you feel pressure as a child competing being pushed into these environments when you were so young?

为了供我接受教育，我母亲不得不打三份工。我的教育主要包括参加各种编程竞赛和教育训练营，在那里学习最优秀的知识，例如某些数据结构和算法等。我想问一个问题：你小时候参加竞赛、进入这样的环境时年纪还很小，会感到压力吗？

Absolutely. Uh but and and and I'm very grateful to my parents that they showed me the path really from that from that early on. Um definitely when you come from this part of the world think about post Soviet countries uh India China like getting to the top of the rankings in any competition in any international competition is the only way to really break out. So by the age of 19 I was able to get to top three in the world in

当然会。不过，我也非常感激父母，是他们从小就为我指明了这条路。对于来自这一地区的人——例如后苏联国家、印度和中国——在任何竞赛或国际赛事中跻身排行榜前列，都是让自己真正脱颖而出的唯一途径。因此，到19岁时，我已经跻身全球竞技编程前三名。不过，后来情况发生了变化。

competitive programming. But then instead of pursuing like um like academical career decided to do startups. [laughter] I'm sure your parents were thrilled. Uh can you take me to that decision? Like this is like the penultimate moment.

我没有选择追求学术生涯，而是决定创办初创公司。[笑声]我相信你父母一定非常高兴。能谈谈你是如何做出这个决定的吗？这是一个至关重要的时刻。你为此努力了19年，而你的父母……

You've worked 19 years for your parents have told you this is like the mother load. This is the thing and you're like I'm going to go and do this really risky thing called a startup at this point like what happens then?

他们一直告诉你，这就是终极目标，也是你应该做的事。然而，你却决定去做一件风险极高的事，也就是创办一家初创公司。当时以及后来发生了什么？那就让我带你回到2014年。

So let me take you back to 2014.

我非常幸运，有机会参与一项研究：

I was very fortunate to work on pre-transformer architecture neural nets and I was primarily just doing optimization make it run faster um parallel across multiple machines and so on and um I was and we actually build state-of-the-art system for language translation from English to Russian and Russian to English apparently talent wars were a real thing even back then a lot of my teammates were hired by Deep

那是 Transformer 出现之前的神经网络架构研究。我的主要工作是进行优化，使其运行得更快，并能够在多台机器上并行运行。我们还构建了当时最先进的语言翻译系统，可以进行英语到俄语以及俄语到英语的互译。即使在当时，人才争夺战也确实存在，我的很多队友都被 DeepMind 和 Meta 挖走了。不过，说到我真正的热情，

Minds and Meta and uh but My passion was actually different. I was very very surprised to learn when I come to to for the first time how quickly Uber actually spread out. And I was thinking if like this app can take over the world so quickly and transform the whole industry, maybe what's going to happen is that mobile phones are going to become the most used devices in the world. Maybe there is going to be a

它其实在别处。第一次来到这里时，我非常惊讶地发现，Uber 竟然扩张得如此之快。我当时就在想，如果这款应用能够如此迅速地席卷全球并改变整个行业，那么接下来会发生什么？也许，手机将成为全世界使用最广泛的设备。未来也许会出现一种人人都身处其中的情形：

version of the future where everyone is going to be spending most of their time in their life watching AI generated videos on the phones cuz I mean who else is going to produce videos for for the phones? Maybe it's going to happen with AI.

人们会把一生中的大部分时间花在手机上，观看人工智能生成的视频。否则，又有谁会专门为手机制作视频呢？也许这会通过人工智能实现。人工智能。

Okay. And so that was the company that we built before that you sold to Snap.

明白了。所以，这就是我们之前创办、后来卖给 Snap 的那家公司。

So yeah, so the company was called a factory. Um was fortunate to meet Mahi 2018. He's co-founder of Higgsfield and he is a like veteran of Silicon Valley went through ups and downs and um sold it to Snap for 100 for million for 166 million and um then I was leading Jenny there pause no offense dude you come from um you know a family of incredibly ambitious parents who push you to do well and you just skipped the moment

是的，那家公司叫 AI Factory。我很幸运在 2018 年遇到了马希。他是 Higgsfield 的联合创始人，也是一位久经沙场、经历过大起大落的硅谷老兵。后来，我们以100……166 million 的价格把公司卖给了 Snap。之后，我在那里负责生成式人工智能。——先停一下。无意冒犯，但你出身于一个父母都极具进取心的家庭，他们一直督促你取得好成绩，而你刚才竟然直接略过了这件事。

where you sell for 166 million It's a lot of money. Um, how did that feel when you did it?

公司以 166 million 卖掉，这可是一大笔钱。交易完成时，你是什么感受？

We both remember these times where the capital for AI companies was not really that much available and when and AI multiples were not like 200 to revenue as they are today but closer to zero cuz AI was not a topic. So there was like severe del dilution which we experienced. So you [laughter] just to calibrate. So can you okay what was around?

我们都还记得，当时人工智能公司其实很难获得多少资本，估值倍数也不像今天这样能达到收入的 200 倍，而是更接近于零，因为人工智能当时根本不受关注。因此，我们经历了非常严重的股权稀释。所以……[笑声]为了让大家有一个直观的概念，你能否介绍一下当时的大致情况？

No look I mean back then rounds like rounds of like$12 million having like$12 million in investments was considered to be really good. Uh but it but it was still an opportunity for me to finally go to the United States. So after the acquisition I permanently moved to uh first to LA and then to Silicon Valley and my dream simply came true.

当时一轮融资的规模如果在 $12 million 左右，能够获得 $12 million 的投资，就已经被认为非常不错。不过，这仍然给了我一个终于能够前往美国的机会。公司被收购后，我便永久移居美国，先住在洛杉矶，之后又去了硅谷。我的梦想就这样实现了。

Was it what you thought it would be?

那里和你想象中的一样吗？

That's a good question. So um as San Francisco is definitely a place where no one judges by race, nationality and so on and that's that's um that's truly phenomenal. There is definitely a meritocracy in a sense that it's possible to meet anyone but in the same time what I see across Silicon Valley investors it's extremely consensus driven. So um I mean I think that last part I expected to be different but then

这是个好问题。旧金山绝对是一个不会有人根据种族、国籍等因素评判你的地方，这一点确实非常了不起。那里确实存在某种意义上的精英制度，你有机会见到任何人。但与此同时，在我接触过的硅谷投资者中，从众倾向极其明显。我原以为最后一点应该有所不同，但后来改变了看法。

I read the book about the law of capital and I realized this is just how the world works.

读完那本讲资本法则的书后，我才意识到，世界就是这样运转的。

So then tell me we have sold to Snap we're now in the US this is the moment you wanted how does Higsfield come to be back then like Snapchat 2020 was uh really growing so so quickly and the face filters which my team has built was driving most of daily new users. What what's important is that um these face filters we were able to manage to run on mobile devices. So it was virtually for free for Snapchat. It's not like current

你接着说。我们已经把公司卖给了 Snap，也来到了美国，这正是你一直期待的时刻。Higgsfield 是如何诞生的？大约在 2020 年，Snapchat 正在飞速增长，而我团队开发的面部滤镜带来了大部分每日新增用户。关键在于，这些面部滤镜可以在移动设备上运行，因此对 Snapchat 来说几乎没有成本，这与现在不同。

LLM tokens cost. Um and but but it and it and it scaled to hundreds of millions of people throughout the world. And it was truly phenomenal to me to build a product which is still probably the most used consumer media AI product. But then um but then what I realized is that there are a lot of unmet needs on advertising sites. Average company cannot figure out how to be relevant on social media. So and this is a major gap

如今，大语言模型的 token 需要付费。当时，这款产品的规模扩展到了全球数亿用户。对我而言，能够打造出这样一款产品确实非同凡响。时至今日，它可能仍是用户数量最多的面向消费者的媒体人工智能产品。但后来我意识到，广告领域还有许多未被满足的需求。普通公司不知道如何在社交媒体上保持影响力，这是一个巨大的市场缺口。

like social media is the main media in the world. A lot of companies are actually able to build direct response advertising so that they can actually sell more. But in the same time, most of the companies in the world cannot simply do that. And basically, no because no one simply can keep up with the pace of production for social media as trends change pretty much every day.

社交媒体已经成为全球最主要的媒体。许多公司其实有能力开展直接反应广告，从而提升销量，但世界上大多数公司根本做不到这一点。归根结底，这是因为没有人能够跟上社交媒体内容的生产节奏，而趋势几乎每天都在变化。

Mhm. And so you were like, hang on a minute, these big brands aren't able to have media houses and so we need to create a tool that lets them. That was the cell.

所以你当时意识到，这些大品牌无法拥有自己的媒体制作团队，因此需要一款工具来帮助它们获得这种能力。这就是产品的卖点。

Yeah. Ex. Absolutely. So where it all really started is that we like there was a tool like to upload set of images and transform them into a slideshow with music.

对，完全正确。真正的起点是，当时有一种工具可以让用户上传一组图片，再将它们转换成带音乐的幻灯片。

It's kind of better than nothing but still pretty bad, right? So another solution was to take long form video and cut them to short vertically oriented videos. This was better but still really not perfect. And it felt to me that um especially 2023 it was absolutely clear that scaling loss finally work. It's not just a concept from science that scaling laws work. Video just takes couple I mean maybe two three years longer than LLMs

这种工具总比什么都没有强，但效果仍然很差。另一种解决方案是把长视频剪成竖屏短视频。这种方式稍好一些，但依然很不理想。尤其到了 2023 年，我清楚地看到规模化缩放定律终于奏效了。它不再只是一个科学概念，缩放定律确实有效。视频只需要更多时间，其发展进程可能会晚于大语言模型等领域。

and coding. Uh but it was clear that uh actually finally scaling loss should work in video as well and I decided just to take a bet. But I just want to go back. I get that in terms of what we see, which is, hey, we want to empower these brands and companies to create amazing media for social media, but it wasn't a hit from day one. And I spoke to Amy at Menllo who mentioned like a couple of pivots before and the

具体而言，视频的发展可能比大语言模型和编程晚两三年。但很明显，缩放定律最终也应该适用于视频，所以我决定赌一把。不过，我想先回到前面的话题。我明白我们现在看到的愿景，是帮助这些品牌和公司创作出色的社交媒体内容，但这项业务并非从第一天起就大获成功。我与 Menllo 的 Amy 交流过，她提到他们之前经历过几次转型，也谈到了我们走过的那些弯路。

meandering that we had. So what happened when we launched? Did we have immediate product market fit? No, actually we spent more than a year in a search of a product which could work. We burned more than 10 million out of 16 million raised in seed fundraising.

那么，我们发布产品时发生了什么？我们是否立刻实现了产品与市场的契合？并没有。实际上，我们花了一年多时间，才找到一款真正行得通的产品。我们在种子轮融资筹得的 16 million 中，烧掉了超过 10 million。

So we felt we have just one attempt left.

因此，我们觉得自己只剩下最后一次机会。

And frankly I feel I I'm responsible cuz I was focusing on the wrong things. I think I just lost the touch with reality back then. I was so much optimizing for what's hype today, what's the right narrative, how we can hijack the attention, all these things really like everything instead of building a good product. So when we had less than 6 million lefts, I guess it was slightly less than five actually, I realized that

坦白说，我认为责任在我，因为我把注意力放错了地方。那时，我已经与现实脱节。我满脑子都在思考当下什么最热门、怎样讲好故事，以及如何抢占关注度。我考虑了所有这些事情，却唯独没有认真打造一款好产品。因此，当资金只剩不到 6 million 时，我记得实际金额略低于5 million，我终于意识到了问题。

the only thing which we can be focused on is to lean into the product PLG and just finally set belief that the best product is going to win. And um so and then we just started to talk to customers. We spoke to eight creative directors about their experience with AI and what's simply missing. Everyone told us that camera control does not exist in AI and camera control is so important to tell a story. So this is a very

我们唯一应该专注的事情，就是全力投入产品和 PLG，并最终坚定一个信念：最好的产品终将胜出。随后，我们开始与客户交流。我们采访了八位创意总监，了解他们使用人工智能的体验，以及当前究竟缺少什么。每个人都告诉我们，人工智能目前还不具备镜头控制能力，而镜头控制对于讲好故事至关重要。因此，这是一个非常关键且必须解决的瓶颈。

important bottleneck to solve. So we released our products uh March 31st last year and since then we are really riding this crazy wave.

于是，我们在去年 3 月 31 日发布了产品。从那以后，我们真正乘上了这股疯狂的浪潮。

Was it immediate product market fit then?

那么，这一次是立刻实现了产品与市场的契合吗？

Like yeah it was immediate. Is product market fit like love? When you know, you know.

对，几乎是立刻。产品与市场的契合就像爱情：当它出现时，你自然会知道。

Um, yes, it's definitely when you know, you know. Like for example, we don't do any paid and like we have we have on the team people who scaled businesses to over like billion and two billion in revenue like other businesses um with paid advertising. Like at Higgsfield, we decided to really make a bet that we don't do paid.

确实，当它出现时，你自然会知道。比如，我们不做任何付费广告。我们的团队中，有人曾在其他企业通过付费广告推动增长，将企业规模扩大到营收超过10亿甚至20亿。但在 Higgsfield，我们决定真正押注于不做付费广告。

We don't do paid. Is influencers not paid?

你们不做付费广告。那么，找网红推广不算付费吗？

That's a good point. So, um with influencers, there is typically there are different types of influencers, but typically there is um some fee for just video production and then like some cost per click like attribution which is like works really well on YouTube. You you guys got into some controversy [laughter] for like I can't remember what it was. you were like pay paying people to promote for you or doing

这个问题问得很好。网红通常分为不同类型，但一般来说，仅制作视频就需要支付一定费用，之后还会产生按点击归因的费用。这种模式在 YouTube 上效果很好。你们之前似乎还因此引发过一些争议（笑声），但具体情况我记不清了。好像是你们花钱请别人推广，或者争议与网红营销有关。

something rogue with influencers.

也可能是你们在网红营销上采取了某些出格的做法。

Was that completely unfair? Was it kind of my bad we did do that? How do how do you respond to that?

这些指责完全不公平，还是你们确实做错了，也确实采取过那些做法？你会如何回应？

The main takeaway from like our experience is that it's very important to own own distribution. Distribution now more important than ever. And like we basically did outsource we had just a team of like two people on creator and customer success sides and we just did outsource to the agency and this was not uh that was not a good experience but uh we are still but but we are still trying to find interesting opportunities to tell

我们从这段经历中得到的主要教训是，掌握自己的分发渠道非常重要。如今，分发比以往任何时候都更加重要。当时，我们基本上把这项工作外包了，内部只有一个大约两人的团队，负责创作者和客户成功两块业务，其他工作则直接外包给了一家代理机构，最终的体验并不好。不过，我们仍在努力寻找有意思的机会，向大家介绍新的媒体形式。

about new media formats. Some of them are rather controversial. So, for example, recently we partnered with Neon, one of the largest streamers in the world, and launched like his own sort of AI generated stream. Um, like no one else did this before cuz this is like real creator making a replica of themselves. A lot of people start to question uh start to question their um like is it really authentic content or

其中一些尝试引发了很大争议。例如，最近我们与全球最大的主播之一 Neon 合作，推出了一场类似于由他专属的人工智能生成的直播。此前从来没有人这样做过，因为这是真人创作者制作自己的数字分身。许多人开始质疑，这究竟算不算真实内容。

not? But in the same time, those creators are under immense pressure. We all know about the story for about from MrBeast about like really how much like there is just pressure to constantly perform. So um and we also know through conversations with many talent agencies a lot of top stars actually want to be able to do more if they could create digital replica. But so what's happening today very frequently is that um those

与此同时，这些创作者也承受着巨大压力。我们都听过 MrBeast 讲述自己的经历，也知道这种持续工作的压力究竟有多大。通过与许多经纪公司交流，我们还了解到，很多顶级明星其实都希望做更多事情，前提是他们可以创建自己的数字分身。不过，如今常见的情况是这样的：

a tier celebrities they simply come up for a recording on like let's say green screen and then there is just a lot of post-prouction which goes on top of it and it feels to me that uh we are we we naturally going to come to the point of time where a AI digital replicas are going to become just one of the ways how creators can monetize.

一线明星通常只需要到场录制，例如在绿幕前完成拍摄，之后再进行大量后期制作。在我看来，我们终将自然地走到这样一个阶段：人工智能数字分身会成为创作者的变现方式之一。

Totally get that. I do just want to go back to part of the story. Where are you at revenue-wise today?

完全理解。我想回到之前的话题。你们目前的营收情况如何？

Uh so today is actually exciting day like when we record just Bloomberg article went out so that we cross 1 billion in annualized revenue. Um if I had a gong here I'd be like hitting the gong. A billion in revenue.

今天其实是个令人振奋的日子。就在我们录制时，Bloomberg 的文章刚刚发布，报道称我们的年化营收突破了 1 billion。如果我手边有一面锣，我一定会敲响它。营收达到了 1 billion。

Yes. Um actually it took us 18 months from 1 million to 1 billion for Corsor it took 24 months. Um so we are probably uh probably like the thirds after open the anthropic 18 months from a million to a billion.

是的。事实上，我们用 18 months 从 1 million 增长到 1 billion，而 Cursor 用了 24 months。因此，我们可能是继 OpenAI 和 Anthropic 之后，第三家用 18 months 从 1 million 增长到 1 billion 的公司。

Yes. How do you calculate revenue? Like it's a controversial topic. Um, how do you help calculate revenue?

是的。你们如何计算营收？毕竟这是一个有争议的话题。你们具体采用什么计算方式？

Absolutely. Uh, by the way, your um, co-host uh, Jason also asked this question in May. [laughter] Luckily, answer didn't change. So, we are at least consistent. So, but let me be transparent on that. What we do is we look um, revenue over the last four weeks and multiply it by 13 from what I know openable all of them use the same methodology.

没问题。你的联合主持人 Jason 在 May 也问过这个问题。[笑声]好在答案没有改变，至少我们的口径始终一致。我可以开诚布公地说明：我们的做法是统计过去 four weeks 的营收，再乘以 13。据我所知，OpenAI 等公司也采用相同的计算方法。

What's very important is that we are we take revenue not sales. So if that's like annual subscription or annual enterprise contract we prorate this across 12 months and take only this uh and only take like a piece which corresponds to one month to 28 days to be uh to be precise. That's the first piece and second it's only live revenue.

非常重要的一点是，我们计算的是营收，而不是销售额。如果是年度订阅或年度企业合同，我们会将其分摊到 12 months 中，只计算其中对应 one month——准确地说是 28 days——的部分。这是第一点。第二，我们只计算已实现的营收。

It's only live revenue. We are not taking like three year enterprise deals and baking into like 1 billion figure.

我们只计算已实现的营收，不会把 three year 的企业合同纳入并计入约 10 亿的数字中。

No, we don't do that.

不，我们不会那样计算。

If you were to break that billion up today into annual contracts, monthly subscriptions and then token spend, what would that be?

如果把这 10 亿拆分为目前的年度合同、月度订阅和代币支出，三者分别是多少？

So, um, videoi is still relatively early in my opinion. Uh, it is still probably two years behind coding in terms of adoption. So on demand usage for leading to coding companies could be over 50%.

在我看来，视频人工智能仍处于相对早期的阶段。就普及程度而言，它可能仍比编程落后两年。因此，头部编程公司的按需使用收入占比可能超过 50%。

And I would be honest for video it's substantially less than that. Um in the same time what's very interesting for us to observe in the business is that there is sub significant revenue expansion. I always love to study stories of the largest customers on the platform. So, one customer started um 6 months ago spending just subscription $99 a month. $99 a month. And now we just signed a deal over 6 million.

坦率地说，视频业务的这一比例要低得多。与此同时，我们在业务中观察到一个很有意思的现象：收入会显著扩张。我一直很喜欢研究平台上最大客户的案例。例如，有一位客户在 6 个月前刚开始使用时，每月只支付 $99 的订阅费。如今，我们刚刚与其签下一份超过 600 万的合同。

6 million.

600 万。

6 million a year. Right. So, yeah. Like this level of acceleration is something which really like mind-blowing to me.

每年 600 万。是的，这种增长速度确实令我大为震撼。

Dude, what are they getting for 6 million a year? that's like a Hollywood content team almost.

他们每年花费 600 万，究竟能得到什么？这差不多相当于一支好莱坞内容团队了。

So there are multiple trends um as and all of them frankly coming from Asia.

这里存在多个趋势，而且它们全都源自亚洲。

So first we're seeing a lot of um direct to consumer e-commerce companies rebuilding their whole go to market to be AI native where they make uh where they just make hundreds of ads if not thousands a week where they can AB test what performs well. But we all know about like short form dramas, right?

首先，我们看到许多面向消费者的电商公司正在重构整套市场进入策略，使其原生适配人工智能。这些公司每周会制作数百条甚至数千条广告，再通过 A/B 测试找出效果较好的广告。短剧则是大家都熟悉的例子。

Like most like short form dramas today is an industry over 10 billion owned primarily by Chinese companies having huge impact both in China, United States, in Europe, everywhere in the world and most of new shows there are made with AI end to end. So look, I think uh like the this adoption obviously is uh coming like bottom up, but um that that's very difficult to refute this new reality.

如今，短剧已经发展成一个规模超过 100 亿的行业，主要由中国公司主导。无论在中国、美国、欧洲还是全球其他地区，短剧都产生了巨大影响，而且其中大多数新剧都全流程使用人工智能制作。这种普及显然是自下而上发生的，但这种新现实已经很难否认。

What percent of revenue is consumer versus enterprise?

消费者业务和企业业务分别占收入的百分之多少？

So that that that's a great question. So um so B business revenue is slightly over 50%.

这是一个很好的问题。B端业务收入略高于 50%。

Wow.

哇。

Yeah, that's impressive.

是的，这很厉害。

Thank you. Um on the consumer side, it's also very important to break it down. So on the consumer sides out of these 50 is around like 10% is pure consumer use cases pure consumer and that's roughly people who use it on mobile. So share of our revenue from mobile is less than 10%. That's why we are we are very different from many other companies and but there are lots of aspiring creators like basically those people who are

谢谢。在消费者端进行细分也非常重要。在消费者端的这 50% 当中，大约有 10% 属于纯消费者使用场景，这部分大致是通过移动端使用产品的人群。因此，我们来自移动端的收入占比不到 10%，这使我们与许多其他公司非常不同。此外，还有许多人希望成为创作者。

freelancers doing social media marketing projects and so on who try to learn video AI so that they can make more money. It's true that their behavior is a little churny. uh within a year most of them actually come back to try again and we do believe that over the time most of them are going to figure stuff out and they're just going to become this new AI native workforce. So it's still important for us to educate them

这类人包括承接社交媒体营销项目的自由职业者。他们会努力学习视频人工智能，以便赚取更多收入。确实，他们的流失率略高，但在一年内，大多数人都会回来再次尝试。我们相信，随着时间推移，他们中的大多数都会逐渐摸索出门道，并成为一支全新的人工智能原生劳动力大军。因此，教育他们对我们依然很重要。

and uh that's why we invest so much in like Higgsfield Academy, YouTube channel and so on. But we also are f fully cognizant that we will never be able to win in a market of subscriptions of $20 a month.

这也是我们大力投入 Higgsfield Academy、YouTube 频道等项目的原因。但我们也非常清楚，我们永远不可能在每月 $20 的订阅市场中取胜。

So why? Because like I think like today Google and OpenAI they pursue like ads so much but fundamentally I think they are going to completely demolish all the consumer subscription markets which is uh $20 a month subscriptions.

为什么？因为如今 Google 和 OpenAI 都在大力发展广告。从根本上说，我认为它们会彻底摧毁所有消费者订阅市场，也就是每月 $20 的订阅市场。

Oh, so you saying that because they provide a horizontal product that's very good, you're just going to not pay for a lot of the verticalized products that you used to pay $ 20 $30 a month for.

所以你的意思是，由于它们提供了非常好用的横向产品，用户不会再为许多垂直产品付费，而过去用户每月会为这些产品支付 $ 20、$30。

Yeah, I do believe that. That's essentially what's going to happen over the time. Um, I know this is a very contrarian bets, but um, at least we can see some of that.

对，我确实这样认为。这基本就是未来会逐渐发生的事。我知道这是一个非常逆势的判断，但至少我们已经看到了一些苗头。

I think it cannibalize Canvas growth if you're honest. A lot of the lowhanging fruit on the consumer design side that Canva used to serve can now be done in open AI in particular.

说实话，我认为这会蚕食 Canva 的增长。消费者设计领域有许多唾手可得的需求，过去由 Canva 满足，现在尤其可以通过 OpenAI 完成。

Is that what you're talking about?

你指的是这个意思吗？

Yeah. And I do believe this is just the most apparent example, but there are couple more which is which is already happening. And I do believe that uh that's why for at Higgsfield what what really matters for us is how we even if we get someone on like $20 a month subscription like how can we show them value how can we make them to upgrade to over to spend over um to over $1,000 a year with us. I can't believe that's 6

对。我相信，这只是最明显的例子，另外几种情况也已经出现。对于 Higgsfield，我们真正关心的是：即使用户订阅的只是每月 $20 的套餐，我们该如何让他们看到价值，并促使他们升级，将每年的消费金额提高到 $1,000 以上。很难相信，我们从 99 美元做到了每年 600 万。那绝对是……

million a year from 99 bucks. That's the best ever slide on a fundraising deck.

融资材料里最出色的一页，也是有史以来最出色的融资材料页面。

[laughter] And all of our customers are going to do the same. Exactly.

而且我们所有客户都会这么做。没错。

Can I ask you mentioned there kind of churn rates when you look at 30-day retention rates for consumers and 90-day retention rates. What are yours and what is good? So, there is um quite massive drop within the first month just simply because people don't fully realize the value and that's a that's a core priority for us to actually get better in that. So, showcasing the value. Is it like half or like

我想问一下，你刚才提到了流失率，包括消费者的 30 天留存率和 90 天留存率。你们的数据是多少？怎样的水平才算好？首月的流失幅度确实相当大，原因很简单：用户没有充分认识到产品的价值。这正是我们目前的核心优先事项，我们要切实改善这方面的问题，把产品价值展示出来。流失比例会达到一半左右，还是……

No, it's uh it's maybe like 30% drop.

不，大约会流失 30%。

Okay.

好。

But then it's it's really flat after that. It's we look obviously at like logo retention.

但在此之后，曲线就会变得非常平稳。我们当然也会关注客户数留存率。

Mhm.

是的。

I wouldn't say it's great but because like we all remember like B2B SAS era like uh retention was expected to be logo retention month one was expected to be over 80%. M um so clearly we have uh we have a lots of work to do on uh user education to get there but some things are truly phenomenal like when I look at the cor at the business segments and NRR at month 12 obviously like you're going to

我不会说这个数字很好，因为大家仍然记得 B2B SaaS 时代对留存率的预期，例如第一个月的客户数留存率通常要超过 80%。显然，我们还需要在用户教育方面做大量工作，才能达到这一水平。但有些数据确实非常惊人。例如，当我查看业务细分群组第 12 个月的 NRR 时，你当然可能会……

argue it's like 18 months old company like what are you talking about but still when I look at the numbers which I have today NR at month 12 is over 300%.

质疑：“这家公司才成立 18 个月，你在说什么？”但即便如此，根据我目前掌握的数据，第 12 个月的 NRR 仍然超过 300%。

just it just never happens in B2B SAS right so um that's why I'm saying that while there is substantial churn in month zero and we have to do better job with user education to address that expansion is unprecedented can we actually just unpack the two different go to markets cuz you got consumer and you got enterprise and I spoke to quite a few of your competitors in all honesty before this show [snorts] and I said hey

这种情况在 B2B SaaS 领域根本不会发生。因此，尽管第 0 个月的流失率很高，我们必须加强用户教育并解决这个问题，但增购表现是前所未有的。我们能否具体拆解一下这两套不同的市场进入策略？你们既面向消费者市场，也开展企业业务。我还采访了不少你们的竞争对手。在本期节目录制前，我问他们：“亚历克斯要来做客了，我们应该问他什么？”

you we've got Alex coming on what should we ask him everyone said the same thing which was an admission of their respect for this particular kind of GTM. They said you've ex executed the most impressive influencer campaign in tech and what I wanted to understand was when you look at the consumer growth what worked what didn't work and how do you reflect on that first and foremost like the goal is to make sure that the

所有人的回答都一样，这也表明了他们对这种特定市场进入策略的认可。他们说，你们执行了科技行业最令人印象深刻的网红营销活动。我想了解的是，回顾消费者业务的增长，哪些方法奏效了，哪些没有奏效？首先，你对此有何反思？我们的目标是确保……

best commercial video content is generated on Higgsfield and we show all the workflows of how to make such uh professionallook videos and we have an in-house team of over 150 creative professionals.

最优质的商业视频内容都在 Higgsfield 上生成，并展示制作这类专业级视频的完整工作流。我们拥有一支由超过 150 名创意专业人士组成的内部团队。

150. It's it's almost half of the whole work workforce frankly. And um they th those people they make product launch videos, they make tutorials like for example we made the first generated movie which is also like obviously um a very um a very sensitive topic but what's important we open sourced all of it and what we learned is that for 90 minutes of uh of like let's say TV quality content it was over 100 hours of

150 人。说实话，这几乎占到整个团队的一半。这些人既会制作产品发布视频，也会制作教程。例如，我们制作了第一部生成式电影。显然，这也是一个非常敏感的话题，但重要的是，我们把所有内容都开源了。我们从中发现，要制作 90 分钟的电视级品质内容，需要从超过 100 小时的……

for yet generated contents. So creative decisioning like picking the right piece is still very important. Um so that's really what's what we are focused on and that's what's driving most of the most of the revenue.

生成内容中进行筛选。因此，选出合适片段等创意决策仍然非常重要。这正是我们真正专注的方向，也是我们绝大部分收入的来源。

So you're saying the the growth in consumer subscription is through own content and distribution.

所以，你的意思是，消费者订阅业务的增长依靠的是你们自己的内容和分发渠道。

Yes. We don't do any paids. Early on you made an interesting architectural decision to have your own models and then you since walked that back. Can you talk me through why did you choose own models and why the walk back?

对。我们不做任何付费推广。早期，你们做了一个很有意思的架构决策，选择使用自研模型，但后来又放弃了。你能否谈谈为什么最初选择自研模型，后来又为什么放弃？

Oh, um yeah, obviously this was [laughter] uh obviously this was my mistake. I'm going to be I'm going to do my best to be um transparent. What I need to admit, we really tried we I at some point of time I really was thinking that chasing benchmarks um is valuable but I don't believe this is just sort of corporate scops frankly.

是的，显然这是我的错误。我会尽最大努力坦诚回答。我必须承认，我们确实认真尝试过。曾经有一段时间，我认为追逐基准测试成绩是有价值的，但现在已经不再这样认为。坦率地说，这不过是某种企业套路。

So um and I was part of the large organization so I know what happens.

我也曾是大型组织的一员，因此知道其中会发生什么。

What happens is that everyone just thinks like we need to show some progress. So we need to have some benchmark but then when I talk to the top researchers from these labs especially larger companies what happens is that they start to put test data into the training.

实际情况是，所有人都会认为自己必须展示一些进展，因此需要拿出一些基准测试成绩。但当我与这些实验室的顶尖研究人员交谈时，尤其是大公司的研究人员，我了解到的实际情况是，他们开始把测试数据放进训练数据中。

They start to kind of use uh leverage test data to use LLM as a judge for training of the models. use all the various tricks to basically gain benchmarks, get get like quarterly bonuses and so on because like who cares, right? So if I make my couple million dollars a year in inside in one of these labs, I can move to another lab easily. So that's unfortunately what's happening in larger organizations. Um

他们开始以某种方式利用测试数据，用大语言模型充当评判者来训练模型。他们会使用各种手段，基本目的就是提升基准测试成绩，获得季度奖金之类的回报，因为没有人在乎这些做法。如果我一年能在其中一家实验室赚到几百万美元，也可以轻松跳槽到另一家实验室。遗憾的是，这正是大型组织里正在发生的事。

and can I just stay on that?

我能就这个话题再追问一下吗？

Yeah.

可以。

What do you mean? You're saying that they are incentivized by benchmarks and so because of that they are doing artificial things to improve their scoring in benchmarks which actually don't increase output efficiently. Yeah.

你是什么意思？你是说，他们受到基准测试成绩的激励，因此会采取一些人为手段提高自己的基准测试得分，但这实际上并没有有效提升产出。对。

Look, I think let's just look at the outcomes which we have today. Out of all the incumbents in the United States, when I look at open router data, the only company which is relevant is Google.

我认为，不妨看看如今的实际结果。在所有美国老牌企业中，根据 OpenRouter 的数据，唯一仍具竞争力的公司是Google。

out of all the incumbents when I look in China where probably obsession with benchmarks probably is less we have 10 cent shyomi Alibaba uh like three incumbents being completely relevant and obviously like by dance obviously trying to catch up as well what's your takeaway from that I just do believe that uh the there is just obviously in the in tech bubble there is a strong obsession over the benchmarks uh which do not uh

纵观中国的所有老牌企业，那里对基准测试的痴迷程度可能没有那么高。Tencent、Xiaomi、Alibaba这三家老牌企业依然极具竞争力，ByteDance 显然也在努力追赶。你从中得出了什么结论？我确实认为，科技圈显然非常痴迷于基准测试，而这些测试

necessarily represent the reality. But I can talk specific specifically in the for video.

未必能反映现实。不过，我可以专门谈谈视频领域。

A lot of benchmarks today for video is really text to video which does not represent actual workflows at all. Um the way to think about video models today, it's just modern rendering engine. It think about this as like Unreal Engine or Unity but just different types of inputs.

如今，很多视频基准测试衡量的其实都是文本生成视频，而这完全无法代表实际的工作流。理解视频模型的一种方式，是将其视为现代渲染引擎。可以把它理解成Unreal Engine 或 Unity，只不过输入类型不同。

And it's virtually impossible to really define a visual output and and direct the execution just through text. If you just go and to our open source projects like this movie which I mentioned average prompt length is over 3,000 words. That's the first thing and like look all these benchmarks which we are talking about they are not like as comprehensive in terms of the details of prompts and people who are labeling they

想要真正仅通过文本定义视觉输出并引导具体执行，几乎是不可能的。以我们的开源项目为例，包括我提到的这部影片，其提示词的平均长度超过3,000词。这是第一点。再看我们讨论的所有这些基准测试，它们所使用的提示词在细节上远没有那么全面，而负责标注的人显然

obviously don't cannot read like 3,000 long word long prompts but also on average there are at least 10 image references for every for every scene. The reason why it's important because it's important to define how the characters look like, how the background looks like, like how actually characters are located to each other in the scene and so on. And so that's why like prompting and like just the workflow is so

也不可能阅读长达3,000词的提示词。此外，每个场景平均至少还有10张参考图。这一点之所以重要，是因为必须明确角色的外貌、背景的样子，以及角色在场景中的相对位置等信息。因此，提示词设计乃至整个工作流才会如此

complex. Benchmarks just don't rep don't represent that.

复杂，而基准测试根本无法体现这些。

So going back to the model selection, why did we decide we're going to do our own and then why walk it back? It's true that like with VFX and camera control, we got very very quickly from like maybe 1 million to 20 million in AR within maybe the first 3 months. Then we released own image model which is really good at um aesthetic photo shoots and product consistency. This is what allowed us to scale then from 20 to 100

回到模型选择这个话题，我们为什么一开始决定开发自己的模型，后来又为什么改变方向？借助 VFX 和相机控制，我们的 AR 很快就从大约1 million 增长到 20 million，可能只用了最初的 3 months。之后，我们发布了自己的图像模型。它非常擅长制作美观的商业大片，并能保持产品一致性。正是这一点，让我们得以从 20 million 进一步增长到 100

million. So help me understand, Alex, why did you decide that you were going to do your own models and why did you abandon them? [snorts] We still do them whenever we see like specific use case like these photo shoots.

million。亚历克斯，请解释一下，你当初为什么决定开发自己的模型，后来又为什么放弃了？我们没有放弃。只要发现特定的使用场景，例如这类商业大片，我们仍然会开发自己的模型。

Uh but but as soon as this is what our customers want. So it's all driven based on the customer feedback, not just by ambition to conquer the worlds and [clears throat] build the best model in the world. Do you think every company will have their own models like we're seeing Harvey, we're seeing Cognition, we're seeing Mccor, Ramp build their own models and we'll see every company have their own models with their own data or

但前提是，这确实是我们的客户想要的。因此，一切都由客户反馈驱动，而不只是源于征服世界、打造全球最佳模型的雄心。你认为每家公司最终都会拥有自己的模型吗？例如，我们现在已经看到 Harvey、Cognition、Mercor 和 Ramp 构建自己的模型。未来，我们会看到每家公司都拥有基于自有数据训练的自有模型，还是

we actually all use a series of providers. So um first of all whenever just to be honest whenever someone says we build our own models very likely what they mean is something what see what's happened with Corsor. We we do remember right a lot of companies they actually take open weights model and just post train on own data.

大家都会使用一系列供应商的模型？首先，坦率地说，每当有人声称“我们构建了自己的模型”时，他们所指的很可能是类似 Cursor 所采用的模式。很多公司其实会采用开放权重模型，再使用自有数据进行后训练。

Mhm.

嗯。

Um and post training can happen in two ways.

后训练可以通过两种方式进行。

Most importance is whenever you have um customer data around like decisions they make like sequence of decisions and you can teach the model to actually take like learn how to compress these 10 steps into one step. like this type of reinforcement learning is the most valuable. So and I think like increasingly more and more companies will have to do that frankly just we see this in the market as well. So the most

最重要的一种方式，是利用与客户决策相关的数据。例如，当你拥有一系列连续决策的数据时，就可以训练模型，使其真正学会如何将这 10个步骤压缩成一个步骤。这类强化学习最有价值。因此，我认为会有越来越多的公司不得不这样做。坦率地说，我们在市场上也看到了这一趋势。当今世界

most of the companies in the world today most of the businesses they don't necessarily need Astra specifically they don't necessarily need the newest fable model and that's why like open router reports that uh share of open source models went from below 30 to over 60 within within this year.

的绝大多数公司和企业，其实不一定特别需要 Astra，也不一定需要最新的前沿模型。这就是为什么 OpenRouter的报告显示，开放源代码模型的占比已从低于 30 上升到超过 60，而且这一变化就发生在今年之内。

What do you think share of open models will be in two years time? Look, I do believe that just because the cap capitalism works, I mean openly ananthropics still are going to have more than 50% of the markets in terms of the dollars generally in terms of the dollars, right? And especially because uh for coding still remains to be very very prolific use case where coders are always jumping to

你认为两年后，开放模型会占据多大份额？我确实相信，只要资本主义继续运转，OpenAI和 Anthropic 仍将占据超过 50% 的市场份额。总体上是按收入计算。对，是按收入计算。尤其是因为编程仍然是一个极其热门的使用场景，程序员总会转向

to to to the recent model over but for our markets we're seeing completely different dynamics. what's actually happening in social media marketing as companies start to print hundreds of create ad creatives um a week they want to have maybe cheapest more steable models cuz like PhD level intelligence is not necessarily needed for to make viral social media video. So um and that's where we actually have

最新的模型。但在我们的市场中，情况截然不同。社交媒体营销领域的实际情况是，随着公司开始每周批量制作数百个广告创意，它们可能更需要成本最低、表现最稳定的模型，因为制作一条在社交媒体上爆火的视频并不需要博士级智能。我们正是在这里

seen that um we get like 80% plus margin whenever we run open-source models like post-trained open source models. Uh but it can be way more cost efficient for our end customer compared to the proprietary models.

发现，只要运行开源模型，例如经过后训练的开源模型，我们就能获得 80% 以上的利润率。对我们的终端客户而言，与闭源专有模型相比，这类模型的成本效益也可能高得多。

What's the comparison on margins between open versus closed for you?

对你们而言，开源模型和闭源模型的利润率分别如何？两者的差距有多大？

The margin on own models and open weights models is over 80%.

自研模型和开放权重模型的利润率超过 80%。

Um, and then it almost doesn't matter.

在这种情况下，其他因素几乎都无关紧要了。

And for closed source models, it's probably between 20 and 30%. And then what becomes important is can we actually steer the traffic. What makes me excited about Higgsfield is that umic grows so quickly and actually for us as companies start to actually create those agentic workflows to make more ads we choose which model we can use. So like we choose what model to use in over 40% cases.

闭源模型的利润率可能在 20% 到 30% 之间。接下来真正重要的是，我们能否真正引导流量。Higgsfield 令我感到兴奋，是因为它增长得非常快。对我们而言，随着企业开始真正构建这些智能体工作流来制作更多广告，我们可以选择使用哪个模型。例如，在超过 40% 的案例中，使用哪个模型是由我们决定的。

In a way model routting becomes a core feature of the business. No.

从某种意义上说，模型路由成为了这项业务的核心功能，对吗？

Yeah. We call we call it tokconomics essentially right as like there is certain amount of work customers want to do um how can we optimize number of tokens which requires and how we can pick the most efficient tokens for them there are actually two incumbents in the United States who figured out models it's not just Google it's also Nvidia why do you think that is what I'm constantly seeing is that um the there

对。我们将其称为 tokconomics。本质上，客户有一定量的工作需要完成，我们需要考虑如何优化所需的 token 数量，以及如何为他们选择最高效的 token。实际上，美国有两家已经站稳脚跟的巨头真正理解了模型：不只有 Google，还有 Nvidia。你觉得这是为什么？我一直观察到的是，模型

is the versions of models so there are these state-of-the-art models which have to be really good in computer use like Astra or in coding. Um but they can be prohibitively expensive and we we're chatting about that like on average at Higgsfield person on the team spends over 10,000 over $10,000 a month on various models and remember like we are split across United States and Asia across so how much do you spend on models per

不同版本的模型各有所长。例如，Astra 等最先进的模型必须非常擅长计算机操作，另一些模型则必须非常擅长编程。但它们的价格可能高得令人望而却步。我们之前也谈到，Higgsfield 团队成员平均每月在各种模型上的花费超过 $10,000。还要考虑到，我们的团队分布在美国和亚洲各地。那么，你们每个月在模型上

month? So internal usage of models a month is over four million.

花费多少？我们内部每月使用模型的费用超过四百万。

Wow. How many people do you have?

你们有多少人？

We have close to 400 people and just want to make sure that the math adds up.

我们有将近 400 人。我只是想确认计算结果是否准确。

Yes, it's um it's definitely over it's definitely over $10,000 per person.

没错，肯定超过了，人均支出肯定超过 $10,000。

How has that changed over time?

这项支出随着时间推移发生了怎样的变化？

That's the best question of the whole show, by the way. Um that's the best question.

顺便说一句，这是整期节目中最好的问题。

What actually started to happen is the creative team started to do VIP coding like the like like this month I was I just caught a guy who spent over 30k in a week on Astra model cuz he was frankly frustrated that some like asset organization workflow and as you said like basically auto editing is still not very good in production and he said, "Oh, I'm just going to do this myself." And just went like five nights,

实际情况是，创意团队开始进行 VIP 编程。就在这个月，我发现有人一周内在 Astra 模型上花费了超过 30k。坦率地说，他对一些素材整理工作流之类的工具很不满意。正如你所说，自动剪辑在实际制作中仍然不太好用，于是他决定：“那我就自己来做。”随后，他连续熬了五个晚上，

five nights straight on Astra and it works.

整整五个晚上都在使用 Astra，最终还真的做出来了。

We learned a lot. I wouldn't say it was production ready, but we learned a lot.

我们从中学到了很多。虽然我不会说它已经达到了生产可用的程度，但我们确实收获颇多。

30,000 in a week.

一周花费 30,000。

Yeah. Yeah. Many people spend over 10,000 in a week.

是的，很多人一周会花费超过 10,000。

Do you mind?

你介意这种情况吗？

Yeah. My finance team will probably say, I don't know if if you know if you ask any of them, but they will probably say that I'm like being too stubborn, too relentless to control the spend cuz sometimes I feel it goes like [laughter] it it really goes out of control like 30k in a week is quite a lot. But we learned this. So this was actually net positive experience.

介意。我的财务团队很可能会说，我在控制支出方面过于固执，甚至不依不饶。我不知道你是否问过他们。因为有时我觉得支出真的会失控。[笑声]一周花费 30k 确实相当多，但我们也从中学到了很多。因此总体而言，这是一次正向的经历。

Okay. So the internal spend 4 million about 10,000 per head. What will that be in 12 months time do you reckon?

所以，内部支出是 4 million，人均大约 10,000。你认为 12 个月后这个数字会是多少？

So that that's very interesting. So across uh the top uh the top engineers and across top creatives I think it's going to keep growing and I do believe we are going to get to to to spend um close to 50k and 100k a month for those who can call 10x engineers 10x creatives unfortunately I also expect that these people will ask for comparable salary raise as well so I think that's just going to correlate at some points um but

这个问题非常有意思。对于顶尖工程师和顶尖创意人才，我认为这个数字会继续增长。我们的支出可能达到每月接近 50k，甚至 100k，而这些费用将用于那些可称为 10x 工程师和 10x 创意人才的人。遗憾的是，我预计这些人也会要求相应幅度的加薪，因此两者发展到某个阶段后可能会形成关联。不过，

also for a lot of other jobs. Let's say we to take legal finance and so on. I think it it really stabilizes around like um $500,000 a month very very quickly. With those 10x engineers, the idea is they have thousands of agents running below them doing a lot of the difficult execution work that took time. Do we just have dramatically smaller teams with those 10x engineers, 10x designers, 10x finance leaders? I can definitely say

对于法务、财务等许多其他岗位，我认为支出很快就会稳定在每月大约 $500,000，而且这个过程会非常快。至于这些 10x 工程师，可以设想，他们手下会有数千个智能体运行，负责大量高难度的执行工作，也就是那些原本需要投入时间才能完成的工作。我们是否只需要规模大幅缩小的团队，再配备效率高出 10 倍的工程师、设计师和财务负责人？我可以肯定地说，

that the I I I had sort of a feeling that legal customer support um is going to be mostly replaced and that's obviously one of the main u mistakes operation which we have done in the company that we didn't ramp these teams quickly. Um what we are seeing today is that like let's say our legal team is like over 10 people our customer success team is over 40 people all of

我此前隐约认为，法务和客户支持的大部分工作都会被取代。这显然也导致了我们在公司运营中犯下的一个主要错误，即没有迅速扩充这些团队。如今，我们的法务团队有 10 多人，客户成功团队有 40 多人，他们全都

them use AI heavily we like at at these professions where I say quite close today I definitely can say that uh there is I don't see any elimination it's true that probably over 60% of customer support requests especially the first line of defense can be handled with AI but when it especially comes to B2B It doesn't like it like AI just doesn't work.

在大量使用人工智能。对于这些我密切关注的职业，我现在可以肯定地说，我没有看到任何岗位被淘汰。的确，超过 60% 的客户支持请求，尤其是第一道防线，可能都可以由人工智能处理。但在涉及 B2B 时，人工智能的效果仍然不理想，甚至根本不起作用。

Revolute has now over 92% resolution rate on customer support for consumers.

Revolute 目前面向消费者的客户支持问题解决率已超过 92%。

Pretty good.

这个成绩相当不错。

It's it's it's pretty good. But obviously they did invest a lot into that [ __ ] ton. A [ __ ] ton.

确实相当不错。但显然，他们在这方面投入了大量资源，多得[ __ ]。

And and but also very important the way how Nick thinks about that uh is in terms of the playbooks. We launch products, new products pretty much every week. So um we have to we have to keep update agents with all the information and so on and just due to the high velocity having um extremely smart coordinated team is is very important.

还有一点也非常重要：Nick 是从操作手册的角度看待这件事的。我们几乎每周都会推出新产品，因此必须不断向智能体更新所有相关信息。由于产品迭代速度很快，拥有一支极其聪明、配合默契的团队非常重要。

That's really interesting how product velocity increases leads to harder customer support for agents.

这一点很有意思：产品迭代速度加快，会让智能体更难做好客户支持。

Of course, cuz uh the agents are as good as context and rules which they have.

当然，因为智能体的表现取决于它们掌握的上下文和规则。

And if context and rules change pretty much twice a week, it gets a little difficult.

如果上下文和规则几乎每周变化两次，事情就会变得有些棘手。

When you look at your engineering team today, what are they on? Are they on cursor? Are they on codeex? Are they on core code? So from a period from March to June, everyone really moved to claude um including the creative team and that's where we actually started to see creative team vibe coding functionality which we don't have in production. But then we started to see that all the coders quickly moved from claude to

看看你们现在的工程团队，他们都在使用什么工具？是 Cursor、Codex，还是 core code？从 3 月到 6 月，所有人基本都转向了 claude，包括创意团队。也正是在那段时间，我们真正开始看到创意团队使用氛围编程功能，尽管这个功能尚未投入生产。随后我们发现，所有程序员都迅速从 Claude 转向了

codeex um as of mid June and um over the time especially 10x creatives moves to codex as well but look I do believe that there it's it's it's cyclical so it's so cyclical my question to you is will we continue to see the velocity of model release that we're seeing now you in 3 years time will It be like, "Oh, Gemini this week. Oh, CL Anthropic this week, OpenAI this week." Or will we see a a reduction in model release rate? I

Codex。这一转变从 6 月中旬开始；一段时间后，尤其是 10x 创意人员也转向了 Codex。不过，我确实认为这种变化具有周期性。周期性实在太明显了。我的问题是：我们现在看到的模型发布速度还会持续下去吗？例如 3 年后，是否仍会是这种情况：“这周是 Gemini，这周是 CL Anthropic，这周是 OpenAI。”还是说，模型发布速度会有所放缓？我

don't think that's going to happen anytime soon. So, I believe like for example, recently OpenAI announced that they basically build OpenAI for law, right? But that's only V0. So over the time they also are going to try to print smaller specialized models for like not like exactly smaller uh but really specialized model for certain use cases.

认为这种情况短期内不会发生。例如，OpenAI 最近宣布，他们基本上打造了法律版 OpenAI。但那只是 V0。随着时间推移，他们也会尝试推出更小、更专业的模型，用于特定场景。也未必一定是更小的模型，而是真正针对特定用例的专业模型。

Um clearly like Astra excels in long-term horizon.

很明显，Astra 擅长处理长期任务。

Do you buy that? Like I look at that GPT for law from Astra and I'm like I'm sorry I think it's complete [ __ ] with the greatest of respects. It is a very deep functionality required to serve some of the biggest law firms in the world. like very very deep and specific functionality. It's very specific according to the different types of law as well. Plus, if you want to sell into these law firms, it's a

你相信这一套吗？例如，我看到 Astra 的法律版 GPT 时，心想：抱歉，恕我直言，我觉得这完全是[ __ ]。要服务全球一些最大的律师事务所，产品必须具备非常深入的功能，才能满足它们的需求。这些功能不仅要非常深入、非常具体，还必须针对不同的法律类型提供极其具体的支持。另外，如果想把产品卖给这些律师事务所，就要经历

multi-year sales cycle with some of the stodgy old lawyers and partnerships.

长达数年的销售周期，还要面对一些古板守旧的老律师和合伙人。

You can't just say, "I'm open AI. Yep.

你不能直接对他们说：“我是 OpenAI。

We've just hacked into the Australian government, by the way, but we're here to serve your law firm."

顺便说一句，我们刚刚黑进了澳大利亚政府，但现在是来为贵所服务的。”

Uh, okay. Yeah. So, first of all, I think uh just uh definitely the ability to switch internal use just for internal teams outside of law firms.

好吧。首先，我认为毫无疑问，这类产品可以转为内部使用，仅供律师事务所以外的内部团队使用。

I think that's I think that's definitely happening. Oh, I think we both investors in company called solve intelligence.

我认为这种情况肯定正在发生。我记得，我们都是一家名为 Solve Intelligence 的公司的投资人。

Love it. Yeah. Very specific. Very specific. And let me try to maybe bring couple examples why like solve intelligence is so special and like where like for example how we learn from this.

太棒了。这是一个非常具体的例子。让我试着再举几个例子，说明软件智能为何如此特别，以及我们能够从中学到什么。

What can happen very often is that a company want to just control the patent workflow even if they outsource the work and that's very valuable just to have one system of records. So whoever can create AI native system of records is going to win. And but going back to Hixel why it's so important for Higgsfield there are so many systems today which are used for just to store assets like some people use Dropbox

一种很常见的情况是，公司希望掌控专利工作流，即使它们把相关工作外包出去也是如此。仅仅拥有一个统一的记录系统，就具有很大价值。因此，无论谁能打造出人工智能原生的记录系统，谁就会胜出。回到 Higgsfield，这一点之所以对 Higgsfield 如此重要，是因为如今有许多系统仅用于存储资产。例如，有些人使用 Dropbox

some people use Google Drive some people are going to try to use Miro some people are going to try to use frame.io like there are many solutions but let's think about what people need.

，有些人使用 Google Drive，还有些人尝试使用 Miro 或 frame.io。市面上有许多解决方案，但我们应当思考人们真正需要什么。

What people need, they want to be able to search contents and and marketers especially want to make sure that content is on brands in terms of the visual identity, but also like if that sort of adheres to certain brand guidelines and that's where like semantic understanding and semantic controls become finally possible. It never existed before. So in our space there are definitely other companies like

人们需要搜索内容。营销人员尤其希望确保内容在视觉识别方面符合品牌调性，同时确认其是否遵循特定的品牌规范。如今，语义理解和语义控制终于能够实现这些需求，而这是以前从未有过的能力。在我们这个领域，当然还有其他公司，例如

Adobe and Canva who builds the best software for the pixel first era where everything was defined with pixels but that's clearly not how the world is going to work in the future. What we're envisioning and that's what everyone wants. They want to just be able to search and um like really work through the library of assets and all the knowledge through natural interfaces. So being able to own this interface and

Adobe 和 Canva。它们打造了像素优先时代最出色的软件；在那个时代，一切都由像素定义，但未来的世界显然不会再以这种方式运作。我们所构想的，也正是每个人想要的：人们只希望能够进行搜索，并通过自然交互界面真正处理整个资产库和全部知识。因此，掌握这一交互界面，并

build the analytics uh like this system of records is important. That's why at Higgsfield we invests it so much in harness so that it improves over the time. And this harness also um allows it basically learns visual style over the time which let's say cloud and open AI cannot necessarily do.

构建分析能力，也就是建立这套记录系统，非常重要。这正是 Higgsfield 在这套框架上投入如此之多的原因：它能够随着时间推移不断改进，还能逐渐学习视觉风格，而这是 Claude 和 OpenAI 未必能够做到的。

Do you believe in moes anymore? you you've been around startups for a long time. We always talked about moes and defensibility. I largely think they're [ __ ] You know, we we saw lovable when I invested. Everyone was like, "Oh, it's a rapper. It's a rapper, you idiot, Harry." And actually, it was a rapper, [laughter] but it's about speed of decision making, product execution, and building value over time very, very

你现在还相信护城河吗？你已经在初创企业圈摸爬滚打了很多年。过去，我们总是在讨论护城河和防御能力。我基本上认为它们都是[ __ ]。我们当时看到了 Lovable。我投资时，所有人都说：“它只是个套壳产品，就是个套壳产品。Harry，你这个白痴。”事实上，它确实是个套壳产品。[笑声]但关键在于决策速度、产品执行力，以及以极快的速度持续

fast. Instinct is a rapper. Of course, it is. It's not that difficult to do an AI assistant today which why there's so many but they're building incredibly quickly very valuable features and you build it over time. Do you believe that moes actually exist really?

创造价值。Instinct 是个套壳产品，当然是。如今，打造一个人工智能助手并没有那么困难，所以市面上才会出现这么多此类产品。但他们正在以惊人的速度打造极具价值的功能，而这些功能可以随着时间推移逐步积累。你真的相信护城河确实存在吗？

[sighs and gasps] I know like you ask this everyone um so um and this is cuz this is on top of everyone minds like how to think about the metrics which matter today and how to think about the modes. So um I think um it's very difficult to figure out where the value occurs in the supply chain. Um we do believe that there are only two like ways of uh modern value creation or modes today. First is when

[叹气并倒吸一口气]似乎每个人都会问这个问题，因为这正是所有人最关心的事情：应该如何看待当今真正重要的指标，又该如何看待护城河。我认为，要弄清楚价值出现在供应链的哪个环节非常困难。我们确实认为，如今创造现代价值或构建护城河只有两种方式。第一种是

you deliver the outcome and for us it's allowing businesses to sell more through AI ads. So that's the first thing and the second thing is network effects.

交付成果。对我们而言，这意味着让企业通过人工智能广告卖出更多产品。这是第一点。第二点是网络效应。

Unfortunately, AI does not replace network effects. And when people talk about swarm of AI agents talking to each other, I'm not sure this is happening in the next five years. So, um, that's why it's so exciting that within Higgsfield, like we really wanted to empower community to create more projects, open source, open source them to really build a snowball where people can capitalize on each other output. This is the reason

遗憾的是，人工智能无法取代网络效应。有人谈到由大量相互交流的人工智能智能体组成的集群，但我不确定这是否会在未来五年内实现。正因如此，Higgsfield 内部的情况才如此令人兴奋。我们确实希望赋能社区，让社区创造更多项目并将其开源，从而真正形成滚雪球效应，使人们能够利用彼此的产出。这正是

why software grows so quickly cuz it's so easy just to go and fork someone's project on GitHub. So, and like we were able to scale from basically like I don't know 10 seeded projects, open source projects like 8 weeks ago to over 10,000 today like seeing these type of network effects I believe can become a mode over the time. When we look at your growth, fundraising is a big part of it.

软件发展如此迅速的原因，因为直接在 GitHub 上复刻别人的项目实在太容易了。我们实现了规模扩张，从大约 10 个种子项目和开源项目，发展到如今超过 10,000 个，而这只用了 8 周。看到这种网络效应后，我相信它可以随着时间推移成为一道护城河。纵观你们的增长历程，融资是其中非常重要的一环。

It costs a lot of money to be able to spend four million on, you know, different aspects of, you know, uh, inference band.

要做到投入四百万，需要花费大量资金，包括用于不同方面的推理频段。

What was the best VC meeting you've ever had?

你经历过最棒的一次风投会面是哪一次？

Obviously, um, Yuri Milner gets gets it.

显然，Yuri Milner 很懂这一套。

How was that meeting? Like, was it was it in person?

那次会面怎么样？是当面交谈的吗？

Yeah, definitely in person. And definitely Yuri stays on top of all the trends. And how was it? Were you nervous?

对，当然是当面交谈。Yuri 始终紧跟所有趋势。 当时感觉如何？你紧张吗？

I I wouldn't say nervous. It was just more uh to see how much of the uh if we see the market the same way and I was truly surprised that Yuri deeply understands this transformation of content first and foremost. Obviously it starts with this direct to consumer AI ads. It starts with short form dramas.

我不会说自己紧张，更多的是想看看我们对市场的看法究竟有多一致。令我非常惊讶的是，Yuri 对这场内容变革有着深刻的理解，而这正是最重要的。这场变革显然始于面向消费者的人工智能广告，也始于短剧。

All these trends come from Asia to the west. And um also fundamentally we believe that most of contents on social and in the world is going to be AI assisted or AI generated and uh the and like this multi- trillion advertisement industry and you know like contextual advertisement is the main business model of the internet. It's all going to be substantially disrupted with video AI.

所有这些趋势都是从亚洲传到西方的。从根本上说，我们相信，社交媒体以及全世界的大多数内容都将由人工智能辅助或生成。规模达数万亿美元的广告行业，以及情境式广告这类互联网的主要商业模式，都将受到视频人工智能的巨大冲击。

This industry still going to be very valuable, but it's never going to be the same.

这个行业依然会非常有价值，但再也不会和过去一样。

Did you know when you left the meeting with Yuri that he was going to write the check?

你与 Yuri 会面结束并离开时，是否已经知道他会开支票投资？

You know, sophisticated investors, they can play games. I had like so many scars like people really shook hands said we do at this price and next day what I learned is that they called other investors and they pulled the syndicates and to invest in 30% lower valuation compared to what we discussed. So like look these things just happen so you never can be sure but it didn't happen with Yuri.

老练的投资者有时会使用一些手段。我已经吃过太多亏。有些人明明已经与我们握手，表示会按商定的价格投资，结果第二天我才知道，他们给其他投资者打了电话，组建了联合投资团，并以比我们讨论的估值低 30% 的价格投资。这种事确实会发生，所以你永远无法确定，但 Yuri 没有这么做。

I think there's a discount placed on Higsfield because you're not Silicon Valley insider. Like let's be clear you're at a billion in revenue now.

我认为 Higgsfield 的估值被打了折扣，因为你不是硅谷圈内人。直白地说，你们现在的营收已经达到 10 亿美元。

Yeah. If you were a Silicon Valley company, that would easily be a $25 billion company growing at the rate that you're growing in 18 months.

对。如果你们是一家硅谷公司，凭借 18 个月内如此迅猛的增长，估值轻松就能达到 $25 billion。

Yeah, you could also argue that's what cognition was valid at 50, right? So there is definitely an upside upper band even more. Yeah, 100%.

对。也可以说，Cognition 估值达到 50 时也是这种情况。因此，估值肯定还有上涨空间。 上限甚至更高。 对，完全同意。

So a couple things which I believe are very important. So first we build for long term. We have seen that direct to consumer space like e-commerce can be disrupted like Shopify is a great example how they become they have become infrastructure to build like direct to consumer businesses and we become infrastructure to essentially build distribution for direct to consumer businesses. That's one

我认为有几点非常重要。首先，我们着眼于长期发展。我们已经看到，面向消费者的领域可以被颠覆，电商就是一例。Shopify 很好地展现了如何成为构建面向消费者业务的基础设施，而我们将成为帮助面向消费者的业务构建分发渠道的基础设施。这是我们的一个愿景。

aspiration and second aspiration is obviously a plain like companies worth over $200 billion. It's insane. So look and as we think long term just this you know like these multiples don't don't matter that much as we know we're building long term we're going to be over 100 billion it's true that most of the people don't get the opportunity that we are going after the biggest industry in the world but I wanted to

第二个愿景与 AppLovin 这类公司的价值超过 $200 billion 有关，这非常惊人。从长期来看，这些倍数其实并不那么重要，因为我们知道自己正在从事一项长期事业，最终会超过 100 billion。确实，大多数人都没有意识到这个机会：我们的目标是全球最大的行业。但我还想再分享一个数字。

drop another another number so when I and I asked the team to double check so it's at least four people on the team who prove so it's not like random fact so I asked um When we look at public companies and we exclude pharma and big tech, spend on sales and marketing is higher than spend on R&D. Like what when it comes to sales and marketing, the goal is to deliver personalized offering which converts the best. A lot of that

我请团队对这个数字进行了复核，团队中至少有四个人验证过，因此这并不是随口一说的数据。我的问题是：如果观察上市公司，并排除制药企业和大型科技公司，就会发现它们在销售和营销方面的支出高于研发支出。销售和营销的目标，是提供转化效果最佳的个性化产品方案。当然，其中很多工作仍然需要人来完成。

is human work of course, but a lot of that is going to be personalized videos in one in some shape or form. So that's why I'm saying that um many people just and that's good for us that many people don't understand the opportunity this large market which we go after.

但也有很多工作将以某种形式转化为个性化视频。因此，我才会说，很多人其实并不了解这个机会——这对我们而言是好事——他们并不明白我们所瞄准的市场机会有多么庞大。

Can I ask you you've mentioned Asia short form dramas a lot. What percent of revenue is from Asia versus the west?

我能问一下吗？你多次提到亚洲的短剧。你们的收入中，亚洲市场和西方市场分别占多大比例？

Um so oh the west makes well over 70% of revenue. well over but just important to say that we learn a lot from trends coming from Asia like Higgsfield does not exist in China for example which is massive market for AI um Higgsfield uh but the largest city by usage is soul in South Korea while the largest country is obviously the United States what's the biggest lesson from Asia that you've learned

西方市场贡献了远超 70% 的收入。但有一点必须说明：我们从亚洲兴起的趋势中学到了很多。比如，Higgsfield 并未进入中国市场，而中国是一个巨大的人工智能市场。就 Higgsfield 的使用量而言，最大的城市是韩国首尔，最大的国家则显然是美国。 你从亚洲学到的最重要的一课是什么？

there is so much IP so many products coming from Asia and they all try to figure out distribution direct to consumer. That's why they lean into the new tooling like video AI which actually helps to achieve that. That's just very different mindset. They feel that they could do they could do way better if they could establish direct relationship with customer instead of having like some other layer. That's why

亚洲拥有大量知识产权，也涌现了大量产品。这些产品都在想方设法打通直达消费者的分销渠道，因此会积极采用视频人工智能之类的新工具，而这些工具确实有助于实现这一目标。这是一种截然不同的思维方式。它们认为，只要能够直接与客户建立联系，而不是在中间隔着其他层级，原本可以做得更好。

they go so many so much direct to consumer rather than using some resale platforms and so on. I sacrifice a lot of life for for the life that I have and the career that I have and I love it. Do you think you will one day regret spending a day with your son in 3 and 1/2 months? Look, this is goes even beyond that because my um from the age of 7 to 12, my mother had to work um three jobs. So, I didn't see her. My father was spending

因此，它们非常重视直达消费者，而不是使用转售平台之类的渠道。 我为现在的生活和事业牺牲了很多，而且我热爱这一切。你觉得自己有一天会不会后悔，在 3 个半月里只陪儿子一天？ 这件事的意义甚至不止于此。从 7 岁到 12 岁，我母亲不得不同时打三份工，所以我见不到她。我父亲则一直陪着我，和我一起参加各种活动。

all the time with me going to all and it was I was basically minor so he had to go to all these camps with me. Um I also play checkers. I was top three in the world. So we went we travel throughout the world and um then I did programming.

他全程陪伴着我。当时我还未成年，所以他必须陪我参加所有训练营。我还下过国际跳棋，世界排名前三。我们去了世界各地，后来我又开始学习编程。

He spent all the time with me like really dedicated his life to me like he did sacrifice and uh since 21st he has Parkinson disease. So um even like having some ability to capital and exits cannot fully change things and um this is something which is um deeply personal obviously.

他一直陪伴着我，真正把一生都奉献给了我，也确实为此做出了牺牲。从 21 年起，他患上了帕金森病。即使拥有了一些资本，也成功退出过，我仍然无法彻底改变这一切。这显然是一件非常私人的事。

Totally.

完全理解。

But you don't need to do what you're doing now. Alex, I didn't [snorts] need to anymore either. [laughter] I still am. I still miss family birthdays. I still miss weddings cuz like mine's about a deep insecurity rooted in me being a fat kid.

但你现在其实已经没有必要再做这些了，亚历克斯。 我也早就没有必要再这样做了。[哼笑][笑声] 但我仍在继续。我依然会错过家人的生日，也依然会错过婚礼。我的动力源自一种根深蒂固的不安全感，而这种不安全感源于我小时候是个胖小孩。

um why are you doing it?

你又为什么这么拼？

So I think Mark and Jason actually described it really well. There are like five archetypes. So obviously for me it's just huge conviction about the technology, about the market, about the opportunity and just huge fear of missing that huge fear of missing that.

我认为 Mark 和 Jason 说得非常到位。大概存在五种原型。对我而言，原因显然是我对这项技术、这个市场和这个机会抱有极强的信念，同时又极度害怕错过，害怕错过这一切。

But remember that um my parents really taught me that um there is a place in the world where technology like good technology products matter. I remember like when I was six there was um like this I guess magazine about Bill Gates like building Microsoft and not being like very like socially accepted everywhere back then and like my mother just told me oh like these examples basically happened in the

不过，我父母确实让我知道，世界上有一个地方非常重视技术，或者说重视优秀的科技产品。我记得六岁时看过一本杂志，内容应该是关于 Bill Gates 创办 Microsoft，以及他当时并未得到社会各界广泛接纳。母亲便以此告诉我：

world. I think she didn't fully understand like San Francisco and Seattle are different cities but still uh this that's still deeply rooted in me.

“世界上的确发生过这样的事。”我想，她并不完全清楚 San Francisco 和 Seattle 是两座不同的城市，但这件事依然深深扎根在我心中。

Childhood shape us a lot.

童年对人的塑造确实非常深刻。

Yeah. What did your parents teach you?

你的父母教会了你什么？

For them, what was important is to just be in merit-based environment sort of um and um that's why getting to uh California felt so important.

他们认为，身处一个以能力为导向的环境非常重要。因此，能够前往 California 对我而言才会如此重要。

What's your biggest lesson on hiring?

你在招聘方面获得的最大经验是什么？

Speaking of a merit-based environment, we see a lot of uh focus on your cognitions of the world who hire mass Olympiads.

谈到以能力为导向的环境，我们发现你非常看重应聘者对世界的认知，也会大量招聘奥赛选手。

Yeah.

当然。

What's your biggest lessons on hiring effectively?

关于如何高效招聘，你获得的最大经验是什么？

I think one of the things why Europe thrives so much like I know that you typically say otherwise but let me just challenge you like who are the most relevant NeoClouds today? It's Nscale, iron and Nobus and Cruso.

我认为，Europe 之所以能够如此蓬勃发展，这是其中一个原因。我知道你通常持不同看法，但我还是想挑战一下你的观点：如今最重要的 NeoClouds 有哪些？包括 Nscale、IREN、Nebius 和 Crusoe。

Mhm. Brusso okay Silicon Valley story I am from Australia and scale from the UK and anobus is UK and Netherlands let's talk about the companies on application layer they that matter I know that you mentioned Merore and you mentioned Harvey but Legora 11 Labs lovable they all deeply matter so if we just go outside of the model layer h cuz then I don't want to go into the mistral topic right but if we go cuz I think like by

Crusoe 是一个硅谷故事；IREN 来自 Australia，Nscale 来自 UK，Nebius 则来自 UK 和 Netherlands。再来看看应用层那些举足轻重的公司。你提到了 Mercor 和 Harvey，但 Legora、ElevenLabs、Lovable 同样举足轻重。因此，只要跳出模型层，情况就会有所不同。我不想讨论 Mistral；但就 Mistral 而言，我认为从

usage they have the numbers are very strong but people for some reason don't don't believe in that so I don't know why but public public data shows that the usage is there but on every other layer Europe is extremely competitive like ASML like without ASML this whole thing just wouldn't happen so I I think fundamentally what's matter is if if like Europe is going to figure out energy but that's goes outside of that's

使用量来看，其数据非常强劲。不知为何，人们始终不相信这一点。公开数据显示，使用需求确实存在，而在其他每一个层面，欧洲都极具竞争力。例如，没有 ASML，整个行业根本无法发展。因此，我认为从根本上说，关键在于欧洲能否解决能源问题，但这超出了我的

above my pay grade right um so very important to say here is that um now there are more opportunities to create company from um different kind of cities from different parts of the world while before it all felt extremely centralized um and we are we are excited uh we are obviously excited about that and um another thing about hiring um is that in Silicon Valley unfortunately what I'm seeing is that people just jump between

能力范围。非常重要的一点是，如今人们有了更多机会，可以在不同类型的城市创办公司，也可以在世界各地创业，而过去的一切都显得高度集中。我们对此非常兴奋。关于招聘，还有一点令人遗憾：我在硅谷看到，人们每隔

jobs every two years that's why um I think Europe can be so competitive because the sense of loyalty matters a lot and that goes sort of a little bit to the childhood. We just discussed that like let's say if you're a Fulham fan you're not going to root for Arsenal just because they won or played in the U Champions League final. But in the United States, uh if uh Lakers are on the top, people are going to say, "Yeah,

两年就会跳槽。因此，我认为欧洲可以极具竞争力，因为忠诚度非常重要，而这在某种程度上也与童年经历有关。正如我们刚才谈到的，如果你是 Fulham 的球迷，就不会仅仅因为 Arsenal 赢了比赛或打进 Champions League 决赛，便转而支持 Arsenal。但在美国，如果 Lakers 正如日中天，人们就会说：“我是

I'm I'm fan of Lakers because it's just makes it easier to start conversation."

Lakers 的球迷，因为这样更容易打开话题。”

You know, when you think about your own CEO style, what's changed most in AI? It's so important to look at actual signals and actual adoption and having access to raw information. Um I was obviously taught the corporate school of management in the United States. Um and when I look at the CEOs whom um whom I'm learn from is obviously Jensen, Elon and

当你审视自己的 CEO 风格时，其中变化最大的是什么？ 在人工智能领域，关注真实信号和实际采用情况，并能够接触原始信息，至关重要。我接受的是美国企业式管理的教育。至于我所学习的 CEO，主要是 Jensen、Elon 和

Nick. Um Nick was on the show. So like obviously like those three are those three they completely abandon all the management principles. They don't necessarily are like fans of like one-on-one and like soft feedback. All of them I think are encouraged like being down to the points knowing the details while it would be called in like corporate America something like micromanagement.

Nick。Nick 也上过这个节目。这三位彻底抛弃了所有传统的管理原则。他们未必热衷于一对一沟通和委婉反馈，而是都提倡直奔主题、掌握细节。在美国企业界，这种做法可能会被称为微观管理。

What management principle do you disregard that many people think is important?

有哪些管理原则是你并不认同，但许多人认为很重要的？

I do believe that it's as simple as hire the best people to do the best work and figure out how to retain them.

我认为事情很简单：招到最优秀的人，让他们做出最好的成果，然后设法留住他们。

Everything else is frankly secondary and people just create so much theory around that and and essentially there is just so many like fake rules uh which are disconnect from reality. It's it's really as simple as hire the best people, empower them to do the best work and just figure out how to establish relationship and retain them.

坦率地说，其他一切都是次要的。人们围绕这件事构建了太多理论，形成了太多脱离现实的虚假规则。归根结底，就是招到最优秀的人，放手让他们做出最好的成果，再设法与他们建立良好关系并留住他们。

A lot of them bluntly are do see dollar signs. We mentioned the transactional nature of America and secondaries are a part of that. How do you think about doing annual tenders to retain people across our team? Um roughly 50 are in um California. Uh maybe we're going to get to roughly 50 remotes and um over 300s in Kazakhstan. So look, I just hope we're going to print uh more dollar millionaires in Kazakhstan, in Central

事实上，他们当中的许多人确实看重金钱。我们提到过美国社会的交易属性，老股交易也是其中一部分。你如何看待通过每年进行要约收购来留住人才？ 我们整个团队中，大约有 50 人在 California。之后可能会有约 50 名远程员工，而哈萨克斯坦会有 300 多人。我希望我们能够造就更多来自哈萨克斯坦、中亚

Asia, in this part of the world uh than any other company.

以及世界这一地区的百万美元富翁，数量超过其他任何公司。

I I I do too. Um what's the labor arbitrage on cost between Kazakhstan and the US?

我也希望如此。哈萨克斯坦和美国之间存在多大的劳动力成本套利空间？

I I know that a lot of people when they look at Hixel, they think about the arbitrage first and foremost like the way is that not true? Look like Kazakhstan is top five in the world in physics.

我知道，许多人关注 Higgsfield 时，首先想到的是成本套利。 难道不是这样吗？哈萨克斯坦的物理水平位居世界前五。

Like you look at the recent international physics olympiad for high schoolers like they're top five in the world on par with like the United States, China, India and this is also like the core of our team are people who won international competitions in math and physics. Um that's the first part.

在最近的中学生国际物理奥林匹克竞赛中，他们位列世界前五，与美国、中国、印度等国处于同一水平。我们团队的核心成员正是那些曾在国际数学和物理竞赛中获奖的人。这是第一点。

The second part is that about Kazakhstan is that they actually took this Soviet school of math but really upgraded with Singaporean principles and Singaporean system of education is considered to be probably the best in the world. At least many people in Silicon Valley believe that. Um and they and the government basically subsidizes for thousand of high schoolers to study abroad and many of these people come back. Um and there

关于哈萨克斯坦的第二点是，他们继承了苏联的数学教育传统，又运用新加坡的理念进行了全面升级。新加坡的教育体系被认为可能是世界上最好的，至少硅谷有很多人持这种看法。此外，政府基本上会资助数千名高中生出国留学，其中许多人后来回到了哈萨克斯坦。

is strong desire just and so the just the density of talents uh definitely got there. It's uh like top 10 largest countries in the world. So over 20 million population and we are also actively hiring bringing their talents from Europe from other countries in Asia and people just enjoy like some benefits like 15% personal income tax. Yeah man, it's like don't even get me started in [ __ ] UK will tax you to breathe. Uh, seriously,

当地有着强烈的引才意愿，因此人才密度确实已经达到了相当高的水平。哈萨克斯坦是全球面积最大的前10个国家之一，人口超过20 million。我们也在积极招聘，从欧洲及亚洲其他国家引进人才。人们也很享受这里的一些优势，比如15%的个人所得税。哥们，这简直就是……别跟我提[ __ ]英国了，那里连呼吸都要收税。真的。

it's in the UK, you get your, you know, paycheck and then it's like, I don't 100,000 and then you get the end and it's kind of like 3,500.

在英国，你拿到工资单后一看，本来是100,000，最后到手差不多只剩3,500了。

But it's also English common law, so it's not like that bad as people think.

但这里采用的也是英美法系，所以实际情况并没有人们想象的那么糟。

Uh, you move it. Let's swap places. Do you have a mega pad in Kazakhstan?

你搬过来吧，我们俩换个地方。你在哈萨克斯坦有一套超级豪宅吗？

No, I don't. I don't own any property.

没有，我没有任何房产。

What? Why?

什么？为什么？

Remember that I come from Asian family.

别忘了，我来自亚洲家庭。

So um whenever we sold the company, I made over a million dollars and I spent all this money buying apartments for my parents, relatives, my wife parents cuz it's just part of the culture and the f like extended family is not small by any means. Uh but look, it's just part of the culture to give back. And then um when it comes to the family, especially to my parents, they obviously sacrificed a lot. So I I felt like I had to give

我们卖掉公司时，我赚了超过一百万美元，随后把这些钱全部用于给父母、亲戚和岳父岳母购买公寓，因为这就是文化的一部分，而且我们的大家庭一点也不小。回馈家人是这种文化的一部分。尤其是我的父母，他们显然牺牲了很多，所以我觉得自己必须有所回报。

back at least at least like things like monetary things which I which I could do but I drive like Tesla Model 3 like and I le so like I I'm not like a guy who's going to just show up with Lamborghini or Porsche.

至少在金钱这类力所能及的方面，我应当回报他们。不过，我开的也只是Tesla Model 3。我不是那种会开着Lamborghini或Porsche招摇过市的人。

Do you invest? We mentioned solve intelligence. um when before I did that but now I spend roughly 90 hours a week 80 90 hours a week on Higgsfield. I try to spend ideally um at least um 3 hours a week with my wife at least 5 hours a week with my son. Um sometimes I do the catch up because when I travel um for a week, for two weeks, for three weeks, then I try to take Sunday off to spend the whole

你会投资吗？我们刚才提到了Solve Intelligence。以前我会投资，但现在我每周大约有90 hours，也就是80 90 hours投入Higgsfield。我希望每周至少能花3 hours陪妻子，至少花5 hours陪儿子。有时我会把缺少的陪伴时间补回来。出差一周、两周或三周后，我会尽量在周日休息一整天。

day with my son. And over the last 3 months, yes, I was able to find one day when I spent like end to end with my son without emails, without talking to without talking to the team members. I get in trouble for this, but I think there's no um shortcut to hard work. The harder I work, the luckier I get. I meet more founders. I find more great companies. I do more shows. I have more success. Do you buy the [ __ ] of the

那一天我会全程陪伴儿子，不看邮件，不与别人交谈，也不与团队成员沟通。过去3 months里，我确实成功挤出了一天。我这么说可能会惹上麻烦，但我认为努力工作没有捷径。我越努力，就越幸运；我会认识更多创始人，发现更多优秀的公司，制作更多节目，并取得更多成功。你相信所谓的平衡那套[ __ ]吗？

balance and uh oh, it's okay. You can leave at 5 and be home for bath time and crush it. This is a good question. So, look, obviously um being an immigrant, I always have that I have to prove like that I belong, right? So, I hope that I feel like now people accept people recognize that Higgsfield is probably a top 10 um application AI companies by revenue, probably number one. But I think when it comes to um hard work like

你相信一个人可以5点下班，回家给孩子洗澡，同时还能大获成功吗？这是一个好问题。作为一名移民，我一直觉得自己必须证明自己属于这里。现在，我想大家已经接受并认可Higgsfield可能是按营收计算排名前10的人工智能应用公司，甚至可能位居第一。但说到努力工作，我认为还需要看看其他人的情况。

the people whom we know in common like we we talked about like let's say Peter Salis like legend in the cons in consumer space obviously Jack look I I spent decent amount of time with them and other product leaders at stamp like the density of product talent and stamp was unprecedented all of them work really hard all of them are smart I I like none of them just uh checks emails for five hours a day and and calls it

比如我们共同认识和谈论过的Peter Sellis。他显然是消费产品领域的传奇人物，Jack也是如此。我与他们以及Snap的其他产品负责人共事了相当长时间。Snap的产品人才密度前所未有。他们每个人都工作得非常努力，也都非常聪明。没有人每天只花五个小时查看邮件，然后就把这称为工作。

work. Each of them is deeply rooted into the recent trends in product product design activation. They know data really well. So yeah, I don't believe that there is any shortcut to hard work.

他们每个人都深入了解产品、产品设计和激活方面的最新趋势，也都非常懂数据。因此，我不相信努力工作存在捷径。努力没有任何捷径。

3 hours a week with your wife. Yeah, I don't know about you, dude. Mine would dump me for 3 hours a week. How do you make marriage [clears throat] work [laughter and gasps] on three hours a week?

每周只陪妻子3小时？哥们，我不知道你是怎么想的。要是我每周只陪妻子3小时，她肯定会甩了我。你是如何让婚姻[清嗓子]在每周只相处3小时的情况下维持下去的？[笑声与惊呼声]真的是每周吗？

Yeah, look, I'm I'm I'm I'm very um I'm I'm very grateful for my wife for being patient, you know. It's also very different if that's like Asian culture.

是的。我真的非常感激妻子一直以来如此有耐心。这可能也与亚洲文化有关，因此情况很不一样。

Uh it's just kind of more natural to try to do sacrifices for each other sort of.

彼此尽量为对方作出牺牲，是一件更自然的事。

Um, and I'm deeply I'm obviously deeply grateful for her for supporting me. But like sometimes at this scale I get invited to parties. I always send her some and don't show up myself. I don't know if I piece people off, but this happens um very frequently.

我当然非常感激她对我的支持。但公司达到现在的规模后，我有时会受邀参加派对。我总是让她代替我出席，自己却不露面。我不知道这是否会惹恼别人，但这种情况经常发生。

So wait, you say yes and then she goes, "Yeah, I say maybe we both can some come together." Then there is always some urgent fire last minutes and my wife just goes. [laughter] What fire was most urgent? What was the Oh [ __ ] Yeah. Look, I think obviously for all the things which we touch base earlier whenever we are not very good in communicating the features or we felt like I mean now it's like team of 40 so

等一下，你会先答应参加，然后她会说：“对，也许我们俩可以一起去。”但到了最后一刻，总会出现一些需要处理的紧急状况，结果只有你的妻子独自前往。[笑声]哪一次的情况最紧急？当时发生了什么？哦，[ __ ]，对。对于我们之前谈到的所有事情，每当我们没能清楚地传达产品功能，或者出现其他问题时，都需要紧急处理。现在团队已经有40人了。

now the life is way better but early days obviously I was involved in all the fires. Um I think recently um all the types of like attacks on AI companies.

因此，现在的日子好过多了；但在早期，所有紧急状况显然都需要我参与处理。最近还出现了各种针对人工智能公司的攻击。

It's crazy. It's like it's like LLMs are being used to hack companies. It's like new types of LLMs to do some frauds, you know, like basically bots using credits and then doing auto refunds. All of that. Look, I like since I have like kind of machine learning background myself, data science backgrounds, I still can move a needle substantially when it comes to statistics and data. So yeah, I have to be involved somehow. But

这种情况太疯狂了。有人开始利用大语言模型入侵公司，也有新型大语言模型被用于实施诈骗。例如，机器人会盗用点数，然后自动申请退款，诸如此类。因为我拥有机器学习和数据科学方面的背景，所以在统计和数据方面依然能产生很大影响。因此，我必须以某种方式参与处理这些问题。

like these LLMs, they they amplify many types of behaviors including various types of attacks and fraud, but and we have to fight against that. Uh we're going to do a quick fire around. So I say a short statement, you give me your immediate thoughts. What have you changed your mind on most in the last 12 months?

这些大语言模型会放大许多类型的行为，包括各种攻击和欺诈行为，而我们必须与之对抗。接下来进行一轮快问快答。我会说一句简短的话，你立刻告诉我你的想法。过去12个月里，你在哪件事上的看法变化最大？

Oh, I was thinking that HubSpot is going to get obsolete. Everyone is going to build their own CRM and but when especially when as we hire and scale B2B go to market team just having familiar interface matters a lot.

我原本以为HubSpot会被淘汰，所有人都会构建自己的CRM。但当我们招聘并扩充B2B市场拓展团队时，我发现使用熟悉的界面确实非常重要。

Wow. I would still say they're going to get [ __ ] You think that just stickiness is there with SMBs?

令人意外。不过，我仍然认为他们会[ __ ]。你觉得那种用户黏性在中小企业中依然存在吗？

Yeah, I I I do think so. And especially I see that when I hire go to market talents.

对，我确实这样认为。尤其是在招聘市场拓展人才时，我对此感受更深。

Wow. Why? Like what is it about hiring them that makes you think that just they're so used to it?

为什么？在招聘过程中，是什么让你觉得他们已经过于习惯使用它了？

I mean like people who are very good in understanding customers and talking to customers they may not just simply accept new interface so quickly and just having HubSpot as a system of records being able if if there is any mismatch going able to just understand where the data flow went wrong. I think that's just still very valuable like the familiarity. What do you believe today that everyone else thinks is [ __ ]

那些非常擅长理解客户并与客户沟通的人，可能不会这么快接受新界面。他们只需将 HubSpot 作为记录系统，一旦出现任何不匹配，就能直接查明数据流究竟在哪里出了问题。我认为这种熟悉感依然非常有价值。你现在相信什么，而其他所有人都觉得这简直是[ __ ]

crazy? I mean, look, I think uh people just still don't fully appreciate that most of the content on social media is going to be AI generated. There are going to be some shows like obviously yours where it's like authentic contents. It's going to be 10 15x higher CPM whatsoever than AI generated contents. So it's going to be it's going to be way less in terms of like content create by but it's going to create way

疯了？我认为，人们仍未充分意识到，社交媒体上的大多数内容都将由人工智能生成。当然，也会有一些节目，例如你的节目，其内容显然非常真实。这类内容的 CPM 会比人工智能生成的内容高出 10 到 15x，甚至更多。因此，这类内容的创作量会少得多，但它创造的

more value uh than a generated content.

价值会远高于生成式内容。

But even when I look into your content specifically like you made multiple very successful shorts millions of views better than anyone else in this space and you do a lot of overlay. While the content is authentic, I think we should do better job so that you use Higgsfield at least for the overlay on top of existing videos. Dude, I would love that. I mean, again, they take 3 hours.

不过，即使只看你的内容，你也制作过许多非常成功的短视频，播放量达到数百万，在这个领域比任何人都做得更好，而且使用了大量叠加画面。虽然内容是真实的，但我认为我们还能做得更好，让你使用 Higgsfield，至少可以用它为现有视频添加叠加画面。我非常愿意尝试。需要再次强调的是，这些内容要花费 3 hours。

So, people don't know this. I spend 2 hours a day just doing Instagram. Now, we decided that Instagram short form is going to be a big new push for us. Um, 2 hours a day just for me. I write the scripts and then I record them. And then it's two people, six hours per one for those three.

大家可能不了解这一点。我每天仅制作 Instagram 内容就要花 2 hours。现在，我们决定将 Instagram 短内容作为接下来重点发展的新方向。仅我本人每天就要投入 2 hours，先写脚本，再进行录制。之后还需要两个人为每条视频各投入六小时，才能完成那三条视频。

And that's extremely smart of you. Like you know like going back to some of the topics is like clipping is like a huge topic and that's like has its own upsides and downsides. But obviously everyone sees this opportunity to win to build massive top of funnel like hundreds of millions of views with short form content. as long as you can have downstream monetization like or value creation like you do.

你这样做非常明智。回到之前讨论的话题，内容切片是一个非常重要的议题，本身既有优势，也有弊端。但所有人显然都看到了借此取胜的机会：通过短内容获得数亿次播放，从而建立庞大的漏斗顶层流量。前提是你能够像现在这样，在下游实现变现或创造价值。

Totally agree with you. What job today does not exist that will be big in 5 years? Okay. So in 5 years people especially in our space creative directors they are going to be talking to computers and generating stories real time and video and AI is going to help to create multiple variations. Today there is no word to really describe that because there is also there are script writers um there are then uh

完全同意。现在尚不存在、但会在 5 年后大行其道的工作是什么？5 年后，人们，尤其是我们这个领域的创意总监，将会讨论如何通过向电脑下达指令，实时生成故事和视频，而人工智能将帮助他们创作多个版本。如今还没有一个词能够准确描述这种角色，因为它既涉及编剧，还有

screenwriters like those who are going to break it down shot by shot then there are people who do that storyboarding then there is like people who person who oversees all of that like movie director and so on. There are so many there are so many there are so many parts of that but eventually taste is going to matter a lot and just having stories to tell there is no word to describe it today.

负责将内容逐个镜头拆解的人、负责制作分镜的人，以及统筹所有环节的人，类似电影导演。这里涉及的环节和组成部分非常多，但最终，品味将变得极其重要，同时还必须有故事可讲。如今还没有一个词能够描述这种角色。

Who do you not have on your board that you would most like to have on your board? Maybe out of like more professional CEOs, I'm definitely Frank Slutman because going back to the point I was just curious all the time, does no [ __ ] culture exist in California or not? Can it allow to scale companies so quickly? Can it is it possible to build successful enterprise go to market motion with no [ __ ] culture? And

目前不在你的董事会里、但你最希望邀请加入董事会的人是谁？如果从更职业化的 CEO 中选择，我肯定会选 Frank Slootman。回到刚才的话题，我一直很好奇：加州究竟有没有“不扯淡”的文化？这种文化能否让公司如此迅速地扩张？有没有可能既打造出成功的企业级市场进入体系，又保持“不扯淡”的文化？

when I read his ampitab book like book called amp it up, I realized it's possible. So like I'm a huge fan. I watched all his interviews.

读完他的《Amp It Up》后，我意识到这是可行的。因此，我是他的铁杆粉丝，也看过他所有的访谈。

The challenge with him, he's amazing.

他的问题在于，尽管他确实非常厉害，但也存在不足。

He's the best leader by far. But the challenge is you can sometimes do it at the sacrifice of product advancement.

到目前为止，他是最优秀的领导者。但问题是，这种做法有时会以牺牲产品进步为代价。

And so he built a GTM machine at Snowflake, but data bricks wiped the floor because they move product as the priority, not GTM. And that was dangerous. I preferred Chad Pets. Do you know Chad Pet?

他在 Snowflake 打造了一台 GTM 机器，但 Databricks 却凭借优先发展产品而非 GTM，将他们打得落花流水。这一点非常危险。我更喜欢 Chad Peets。你认识 Chad Peets 吗？

Oh, dude. This guy is no [ __ ] I'll introduce you afterwards. He's the best sales leader in the world. Um, and he is no [ __ ] [ __ ] Unbelievable.

这个人绝不扯淡。我稍后介绍你们认识。他是全世界最优秀的销售领导者，而且从不扯淡，简直令人难以置信。

And we probably should have him on board. [laughter] Oh my god. I find any way to have him on board. He is terrifyingly good. Um, so what's the biggest lesson from Snap?

那我们或许应该邀请他加入董事会。[笑声]天哪，我会想尽一切办法让他加入董事会。他强得吓人。那么，你从 Snap 学到的最大教训是什么？

The momentum doesn't last forever. Um, like today, Snap market cap is is below 15 billion. There are a lot of mimis on the internet, but this is a great company. cares so much about trust and safety and experience and it puts it first.

势头不会永远持续下去。如今，Snap 的市值已经低于 15 billion。网上有很多梗在嘲讽它，但它仍是一家很棒的公司，非常重视“信任与安全”以及体验，并将这些放在首位。

Do you think it is a great company? No offense. It's like it's been mismanaged as [ __ ] Its SBC is through the roof.

你认为这是一家很棒的公司吗？无意冒犯，但感觉它一直管理得像[ __ ]一样。它的 SBC 高得离谱。

It's tough to say it's a good company.

很难说它是一家好公司。

That's why I say that momentum doesn't last forever. When Snapchat was worth eight $80 billion and the gap with Meta was less than 10x, then it felt, oh, we just go explore. we we just we just really must lean in. Um but momentum doesn't last forever and that's my core learning. So that's why like while we do have the positive momentum, we don't we do not take this for granted. Clearly um like the nature of uh capitalism is

所以我才说，势头不会永远持续下去。当 Snapchat 的估值达到eight $80 billion，而且与 Meta 的差距不到 10x 时，当时感觉就是，哦，我们只管放手探索。我们，我们就是，我们就是真的必须全力投入。嗯，但势头不会永远持续下去，这是我最核心的体会。所以说，趁着我们现在势头正盛，我们绝不能把这一切视为理所当然。显然，嗯，资本主义的本质就是

there are ups and downs and since we're building long-term, we just should capitalize on the opportunity like with the fundraising and just keep pushing progress every day. What is the reason why the divergence between Meta's market cap and Snap's market cap has increased so significantly if there was one reason? Just maybe saying this trait, a lot of public companies did not figure out their AI story.

有起有落。既然我们着眼于长期发展，就应该抓住机会，例如通过融资，并且每天持续推动进步。如果只能归结为一个原因，为什么 Meta 与 Snap 之间的市值差距会扩大这么多？也许可以归结为一个共性：许多上市公司都没能讲清楚自己的人工智能故事。

Um, Snap unfortunately is part of that.

很遗憾，Snap 也是其中之一。

We have seen other great companies like Figma trying to tell their story. You mentioned Canva. It's not necessarily easy to be successful in private markets and public markets. And Zach is one of the best CEOs of all time cuz he managed that. He's such a [ __ ] beast. He's such a beast. You watch him last night with the event and you're just like, "Ah, [sighs] now I get it." Like that totally makes

我们也看到过 Figma 这样的优秀公司努力讲好自己的故事。你还提到了 Canva。要同时在私募市场和公开市场取得成功并不容易，而 Zach 是有史以来最优秀的 CEO 之一，因为他做到了。他简直是个[ __ ]猛人，实力太强了。昨晚看到他出席那场活动时，你会产生一种“现在我懂了”的感觉，一切突然都

sense. And you know what? Scale with Alex Wang. I was one who was like really like what's gonna he he he basically acquired a second CEO you know Alex is now the CEO of Muse and he's crushed it crushed it. What an effective buy for 0.5% of your market cap. Do you know what I mean?

说得通了。还有 Scale 和亚历克斯·王那件事。我当时非常怀疑，不知道事情会如何发展。他基本上相当于收购了第二位 CEO。亚历克斯现在是 Muse 的 CEO，而且他做得非常成功。这笔收购太划算了，只花了你市值的 0.5%。

Yeah. Look, but this happens with Instagram with WhatsApp. That's why I'm saying that we just maybe should put Meta a little bit in its own league.

是的，Instagram 和 WhatsApp 的收购也是如此。因此，我才认为或许应该将 Meta 单独归为一个级别。

Yeah, but he got rid of Cyrum and Kger.

是的，但他把 Cyrum 和 Kger 都淘汰了。

Here he's been like, "No, no, no. You, Alex Wang, are my guy."

他一直强调：“不，你，Alex Wang，才是我要找的人。”

Do you see what I mean?

你明白我的意思吗？

Yeah. The best talent hire.

明白。他招到了最优秀的人才。

Look, I I do believe that it's a little bit early to look at whole Meta AI initiatives. We probably need to see like year of like successful launches and so on to and then we can look back and see what was good, what was not good. But at least the consistency of storytelling and explaining what he is doing to public investors being able to articulate why Muse is so different um is phenomenal.

我确实认为，现在评价整个 Meta AI 计划还为时过早。我们可能需要观察一年左右，看看有哪些项目成功发布，然后再回过头判断哪些方面做得好，哪些方面做得不好。但至少，他在叙事以及向公开市场投资者解释自己在做什么这两方面始终保持一致，也能够清楚阐述 Muse 为何如此与众不同，这一点非常了不起。

Okay.

好。

Revenues are a billion. What are the revenues in 12 months time?

目前的营收是 10 亿。12 个月后，营收会达到多少？

Our current business model uh projects uh 4.5.

我们目前的业务模型预测是 4.5。

It says by the end of the next year but this basically involves substantial deceleration and that's what like just my finance team like there are couple strong quant people they told me that's just how the business works but look we are still pushing to grow at least 30% month over month what do you think it is they said 4.5 what do you think it is this is me to you not me to your finance team over 10

模型显示，这一数字将在明年年底前达到，但该预测实际上假设增长会大幅放缓。我的财务团队也是这样认为的。团队中有几位很优秀的量化人才，他们告诉我，业务规律就是如此。不过，我们仍在努力保持至少 30% 的环比增长。你认为会达到多少？他们预测是 4.5，但我现在问的是你，而不是你的财务团队。答：超过 10。

over 10 let me tell you why like in a lot of adoption in creative AI space is driven by monetization like all these direct to consumer brands making more ads and also having like the aspirational um cinematic AI content as this inspires creatives to explore the tooling.

超过 10。我来说明原因。创意人工智能领域的许多应用都由变现推动。例如，所有这些直面消费者的品牌都在制作更多广告；与此同时，那些令人向往的电影级人工智能内容也在激励创意人士探索这些工具。

I it feels to me that Hollywood starts to embrace AI mostly today as a way to as a tool for hybrid production as a just new form of CGI.

我认为好莱坞已经开始接纳人工智能，目前主要将其作为混合制作工具，视为一种全新的计算机生成影像形式。

But the sentiment really shifted from like strictly negative to neutral to slightly negative. And in private conversations, yes, there are maybe more than half of sier talents who is going to say we're anti-AI forever.

但人们对人工智能的态度确实已经从完全负面转向中性偏负面。在私下交流中，可能仍有超过一半的资深人才表示，他们永远反对人工智能。

But increasingly there are more and more people who are actually asking a question. Can we tell more stories with AI? Can we overcome certain budget limitations which existed before? And maybe AI can help to tell new stories which we couldn't tell before. And I do believe this just change in perception that's at least comes from my conversations is extremely is extremely is extremely positive. If you are in a

但现在确实有越来越多的人开始思考：我们能否借助人工智能讲述更多故事？能否突破过去的某些预算限制？也许人工智能能够帮助我们讲述新的故事。这是我们此前无法判断的。从我与他人的交流来看，我确实相信，这种认知上的转变极其积极。如果你目前的营收是 10 亿，

billion today, 10 billion in 12 months.

那么 12 个月后就会达到 100 亿。

Where do you peg the next fund raise?

你认为下一轮融资的估值会是多少？

You know, if you're in a billion, say conservative multiple, you'd be like, you know, 15.

如果目前是 10 亿，按照保守的倍数计算，估值大约会是 15。

Um but if you're hitting 10 next year you're like paying end of year it's like 80. Look we are not chasing just the valuation cuz again the goal is just to make sure that the company can be uh sustainable over the time in public market. So there is a lot of company building to be done beyond just uh chasing the revenue. But I just do believe do you want to be public? Huh?

但如果明年能做到 100 亿，那么到年底时，估值可能会达到 800 亿。我们并不只追求估值，因为归根结底，我们的目标只是确保公司能够长期在公开市场上稳健发展。因此，除了追求营收，公司建设方面还有许多工作需要完成。但我确实相信这一点。你想上市吗？

Do you want to be public at some point?

你希望公司有一天能够上市吗？

Yeah, I do believe that Higgsfield has great potential to be bigger than Applain and Shopify because fundamentally like building is one part of that Shopify one layer of infrastructure. Then for coding there is obviously like a cloud, there is codeex but what matters is distribution over the time but distribution matters. You know that this better than any other business that's why we do what we do.

是的。我相信 Higgsfield 非常有潜力成长为一家规模超过 AppLovin 和 Shopify 的公司，因为从根本上说，构建只是其中一部分，Shopify 是基础设施中的一层。在编程领域，显然既有云服务，也有 Codex，但从长期来看，真正重要的是分发。分发至关重要，这一点你比其他任何企业都更清楚，因此我们才会做现在正在做的事。

Yeah, exactly.

是的，没错。

Dude, I cannot thank you enough for being so amazing on the show. You've been fantastic. I've loved doing it, you can tell. And you've been an amazing guest. So, I really appreciate you joining me today.

非常感谢你在节目中的精彩表现。你确实非常出色，我也非常享受这次访谈，这一点你应该能够看出来。你是一位非常优秀的嘉宾，非常感谢你今天来参加我的节目。

Thank you so much. It's a pleasure.

非常感谢，这是我的荣幸。