本集简介
双语字幕
仅展示文本字幕，不包含中文音频；想边听边看，请使用 Bayt 播客 App。
如果你问题的前提是，这件事能不能成功，或者那件事能不能成功，那么在AI领域，答案很可能是两者都能成功。
If the premise of your question is, is this going to be successful or that thing going to be successful, the answer in AI is probably and.
现在是投资科技市场最令人兴奋的时刻。
It's just the most exciting time to ever be an investor in the technology market.
要跟上它的步伐会很难。
And keeping up with it is going to be hard.
跟上它的步伐会很难。
Keeping up with it is going to be hard.
前沿模型还是开源，实验室还是应用。
Frontier models or open source, labs or applications.
大卫·乔治认为，AI领域许多重大的二选一问题，答案其实就是两者兼顾。
David George thinks the answer to many of AI's biggest either/or questions is simply and.
在本期杰克·奥特曼的《Uncapped》节目中，杰克与a16z的普通合伙人大卫·乔治坐下来，讨论为什么他认为AI在经济中的普及仍处于早期阶段。
In this episode from Jack Altman's Uncapped, Jack sits down with a16z General Partner David George to discuss why he believes we're still early in the diffusion of AI across the economy.
他们深入探讨了为什么如今AI的收入仍然集中在相对少数重度用户手中，当AI触及更广泛的知识工作者时会发生什么，以及为什么大卫预期前沿模型、开源模型和应用公司会共同成长。
They get into why today's AI revenues are still concentrated among a relatively small number of heavy users, what happens as AI reaches the broader knowledge workforce, and why David expects frontier models, open models, and application companies to grow alongside one another.
他们还把目光投向了AI之外的自动驾驶和机器人技术。
They also look beyond AI to autonomy and robotics.
为什么消费级AI最终可能成为最大的机会之一，以及当公司和它们进入的市场都可能变得比预期大得多时，这对投资意味着什么。
Why consumer AI could ultimately be one of the biggest opportunities, and what all of this means for investing when both the companies and the markets they're entering can become much larger than expected.
好的，大卫，我真的很期待做这次对话。
All right, David, I'm really excited to do this.
我一直很期待这次交流。
I've been really looking forward to this.
所以谢谢你过来。
So thanks for coming by.
很高兴和你聊聊。
Great to hang with you.
最近我一直想聊的一个话题，我觉得你在这方面很有见解，就是我的合伙人埃里克·维什里亚最近在《像最优秀的人一样投资》播客上说的，基本上就是他说：“一切都会顺利的。”
So one of the topics that's been on my mind a lot that I think you've got really good thoughts on is something that my partner, Eric Vishria, said recently on the "Invest Like The Best" podcast, and it was basically, he was like, "It's all gonna work."
那是在人们质疑到底是开源还是前沿模型的时候说的。
And it was in the context of people question, is it gonna be open source or frontier?
他说：‘开源会成功的。’
And he's like, "The open source is gonna work.
‘前沿模型也会成功的。’
The frontier is gonna work."
会是英伟达，还是新的芯片公司？
Is it gonna be Nvidia, or is it gonna be new chips companies?
他说：‘两者都会有。’
And he's like, "There's gonna be both."
你可以沿着AI的整个技术栈一直问下去，而Eric的核心观点是，人们问错了问题，总是在问是A还是B。
And you can go on and on through the stack in AI, and Eric's point was basically, people are asking the wrong question of just, is it this or it's that?
他说：‘答案是两者都有。’
And he's like, "It's both."
我很好奇，你在Andreessen，坐在你这个位置上，是怎么看待这个问题的？
And I'm curious how you think about this at Andreessen and in your seat.
对，当然。
Yeah, of course.
所以在这个问题上，我和埃里克的观点非常一致，就是一切都会顺利发展。
So I am very closely aligned with Eric on this point that it's all gonna work.
我其实很喜欢他那种表述方式。
I actually like the way that he framed it up.
我的表述方式稍微有点不同，就是所有这些问题的答案都是“和”。
I framed it up slightly differently, which is just like the answer to all this is and.
所以如果你问题的前提是，这个会不会成功，或者那个会不会成功，在AI领域，答案很可能是“和”。
And so if the premise of your question is, is this gonna be successful or that thing gonna be successful, the answer in AI is probably and.
我觉得先看看我们现在处于什么阶段会很有帮助。
I think it's helpful to start just like, where are we?
现在，从基础设施建设的角度来看，我们已经超过了铁路时代。
So, right now, from an infrastructure build-out standpoint, we just surpassed the railroads.
占GDP的比例，这一点已经被广泛报道过了。
As a percentage of GDP, which has been well covered.
现在大概是3%左右吧。
It's like 3% now or something like that.
对，没错。
Yeah, exactly.
而且你看，以后建数据中心可能会越来越难，这个我们回头可以聊，但我觉得这个投入会翻十倍。
And look, it's probably gonna get harder and harder to build data centers, which we could talk about later, but I think that's gonna 10X.
所以现在的情况基本上就是——
So right now, what's happening is basically-
你觉得会占GDP的多少？
As a percentage of GDP, you think?
我觉得希望这能带动GDP增长。
I think hopefully this will drive GDP growth.
但这需要很多年。
And this is gonna take many years.
所以不会一年就翻十倍，也不会达到30%、40%。
So it won't 10X in one year, so it won't be 30, 40%.
但我们会花好几万亿美元。
But we will spend many trillions.
但我们确实会花费数万亿美元，对吧？
But we will spend many trillions, right?
所以所有这些成功，是我们在商业史上见过的最伟大、最快的成功，基本上就是OpenAI和Anthropic加在一起，再加上SpaceX AI，现在还有Cursor，加起来大概1200亿美元的收入，而且他们每月新增的收入比除了亚马逊之外的所有超大规模云服务商都要多。
So all of this success, which is the greatest, fastest success we've ever seen in business, which is basically OpenAI, Anthropic together, and then if you add in SpaceX AI, now with Cursor, call it 120 billion of revenue, and they add more revenue per month than all the hyperscalers except for Amazon.
这太疯狂了。
That's crazy.
然而，这一切其实只是靠一小部分人在付费支撑起来的。
Now, all of that is on the back of just actually a small amount of people paying for stuff.
所以，全球大约3000万程序员贡献了绝大部分收入。
So it's like 30 million coders in the world who are the vast majority of the revenue.
全球有15亿知识工作者。
There's 1.5 billion knowledge workers.
但基本上所有这些，正在增加的——
But basically all of this, which is adding-
……比历史上最成功的企业还要多的收入，其实只是靠一小群人撑起来的。
... more revenue than the best businesses that have ever been created is on the back of a small group of people.
这其实是个很有意思的观点，这些产品有超过十亿用户，我们通常也是这么想的。
That's actually a really interesting point, that it's over a billion users of these things, and we think about it that way.
但实际上，你是说只有三千万人在花大钱。
But it's actually, you're saying it's 30 million are spending big bill money.
就是三千万。
It's 30 million.
所以有十亿消费者用户，顺便说一句，我们得聊聊消费者市场，因为虽然我们有十亿用户，但在消费者领域我们目前几乎没什么存在感。
So there's a billion consumer users, and by the way, we should talk about consumer because we're nowhere in consumer right now, even though there's a billion users.
但在企业市场，基本上绝大部分收入都来自一小部分程序员。
But in the enterprise, basically, the vast majority of the revenue is coming from just a small amount of coders.
这就像幂律分布一样——
And this is like in power law terms-
而且就算在那三千万人里，我敢肯定也是幂律分布。
And even within that 30 million, I'm sure it's a power law.
那是个巨大的幂律分布。
It's a massive power law.
我相信你已经和你的公司聊过这些了。
I'm sure you've talked to your companies and all that stuff.
那3000万里的前100万人在花费——
The top million of the 30 are spending-
那3000万里的前100万人在花费巨额的金额。
The top million of the 30 are spending just a massive amount.
所以如果你乐观地看待这个现象，潜力远远大于目前已经扩散到那100万人之外的部分。
And so if you take an optimistic view of what that means, the potential is so much greater than what has actually diffused into everything beyond those million people.
再说一次，有15亿知识工作者，而我们只是在谈论B2B。
And so again, there's 1.5 billion knowledge workers, and we're just talking about B2B.
所以接下来会发生的事情规模巨大——
And so it's just a massive amount that is gonna happen-
……我认为会在未来五年内发生。
... in what I think is gonna be the next five years.
是的。
Yes.
现在需求完全是供不应求，对吧？
Right now, demand is totally insatiable, right?
在企业里，目前真正见效的，就是写代码、写文档这些。
In the enterprise, basically what's working, it's like coding, writing stuff.
你开始看到一些其他职能领域出现早期迹象了。
You're starting to see early signs in some functions beyond that.
法律就是一个我们已经看到开始起飞的领域。
Legal is one where we've kind of seen this take off.
我觉得大概落后12个月吧。
I think it's 12 months behind.
但对于广泛的脑力劳动，对于全球15亿的知识工作者来说，我们才刚起步。
But for broad knowledge work, for the 1.5 billion knowledge workers in the world, we're nowhere.
所以你就以这个为前提，那就是，我们已经看到了前所未有的规模，而扩散才刚开始触及表面。
So you just start with that as the premise, which is, we're already seeing scale that's never been seen before, and we're scratching the surface in diffusion.
在B2B经济中，渗透率还不到5%。
It's less than 5% diffused into the B2B economy.
这意味着我们会变得非常大——
That means we're gonna get really big-
……如果我们对它的普及和广泛应用判断正确的话。
... if we're right about this sort of diffusing and being applicable broadly.
所以这是其中一方面。
So that's one piece.
在供应端，我们受到极大的限制。
On the supply side, we are massively constrained.
极大的限制。
Massively.
基本上供应链的每个环节都是如此。
Like basically in everything in the supply chain.
对。
Yep.
所以基本上，你直到2028年才能获得数据中心容量，因为供应链的每一步都存在瓶颈。
So basically, you can't get data center capacity until 2028 because basically there's bottlenecks in every step of the supply chain.
所以这就是我们目前供需的现状，我觉得有必要先把这个基础讲清楚，因为这是我认为一切都会顺利运转的关键所在。
So that is where we are in the supply and demand, and I think it's important to lay that groundwork because it is the crux of why I think everything is gonna work.
是的。
Yes.
整个链条上的每一步都会顺利推进。
All of the sort of steps in the ladder are gonna work.
先说说供应端，目前是四吉瓦。
Talking about the supply side for a second, so it's four gigawatts right now.
你现在的看法是不是觉得，如果我们有八吉瓦在线运行，这些电力都会被充分利用起来？
Is sort of your view at the moment that if we had eight gigawatts live and online, it would all get used well?
对，会被用掉的。
Yeah, it would get used.
是的，我觉得都会被用掉。
Yeah, I think it would all get used.
都会被用掉。
It would all get used.
而且你看，可能价格会不一样什么的。
And look, maybe it would be at different prices or something.
我们可以聊聊代币现在的定价情况。
We could talk about pricing of what's happening for tokens.
对。
Yes.
但是是的，我觉得如果OpenAI、Anthropic或者SpaceX AI的算力上线更多的话——
But yeah, I think if you get more OpenAI and Anthropic or SpaceX AI capacity live-
……通常都会被用掉的。
... it generally gets used.
所以我想一路验证一下这些前提，显然，你也知道，我完全同意你说的。
So just to check the premises along the way, and obviously, as you know, I fully agree with what you're saying.
但为了验证一下，如果要出现那种情况——就是随着千兆瓦上线，代币没有被全部消耗——那需要发生什么？
But just to check it along the way, what would have to happen for it not to be the case that as gigawatts come online, all the tokens keep getting consumed?
有没有什么能让我们改变看法的，比如从4千兆瓦变成20千兆瓦，但代币并没有以相同甚至更高的价格被全部用掉？
Is there anything that could update our priors, that four gigawatts becomes 20, and they don't all get used at the same or even higher prices?
我觉得有两种情况可能会发生。
I think there's two forms where this could take hold.
一种是，编程领域的情况相对于B2B或知识工作的其他领域来说是个特例，而且由于公司很难从知识工作中直接产生商业价值，所以其他白领工作其实不会受到太大影响。
One is like What happened with coding is anomalous relative to everything else in B2B or knowledge work, and because of the difficulty of injecting commerce from knowledge work from companies, the rest of white-collar work just doesn't really have major impact.
我觉得这种可能性非常小。
I think that is extremely unlikely.
我觉得这更像是一个数据收集的问题——
I think it's sort of like a data collection problem-
确实很难相信
It's hard to believe
……以及模型应用的问题。
... and an application of the models problem.
另一种情况是，我们投资的一些团队正在研究算法突破，对吧？
The other would be, we do have teams that we've backed that are working on algorithmic breakthroughs, right?
所以，也许会出现一种算法突破，让我们不必再消耗——
And so, perhaps you would get an algorithmic breakthrough where, rather than having to consume-
哦，就是效率大幅提升。
Oh, just efficiency goes way up
……巨大的提升。
... massive.
对，效率大幅提升，无论是在训练还是推理方面。
Yeah, efficiency goes way up, both in training and in inference.
我觉得这不太可能。
I think that's unlikely.
我认为在训练方面取得突破是有可能的。
I think it's possible we could get breakthroughs in training.
我不知道，你可能看过我们构建知识库所需的数据统计。
I don't know, you've probably seen the stats on what it took you and I to build our knowledge base.
大概比模型实际训练所用的数据少两到三个数量级。
It's like, whatever, two or three orders of magnitude less than what the models actually train on.
所以那里似乎有机会。
So there seems to be an opportunity there.
不过推理仍然会存在，而且推理市场绝对巨大，无论如何都会比训练市场大得多。
You still will then have inference, which is an absolutely massive market that's going to be much bigger than training anyways.
所以，是的，我们可以在推理方面获得效率提升，但我觉得到目前为止我们学到的是，你在推理时思考得越多——
And so, yeah, we could get efficiency gains in inference, but I think what we've learned so far is the more you reason at the model-
——推理时思考得越多，结果就越好。
... at the time of inference, the better the results get.
所以我认为这在广泛的知识工作中会继续成立。
And so I think that'll continue to hold in knowledge work broadly.
所以我很难找到一种方式——
So I struggle to find a way-
——让这件事不成功。
... for this not to work.
现在，如果我们很快从五跳到一百——
Now, if we went to from five to 100 in a quick amount of time-
当然
Sure
而且产品本身也没准备好真正承担这些。
... and the products aren't built to actually take that on.
对。
Sure.
但我们不会遇到这个问题，因为要让这些东西上线真的很难。
But we're not going to have that problem, because it's going to be really hard to get this stuff online.
我很好奇你怎么看这个。
I'm curious what you think about this.
我也很难直接看到，但我听过最好的说法是和我一个朋友的对话，我们基本上在说，所有这些都上线了，所以一开始，几年前，有大规模的建设，当时你看到我们投入了数千亿美元。
So I also can't easily see it, but the best factor that I've heard was a conversation I had with a friend where we were basically saying, all this stuff has come online, and so at first, a few years ago, there was the big build-out, and you saw we're putting hundreds of billions in at the time.
然后模型变得好一点，你开始部署它们，然后你继续推进整个流程。
And then the models get a little bit better, and you start deploying them, and you keep moving through the pipeline.
现在你看到收入在人工智能的推动下疯狂增长。
And now you see revenues growing like crazy on the back of AI.
但那次对话的重点是，我们谈论的大部分收入来自卖token的人，而不是买token的人。
But the point in this conversation was most of the revenues that we're talking about are sellers of tokens, not buyers of tokens.
所以接下来我们需要看到的是，购买算力的客户确实获得了巨大的效率提升的数据。
And so the next thing we need to see is data around the buyers of tokens actually getting these huge efficiency gains.
换句话说，就是超大规模云服务商或者——
In other words, a hyperscaler or-
对，就像企业层面的投资回报率怎么样？
Yeah, like what's the ROI at the enterprise level?
是的。
Yes.
用户，写了这么多代码之后。
Users, of writing all this more code.
是的。
Yes.
你的业务因为写这些代码变得更好吗？
Is your business getting better by writing all this code?
或者换个说法，如果卖算力的赚了很多钱，而买算力的没赚到钱，那结果就不好了——
Or said another way, it would be a bad outcome if the sellers of tokens were making lots of revenue and the buyers of tokens were not making lots of revenue-
……因为买了大量代币。
... because buying lots of tokens.
听着，我对这一点非常乐观。
Look, I'm very optimistic about this point.
我觉得整个论点都建立在这样一个前提上：写更多代码不会带来实际的生产性用途——
I think that whole argument is predicated on the fact that there won't actually be tangible productive use of writing more code-
……更高效地写代码，而且效果更好。
... far more efficiently that is better.
我只是觉得这从根本上就是错的。
I just think that's fundamentally wrong.
对，我还没看过相关数据。
Yeah, and I haven't looked.
我知道一年前有一项研究，可能时间更长一点，但差不多一年前，有一项研究发现，在AI编程中，使用AI代码实际上效率降低了20%。
I know a year ago there was a study, or it may have even been a little more, but something like a year ago, there was a study that in AI coding, it was actually you got 20% less effective by using AI code.
但这是在……之前。
But this was before-
那是在四大、五大AI突破之前，对吧？
This was before the big four, five AI breakthrough, right?
我猜如果重新做那个研究，结果就不会是这样了。
My guess is if you reran that study, you would not see that anymore.
对，而且世界上最顶尖的程序员已经完全转变了看法。
Yeah, and the best coders in the world have fully flipped over on this.
如果你跟那些前沿公司的CEO聊，我觉得他们会告诉你，他们感觉效率高多了。
And if you talk to the CEOs at the cutting edge companies, I think they would tell you that they feel much more productive.
跟Stripe合作的时候，他们觉得回报非常高——
Spending time with Stripe, they feel like there's very high returns-
——因为他们写代码的效率能大幅提升。
... to them being able to be a lot more efficient in writing code.
宝洁公司现在在AI上的投入，有没有带来很大的效率提升？
Is Procter & Gamble getting big efficiency gains from what they're spending on AI today?
可能还没有。
Probably not yet.
但我觉得他们会。
But I think they will.
他们可能需要很多指导。
They probably need a lot of hand-holding.
答案可能就是——
That's probably the answer-
——我觉得像Labs、Databricks和Palantir这样的公司，他们可能想得对，你需要做很多指导工作才能产生价值，但你会为这些公司创造很多价值。
... is I think the Labs and Databricks and Palantir, companies like that are probably thinking about it the right way, where you have to do a lot of hand-holding to produce value, but you will produce a lot of value for those companies.
然后还有像Stripe这样的公司。
And then there's companies like Stripe.
我不知道，我们现在想得很多的一件事是，这些努力到底用在了哪里？
I don't know, one of the things that we're thinking a lot about right now is where is that effort actually going to?
是用在了前端，开发更多产品，还是用在了后端，让业务更高效？
Is it going to front end, building more products, or is it going to back end, trying to make your business more efficient?
现在世界上最好的公司正在——
And the best companies in the world right now are-
前端。
Front end
……全都集中在做前端产品上。
... singularly focused on front end.
百分之百。
100%.
如果你把下一个产品做好了，收入潜力可能提升十倍甚至百倍。
You can produce 10X, 100X more revenue potentially if you nail the next product.
对。
Yes.
但在成本端，效率提升的空间是有限的。
Whereas on the cost side, you can only get so much more efficient.
而且那个机会一直都在那里。
And that opportunity will kind of always be there.
只是还没被挖掘出来。
It's latent.
所以如果你只盯着成本，那就像是在透支自己的未来。
So it's like shorting your own future if you are just focused on the cost side.
而如果你从更广阔的视角看待自己能成为什么——
Whereas you're taking an expansive view of what you can become-
完全同意
Totally
——如果你专注于前端的话。
... if you're focused on the front end side.
顺便说一句，OpenRadar那个想法在那里也非常棒。
By the way, the OpenRadar idea was very good there, too.
就像是在说，‘我们也要在AI层面掌控这一切。’
Just to be like, "We're going to own this at the AI layer as well."
这非常聪明。
That's very smart.
你看，我觉得他们说得对，经济会在美元和代币之间来回摆动——
Look, and I think they're right that the economy is going to oscillate between using dollars and tokens-
……而且它会成为一种新的货币形式。
... and it's going to be a new form of currency.
完全同意。
Totally.
所以是的，我们显然非常看好那些人，我觉得这是好事。
So yeah, we're obviously big fans of those guys, and I think it's good.
另外，关于一切都会顺利运转这一点。
Also, to the point of it's all going to work.
我们经常在谈到前沿实验室时提到开源，我想关键点在于，如果总的代币使用量增长100倍，那这些事物都会变得非常庞大。
We talk about open source a lot in relation to the Frontier Labs, and I guess the point there is if the total token usage just goes up by 100X, it's just like these things are all going to get so big.
它们都会变得非常庞大。
They're all going to get so big.
所以我想先打好基础，然后再论证技术栈中哪些具体层面会成功，因为我觉得这描绘了一个画面：未来十年，我们用这些东西完成的工作量将增长20倍或50倍。
And so I wanted to lay the groundwork first before making the case for which of the individual layers of the stack are going to be successful because I think it paints a picture where we're going to 20X or 50X over the next 10 years in just how much work is being done with this stuff.
顺便说一句，我们现在只是在讨论企业领域。
And by the way, we're just talking about enterprise right now.
我们该聊聊消费者市场，但先只谈企业这块。
We should talk about consumer, but just in enterprise.
所以你看前沿实验室，现在它们占了大概95%以上的资金。
So if you look at the Frontier Labs, today, the Frontier Labs are like 95% plus of dollars.
三家加起来大约1250亿美元。
The 125 billion or so across the three of them.
基本上，它们把市场上的所有收入都吃掉了。
It's like they're consuming all of the revenue in the market, basically.
其他所有东西都只是误差范围。
Everything else is surrounding error.
它们筹到的钱比整个下游生态系统加起来还多。
They have raised more than the entire downstream ecosystem.
光是OpenAI和Anthropic就筹了3500亿美元。
Just OpenAI and Anthropic have raised $350 billion.
所以它们确实有规模。
So they do have scale.
哇。
Wow.
他们确实有巨大的规模优势。
They have massive scale benefits.
而现在，从用户的偏好来看，大家更倾向于使用前沿模型，我觉得这背后有很多原因。
And right now, what's revealed in the preferences of users is users are opting to use Frontier models, and I think there's a bunch of reasons why that's the case.
目前，人们愿意为前沿模型的令牌多付10倍的价格。
Right now, people are willing to pay 10X more for Frontier tokens.
我觉得在编程领域，这种情况肯定会持续一段时间。
I think in coding, that's definitely going to persist for some period of time.
有理由认为，人们可能会进行成本优化——
There's reasons why you could see a case that people will cost optimize-
……并随着时间的推移，使用低于前沿水平的模型。
... and use below the Frontier over time.
是的，但我们更倾向于假设……对。
Yeah, but we're much more to assume that the... Yeah.
从实际选择来看，每个人——
The revealed preference is everyone-
——每个重度使用这些工具的人，都想用最前沿的模型。
... everyone who is a heavy user of this stuff wants to use the Frontier.
我觉得有——
I think there's-
嗯，而且在我看来，如果真是这样，前沿模型应该用在更有价值的任务上，那些任务的价值会更高。
Well, and also it seems to me like if that happens The frontier should be at even more valuable tasks that are worth even more.
确实，比如治愈癌症。
Even more so like, yeah, curing cancer.
如果写代码值X，那治愈癌症值多少？
If coding was worth X, what's curing cancer?
治愈癌症的价值——
Curing cancer is worth-
比X还高。
More than X
……一些宇航员级别的钱。
... some astronaut money.
是的，当然。
Yes, of course.
所以现在，我觉得人们用前沿产品有几个原因。
And so right now, I think there's a couple reasons why people are using the Frontier stuff.
一个就是，让这些工作替你完成，带来的用户或消费者剩余太大了，所以你愿意为那个前沿多付一点钱。
One is just there's so much user or consumer surplus to having this work done for you that you're willing to pay the incremental amount for that frontier.
你明白吗？
Does that make sense?
明白。
Yes.
第二，他们在自家产品上做得相当不错。
Secondly, they've done a pretty good job with first-party products.
所以大家真的很喜欢用Codex、Claude Code和Cursor，而且我觉得以后会有越来越多的人用像Rock Bot这样的东西。
So people really like using Codex and Claude Code and Cursor and, I think increasingly we'll use stuff like Rock Bot.
使用第一方产品真的会很有价值。
It's going to be really valuable to use the first-party stuff.
这其中有几个原因。
There's a bunch of reasons why that's the case.
他们在产品上做得很好。
They've done a good job with product.
我还觉得他们把工具和模型结合得非常紧密——
I also think they have very tightly coupled their harnesses with the models-
所以你在使用这些产品时的体验真的非常非常好。
... so that the experience you have with those products is actually very, very good.
所以现在没有人会去换用其他模型，因为那些模型没有配套的工具、产品和界面。
And so subbing out some other model where you don't have the harness and you don't have the product and don't have the interface, no one's doing that today.
我很好奇你怎么看，虽然这很难说，但你觉得他们在开发更多第一方产品方面的激励应该是什么样的？
I'm curious your thoughts on, and obviously it's very hard to know, but how do you mentally picture what their incentive should be on making more first-party products?
我问这个是因为，当你考虑在成长阶段投资应用时，你会想哪些应用值得他们去开发第一方产品，这会不会有影响，还是说最终都会赢？
And I'm asking this in relation to, as you think about investing in applications at the growth stage, and you think like what's going to be worth it for them to build a first-party app in, and will that matter, or again, does it all win?
你怎么看待他们的激励措施，以及这跟你投资应用有什么关系？
How do you think about their incentives and how, if at all, it relates to your application investing?
我们确实做了很多应用方面的投资。
So we've done a lot of application investing.
所以，这又回到了那个主题，对吧？
So again, this is like in the theme of and, right?
我们觉得很多事情都会成功。
We think a lot of things are going to work.
在应用方面，我的看法是，大型实验室会非常专注于第一方产品，尤其是那些明显与编程相关，或者在编程影响范围内的产品。
On the application side, my framework is the big labs are going to be very focused on first-party products that are obviously in coding and within the blast radius of coding.
所以你可以按自己的理解来界定编程的影响范围——
And so you can interpret the blast radius of coding however you'd like-
但可能也就是周边的一些东西。
... but it's probably some things around it.
这是其中一部分。
That's one piece.
其次，他们会专注于那些能横向吸引所有15亿知识工作者的第一方产品。
And then secondly, they're going to be focused on first-party products that appeal horizontally to all 1.5 billion knowledge workers.
就像微软Office和谷歌应用那样。
So think like Microsoft Office and Google Apps.
他们希望这些产品能成为自己的第一方产品。
And they would like to have those as first-party products.
我们从他们与数据供应商的合作中知道，他们非常关注这些领域，对吧？
We know from the work that they do with data vendors, they are very focused in those areas, right?
想想所有微软Office产品。
So think all the Microsoft Office products.
那么，除此之外，问题就是这些对他们来说优先级有多高。
And so, outside of that, there's a question of what number priority is it for them?
我们是Harvey的大投资者。
So, we're large investors in Harvey.
所以，Harvey，我觉得法律领域，怎么说呢，在X上对人们来说很有趣。但这是一个绝对在起飞的市场。
And so, Harvey, I feel like legal is, I don't know, it's very interesting to people on X. But it's a market that's absolutely in takeoff.
所以如果你乐观地看，虽然有很多不同之处我们可以讨论，但它大概比编程落后12个月。
So if you take an optimistic view of it, there's many differences which we could talk about, but it's probably 12 months behind coding.
就劳动力市场的普及程度而言。
Just in terms of diffusion into the workforce.
现在已经到了终端客户要求他们的律师事务所使用Harvey的地步。
And it's now gotten to the point where end clients are demanding that their law firms use Harvey.
他们说，'我们看重它的产品，也看重它的成本。
So they're like, "We care about it for product, and we care about it for cost.
所以你必须用它。'
And so you must use it."
以前人们总是谈论幻觉问题，现在呢，大家都用上了。
So it used to be this thing where people would talk about hallucinations, and then now it's just like, no, everyone has it.
所以，它大概在优先级列表的第六到第十五位之间——
And so, it's probably somewhere between number six through 15-
……在优先级列表上。
... on the priority list.
顺便说一句，那个真的很有意思，在X上特别能引起大家的讨论。我们当然也知道这一点。
By the way, that one, it really is funny how much it gets people on X. And we know this too, obviously.
对，来自Agora。
Yeah, from Agora.
我觉得大家，第一，很惊讶这些公司在法律领域能发展这么快，因为法律本来不像是会早早采用新技术的领域。
And it's like, I think people, A, were surprised that these companies can grow this fast in legal, which you would not have thought would be an early adopter.
然后很多人就开始怀疑，因为他们会说：“哦，它之前不是会胡编乱造吗？”
And then you got a lot of doubt, because they're like, "Oh, well, wasn't it hallucinating and stuff?"
那都是18个月前的事了。
It's like, yeah, 18 months ago.
对，18个月前。
Yeah, 18 months ago.
他们做了一个明智的赌注，那就是当模型性能提升100倍时会发生什么？
And they made a smart bet, which is what happens when models get 100X better?
然后——
And like-
而且它会成功的。
And it's going to work.
而且它会成功的。
And it's going to work.
所以最大的——
And so the big-
法律其实也是个完美的领域，因为它虽然涉及语言，但很数学化。
And legal is actually just a perfect thing too, because it's language, but it's like mathematical.
对，完全正确。
Yes, exactly.
它非常讲究逻辑。
It's extremely logical.
是的，非常讲究逻辑。
Yeah, it's extremely logical.
而且文档记录非常完善——
It's extremely well documented-
……在判例法方面。
... in terms of case law.
对。
Yep.
我们看到的重大突破，我相信Agora的情况也一样，就是推理模型。
And the big breakthrough that we saw, which I'm sure is the same case with Agora, is reasoning models.
完全同意。
Totally.
所以律师工作本质上就是推理。
And so lawyering is reasoning.
所以，是的，这说得通。
And so, yes, it makes sense.
不过，我同意。
But no, agreed.
但这似乎不是下一步……因为我觉得这里也有个权衡，如果他们开始随意挑选公司，那就会失去信任，然后你就得——
That doesn't seem like the next... Because there's a trade-off too, I think, where if they start picking off random companies, then you lose the trust and then you have to-
对，那整个生态圈。
Yeah, then the ecosystem.
对，当然。
Yeah, of course.
所以，是的。
So, yeah.
我觉得所有大平台，微软肯定也遇到过这种情况，而且——
I think all the big platforms, I'm sure Microsoft had this and-
Salesforce也遇到过。
Salesforce had it.
……他们都得决定，为了去吞并某个小东西，值不值得把所有人都得罪了。
... they all had to decide, is it worth pissing everybody off to go eat some random thing?
对，微软那套老理论，一个平台只有在它上面构建的所有东西创造的收入超过它本身时，才算真正的平台。
Yeah, the old Microsoft thing, a platform is only a platform if all of the things built on top of it generate more revenue than it.
没错。
That's right.
没错。
That's right.
所以我觉得实验室里并没有完全丢掉这个理念。
And so I don't think that that's totally lost in the labs.
不过，它可能还是排在第六位。
Now, it might still be priority number six.
对。
Yes.
而且有一个重大的验证事件——
And there was a major validation event-
那就是柯克兰出来宣布，他们要花五亿美元来——
... which was when Kirkland came out and said they were going to spend 500 million bucks to-
打造自己的技术栈。
... build their own technology stack.
是的。
Yes.
大家当时都说：“天哪，这对Harvey来说意味着什么？”
People were like, "Oh man, what's that mean for Harvey?"
对，所以你就有了那些泡沫型的，它们——
Yeah, so then you've got these bubble ones where they're-
但这很好。
But it's great
……它不是小众市场，但也不是通往AGI的道路。
... it's not a niche, but it is not the path to AGI.
那些应用。
Those applications.
对，没错。
Yeah, exactly.
但它会价值数十亿甚至更多。
But it is going to be worth many, many billions.
关于这一点，有两件事非常重要。
And there's two things that are really important about it.
第一，产品的细节对用户来说真的很重要。
One, the last details of the product really matter to the users.
第二，这需要真正的市场推广和实地努力。
And then two, it requires real go-to-market and on-the-ground efforts.
对。
Yes.
所以，这两点，大家都在讨论现在软件或应用的护城河是什么。
And so, those two things, everyone talks about what are now the moats of software or applications.
实地部署。
FDs.
这些在地面上的事情——
Those things like in the ground-
确保它们能正常工作——
... making sure they work-
是很重要的一件事。
... is a big thing.
那么，实验室会以垂直化的方式去做这件事吗？
And so, will the labs go do that in a verticalized way?
在法律领域，可能不会。
At law, probably not.
不管怎样，确实有个叫Claude Legal的产品。
Whatever, there's a Claude Legal product.
对，他们不可能随时随地都覆盖到。
Yeah, they just can't be everywhere all the time.
其实没什么人真正在用。
No one really uses it.
他们不可能把所有事情都做了。
They can't do all of the things.
是的。
Yes.
所以，我们的投资决策也是基于这一点。
And so, our investments are informed by that.
这并不是说我们不会在那个影响范围内，或者那个知识工作者横向领域里做事情——
It's not to say that we won't do things in that blast radius, or in that knowledge worker horizontal-
……这个类别。
... category.
但我们确实会更多地问这个问题：产品路线图是什么？
But we do ask the question a lot more about what is the product roadmap?
凭什么能赢？
What's the right to win?
在这些应用之外，我觉得会有百花齐放的局面。
In the applications beyond those, I think there's going to be hundreds of flowers that bloom.
因为那将是一个巨大的解锁。
Because it's going to be a massive unlock.
所以——
And so-
对，但回到那个一切都会奏效的观点，我觉得即使是影响范围内的那些东西
Yeah, but back to the point of it's all going to work, I think even the stuff in the blast radius
展开剩余字幕（还有 480 条）
这大概能行。
It's probably going to work.
这大概能行。
It's probably going to work.
我们是大型投资者。
We're large investors.
我们是Replit最大的投资者。
We're the biggest investors in Replit.
它运行得很好。
It's working very well.
对，Replit和Lovable。
Yeah, Replit and Lovable.
你可能会想，"那Codex和Claude Code难道不能做个网站吗？"
You would think that you're like, "Well, aren't Codex and Claude Code going to be able to make a website?"
答案是，是的，它们可以。
And the answer is, yeah, they can.
而且，这些东西简直不可思议。
And also, these things are unbelievable.
它们非常好。
They're very good.
还有 Cursor 也是，它正好处在冲击波的中心。
And Cursor also, and that is dead in the center of the blast radius.
完全同意。
Totally.
而且有大量的人在使用它。
And there's a huge market of people who use it.
所以，是的，一切都会顺利的。
And so, yeah, it's all going to work.
我觉得，有了那个模型，对前沿技术的渴望，我认为这是毋庸置疑的。
I think, with the model there, the appetite for frontier technology, I just think it's undisputed.
它肯定会一直存在。
It's just going to be there.
对N-1或开源技术的需求也会非常巨大。
The appetite for N-1 or open source technology is also going to be huge.
这会非常庞大，因为如果使用量增加50倍，人们就会关心成本。
It's going to be massive because if you 50x the usage that happens, people will care about cost.
而且也许会有那么一个时刻，如果前沿技术的成本高出100倍——
And maybe there will be a point where if cost is 100x higher for frontier-
……人们真的会开始关注这一点。
... people really start paying attention to it.
嗯，而且N-1很快就会变得非常好。
Well, and N-1 is going to be really good soon.
它会变得非常好。
It's going to be very good.
N-1会比5、6更好。
N-1 will be better than 5, 6.
到某个时候，我不知道具体什么时候，但很快。
By the end of, I don't know when, but by soon.
所以你可以看到，在应用公司的最前沿，他们很在意自己的成本，因为这直接关系到他们的毛利率。
And so you can see it actually at the cutting edge of the application companies, right, who care about their own cost because that's their gross margin.
那么，从什么时候开始，大多数客户支持工单不再需要最高端的模型了？
And so, how long has it been since most customer support tickets required the highest end model?
大概三年前——
Like three years-
……或者差不多那时候，对吧？
... or something, right?
所以你可以看到，抽象层会随着所需的价值——
And so you can see a path to the sort of abstraction layer following how much-
对
Yes
……而变化。
... value is required.
而且会有很多人争夺这个抽象层，比如应用公司，显然还有那些有自家产品的实验室，像Databricks、Palantir或者Snowflake这样的公司，他们都会来争这个位置。
And there's going to be a lot of people vying for the abstraction layer, like the application companies, obviously the labs with their first-party products, companies like Databricks or Palantir or Snowflake, they're gonna vie for it.
但我觉得人们会尝试各种不同的方式。
But I think people are gonna use every variation of that.
而且都会有效。
And it's all gonna work.
顺便说一句，客户支持这个例子对我来说特别能说明扩散需要多长时间，因为技术已经足够好有一段时间了，但我们仍然，我不知道它的渗透率是多少——
By the way, the customer support example is such a good one to me of how long diffusion's gonna take because the technology has been more than good enough for a little while, and we still, I don't know what percent penetration it has-
很小。
Tiny
……但很小。
... but tiny.
这对我来说是最清楚的例子，说明技术已经完全到位了。
And that's the clearest example to me of the technology is fully there.
只是需要很多年才能实现。
It's just gonna take many years to get there.
对，就是企业界常说的更好、更快、更便宜。
Yeah, the old enterprise thing of better, faster, cheaper.
它已经变得更好、更快、更便宜了——
It's been better, faster, and cheaper-
它什么都占全了。
It's one on everything
……已经两年了。
... for two years.
而且就像——
And like-
而且这些公司都做得很好。
And the companies are doing great
……这些公司比之前的任何应用公司都发展得更快。
... the companies are growing faster than any other application companies before them.
但这里还是有些物理定律的。
There's still some laws of physics here.
但没错，你得去推销它，还得去实施它——
But yeah, you got to sell it and you got to implement it-
而且有很多复杂的集成工作要做，
... and it's like there's a lot of messy integrations you have to do-
对
Yes
还有要遵守的规则，品牌指南之类的。
... and rules that you have to follow, brand guidelines.
再说一遍，实验室不会去构建那些动作。
Again, the labs aren't gonna build that motion.
会有公司在它们之上建立业务。
There will be companies built on top of them.
所以这是一个经典的“和”的情况。
So that's a classic and.
所以这些都是企业级的。
So this is all enterprise.
那消费者方面呢？
What about consumer?
所以，我觉得，显然ChatGPT是第一个面向消费者的时刻，也是整个AI的时刻，但第一个AI时刻是消费者时刻，正是这一点吸引了所有人的注意，因为人们第一次开始使用这个东西，而他们之前已经听说AI大约十年了。
So, I think, obviously ChatGPT was sort of the first consumer moment and it was also the whole AI moment, but the first AI moment was a consumer moment and that's what caught everybody's attention was for the first time people were using this thing, and they had been hearing about AI for, I don't know, 10 years.
然后突然之间你有了这个东西，即使是第一版ChatGPT，我们也会说：“哦，我以前从没见过这样的东西。”
And then all of a sudden you have this thing, and even the first version of ChatGPT, we were like, "Oh, I've never seen that before."
但现在，四年后或者不管过了多久，你跟我聊聊消费者这边的情况。
But now talk to me four years later or whatever it is.
消费者方面进展如何，你怎么看？
What's the progress of consumer and how do you see it?
有十亿人在用它，对吧？
There's a billion people using it, right?
而且不止十亿人在用——
And there's more than a billion people using it-
痴迷地
Obsessively
……痴迷地。
... obsessively.
但是产品的具体形态和使用场景其实非常基础，对吧？
But the instantiation of the product and the use case is very basic, right?
绝大多数用户还停留在拟物化的阶段，也就是说他们只是在ChatGPT里做以前在搜索引擎上做的事情。
We're in skeuomorphic mode for the vast majority of consumers, which is they are doing what they used to do on a search engine inside ChatGPT.
而这就是绝大多数用户的使用范围了——
And that's the extent of the use case of the vast majority of those-
除了那些孩子们。
Except for the kids.
孩子们用得比那好多了。
The kids use it better than that.
我儿子，顺便说一句，一个10岁的小孩在写完整的说唱歌曲——
My son, total aside, a 10-year-old is writing full-on rap songs-
还在创作音乐。
... and creating music.
不，这可不是顺便说说的。
No, it's not an aside.
这就是未来。
This is the future.
这就是未来。
This is the future.
但这是——
But this is-
而且这太不可思议了。
And it's incredible.
这说明了区别，就像有一大群人可能永远都会用谷歌，把它当作谷歌的替代品。
It tells you the difference where it's like there's a large cohort of people who are gonna probably always Google, use this as a Google replacement.
但孩子们不会这么用。
But that's not how the kids are gonna use it.
孩子们会完全自然地使用它。
The kids are gonna be totally native.
对。
Yes.
现在你从孩子们身上就能看出来。
And you can see it with the kids now.
他们从来没受过训练。
They never were trained.
我的孩子们从来没学过用搜索引擎——
My kids were never trained on search engines-
……现在完全是另一回事了，他们让它回答任何问题，给它任何任务都能直接产出成果。
... and now it's a whole new thing where they get it to give them answers to anything and give them work product of anything.
最大的转变会是从被动响应变成主动行动，而我们现在才刚摸到皮毛。
The big shift is gonna be when it goes from being reactive to proactive, and we're just scratching the surface on that.
它会变成多模态的，不会只是坐在那儿干这个。
It's gonna be multimodal, so it's not gonna be just sitting doing this.
现在语音功能其实已经很好了。
The voice stuff actually is very good now.
之前有点虚晃一枪，因为OpenAI的语音功能有一段时间不太好用，还得靠一个更笨的模型撑着。
There was a bit of a head fake because the OpenAI voice stuff was not very good for a period of time, and it had to rely on a dumber model.
现在它不会了，而且它能处理打断之类的情况。
And now it doesn't, and it can handle interruptions and things like that.
所以，自然的人类交互形式现在已经存在了。
So, a natural human interaction form now exists.
而且替你主动做事、采取行动的能力现在也具备了，对吧？
And capabilities for doing proactive work on your behalf and action-taking is just now there, right?
OpenClau是第一个关键时刻。
OpenClau was the first moment.
我觉得GrokBot现在也是一个关键时刻。
I think GrokBot is a moment now.
我认为这将是消费者使用的未来。
And I think that'll be the future of consumer use.
让我觉得好笑的是，现在没人谈论消费者。
It's so funny to me that no one talks about consumer right now.
大家都痴迷于编程，这也很合理——
Everyone is obsessed with coding, rightfully so-
考虑到这些公司的收入增长情况，确实如此。
... given what's happened with the revenue ramps of the companies.
但如果以历史上科技市场的情况为参考的话——
But if historical technology markets are any guide-
对
Yes
最大的赢家将是那些杀手级消费产品，因为它们能触达数十亿用户。
... the big prize is gonna be the killer consumer products because you can address billions of people with them.
所以我觉得这些产品会越来越好。
So I think those are gonna get a lot better.
我非常乐观。
I'm very optimistic.
而且这些产品还会带来世界上最好的商业模式。
And those are gonna come with the world's best business models, too.
将会创造巨大的价值。
There's gonna be a lot of value generated.
会有大量的消费者剩余。
There's gonna be a ton of consumer surplus.
那会是什么样子呢？
What would that look like?
我觉得会是广告，或者某种更高级的广告形式。
I think it'll be advertising or some greater form of advertising.
我想说的是，如果你试图向报纸分类广告时代的人描述信息流的形态，他们根本不会理解。
The thing I would say is if you tried to describe the form factor of the feed to people during classifieds at newspapers, it would make no sense.
所以我们还没搞清楚原生的消费模式会是什么，以及随之而来的原生广告格式会是什么。
So we have yet to figure out what the native consumption model is gonna be and then what the native ad format is gonna be that follows that.
你觉得会是广告加订阅吗？
Do you think it's ads and subscriptions?
对，肯定是。
Yeah, definitely.
对，我觉得两者都有。
Yeah, I think it's both.
你觉得到时候会不会有数亿人愿意为这个付钱，而且付的钱还挺多的？
You think that it'll be a situation where hundreds of millions of people might be willing to pay today surprising amounts for this then?
是的，绝对会。
Yes, absolutely.
而且，我觉得——
And again, I think-
那就是当它成为每个人的个人助理的时候。
And that's when it's like a personal assistant for everybody
……那就是当它成为每个人的个人助理的时候。
... that's when it's a personal assistant for everybody.
所以，我还是觉得我的框架是，创造的价值会是用户剩余价值的十倍。
And so, again, I think my framework is 10x more value will just accrete to the surplus of the users.
但也会为那个创造很多价值——
But there's going to be a lot of value created for that-
对。
Yes
……那10%的部分。
... for that 10%.
这部分价值会累积到公司身上，所以会以订阅的形式体现。
And so that accretes to the companies, and so it'll be subscriptions.
也可能是广告，或者某种我们现在还想不到的、可能还没被发明出来的新形式。
It'll be ads or some new form that isn't quite what we think of today as ads, that hasn't probably yet been invented.
而且这会是一种全新的消费方式。
And it's going to be a new form of consumption.
对。
Yes.
所以我对这一点非常乐观。
And so I'm very optimistic about that.
你最兴奋的消费领域的东西，就是能成为每个人助理的那种东西。
The thing that you're most excited about in consumer is something that gets to an assistant for everybody.
是的。
Yes.
是的，并且以可信赖的方式代表你采取行动。
Yeah, and take action on your behalf in a trustworthy way.
我不知道。
I don't know.
我已经开始为自己构建所有这些自动化，而且可以说，我正在积极尝试让自己更适应这种新方式。
I've started building all these automations for myself and, I'd say, I'm aggressively trying to make myself more native.
其中十个里有两个确实给我带来了实际价值。
And I get real value out of two out of 10 of them.
但还有更多——
But there's a lot more-
……我认为将来会实现的。
... that I think will be done.
而且未来五到十年，模型的能力会提升十倍到一百倍。
And the models will get 10X to 100X better over the next five to 10 years.
所以想想那会带来什么样的结果。
And so think about what that's going to resolve in.
所以，我完全同意这一点，这是显而易见的。
So I fully agree with this, obviously.
另一个我经常思考的关于个人消费自主权和AI的重要话题就是自动驾驶汽车，我觉得……
One of the other big personal consumer autonomy, AI things I think about a lot is just self-driving cars and autonomy is, I just feel like...
我最近第一次买了辆特斯拉，然后……
I just got a Tesla for the first time and like- ...
我明白了。
I got it.
这太不可思议了。
It is insane.
它就像是最好的产品，因为你坐进车里，输入你想去的地方，它就会带你去。
It is like the best product because you get in the car, you type in where you want to go, and it takes you.
你什么都不用想——
You don't have to think about anything-
……其他的事。
... else.
整个行程中，它都能帮你摆脱任何麻烦。
The whole ride, it gets you out of any weird thing.
它能在任何地方停车，我觉得大多数人还没意识到自动驾驶有多好，除非你有一辆带自动驾驶的特斯拉。
It can park anywhere, and I don't think most people have realized how good self-driving is yet, unless you have a self-driving Tesla.
对，或者你住的地方能打到Waymo。
Yeah, or if you live in a place where you can get a Waymo.
就像Facebook广告——
Like Facebook ads-
或者，对
Or, yeah
……这只是个小东西——
... it's a small thing-
当然
Of course
……对大多数人来说。
... for most people.
所以现在的情况是，有些公司已经能做到这一点了。
And so it's like you've got these companies that can do it now.
Waymo和特斯拉。
Waymo and Tesla.
但这项技术还没有普及开来。
But it has not diffused yet.
所以，第一，我觉得那些已经做到的人可能没意识到，但我觉得大家也都没完全意识到现在的自动驾驶有多厉害。
And so, A, I don't think the people who have done it realize, but I don't think everybody realizes how good autonomy is right now.
但另一个更重要的问题是，我觉得它还没有转化成它应该有的商业模式。
But then the other most important thing is, I don't think that it has turned into the business model that it needs to turn into yet.
所以我很想听听你的看法，我知道你是投资人，而且你在这方面想了很多，但是——
And so I'm curious your take on this is, I know you're an investor, and you have thought a lot about this, but like-
我们现在处在自动驾驶发展的哪个阶段？
... where are we on the autonomy curve?
想到这些正在发生的其他重大科技趋势，真的很有意思，因为AI简直压倒一切——
It's so funny to think about these other major technological tailwinds that are happening because, AI just supersedes everything-
……在我们现在的脑海里。
... in our minds right now.
完全同意。
Totally.
我总觉得退一步看问题很有帮助，比如说，产品周期驱动我们的业务，那我们现在正在抓住的大浪潮是什么？
I always find it helpful to just take a step back and just say, like, okay, product cycles drive our business, and so what are the big waves that we're catching right now?
自动驾驶就是其中之一。
Autonomous driving is one.
但上一代的大浪潮是手机加社交、电商、SaaS、云，这些差不多同时出现。
But the previous generation of these was mobile phones plus social, e-commerce, SaaS, cloud, kind of all coinciding.
对。
Yep.
那创造了25万亿美元的市值。
And that produced 25 trillion of market cap.
那可真不少。
That's a lot.
那是很多，对吧？
That's a lot, right?
那我们现在眼前摆着的是什么？
And what do we have staring at us in the face right now?
我们有AI。
We've got AI.
我们有自动驾驶。
We've got autonomy.
我们有机器人技术。
We've got robotics.
我们有AI在医疗领域的应用。
We've got AI's application to health.
对，我正想说生物科技。
Yeah, I was going to say bio.
我们还有美国活力这个类别，就像——
We have the American dynamism category, which is like-
……国防现代化。
... modernization of defense.
对。
Yep.
所有这些加在一起，我……
The combination of all of those, I'm-
肯定不止25个。
Got to be more than 25.
远远超过25个。
It's way more than 25.
这是成为投资者和投身科技市场最激动人心的时代。
It's just the most exciting time to ever be an investor and in the technology markets.
特别是自动驾驶，我觉得它被严重低估了。
Autonomy specifically, I think it's just very underappreciated.
汽车行业和人们开车，显然是世界上最大的行业之一。
The auto industry and people driving cars is one of the biggest industries in the world, obviously.
现在，好吧，如果你一路回想到出租车时代，对吧？
Right now, okay, so if you rewind all the way to taxi driving, right?
那在美国总行驶里程中占比不到0.1%。
That was less than 0.1% of the overall miles traveled in the US.
之所以这样，有很多原因。
And there were a bunch of reasons why that was the case.
后来网约车公司出现了，我当时是优步的投资者。
Then the rideshare companies came along, and I was an investor in Uber.
显然，你们也是。
Obviously, you guys were, too.
当然，最大的变化是更好的商业模式，但真正重要的是市场规模扩大了十倍。
The big thing that happened, obviously, was a better business model, but it was really the 10X-ing of the market size.
所以旧金山的出租车市场是一亿美元，优步和来福车推出三年后就变成了十亿美元。
So taxis were 100 million bucks in San Francisco, and then they became a billion dollars three years into Uber and Lyft.
所以市场规模扩大了十倍，因为你提供了更好的产品。
And so it 10X the market size because you delivered a much better product.
自动驾驶，十年后我们回头看会说，"哦，那又是一个明显的十倍增长机会。"
Autonomous driving, we're going to look back 10 years from now and say, like, "Oh, there was another obvious 10X-ing that would happen there."
至少是这样。
At least.
至少光是在网约车这一块。
At least just on the ride-hail side.
所以如果你有一个比人类司机安全14倍的产品——
So if you have a product that is 14 times safer than a human driver-
那人类开车就显得很野蛮了。
Then it's barbaric to drive
……
...
实际上，这就是我现在对城市的看法——
It's actually so, this is my view now on the cities-
你不觉得吗，比如说，每年有多少人死于人类司机之手？
Don't you think if you're like, how many people die a year from a human driver?
如果它更安全，那岂不是——
If it's safer, isn't that-
你知道——
You know-
因为现在数据不是已经很充足了吗？
Because isn't the data pretty abundant now?
Waymo已经行驶了数百万英里，而且它们比人类驾驶员安全10到14倍。
It's millions of miles traveled in Waymos, and they're between 10 and 14 times safer than a human driver.
所以，如果不让这些技术真正融入经济，那才是野蛮的。
So it is barbaric to not actually allow for these things to diffuse into your economy.
所以，我觉得历史会严厉评判那些不让它发生的人。
So, I think history will look back on people and judge them very poorly if they don't let it happen.
但对消费者来说，先从这一点开始。
But for the consumer side, just start with that.
如果你能乘坐比人类驾驶安全10到14倍的车，你肯定愿意付高得多的价钱。
If you can drive in something that's 10 to 14 times safer, you would pay a massive premium.
我觉得很明显有办法让成本降到比现在优步和来福车还低。
I think there's a very clear case for how you actually get the cost below the cost of Uber and Lyft today.
现在，自己拥有一辆车，每英里成本大概是80美分，包括折旧、保险、油钱，所有都算上。
So right now, owning a consumer car equates to 80 cents per mile, like fully loaded, depreciation, insurance, gas, all that stuff.
这就是你每多开一英里的成本。
That's your cost of an incremental mile.
坐优步或来福车大概每英里要两美元多。
Riding in an Uber or Lyft is probably two bucks and change per mile.
这个市场的弹性非常大。
This market is massively elastic.
所以如果你能把——
So if you can get-
成本降下来
The cost down
——把威摩或者特斯拉或者其他叫车服务的成本——
... the cost of a Waymo or a Tesla or whatever ride-hail service-
一切
Everything
我觉得还会有其他的公司——
... which I think there'll be others-
……
...
低于这个价格，需求会爆炸式增长。
down below that, there's going to be an explosion of demand.
所以它安全了14倍。
So it's 14 times safer.
顺便说一句，它只会变得更好。
It's only going to get better, by the way.
它会变得更安全。
It's going to get safer.
很快它就会安全20倍。
It's going to be 20 times safer soon.
而且成本还会更低。
And it's going to be less than the cost.
你会看到十倍以上的增长。
You will see a 10X plus.
然后再说说个人拥有汽车这事儿。
Then you talk about personal car ownership.
美国每年新车销量是一千七百万辆。
And so there's 17 million cars sold a year in the US, new.
顺便说一句，美国大概有两亿五千万辆车。
And by the way, there's like 250 million cars in the US.
就美国来说，你愿意为自己拥有的乘用车支付多少钱来获得完全自动驾驶功能？
Just in the US, how much would you pay for full autonomous features for a passenger car that you owned?
很多。
A lot.
大概要翻倍。
It's like double.
一万。
10,000.
对，至少一万美元。
Yeah, minimum $10,000.
所以——
And so-
我的意思是，一旦你习惯了，那真是不得了。
I mean, once you're used to it, it is extreme.
对，你没法没有它。
Yeah, you can't not have it.
真的很难回到过去。
It's really hard to go back.
所以你可以一次性卖，也可以订阅，但就算每辆车最低一万美元，一千七百万辆车。
So you could sell it upfront, you could do a subscription, but call it minimum 10,000 bucks a car, 17 million cars.
然后还有整个西方世界，最终成本会低到足以进入发展中市场。
Then you have all of the Western world, and then eventually it'll get cost compelling enough to reach developing markets.
所以这会是那种五十年后的事情——
So it's going to be one of these things where 50 years from now-
……整个——
... the whole-
就像没人会再开车了
Like no one will be driving
……构成。
... composition.
对，没人会再开车了。
Yeah, no one will be driving.
所有东西的构成都会变得不一样。
The composition of everything looks different.
这需要时间，因为你得先造出这些车。
It takes time because you got to produce the cars.
人们总是惊讶于Waymo的车有多少。
People are always struck how many Waymos there are.
比如Waymo，如果你在他们运营的城市里，你会觉得它们到处都是。
Like Waymos, if you're in one of the cities where they are, they feel ubiquitous.
在美国还不到一万辆——
There's less than 10,000-
太疯狂了
It's crazy
……
... in the US.
在整个美国范围内。
In the whole US.
所以，在这个领域里，有Waymo，也有特斯拉。
And so, it's one of those areas where there's Waymo, there's Tesla.
我觉得还会有其他参与者。
I think there'll be other players.
我们投资了几家公司，我觉得它们真的有机会脱颖而出。
We've backed a couple that I think have a real shot at it.
而且这个市场不会只有十家玩家，它会是一个巨大的市场。
And it's not going to be 10 players, and it's going to be a massive market.
所以像这样的市场很容易被忽视，因为就像——
And so it's easy to lose sight of markets like that because it's like-
完全同意。
Totally
——AI的增长一样。
... the growth of AI.
但那些将会是巨大的市场，我们对此非常兴奋。
But those are going to be huge markets that we're super excited about.
然后，相关的，还有机器人技术，这将会是巨大的——虽然现在还没有什么实质性的进展，但你已经能看到一些苗头了。
And then, related, there's robotics, which is going to be enorm-- There's nothing quite going yet, but you see the glimmers of it now.
就像你走进那些正在研究这个的实验室和正在做这个的初创公司。
And it's like you go into the labs that are working on it and the startups that are doing it.
显然，特斯拉在这方面非常擅长。
Obviously, Tesla's very good at it.
但这显然会成功的。
But it's clearly going to work.
对，显然会成功的。
It's clearly going to work, yeah.
而且一旦成功，它的市场会比语言模型还大，对吧？
And it'll be a bigger market than language, right?
等它真正实现的时候。
When it does work.
因为它会有企业级的应用，也会有面向消费者的应用。
Because it will have B2B applications, it'll have consumer applications.
我觉得我们可能还差得远——
I think we're probably pretty far off-
离真正实现——
... from having-
确实感觉我们还没到那一步。
It does feel like we're not
……机器人进我们家。
... robots in our house.
但对于——
But for-
但对于像——
But for like-
特定用途——
Defined use-
在工厂里。
In a factory.
对，没错。
Yeah, exactly.
对，在工厂里。
Yeah, in a factory.
比如相对明确的用途，任务重复，环境安全。
Like relatively defined use cases with redundant tasks, with safe environments.
我们看到了。
We see it.
我们投了资，我的一个合伙人牵头投资了Mind Robotics。
We invested in, one of my partners led an investment in Mind Robotics.
所以Sarah促成了这笔交易，这位创始人是Rivian的创始人，他首先给人的印象就是，他是一位非常出色的创业者。
And so, Sarah did this deal, and it's the founder of Rivian, and he basically was like, first of all, he's an exceptional founder.
并且打造了一家拥有优秀产品的卓越公司。
And has produced a great company with great products.
但他会把机器人放到工厂车间里，让它们负责 Rivian 的组装工作。
But he's going to put robots on the factory floor, and so they'll do manufacturing assembly work on Rivians.
这类应用场景相对明确，而且现在的投资回报率相当高。
It's relatively defined use cases with pretty high ROI today.
然后你就成了嵌入式客户，既有了稳定的收入来源，还能让模型从你的实际工作中不断学习。
And then it's an embedded customer, you kind of have it, and you can not only get revenue from that, but you can also have the models learn from the work that you're doing.
这样你就在现场形成了一个可以实际优化的反馈闭环。
And so you have this feedback loop in the field that you can actually work on.
还有其他一些我们很看好的周边配套项目，但机器人领域一定会迎来一个ChatGPT式的爆发时刻。
There's other stuff that is kind of picks and shovels stuff that we're excited about, but we will have a ChatGPT moment for robotics.
我不知道具体什么时候会来。
I don't know when it's going to be.
我觉得希望是在五年之内。
I think it's within five years, hopefully.
现在有很多非常聪明的团队在搞这个。
There's very smart teams working on it.
而且我认为它会成为最大的市场之一，甚至比语言市场还要大。
And I think it's going to be one of the biggest markets, so it'll be bigger than language.
想想它的应用场景，这几乎是必然的。
And it almost has to be when you think about what it's going to do.
想想它对工业的影响，对吧？
Think about the industrial impact of it, right?
这件事的关键在于，物理规律并不阻碍它实现，而且我们已经能看到苗头了。
And this is one of these things where it's like the physics doesn't prevent it from happening, and you can see the glimmer.
所以这确实需要时间，我同意你的看法，要完全实现这一点恐怕得花十年以上，但我实在想不通为什么它不会成功。
So this'll take, I agree with you, this is going to take more than 10 years to get all the way there, but I just don't see why it wouldn't all work.
我觉得这事儿能成。
I think it's going to work.
我觉得能成，而且到时候它们会无处不在。
I think it's going to work, and I think we'll have them everywhere.
顺便说一句，我认为随之而来的副产品将是生产力的巨大飞跃。
And I think, by the way, the byproduct of this is that we'll have massive productivity booms.
我觉得这会带来巨大的经济影响，生活质量也会提升。
I think that's going to be the economic impact, and quality of life will go up.
我们得认真思考对就业的影响——
We'll have to think hard about job impacts-
还有类似的问题。
... and stuff like that.
工人们是否做好了转型的准备，尤其是如果变化发生得很快的话。
And are workers trained to make a transition, especially if it happens fast.
但回顾历次技术突破的历史，我们发现人类总能找到更高级别的任务去做。
But the history of all these technology breakthroughs is that we find higher order things to do.
不管怎么说，如今我们拥有的工作中，有60%在70年前根本不存在。
Whatever, 60% of jobs didn't exist 70 years ago that we have today.
所以，我们会经历一个类似的过渡期。
So, we'll have a similar transition period that happens.
但我认为这将带来巨大的生产力飞跃。
But I think it's going to be a massive productivity boom.
那肯定会很棒。
It's going to be awesome.
我想聊聊资本和融资方面的问题。
So I want to talk about capital and financing a little bit.
我们Benchmark刚刚历史上首次完成了成长期基金的募集，其基准设定是……
We at Benchmark just raised our first growth fund in history, and the baseline for that was something.
欢迎来到黑暗面。
Welcome to the dark side.
对，没错。
Yeah, exactly.
就像禁果一样。
It's like the forbidden fruit.
但我们背后的想法是，现在风险投资的回报往往发生在增长阶段，而以前不是这样。
But the thinking for us behind it was something along the lines of venture returns now happen at growth in a way that they didn't used to.
我们刚才的对话正好印证了这一点，因为结果就是那么大，真正跑起来的项目会一直跑，一切都会顺利，诸如此类。
And this conversation that we've just had really supports it because the outcomes are that big, and the things that run really run, and it's all going to work and all of these things.
这让我们想到，10倍、100倍甚至1000倍的巨大倍数，其实可以在整个过程中随时出现。
And that sort of led us to the thinking that the sort of 10 to 100x or 1,000x whatever, the really big multiples can happen all along the way.
我知道你显然也这么认为——
Which I know you believe that obviously as-
你这是在跟行家说行话。
You're preaching to the choir.
不过我想把这一点联系起来。
I want to connect this, though.
所以，我想听听你对这个问题的看法，同时，我们刚才聊了产品周期和我们现在所处的位置，但还有一个相关但又不完全相同的东西，就是资本周期。
So, I would like to hear your thinking on that, and I would also like to hear, we just talked about the product cycle and where we are, but there's also this related but not exactly the same thing of the capital cycle.
资本跟着产品走，投资在它们背后，所有这些事情。
And the capital following the products and investing behind them and all of those things.
所以我很好奇，从资本的角度，从成长投资的角度，你怎么看我们现在所处的位置。
So I'd just be curious to hear your take on where we're at from a capital perspective, from a growth investing perspective.
好的，当然。
Yeah, of course.
那么，关于为什么要做成长投资，以及它是不是一个有吸引力的类别？
So, as it relates to why go do growth and is it a compelling category?
显然，你已经看到了光明。
Obviously, you have seen the light.
但是——
But-
黑暗中的光明。
The dark light
……黑暗之光。
... the dark light.
这很棒。
It's great.
我们所在的市场是一个大约5万亿美元的市场。
The market that we're in is like a $5 trillion market.
所以，差不多是标普500的20%。
So it's like, whatever, it's like 20% of the S&P 500.
所以如果你不在私募市场中全面布局，你可能会错失很多机会。
So if you're not actually playing the full gamut of that in the private markets, you're probably leaving a lot of money on the table.
我们很久以前做过一个分析，基本上显示私募市场一半的回报来自C轮到B轮之间，另一半来自C+轮及以后。
We did this analysis a long time ago that basically showed half of private market returns get generated between the C and the B, and then half of returns get generated from the C+.
所以如果你不积极布局C+轮及以后——
And so if you're not aggressively playing from the C+ -
你就会错过一半。
You miss half
那你肯定会错过一半。
... you're definitely leaving off half.
而且顺便说一句，那还是在公司更早上市的时代。
And by the way, that was during an era when companies went public earlier.
所以现在公司私有化的时间更长了，如果你快进五年，这个比例会更加极端。
So now that companies are staying private longer, if you could fast-forward five years, the ratio is going to be even more extreme.
可能变成七三开，甚至更偏向后期。
It'd probably be 70/30 or something in favor of the late stuff.
你还可以考虑风投的结果，因为现在的幂律分布太强了，对吧？
You can also give venture outcomes because the power law is so crazy strong today, right?
所以一旦你押对了，回报会非常可观。
And so, when you get it right, it can really go right.
我们现在所处的这个动态市场，幂律分布比过去更极端，我觉得这有几个原因。
And we're in this dynamic market where the power law is more extreme than in the past, and I think it's a result of a few things.
第一，我们处于早期周期，所以早期公司能更早确立领导地位。
One, is just we're early cycle, and so the early companies can establish themselves as leaders earlier.
第二，AI这个领域很特别，因为如果你投入更多资金，它就能直接扩大规模——
Secondly, this one in AI is unique because if you throw more money at it It can just scale up-
然后变得更好，对吧？
... and get better, right?
所以这和之前的周期不一样，对吧？
So that is different than previous cycles, right?
如果在SaaS时代，你往ServiceNow、Workday或Salesforce这些公司砸无限的钱，它们会搞得一团糟。
If you threw endless amounts of money at ServiceNow or Workday or Salesforce, during the SaaS era, they would get all messed up.
我们见过Vision Fund这个实验，对吧？
We saw this experiment with the Vision Fund, right?
你往这些公司砸钱，结果它们反而被搞砸了，对吧？
You throw money at these things, and the companies get screwed up, right?
或者如果是消费类公司，它们会把钱花在Google和Facebook上，那也不是什么好事。
Or if it was a consumer company, they would spend it on Google and Facebook, and that's not a good thing either.
那就是Vision Fund之所以……的原因吗？
Was that the mechanism for why the Vision Fund was...
是因为往不需要钱的公司里砸了太多钱，还是因为时机太晚，还是两者都有？
Was it just dumping too much money in companies that didn't need it, or was it hitting the cycle at too late of a time, or was it both?
我觉得两者都有。
I think it was both.
首先，我不认为那是个坏主意。
Well, first of all, I don't think it was a bad idea.
我觉得只是时机不对。
I think it was just wrong time of the cycle.
而且当时的环境不是那种你投入更多钱就能变得更好的情况。
And then we were not in an environment where if you put more money into something, it got better.
没有规模限制，自动变好。
Just automatically, with no scale limitations.
AI就是这样。
That is what happens with AI.
对。
Yes.
以前，如果你给Salesforce投400亿美元，比如——
It used to be, if you put a $40 billion fundraise into Salesforce, like-
那反而会变得更糟。
It becomes worse
……他们能干什么？多招一堆销售和工程师？
... what are they going to do, go hire a bunch of sales reps and engineers?
根本行不通。
It just wouldn't work.
但如果你拿400亿美元去训练这些模型——
Whereas if you put $40 billion towards training these models-
对。
Yes
……它们会变得更好。
... they get much better.
所以，这就是我们现在所处的这个时代的独特之处。
And so that is very unique in the moment that we're in right now.
我是从产品周期和资本周期的角度来看这个世界的，我跟团队说，理想状态是你在产品周期上处于有利位置，一到十分，十分好，一分不好。
I think about the world as product cycles and capital cycles, and I tell my team, the ideal state of the world would be you're in a great position on the product cycle, scale of one to 10, 10 being good, one being not good.
2021年的时候，我们处于一分的位置。
In 2021, we were at a one.
当时我们并没有意识到这一点。
Which is, we didn't realize it at the time.
但那已经是SaaS的末期了——
But that was late SaaS-
已经到头了。
It had played out
——云计算的末期。
... late cloud.
手机增长期已经过去了。
It was post-mobile phone growth.
所有新增的消费者时间都已经流向了大型科技公司。
All of the incremental consumer time spent was going to the major technology companies already.
所以没有空间容纳新事物了。
So there was no room for something new.
所以产品周期只有十分之一。
So product cycle was a one out of 10.
而在2010年，那是十分之八。
And in 2010, it was an eight out of 10.
非常好。
Very good.
所以现在，产品周期是九分或十分。
So right now, product cycle is a nine or a 10 out of 10.
当然。
Sure.
基于我列出的所有原因，比如人工智能、自主性、美国活力、机器人技术、生物健康。
For all those reasons I laid out, like AI, autonomy, American dynamism, robotics, biohealth.
所有这些事情将在未来十年内发生。
All these things are going to happen over the next 10 years.
资本周期，理想的情况是产品周期满分，资本周期也满分，也就是说估值非常低。
Capital cycle, your ideal would be a 10 on product cycle and a 10 on capital cycle, meaning valuations are really low.
这两者几乎不可能同时出现——
They almost never coincide-
……因为大家对未来充满期待。
... because there's excitement about what's ahead.
对，那就像是产品周期最好的时候，而且只有你看到了，别人都没发现。
Yeah, it'd be like the best part of the product cycle ever, and no one else saw it but you.
2023年，AI刚开始发挥作用的那段早期——
2023, early days of when AI started working-
那段时间应该挺不错的。
That was probably pretty good
……而且资本市场相对温和，那些疯狂的事情和暴涨还没发生。
... and the capital markets were a little more benign, and some of the crazy stuff and run-ups hadn't happened yet.
那大概，我也说不好，能打个七分吧。
That was, I don't know, a seven.
2010年的时候，你可以用四倍营收或者更低的价格买到SaaS公司和B2B市场平台公司。
And 2010, you could buy SaaS companies and B2B marketplace companies for whatever, four times revenue or something.
那时候也挺不错的。
That was pretty good, too.
现在从资本周期的角度看，大家都看到了兴奋点，你仍然能找到好的投资机会，而且这些公司增长这么快，这么有活力，所以还好。
Right now, capital cycle-wise, people see the excitement, and you can still find ways to do good investments, and the companies are growing so fast, and they're so dynamic, it's fine.
但我们现在可能也就是六分左右。
But we're probably at a six or something.
你会觉得，还行，但不是特别好。
You're like, it's okay, but not great.
但真正驱动我们业务十年长期发展的，不管是风投还是成长型投资，都是产品周期。
But the thing that drives our business over 10 years, whether it's venture or growth, is the product cycles.
而现在产品周期大概是九分或十分。
And that's like a nine or 10 out of 10 right now.
这有点像——我不记得是不是先从Thrive那里听到的——就是说，一种逆向思维的方式是相信一个别人都不相信的事情。
It's sort of like, I can't remember if this is, I heard this first from Thrive or somebody, but it was like one way to be contrarian is to believe a thing nobody else believes.
另一种方式是相信大家都相信的事情，但相信的程度要高出十倍左右。
The other way is to believe the same thing everybody else believes, but believe it 10 times more or something like that.
这有点像我们公司的座右铭，尤其是在成长阶段，就是逐步相信事情可能往好的方向发展。
And that's a little bit our motto as a firm, I would say, is just, especially at the growth stage, incrementally believe what could go right.
把时间花在思考可能往好的方向发展的事情上。
And spend the time thinking about what could go right.
我常说我们的业务就是用合理的价格买下优秀的公司。
I always say our business is we pay fair prices for great companies.
你获得的超额收益总是来自收入增长这一块。
And the alpha you get is always on the revenue side.
而不是纠结于这家公司最终会不会成为利润率更高的企业。
It's not like, is this going to end up a higher margin business than not?
所以这需要市场洞察力和产品洞察力。
And so that's market insights, product insights.
而市场有时并不欣赏那些能够长期高速增长的事物。
And the market doesn't appreciate sometimes the things that can really, really grow fast for a long period of time.
Databricks就是一个例子。
Databricks is an example of this.
你知道吗？
You know?
顺便说一句，我最近和Ali吃了顿饭，天哪，他真是个了不起的CEO。
By the way, I got to have dinner with Ali recently, and oh my God, what an amazing CEO.
所以我发明了“Model Busters”这个说法，基本上就是我们第一次从成长基金里投资的时候。
So I coined this phrase Model Busters or whatever, which is basically the first time we invested out of the growth fund.
显然，Ben做了A轮，然后我们基本上每一轮都跟了，现在应该到M轮了。
Obviously, Ben did the Series A, and then we did basically every letter, and I think we're on M now.
我跟Ali开玩笑说，我们都要坐到体育场座位了，快变成双A了。
I joke with Ali that we're going to be stadium seating and get to double A.
那挺好的。
That's fine.
你不得不这样。
You have to.
我们打算——
We're going to-
那就永远保持私有吧。
Just stay private forever.
我觉得不会。
I don't think so.
但我们在增长基金做的第一笔投资，是在基金刚启动的时候，当时我们投了一轮，估值是60亿美元。
But the first one that we did out of the growth fund was right when we started the growth fund, and we did a round at $6 billion.
而技术终结者阿里，七年后发现他的收入加速增长了。
And the technical terminator, Ali, has figured out seven years later, his revenue has accelerated.
他还推出了更多新产品。
He has figured out more new products.
我们的观察是，当事情真正顺利时，往往是创始人自己找到了下一步的方向。
And our observation when things really go right, it's often the founder that figures out the next thing.
现在他的业务正好能受益于人工智能，所以才会增长这么快。
And his business is now positioned to benefit from AI, that's why it's growing so fast.
我们的赌注是，他会想出下一个点子，再下一个点子——
And our bet is he'll figure out the next thing and the next thing-
……再下一个点子。
... and the next thing.
而且他非常出色。
And he's great.
像这样的创始人，正如你们所拥有的成长型业务、晚期风险投资业务，这些创始人本身就是资产类别。
And those founders like that, as you guys have a growth business, late stage venture business, those founders are the asset class.
如果你支持他们，他们会想出所有这些令人惊叹的事情来做。
If you back them, they're going to figure out all these amazing things to do.
当我们投资SpaceX时，Starlink甚至还没有正式可用。
When we did SpaceX, Starlink wasn't even GA yet.
他们当时有一些理论，但经济性很差，而且没有成功。
And they had theories as to, and the economics were terrible, and it didn't work.
而现在，这成了有史以来最好的业务领域之一。
And here we are, that is one of the best business segments ever created.
而且他还在上面搞了个AI业务，接下来还会有轨道计算，这个也会成功的。
And he's got an AI business on top of it, and he's going to have orbital compute, which is also going to work.
所以，这些创始人就是你必须要押注的对象，而公司的成果可能会远超你任何财务模型里的预测。
And so those founders are the things that you have to bet on, and the outcomes of the companies can be much larger than you would ever put in a financial model.
这就是我们的投资理念。
And that is the thesis that we have.
我非常敬佩安德森的一点是，我觉得他从一开始就抓住了叙事的重要性。
One of the things that I really look up to about Andreessen is I think you got this sort of importance of narrative right from the beginning.
我想，即使是在公司刚成立的时候，他们就已经深刻理解到，当时媒体非常重要，但他们明白故事很重要，呈现方式也很重要，而且要高调参与对话，这真的很关键。
I think about even when the firm was started, there was such an understanding that, at the time, media was very important, but there was just an understanding that stories matter and that the way you present matters, and being loud and in the conversation is really important.
而且我觉得，显然从，嗯，2009年左右开始？
And I feel like obviously from, what, 2009 or something?
2009年，对。
2009, yeah.
从那以后，显然变化很大，而且我觉得一个很令人高兴的更新是，我们看到了某种“露露化”的现象，人们直接面向受众，掌控自己的叙事，这样能获得更多的真实性。
Since then, obviously, things have changed a lot and I think in a very happy update, we've had sort of the Lulu-ification of everything, where people are going direct and owning their own narrative, and you get more authenticity.
你们现在搞了很多新媒体，还有其他各种东西，但我觉得 Andreessen 一直都很清楚叙事的重要性。
You guys have a lot of new media going and all these other things, but I think Andreessen has understood the importance of narrative.
在我看来，现在对公司的叙事比以往任何时候都更重要，而且贯穿了各个发展阶段。
And it seems to me that narrative is louder than ever right now for companies, and it's at all stages.
就像公司刚成立时，你会看到那些精美的宣传视频；到了后期阶段，投资者也会以各种方式帮他们讲故事，同时公司自己也掌控着媒体渠道，整个氛围就是非常喧闹、高调。
It's like companies launch, and you see these beautiful videos, and companies at the later stages have their investors kind of creating stories for them in various ways, and they're owning media, and it's all just loud.
不过我觉得这很重要。我很好奇，在当下这个周期里，我们都活在 X（推特）上，每天被新闻包围，你是怎么看待这一切的？
But I think it's important, and I'm curious how you think about narrative now at this point in the cycle when we're all living on X, and we're living in the news, how do you think about this all?
你看，这其实就是氛围感的问题，对吧？
Look, this is just like the vibes, right?
氛围感真的特别重要。
The vibes matter tremendously.
但确实很重要。
But they matter.
重要性非常大。
They matter a lot.
其实我有个有趣的故事，是2018年和Ben开会时发生的。那时候，大概是我加入前三四个月吧，我们正在讨论我们的策略应该是什么。
I actually have a funny story from a meeting that I was having with Ben in 2018, so this was, I don't know, three or four months before I joined, and we were kind of hashing out what is our strategy going to be.
我当时有很多想法，比如我们需要组建哪些新团队，投资团队应该是什么样子。然后他就问：“那你的营销策略是什么？”
And so I had all these ideas of these are the new teams we need to build, and this is what the investment team looks like, and he's like, "Well, what's your marketing strategy?"
我就说：“我们其实不太需要营销策略。”
And I was like, "We don't really need a marketing strategy.
“在增长领域，这不算什么大事。”
It's not really a thing in growth."
他就看着我，说：“这是我听过的最蠢的主意。”
And he just looked at me and said, "That's the dumbest fucking idea I've ever heard.
“以前风投公司都不做自我营销，然后我们做了，效果非常好。那你为什么不在这里做呢？”
And venture firms didn't market themselves, and then we did, and it worked really well, so why would you not do it here?"
所以我就加入了。
And so I came along.
我之前待的地方比较保守，所以现在算是开了眼界。
I came from a place that was a lot more conservative, and so I've sort of seen the light.
氛围很重要。
Vibes matter.
我们来聊聊为什么氛围很重要。
Let's just talk about why the vibes matter.
氛围之所以重要，是因为它影响融资、估值、留住员工和招聘新员工。
So the vibes matter because of fundraising, of your valuation, for retaining your employees, and for hiring new employees.
还有其他原因。
Among other reasons.
在某些市场，氛围对客户也很重要。
And in some markets, the vibes matter for your customers, too.
但至少，就是这些方面。
But at a minimum, it's those things.
你看，这很基本，但为什么更高的估值很重要？
If you look at, and this is basic, but why does a higher valuation matter?
更高的估值意味着更少的稀释，更强的融资能力，而且你可以把它当作对抗竞争对手的武器。
A higher valuation means less dilution, more ability to raise capital, and you can use that as a weapon against competitors.
在一个资本真正能让公司变得更好的世界里，
In a world where capital actually makes your company better.
它就更有价值。
It's worth more.
没错。
Exactly.
它更有价值。
It's worth more.
当资本不是关键因素时，它的价值就低了。
It was worth less when it was not the thing.
现在每个投资者、每个员工和潜在员工都在X上，
And now every investor and every employee and potential employee is living on X-
……并吸收这些氛围。
... and consuming the vibes.
所以你看公开市场，那些在这方面做得好的公司，它们的估值直接受益，而且估值越高，日子就越轻松，对吧？
So if you look in the public markets, the companies that have done a great job of this, there's a direct line to their valuation, and their life gets a lot easier with a higher valuation, right?
现在最出名的就是Palantir和Karp，而且——
So famously right now, it's Palantir and Karp, and-
……他有一群狂热的追随者。
... he's got a cult following.
你知道吗，让我觉得特别神奇的是，Palantir在最近这波增长之前就已经有很高的估值倍数了。不过现在我觉得增长确实加速了，对吧？
You know what's amazing to me on that one is that he, I mean, Palantir had a huge multiple before the recent-- Now, I think the growth has inflected recently, right?
对，确实如此。
Yeah, it has.
对，增长大幅加速了。
Yeah, it's massively accelerated.
大幅加速了。
Massively accelerated.
但在增长加速之前，它就已经有了这个溢价。
But it had the premium before it had that.
是的。
Yes.
关于 Bayt 播客
Bayt 提供中文+原文双语音频和字幕，帮助你打破语言障碍，轻松听懂全球优质播客。