人工智能如何创造了史上最快的产品周期

是的。

Yeah.

嗯，我认为有两方面原因。

Well, I think it's two things.

我认为，自上而下的商业模式质量确实很高。

I think, you know, top down business model quality is really high.

所以作为消费领域的创始人，你必须自问：我每天该如何分配时间？

So you can just as a consumer founder, you have to ask yourself, like, how am I gonna spend every day?

我认为，作为2014至2023年间的一位消费领域创始人，你日常的工作重心可能都放在了营销上。

And I think as a consumer founder from 2014 to 2023, the way you probably spent your day with marketing.

你知道吗？

You know?

但真正能从消费品营销和客户获取中获得乐趣的人其实寥寥无几。

And that's just very few of us are people that get joy from consumer product sort of marketing and customer acquisition.

而如今，作为消费领域的创始人，你每天的时间都花在了产品打造上。

Whereas if you look at how you spend your day today as the consumer founder, you're spending it building product.

我认为如今消费品公司不存在营销问题。

I would argue that there are no marketing problems today for consumer companies.

只有产品问题。

There are only product problems.

如果你没有获得市场覆盖，很可能是因为你在产品上还不够有野心。

And if you're not getting distribution, you probably haven't been ambitious enough in product.

所以我觉得这是一个非常有趣且神奇的时代。

So I just think that this is a really fun and magical time.

你知道，商业模式的质量很高。

You know, business model quality is high.

消费者愿意付费。

Consumers are willing to pay.

他们想尝试新事物，作为消费品创始人，你有太多可以发挥的空间。

They wanna try new things, and then you have so much to work with as a consumer founder.

明白吗？

You know?

你们能够打造出富有情感、能够捕捉和理解情感体验的产品，比如poke这样的产品。

You're able to build products that are emotional and sort of capture and understand emotional experiences, products like poke.

你知道吗？

You know?

你们可以利用语音这类新型基础元素进行创新。

You're able to do things with new sort of primitives like voice.

我的意思是，语音是人类最原始的交流方式，但至今却几乎没有围绕它开发的技术。

I mean, voice is the original form of human communication, and yet there's really been zero tech built around it ever.

比如，语音和界面的未来会是什么样子？

Like, what will the future of voice and interface look like?

我不知道，但你今天就可以某种程度上创造并想象它。

I don't know, but you can kind of invent it and imagine it today.

所以感觉就像你正在享受的乐趣，以及成功和获得这种乐趣所带来的真正令人兴奋的成果，是我们自2010年以来未曾见过的。

So it just feels like the amount of fun you're having and, like, the really exciting consequences of being successful and having that fun are something that we haven't seen since 2010.

另外，你知道真正有趣的是什么吗？消费者愿意为这些产品付费。

Also, you know, what's really interesting, consumers are willing to pay for these products.

以前，可能只是为Spotify和Netflix支付10到20美元，或者一些非常特定的订阅服务。

Before, may have been $10.20 dollars for Spotify and Netflix or very, you know, specific subscriptions.

哦，我不确定为什么我要为此付费。

Oh, I'm not sure why I wanna pay for it.

现在为Lovable支付25美元？没问题。

$25 now for Lovable is like, sure.

我明天就注册。

I'll sign up tomorrow.

嗯，是的。

Well, yeah.

实际上，200美元的价格对很多人来说也是可以接受的。

And actually think $200 is like, sure.

很多人明天就会注册。

I'll sign up tomorrow for many people.

这些就是价格点。

And those are the price points.

你看，Gemini Ultra，Google Ultra，每月250美元。

If you look at, you know, Gemini Ultra, Google Ultra, that's 250 a month.

ChatGPT的最高级套餐是200美元。

ChatGPT's top SKU is 200.

Grok Heavy要300美元。

Grok Heavy is 300.

人们愿意付这个钱。

People are paying it.

你知道吗？

You know?

我是说，如果你看看Cursor上的消费，我很想知道业余爱好者在Cursor上花了多少钱。

I mean, if you look at the spend on Cursor, I'd love to see what hobbyist spend is on Cursor.

我打赌数额相当可观。

I bet it's really significant.

我觉得看他们的收入规模很容易就会说，嘿。

I think it's easy to look at their revenue scale and say, hey.

这是企业级市场。

It's it's enterprise.

我很清楚这一点。

And I know it.

我确信其中很大一部分是企业用户，但肯定也有不少是业余爱好者。

I'm sure a significant part of it is enterprise, but I'm sure a lot of it is also hobbyist.

另一个很酷的亮点是，这是史上首次出现了面向消费者的按量付费模式。

The other cool thing is for the first time ever, you've got consumption revenue for consumers.

对吧？

Right?

过去消费者通常只能订阅产品，但你知道，没人会每月给Spotify支付100美元。

So, typically, in the past, a consumer could subscribe to a product, but, you know, you're never gonna pay Spotify a $100.

如果你每月支付20美元订阅费，即使你听了五倍多的音乐，也无法在一个月内支付100美元。

If you pay 20 a month in subscription, there wasn't a way if you listen to five x more music for you to pay them a $100 in a given month.

而现在使用Cria这样的创意工具或Cursor这样的编程工具，理论上你可以支付远超固定订阅费的费用给这些公司。

Whereas now if you're using a creative tool like Cria, a coding tool like Cursor, you can conceptually pay that company a lot more than the fixed subscription rate.

这对面向消费者的创业者来说是一种新的商业模式改进。

And, that's kind of a new business model improvement for consumer founders.

你之前提到过这种新型的分发模式。

You mentioned earlier around this new distribution type of model.

对于新兴的消费级AI公司进入市场，这会是什么样子？

What does that look like for new consumer AI companies coming out to market?

是的。

Yeah.

这个话题非常棒。

This is such a great topic.

我认为这正是我们一直在寻找的东西。

And I think this is the one thing we've all looked for.

你知道，那是在2005、2006年的时候。

You know, it was in 2005, 2006.

当时是Facebook这类社交API加上Facebook通知渠道的开放。

It was the Facebook sort of social APIs plus access to Facebook notifications as a channel.

如果你回顾2009年，那就是应用商店的诞生。

If you look at 2009, it was the App Store.

而现在我们实际上正迎来这个时刻，我认为在2026年我们将看到类似的情况。

And now we actually have this moment, I think, in 2026 where we're gonna see something similar.

所以我首先关注的是OpenAI大约30到45天前宣布的Apps SDK，它允许你将体验嵌入ChatGPT并获得分发渠道。

So the first thing I'm paying attention to is the Apps SDK, which was announced, I don't know, thirty or forty five days ago by OpenAI where you can embed experiences within ChatGPT and get distribution.

他们将建立发现机制和应用商店，某种形式的应用商店。

They're going to have discovery and an app store, some sort of an app store.

我关注的第二件事是迷你应用。

The second thing I'm paying attention to is Mini apps.

这很酷，因为Wabi在迷你应用方面走在了前面，我没想到苹果会完全接纳它，但他们确实这么做了。

And this is it's cool because Wabi was ahead on the Mini apps, and I didn't expect Apple to fully embrace it, but they have.

他们宣布将支持迷你应用生态系统。

They announced that they're gonna support a Mini apps ecosystem.

重要的是，他们已将迷你应用的抽成比例从30%降至15%。

Importantly, they've reduced their take rate for Mini apps from 30 to 15%.

所以这对平台和迷你应用创作者来说是个相当不错的经济利好。

So that's a pretty cool economic benefit to the platforms and the mini app creators.

你知道吗？

You know?

最后，我们看到ChatGPT中围绕群聊功能有很多热议。

And then finally, we saw a bunch of hubbub around group chat in ChatGPT.

最初在新西兰和一系列其他国家推出。

Launching it initially in New Zealand and a bunch of other countries.

当然，它也会进入北美市场。

It's, of course, gonna come to North America.

我猜测这意味着OpenAI将在群聊中实现应用发现和消费功能，而Meta等公司将不得不效仿。

I suspect what that means is that OpenAI will have app discovery and consumption within group chats, and then Meta and others will be forced to replicate that.

所以作为消费者创业者，你看，你有这三个渠道：AI、应用SDK、迷你应用和群聊，这些都可能变得非常非常重要，这意味着2026年将是一场激烈的竞赛。

So as a consumer founder, you know, you've got these three channels, AI, Apps SDK, MiniApps, and group chats, which could be really, really big, which means 2026 is gonna be a foot race.

明白吗？

You know?

这可不是胆小者能参与的游戏，但到明年我们再次讨论时，应该会涌现出多家用户过亿的消费级公司。

It's not gonna be the for the faint of heart, but the consequences should be multiple consumer companies with a 100,000,000 users when we have this conversation next year.

让我们实现它。

Let's make it happen.

我认为另一个非常有趣的消费趋势正在形成，那就是语音技术的未来。

Well, I think another really interesting consumer trend, but also trend happening is, like, this whole future of voice.

语音正在成为潜在的新界面，提升生产力，让我们在工作中表现得更好。

Voice is now becoming, you know, potentially the new interface and productivity and, you know, making us, you know, just better at all our jobs.

语音领域正在发生什么？为什么现在如此令人兴奋？

What's happening in voice, and why is now so exciting?

语音技术很酷。

Voice is cool.

你知道，过去几年我们一直痴迷于语音技术并密切关注其发展。

You know, we've been obsessed with voice and covering it really closely for the last few years.

我认为部分灵感来源于一个梦幻般的愿景——不知道你是否读过《安德的游戏》，书中有个叫简的语音助手。

I think part of it started from a little bit of a dreamy vision of I don't know if you've read Ender's Game, but they've got a voice assistant named Jane in that.

当然，关于她的讨论已经多到让人厌烦了。

And, you know, of course, her has been discussed to death.

但我们一直对语音界面和技术抱有这种奇幻的想象。

But we've had these sort of fantastical imaginations of a voice interface and technology.

只是我们一直缺乏实现它的技术。

We've just never had the technology to make it happen.

所以在2023年我们做着这些梦，做出一些预测。

So we were sort of dreaming those dreams in 2023 and making some predictions.

就像我之前说的，对于AI代码和法律科技，我们低估了它们的重要性和规模。

And as I as I said, for AI code and legal tech, we under predicted how important, how big it was going to be.

语音并不是一个市场。

Like, voice is not a market.

语音并不是一个原始功能。

Voice is not a primitive.

语音是一场产业变革。

Voice is an industry change.

语音技术正在成为AI进入企业的切入点，因为这是企业已经在做的事情。

What's happening with voice is that it's turning out to be the insertion point for AI into the enterprise because it's something that the enterprise already does.

全球每家企业都有电话和语音业务。

Every enterprise in the world has phone calls and voice.

语音技术包含两个部分。

There's two parts of voice.

一部分是广义的笔记记录，另一部分是电话业务和我们所说的代理服务。

There's sort of scribes, which is note taking broadly, and then there's phone calls and what we call agents.

事实证明，目前代理服务运作得更好，规模也比笔记记录更大。

It's turned out so far that the agents have worked, you know, better and are at higher scale than the scribes.

那些笔记记录功能也很有意思且重要。

Those scribes are also very interesting and important.

结果证明这个市场比我们任何人预期的都要大得多。

And it's just it's turning out to be a lot bigger than any of us expected.

所以我认为将会有数百家公司以语音技术为切入点，在各个垂直领域实现规模化。

So I think there's going to be hundreds of companies that get to scale in verticals with voice as their wedge.

如今语音技术最有趣的地方在于，它过去听起来非常机械。

And what's so interesting about voice now is it used to be very robotic.

对吧？

Right?

现在则更接近人类的感觉。

Now it's more human ish.

那些‘嗯’、‘啊’之类的语气词就很像真人。

The ums, ars, ums are kind of like that.

也许情感连接还没完全到位，但我们仍处于非常早期的阶段。

Maybe the personal connection is not quite there yet, but we're still so early.

我认为情感连接已经存在了。

I think the personal connection is there.

你知道吗？

You know?

而且我认为语音技术还有几个特别有趣的特点。

And I think that it and there's there's a few interesting things about voice.

首先，我们都被电话菜单系统折磨出心理阴影了，所以一听到语音AI就会想‘天啊’

So one is we've all been so scarred by phone trees that whenever we hear voice AI, we're like, oh my god.

肯定又是按2、按3、按4那一套

It's gonna be another press 2, press 3, press 4.

但完全不是这么回事

It's not that at all.

我认为第二点是，语音AI公司和基础技术提供商根本没想欺骗电话那头的人

I think the second is that the voice AI companies and just sort of primitives, they're not trying to trick the person on the other side.

他们会直接告诉对方‘嘿’

Like, they tell the other person, hey.

‘我是AI’

I'm an AI.

第三点是，尽管有第二点存在，人们仍会建立情感联系——即便知道自己在和AI对话

The third is that the relationships are being formed despite point two, that people know they're talking to an AI.

因为我们原始大脑的机制就是如此：一旦听到类似人类的声音，就会本能地产生反应并开始形成情感连接

Because we're so our sort of primitive brain is so trained to react and start to form an emotional connection once it hears something that mimics a human voice on the other side.

尽管知道它是AI、是机器，你还是会逐渐对它产生好感。

That despite knowing that it's an AI and it's a machine, you you start to warm up to it.

正因如此，AI在谈判、说服和建立友谊方面才如此高效。

And and that's why the AI has been so effective in negotiation, persuasion, building friendships.

这不仅仅是低层次、无关紧要的信息收集电话。

This is not just low level unimportant, you know, information gathering phone calls.

我认为企业每年最重要的电话应该且将会由AI来处理。

I think the most important phone call that happens in a business in a given year should be and will be handled by AI.

完全同意。

Totally.

我们正越来越多地看到这种现象。

And we're seeing this more and more.

对吧？

Right?

几周前我和一位创始人聊过，他正在开发语音版的Gmail。

I spoke to a founder a few weeks ago who's building Gmail for voice.

所以如果你在跑步或在车里，你其实可以通过语音回复邮件。

So if you're on a run or you're in the car, you can actually respond to emails by voice.

我觉得这挺有意思的。

I think that's quite interesting.

是的。

Yes.

处理比如医疗预约、牙医、试图向航空公司退款这些我们都讨厌的事情。

Dealing with, like, medical appointments, dentists, trying to get a refund from a airline, which we all hate.

我们都经历过这种情况。

We've all been there.

你知道吗？

You know?

是的。

Yes.

做语音处理，我认为会非常有趣。

Doing doing voice, I think, could be really interesting.

你认为机会在哪里？

Where do you think the opportunities are?

这更像是消费级产品还是企业级产品？

Is this more gonna be a consumer product, an enterprise product?

这个领域谁会胜出？

Who wins in this space?

我认为不会有单一的赢家。

I don't think there's one winner.

我觉得这是基础设施层面的技术。

I think it's a primitive.

你知道，每个人都会以不同方式将其产品化。

You know, it's something that everybody's going to productize in different ways.

到目前为止，这更多是企业级应用而非消费级应用。

So far, it's been more of an enterprise story than a consumer story.

但你知道，它们没有理由不能同时触达消费者市场。

But, you know, I there's no reason they can't reach into the consumer as well.

我有个习惯，有时会散步时和ChatGPT对话，进行语音交流。

I do this thing where I'll sometimes go for walks and talk to ChatGPT, just have a voice conversation.

当我想探索话题、提问或了解某事物的历史时，我也会在车里这么做。

Where I wanna explore topics or ask questions or learn about the history of something, I do it in the car as well.

所以我认为语音正成为继文字之后消费者接触模型的第二界面，没有理由它不会比现在普及得多。

So I think we're starting to see voice as the second interface to the models for consumers after text, and there's no reason it won't get to a lot more scale than it is today.

顺便说一句，我很喜欢Gmail这个创意。

I love the Gmail idea, by the way.

我觉得这是个很棒的主意。

I think that's a great idea.

是啊。

Yeah.

这是家相当有趣的公司。

It's a pretty interesting company.

看看语音是否会进入企业端，让我们都在Teams里使用语音，这会很有意思。

It's gonna be interesting to see if voice goes to the enterprise side and we're all using voice within Teams.

你能想象走进办公室，所有人都在跟这些语音助手交流想法或回复邮件吗？

Can you imagine going into an office and everyone's talking to these voice agents for ideas or responding to emails?

我觉得这行不通，对吧？

I don't think that's gonna work, is it?

我是说想象一下。

I think imagine.

我...我觉得，试着想象一下我们如今的工作方式被这些能力重新混合，确实感觉有点奇怪。

I I think it you know, it try to imagine the way that we work today, you know, remixed by these capabilities does feel a little bit strange.

但我同时也认为，企业的设计本就是围绕人类喜欢的工作方式、学习方式和专业化方式。

But I I also just think that we've designed the enterprise to build around the ways that humans like to work, the way that humans like to learn, the way that humans like to specialize.

举个例子，如果是呼叫中心，可能会有一组人负责客服，另一组人负责销售。

So for example, if you have a call center, you might have a set of people that handle support and a set of people that handle sales.

擅长客服可能意味着你非常有同理心。

And maybe being great at support means you're very empathetic.

你非常耐心。

You're very patient.

你非常了解产品。

You really know the products.

你自己就是热情的用户。

You're your user passionate user yourself.

擅长销售的人可能更像是个狠角色。

Maybe being great at sales is you're more of a killer.

你是个猎手。

You're a hunter.

你擅长说服他人。

You're good at persuasion.

你性格非常开朗。

You've got a really bright personality.

你更善于说而非听。

You're more of a talker than a listener.

这些只是两种不同的人类原型，所以我们为它们设置了两种不同的工作职能。

Like, these are just two different human archetypes, which is why we have two different job functions for them.

但是否存在这样一种可能：客服与销售由同一名代理完成？或者是否存在这样一种场景：你在销售过程中提供客服支持，在客服过程中进行销售？

But is there a world in which support and sales are handled by the same agent, or is there a world where you do some support while you're doing sales, some sales while you do support?

因此我认为这就是企业配置将发生根本性变革的原因。

So I think this is why the kind of configuration enterprise is gonna dramatically change.

我们不再受限于人类必须遵循的工作方式，可以从基本原理出发，围绕模型能力重新构建这些体系。

We're no longer constrained to the ways that humans need to work, and we can sort of build these things from first principles around what the models can do.

我认为对语音技术而言，创作者经济领域同样充满机遇。

I think what's really interesting for Voice as well is around the creator economy.

记得我们节目前提到的11 Labs吗？在新冠疫情前或这些AI模型出现前，创作者们只能用母语发布口语内容。

You know, we mentioned 11 Labs earlier, right, before the show and it's, you know, creators before COVID or before these AI models could only, you know, publish content in their spoken language, native language.

现在有了这些工具，我们作为创作者可以让内容被翻译成全球各种语言。

Now there are tools available where you as a creator, us as a creator, can have our content translated into every language globally.

这太棒了。

Amazing.

确实令人惊叹。

Which is amazing.

对吧？

Right?

你知道，我们提到现在每月在所有渠道上大约有百万浏览量。

You know, we mentioned we do around a million views a month now across all of our channels.

如果我们开始把内容翻译成中文、印度语、日语，很快就能达到五百万甚至千万。

If we start translating our content to Chinese, Indian, Japanese, we can get, like, five, ten million very quickly.

这正是我想说的。

Well, this is what I mean.

我是说，语音模型的多语言能力实在太惊人了。

I mean, the multilingual capabilities of the voice models are so amazing.

这也意味着对世界的一个重大影响就是信息传播更广泛了，这很棒。

It also just means one that the big implication for the world is just more information diffusion, which is great.

你明白吗？

You know?

我们需要更多这样的知识。

We need more of this knowledge.

这些知识曾经非常专业化。

The knowledge was so specialized.

二十年前，你必须身处硅谷的某个房间才能获取那些知识，比如如何打造一个'为工具而来，为网络而留'的商业模式。

Twenty years ago, you had to be in one room in Silicon Valley to get the knowledge of, like, how to build a you know, come for the tool, stay for the network, for example.

你知道克里斯那篇著名的文章。

You know, Chris's famous essay.

这种行业秘诀当时可能只有五到二十个人真正理解。

And that was this sort of lore that maybe five, seven, ten, twenty people understood.

后来克里斯写了博客文章，但文章是英文的，而且你得知道去哪里找它。

And then over time, you know, Chris wrote the blog post, but the blog post is in English, and you had to know to look for it.

渐渐地，数百、数千乃至数百万人掌握了这个框架。

And, you know, hundreds or thousands and eventually millions of people became trained on that framework.

但未来还有数亿甚至数十亿有抱负的创业者没接触过这些，语言障碍正是阻碍他们的因素之一。

But there are hundreds of millions or billions of aspiring founders in the future that have not been trained on it, and one of the things that holds them back is the multilingual necessity.

所以我认为所有这些都在推动信息扩散，这是件极其强大的事情。

So I just think that all of these things are pushing to information diffusion, which is a very, very powerful thing.

这也意味着产品更具吸引力。

It also means that products are more compelling.

你知道吗？

You know?

如果你看看企业语音应用场景，比如催收业务，当你能用对方的母语沟通时，效率会大幅提升。

If you look at some of these enterprise voice use cases like collections, when you're able to call somebody and speak to them in their native language, you're just able to be a lot more effective.

你还能展现出更强的共情能力。

You'll be able to be a lot more empathetic.

所以我认为，对于像我这样在双语家庭长大的人来说，能用母语交流的重要性怎么强调都不为过。

So I think for for folks who grew up in a household like I did where my parents spoke English but also other languages, for them to be able to connect in their native language is it it kinda can't be overstated how important it is.

完全同意。

Totally.

我们将见证这个领域的巨变，说到这个，现在正好可以聊聊创作者经济。

We're gonna see a big change in this, and, you know, this is a good segue talking about creators now, the creator economy.

感觉创作者经济在疫情期间确实是个大热点。

It kind of feels, you know, the creator economy, this was a big thing during COVID.

对吧？

Right?

我们每年都会发布创作者报告，感觉创作者经济非常令人兴奋，但至今还未出现能与AI规模相匹敌的头部公司。

We publish these creator reports every year, and it it kind of feels the creator economy is is super exciting, but we haven't seen any outlier companies really at the scale versus AI yet.

是的。

Yes.

是的。

Yes.

创意经济是否依然充满吸引力？你认为AI将对创作者乃至整个创意领域产生怎样的影响？

Is is the creative economy still exciting, and what do you think AI's impact is going to have on creators more generally, but also creativity?

在我看来，创作者是我们整个社会中最了不起的一群人。

I mean, creators are some of the most tremendous people, I think, in our entire society.

我特别反感那种论调，说什么'现在的孩子不想当宇航员了，只想当YouTuber'。

And I I really hate the conversation around, like, oh, the kids used to wanna be astronauts, but now they wanna be YouTubers.

没错，他们确实想当。

Like, yes, they do.

我认为这句话背后的真正含义是，他们想成为企业家。

And and what what is meant by that, I think the underlying desire is they wanna be entrepreneurs.

如果你是个懂技术的人——要知道真正懂技术的人并不多。

And if you're an individual who's done technical, not that many people are technical.

对吧？

Right?

再说一次，全球只有2000万程序员。

Again, 20,000,000 programmers in the world.

成为数字创业者的最佳途径可能就是当一名主播。

The best path to being a digital entrepreneur was probably being a streamer.

所以我认为这就是当人们说想成为创作者或内容创作者时的核心诉求，尽管他们唯一能创造的就是内容。

So I think that is the kind of heart of what people say when they say they wanna be a creator or a content creator, but the only thing they've ever been able to create is content.

他们在这个领域取得的成就是惊人的。

What they've done with it is tremendous.

YouTube这个5500亿美元的商业帝国，还有像你这样的节目，获得了巨大的影响力并帮助了无数人。

YouTube, $550,000,000,000 enterprise, shows like yours, which get incredible reach and help so many people.

现在创作者的工具箱里多了两样新工具。

Now creators have two other tools in the toolbox.

明白吗？

Okay?

其一，除了内容创作，他们还能开发软件。

So one is in addition to content, they can create software.

因此，使用Wabi等平台的创作者现在有了开发软件的途径。

So creators using platforms like Wabi now have a place to create sort of software.

软件与内容截然不同，其价值会随时间复利增长，而内容价值往往趋于平缓或衰减。

And a software is very different from content in that the value compounds over time, where the value of content can tend to plateau or decay.

所以创作者们有了这个新创作领域，他们将会在这个新画布上迸发惊人的创造力。

So you've got this new thing that creators can make, and, you know, they will be wildly creative with this canvas that they've been given.

另一个有趣的新创作方向是模型开发。

The other interesting thing that creators can make is models.

我们在创意工具领域经常看到这种现象，比如Civet网站上的创作者们通过微调模型，使其指向特定的艺术风格方向。

And we've seen this a lot in the in the creative tool space where you've got all these folks on sites like Civet who have fine tuned these specialized models to point in a specific artistic direction.

这些都是个人在做的事情。

And these are individuals doing it.

很多时候都是些无名之辈。

There are nones a lot of the time.

他们是业余爱好者。

They're hobbyists.

他们为此收费。

They're charging money for it.

所以我认为现在给创作者更多工具——软件模型加上内容，会让他们比以往任何时候都更具吸引力，我预计我们会投资更多像Wabi这样为他们提供平台的公司。

So I think now creators giving them more tools, software models, plus content is gonna make them more compelling than ever, and I expect we'll invest in more companies like Wabi that sort of provide a platform for them.

是啊。

Yeah.

还有，你知道，我们现在看到这个趋势，完全由AI生成的电影，这真的会成为现实吗？

Also, you know, we're seeing this trend now where fully AI generated movies, is this really gonna be a thing?

没错。

Yeah.

嗯，你知道，这很有趣。

Well, you know, it's funny.

如果你回顾电影史，那些在电影摄影机发明后不久问世的电影，看起来就像舞台剧。

So if you look at the history of film and, you know, movies so the the the movies that came out soon after the the sort of film camera was invented, they looked like plays.

明白吗？

Okay?

它们会有固定场景，演员的表演方式也和舞台剧一样。

So they'd have set pieces, and people would act the way they act in stage plays.

当然，后来电影发展出了自己的语法体系。

And then, of course, what happened was film developed its own grammar.

比如方法派表演。

So things like method acting.

对吧？

Right?

方法派表演实际上是电影这种媒介发明后相对较晚才出现的创新。

Method acting by a relatively an innovation that came later for after the invention of the medium, which is, of course, film and film cameras.

我认为同样的事情也会发生在AI身上。

It's the same thing I think is going to happen with AI.

所以今天我们谈论AI电影，但这究竟意味着什么？

So today, we say things like AI movies, but what does that really mean?

我认为这有两层含义。

I I think it means two things.

首先，现在制作电影的人。

I think, one, you're gonna have people that make movies today.

他们将能够实现更具野心的创作，天马行空发挥创意，而不必经过好莱坞层层审批，向投资人申请执行创意构想。

They're able to, like, do much more ambitious things, be wildly creative, and not have to go through the Hollywood gauntlet of getting permission from a financier to execute on your creative vision.

但另一方面，我认为我们将看到AI原生的新形式，比如可能是一次性的微电影。

But I think the other thing is we're going to see new formats that are native to AI, things like microfilms that might even be disposable.

就像五年前你根本不会考虑制作一次性电影。

Like, you would never make a disposable film five years ago.

因为成本实在太高了。

It's just too expensive.

但现在你可以以非常、非常低廉的成本制作五分钟、七分钟、十分钟甚至十五分钟的影片，而且不会有同样的代价。

But now you can make a five, seven, ten, fifteen minute film at a very, very cheap price with, you know, without the same consequences.

所以，是的，我不知道'AI电影'这个概念是否是我们十年后还会讨论的话题。

So so, yeah, I I don't know if AI movies is like a concept we're talking about in ten years.

我认为我们会讨论由AI辅助制作的电影，以及像微电影这样AI原生的格式。

I think we talk about movies that are assisted by AI plus AI native formats like microfilms.

我觉得这个观点很好。

I think that's a good take.

我已经看到几个YouTube频道，它们实际上才开了12个月。

I've seen a couple of YouTube channels now which literally started 12 ago.

他们已经拥有四、五百万订阅者，内容完全是AI生成的，而且故事讲述得非常精彩。

They have five or five four four five million subscribers, and there's AI generated content with amazing storytelling.

我认为这类频道如果专注于某个特定领域，可能会获得非常有趣的发展机会。

I think those types of channels, you know, potentially have really interesting opportunities if they focus on a particular niche.

这太神奇了。

It's amazing.

我是说，我认为这背后的创作愿景是——如果每个故事都能被讲述会怎样？

I mean, I think that the the the creative vision for this is what if every story was told?

你知道有多少故事就这样被埋没了吗？

You know, how many stories are just left untold?

有多少故事停留在剧本阶段？

How many of them are, you know, where they're left in a script?

有些电影甚至已经拍完却未发行

Or the movie's even made, but it's not published.

有些电影拍出来了但不够大胆创新

Or the movie's made, but it's not sufficiently ambitious.

有些人有创意却从未写成剧本

Or the person has the idea, but they don't write the script.

我们将生活在一个所有故事都能被讲述的世界，这感觉会是一个充满创意繁荣的天地，令人期待而非恐惧

Like, we're gonna live in a world where all stories are told, and I just it feels like that'll be a world of creative abundance that we're excited about, not, you know, in dread of.

机会真是太多了

So many opportunities.

对吧？

Right?

是的。

Yes.

我们或许可以接着谈谈AI封装器。

Let's maybe jump on to AI wrappers.

我们之前讨论过这个，大概就是这样。

We talked about this earlier, and it's around okay.

很多AI模型都向初创企业提供API接口。

A lot of these AI models provide APIs to startups.

我之前和一位叫PJ的人聊过。

And I was talking to someone called PJ.

他是欧洲一家名为North Zone的风投公司合伙人。

He's a partner at a VC firm called North Zone in Europe.

简单来说，如果这些AI模型向初创企业提供API，它们就能完全掌握这些初创企业的业务动态和运营状况，结果两个月后它们就推出了自己的版本。

And if these if these AI models in simple terms provide APIs to start ups, they have full visibility on what these start ups are doing, how well they're doing, and all of a sudden, two months later, they launch their own version.

它消灭了所有公司。

It wipes out all the companies.

我们会看到更多这样的情况吗？这公平吗？

Are we gonna see more of this, and is it fair?

是的。

Yeah.

这是个有趣的话题。

This is this is a fun topic.

所以我完全不担心AI包装器的讨论，让我告诉你几个原因。

So I don't I don't worry at all about the AI wrapper conversation, and let me give you a couple of reasons why.

首先，我认为市场上有一些领域，初创企业具有明显优势。

So I think first of all, there are parts of the market where you just you're very advantaged being a startup.

市场的一个重要部分是那些客户使用多模型比单模型能获得更好体验的领域。

And an important part of the market is any category where the customer gets a better experience when they use multi model instead of a single model.

所以如果你在OpenAI工作，你只能发布使用OpenAI模型的产品。

So if you're at OpenAI, you're only ever gonna be able to ship products with OpenAI models.

如果你在谷歌工作，你也只会使用谷歌的模型，诸如此类。

If you're at Google, you're only gonna ship with Google models, etcetera.

但如果你看看像Cursor或Kriya这样的产品，这些产品确实需要接入所有模型和所有模型提供商，从定义上来说，实验室和大科技公司是无法做到这一点的。

But if you look at a product like a Cursor or a Kriya, these are products where you really want access to every model and every model provider and sort of definitionally, the labs and big tech are not gonna be able to do that.

所以多模型是一个真正有趣的领域。

So multimodel is one really interesting area.

我认为第二点是，如果你把模型不作为产品来看待，这是一个重要的区别。

I think the second is if you look at models not being products, that's an important distinction.

对吧？

Right?

所以模型可以做的事情开始暗示产品的可能性。

So models can do things that start to hint at products.

但是，你知道，即使OpenAI复制了Granolah的会议记录功能（他们确实已经复制了），我认为Granolah将要构建的是一个完整的AI原生生产力套件，包括电子表格、文档和演示文稿。

But, you know, even if OpenAI replicates Granolah's meeting recorder feature, which they have replicated, I think what Granolah is going to build, we're not an investor, is an entire AI native productivity suite, you know, with spreadsheets and documents and presentations.

而对于OpenAI来说，要复制所有这些功能并全面考虑实在是太复杂了。

And it's just so much feature service for an OpenAI to replicate all of that and think it all through.

这对他们来说确实是个棘手的优先级排序问题。

It's just a really hard prioritization problem for them.

我认为针对AI外壳担忧的最后一点反驳是，现在可以通过反馈对模型进行微调，无论是通过数据的强化学习还是人类反馈的强化学习，这意味着当这些初创公司发展到一定规模时，它们能够提供更优质的服务，因为它们收集的数据会反馈到只有它们能使用的定制版模型中。

I think the last sort of counterpoint to the AI wrapper concern is that, you know, you now have fine tuning of models with feedback, either reinforcement learning through data or reinforcement learning through human feedback, which means that as some of these startups start to get to scale, they're able to deliver a better offering because the data they're collecting is feeding back into a specialized version of the model that is only available to them.

所以我不认为外壳问题是个大问题。

So I don't I don't think the wrapper concern is a big one.

我认为如果你开发的产品与实验室当前的工作直接重叠，那确实存在风险。

I think that if you're, you know, if you're if you're building a product that is directly overlapping with something the labs do today, then sure.

可能有些风险。

Maybe there's some risk.

可能很难...比如说在云端代码领域超越云端代码。

Maybe it's difficult to, you know, out cloud code cloud code.

但看看像Cursor这样的产品，尽管从表面看可能会被复制，他们还是取得了巨大成功。

But if you look at something like Cursor, they've had tremendous success despite from afar saying, hey.

那是像Anthropic或OpenAI这样的公司可能会复制的功能。

That's something that, you know, an Anthropic or an OpenAI might replicate.

确实。

True.

这有点像那种必然会发生的情况，有些初创公司会遭遇这种事，但初创公司应该专注于自己的目标并坚决执行。

It's kind of one of those things where it's mean, it's gonna happen to some startups, but startups should just be focused on whatever they get at and just execute.

完全同意。

Totally.

完全同意，老兄。

Totally, man.

伙计，我还觉得其中一些产品真的很奇怪。

Dude, I also just think that there are some some of these products are really weird.

你懂我意思吗？

You know?

我经常开玩笑说，大科技公司的优先级都是在晋升委员会里定的。

And I always joke that the, you know, the company's priorities get set in promo committee for big tech.

所以如果你在谷歌工作，作为一个产品经理，你会这么想，嘿。

So if you're at Google, you know, you're thinking as a PM there, like, hey.

我怎样才能获得晋升？

How do I get promoted?

而获得晋升的方式并不是通过推出一些过于标新立异甚至冒险的新产品。

And the way you get promoted is not by launching some wildly creative sort of risque new product.

你知道的，你不会说，嘿。

You know, you don't say, hey.

我要创建一个，怎么说呢，有点暴躁的、个性鲜明的电子邮件界面。

I'm gonna create this, like, sort of grumpy personality driven interface to email.

比如，这可不是获得晋升的好方法。

Like, that is not a good way to get promoted.

听起来风险很大。

That sounds very risky.

你会说，哦，不行。

You say, oh, no.

不行。

No.

我要给Gmail添加AI生成的邮件草稿功能。

I'm gonna add, you know, AI generated email drafts to Gmail.

这感觉既稳妥又安全，我可以去找经理谈升职的事。

That feels good and safe, and I can go talk to my my manager about getting promoted based on that.

所以我认为，在大型科技公司里被鼓励追求的那种职业野心方向是截然不同的。

So I just think that sort of direction of ambition you're incentivized to have at Big Tech is really different.

而当你是一名创始人时，你会想：听着，

Whereas when you're a founder, you're like, look.

我必须与众不同。

I need to distinguish myself.

我要做些疯狂的事情。

I wanna do something crazy.

这里没人能对我说不。

Nobody is here to tell me no.

这就是为什么最终会做出像Poke这样的产品，而不是在谷歌或Facebook当产品经理。

That's how you end up with a product like Poke, not being a PM at Google or Facebook or wherever else.

确实如此。

Exactly.

你在这些公司的经历可能让你收获了不少宝贵经验。

You probably have some great lessons from being, you know, in all of these companies.

所以，如果你是粉丝，就专注于你擅长的领域。

So, you know, if you're a you're a fan, just focus on what you're good at.

是的。

Yes.

我们为什么不聚焦最后一个趋势——社交媒体平台呢？

Why don't we focus on the last trend, which is around social media platforms?

这个话题虽然已经讨论过一些，但我们确实没看到真正的新社交媒体平台出现。

You know, I think this has been talked about a little bit, but we haven't seen any new social media platforms really.

Substack可能算一个。

Substack is probably one.

我知道你们和投资方对Substack都非常满意。

I mean, I know you guys and investor were very happy on on Substack.

很好。

Good.

你觉得我们还会看到更多这类平台出现吗？

You know, are we gonna see more of those types of platforms?

为什么过去五六年都没有新的社交媒体平台崛起？

Why haven't we seen the new social media platform emerge in the last five, six years ish?

是的。

Yeah.

这是个很棒的话题。

It's a great topic.

我的意思是，我们总是习惯参照上一个周期，并假设下一个周期的成功模式会与上一个周期相似。

I mean, I again, I I think we're so trained to look at the last cycle and assume that what success will look like in the next cycle will be similar to the last cycle.

所以大家都在寻找类似新一代Instagram会像上一代Instagram那样的东西，但我认为它会完全不同。

So everybody is looking for you know, the next gen Instagram will look like the last gen Instagram, and I think it'll be totally different.

所以当我再次审视像Wabi这样的产品时，我会说，看。

So when I look, again, at a product like a Wabi and I say, look.

要知道，这可能就是下一代AI原生网络的模样。

You know, maybe that's what the next AI native network looks like.

在这种网络中，地位游戏将聚焦于人们创造出最酷、最有趣的软件，而非最酷、最有趣的内容。

Something where the status game is focused on people making the coolest, most interesting software instead of the coolest, most interesting content.

我认为另一个问题在于：什么是与新技术原生的媒体形式？

I think another question is just around, you know, what is a media format that's native to the new technology?

视频、照片等所有现有形式，都是移动互联网原生的产物，这很棒。

So, you know, video, photograph, all this other stuff, that's all native to mobile, and that's great.

但我认为AI原生的社交网络不会基于上一代的媒体形式。

But I just don't know that the AI native social network will be based around media from the last generation.

新一代的媒体将是模型、软件这类事物。

The media of the new generation is things like models, things like software.

所以我确实认为存在新社交产品或广义上新网络的发展空间。

So I I do think there's space for a new social product or a new network broadly.

只是它看起来会截然不同。

I just think it's going to look so different.

它不会是Instagram的简单升级版。

It's not gonna be Instagram plus plus.

从定义上来说就不可能是，因为Instagram自己会做那个升级版。

It just sort of definitionally can't be because Insta's gonna build that.

我认为目前对AI的主要批评是，嘿。

I think the critique of a lot of AI so far has been, hey.

已经有很多工具了，但网络平台还不多。

There's been a lot of a lot of tools, not a lot of networks.

不过，再回到你说的千日理论，如果对比移动互联网的发展阶段，我们现在处于什么位置？

But, again, to your point about a thousand days, like, where are we if we overlay where we are today in comparison to mobile?

我们相当于2011年。

You know, we're 2011.

2011年时Airbnb、Uber、WhatsApp都还没形成规模。

2011 was really before Airbnb, Uber, WhatsApp, everything got to scale.

所以现在还非常非常早期，很难预测最终会是什么样子。

So it's still very, very early days, and it's it's hard to predict what it's gonna look like.

你认为新的社交媒体平台会聚焦在哪些方面？

Where do you think the new social media platforms focus?

这些平台会是高度垂直化的语音类平台吗？

Are these gonna be highly focused type of, you know, voice class types of platforms?

它们会完全向所有人开放吗？

Are they gonna be completely open to everyone?

当我想到几个新兴平台时，确实存在一些非常优秀的健康养生类平台。

You know, when I think about a couple of Wish of Emerge, there are a couple of, you know, very good health and wellness type of platforms.

看看类似Time Left这样的平台，虽然不算严格意义上的社交平台，但更像是为了解决孤独感而存在。

You look at the you look at some of the likes of Time Left, which is not really a social platform, but it's more around kind of, like, solving loneliness.

不知道你是否了解TimeLabs，它让你可以和陌生人共进晚餐。

I don't know if you know TimeLabs, but you can have dinner with strangers.

这些平台大约两年前才出现。

These guys are two years ago.

我几个月前采访过他们的创始人。

I interviewed the founder a few months ago.

他们仅用两年时间通过连接陌生人就实现了1600万美元的净收入。

$16,000,000 in net revenue they make in just two years by connecting strangers together.

这确实相当有趣。

That's something quite interesting.

但我们可能不会看到另一个Instagram或Facebook级别的平台崛起了。

But when are we gonna we're probably not gonna see another Instagram, Facebook, possibly emerge.

我不知道。

I don't know.

我是说，Sora是个有趣的尝试。

I mean, Sora was an interesting effort.

对吧？

Right?

也许Sora会成功。

And maybe Sora will work.

Sora最酷的地方在于——我注意到的是，首先它的社交地位游戏设计就与众不同。

Like, the cool thing about Sora, what I paid attention to was, one, the status game was very different.

要知道，我认为在Acts上的地位游戏就是看谁最聪明，或者可能是那种发垃圾帖的行为。

You know, I think the status game on on acts is be the smartest or, you know, maybe it maybe shit posting something like that.

Instagram上的地位游戏，大概是看谁最火辣吧。

The status game on Instagram is, like, be the hottest, I guess.

Sora上的地位游戏似乎聚焦于喜剧效果，看谁最搞笑。

The status game on Sora seemed like it was focused on comedy, be the funniest.

比如，看谁能生成最搞笑、最奇怪的内容？

Like, who could generate the funniest, strangest things?

我朋友Kevin Rose做了些超搞笑的视频，效果非常好。

My friend Kevin Rose did some hilarious videos that did really well.

所以第一，这是个不同的地位游戏。

So one, it was a different status game.

第二，创作者的数量——

Two, the the number of people creating.

这并非典型的90-10或99-1分布。

It wasn't a typical ninety ten, ninety nine one.

实际上想想，大多数人都至少创作了一个视频，这与当今大多数社交产品截然不同。

Think, actually, a majority of people were creating at least one video, which is super different from most social products today.

所以这些尝试非常非常有趣，当它们成功时，其本质看起来与过去的社交产品大不相同。

So these efforts are are very, very interesting, and the kind of nature of them when they work look really different from the past social products.

所以，谁知道会发生什么，但我想我们一年后会再讨论这个话题。

So, you know, who knows what'll happen, but I think we'll we'll have a conversation a year.

我认为我们将看到一些新型原生社交网络的雏形。

I think we'll see the beginnings of some new native networks.

太棒了。

Awesome.

凯文·罗是另一位很厉害的企业家。

Kevin Rowe is another, like, awesome entrepreneur.

我想你们现在交情很深了。

I think you guys go way back now.

是的。

Yes.

是啊。

Yeah.

没错。

Yeah.

对。

Yeah.

我们早在谷歌时期就认识了。

We've we've known each other since our our days at Google.

是的。

Yeah.

他也是个非常了不起的人。

He's he's and he's a tremendous person as well.

他思想非常开放。

He's very open minded.

他之所以能如此准确地预测未来趋势，正是因为他不会试图用先入为主的观念来框定世界。

He's so much of his ability to predict what comes next is that he doesn't have preconceived notions that he's trying to back the world into.

他保持着初学者的心态，而且，你知道，他有几次都早早预测正确。

He's just got the beginner's mind, and, you know, and he's been early and right a few times.

是啊。

Yeah.

看，我们为所有这些节目都尽量准备充分。

See, we come try and come well prepared for all of these shows.

我喜欢这样。

I love it.

我喜欢这样，阿里。

I love it, Ali.

你功课做得很到位。

You've done your work.

就是这样。

There you go.

对吧？

Right?

还剩几分钟。

A few minutes left.

我们来聊聊创始人的融资问题。

Let's talk about fundraising for founders.

我认为很多创始人会有些创始人会非常焦虑。

I think a lot of founders will be some founders will be freaking out.

我必须尽可能快地筹集尽可能多的资金。

I have to raise as much as I can as quickly as I can.

有些人则会非常冷静，你知道，对此很放松。

Some will be very chilled and quite, you know, relaxed about it.

创始人现在应该如何看待融资问题？

How should founders be thinking about fundraising today?

他们应该专注于尽可能多地筹集资金，还是应该真正专注于筹集实际可行的轮次？

Should they be focused on raising as much as they can or just be really focused on raising realistic rounds?

是的。

Yeah.

这是个非常复杂的问题。

It's it's such a complicated question.

所以，如果你回归第一性原理思考，为什么创始人不应该在创业第一天就筹集公司整个生命周期所需的所有资金？

So, you know, if you think back to first principles, you know, why why is it not optimal for founders to raise all the money they need in the lifetime of a company on day zero?

如果你是一位创始人，计划先融资500万，再2000万，然后5000万等等，为什么不干脆在第一天就融1.25亿呢？

So if you're a founder and you're gonna raise, you know, five and then 20 and then 50, etcetera, why don't you just raise a 125 on day zero?

这个做法行不通的原因其实很有意思。

The reason that that doesn't work well is actually interesting.

问题不在于你无法获得1.25亿美元的资金。

It's not it's not because you can't get a 125,000,000.

关键在于当创始人拥有过多资金时会发生什么——我自己也经历过这种情况——通常你会把稀缺资源（不是资金，而是人才）分散到太多项目上。

It's because what happens when founders have too much money, and and I've been through this myself, is you typically take your scarce resource, which is not money, which is talent, and you spread it across too many efforts.

所以你真正需要的是人才集中，并在特定时间内只专注于一件事。

So what you really need is concentration of talent and focus on just one thing at a given time.

而筹集适当数额的资金恰恰能推动这种集中。

And raising the right amount of money sort of drives that concentration.

每个人都这么想，我当创始人时也这么想过，比如，嘿。

And everybody thinks I thought it myself when I was a founder, like, hey.

我不会分心的。

I won't get distracted.

我会优先处理，但你就是做不到。

I'll prioritize, but you just don't.

要同时不做很多事情是很难的。

It's hard to, like, not not do many things at once.

所以我认为筹集适量资金非常重要，因为它能强制形成你原本不会有的纪律性。

So I think raising the right amount is really important because it forces a discipline you wouldn't otherwise have.

我认为对于竞争激烈的市场，多筹集一些资金总是有道理的，因为你想保持竞争力。

And I think for for markets that are highly competed, raising a little bit more always makes sense because you wanna sort of keep up.

对于竞争较少的市场，比如我认为Shopify就是个很好的例子，它是在相对默默无闻中建立起来并发展到巨大规模的，就不需要筹集那么多资金。

I think for markets that are less competed, like, I think Shopify is a great example of just something that was built in relative obscurity and compounded to enormous scale, you don't have to raise as much.

但我认为现在的神奇之处在于，你可以用产品而非营销资金来引领市场。

But I think that the magic right now is that you can lead with products instead of marketing dollars.

所以我们所处的2021年的世界是这样的，听着。

So the world we were in in 2021, which is, look.

我需要筹集1亿美元，然后在谷歌和脸书上花掉1亿美元。

I to raise a 100,000,000 to spend a 100,000,000 on Google and Facebook.

那样的世界已经不复存在了。

That world doesn't exist anymore.

如果你需要筹集1亿美元来训练更好的模型或打造更好的产品，那很好。

It's like if you have to spend raise a 100,000,000 to train a better model or build a better product, great.

因为我认为这正是投资者愿意投入、创始人希望筹集的那种风险资本，与三四年前的情况截然不同。

Because that's exactly, I think, the kind of, you know, risk capital that investors wanna invest and founders wanna raise as opposed to where we were three or four years ago.

我知道这不是一个完美的答案，但或许最实际的做法是以你能获得的最佳条件筹集24个月的资金，但价格不要高到让下一轮融资变得困难。

So I I know that's not a perfect answer, but I think, you know, maybe the brass tacks is raised for twenty four months at the best terms you can, but not, you know, such a high price that it makes the next round difficult.

完全正确。

Totally right.

而且，当创始人在考虑如何让融资过程顺利进行时，是的。

And, also, you know, when founders are thinking about how to run a super smooth fundraising process Yeah.

你觉得最佳方法是什么？

What have you found is the best approach?

是集中两周时间尽可能多见投资人、获取多份投资意向书，还是分散在一个月内零零散散地进行？

Is it just to focus two weeks going meet as many investors as you can and get as many term sheets, or is it a very scatty end over a month?

我认为有两方面。

I think it's two things.

我觉得，随着时间的推移建立关系很重要，因为我常说当你需要信任时，再去建立就为时已晚了。

I think, you know, over the course of time building relationships, because I always say when you need trust, it's too late to build it.

所以我认为至少与你可能想要融资的对象见过一面，或许每季度喝一次咖啡，持续一年，这不是个坏主意。

So I think having met people that you might wanna raise from at least once, maybe having had a coffee, you know, once a quarter over the course of a year is not a bad idea.

然后，当你准备好融资时，你应该计划两周内不做其他任何事情。

And then, yeah, when you're ready to fundraise, you should plan to get nothing else done for two weeks.

专心致志。

Just put your head down.

让我们全力以赴去做这件事。

Let's go do the work.

让我们去和所有已经建立关系的人谈谈。

Let's go talk to all the folks that we already have relationships with.

然后，听着，我认为如果你在前两三次会议中得到的反响不温不火，那是一个非常非常重要的信号。

And then, look, I think that if you're getting lukewarm reception in the first two or three meetings, you that's, like, a very, very important signal.

我觉得那种关于投资者更倾向于从众依赖的说法对创业者来说是不健康的。

I I think that the the sort of narrative around investors being more circularly dependent is an unhealthy one for founders.

我会直接从第一性原理出发。

I would just start from first principles.

你有一个很棒的产品。

You have a great product.

你有支持你假设的数据指标吗？

Do you have metrics to support your assumptions?

然后就是与投资者的对话——你们之间的氛围是融洽，还是冷淡？

And then is the investor conversation you know, you are the vibes good, or are they lukewarm?

如果反应冷淡，可能还需要做些工作。

And if they're lukewarm, there might be some more work to do.

我认为投资者除了给你钱之外，最有用的就是给你实话实说。

I think the most useful thing an investor can do other than giving you money is just giving you the real talk.

所以有时候你得稍微‘折磨’一下投资者，特别是那些自己没创过业的人。

So think sometimes you've gotta torture the investors a little bit, especially if they haven't been a founder themselves.

如果他们创过业，通常就会明白，你看——

If they have, then they often know that, like, look.

虽然我没法给你投资，但让我告诉你问题出在哪。

Like, I can't give you the money, but, like, let me tell you what's up.

让我提醒你鞋底粘了口香糖，你可以先处理掉，咱们三个月后再谈。

Let me tell you, like, that you've got gum on your shoe so you can go take that off, and let's have another conversation in three months.

我喜欢这个说法。

I love it.

你接触过这么多创业者，肯定积累了不少绝妙点子吧。

So, you know, you've seen and met many founders, and you probably got so many cool ideas.

说不定哪天你自己也想创业呢。

You would like to start maybe one day.

如果你现在不做投资，你会立刻创办什么样的公司？

If you weren't investing today, what company would you build right now?

是啊。

Yeah.

我不知道。

I don't know.

我的意思是，我肯定不会选择去大科技公司工作或做规模化的事情。

I mean, the one thing I would not do is is sort of go work at big tech or scale.

你懂吗？

You know?

因为现在感觉正是处于前沿领域的最佳时机。

Because it it just feels like right now is the moment to be at the edge.

当然，大公司里可能也有一些非常令人兴奋但并非前沿的团队，那属于例外情况。

So and maybe there are teams within big tech that are really exciting that aren't at the edge, and that's the exception.

但我总觉得你应该致力于最重要的事情，那些你能做出最大贡献的领域。

But I just feel like you kinda wanna be working on the most important thing, the thing that you can contribute the most to.

而对我来说，我认为那会是成为一名创始人，或者在大科技公司内部或最重要的实验室之一从事前沿工作。

And I think for me, that would be being a founder or sort of working at the edge either within big tech or one of the most important labs.

我个人投入最多精力的领域是AI编程。

The area that I'm personally spending the most energy on is AI code.

我之所以如此兴奋，主要是因为我一直是个工程师，但水平一般，而现在多亏了这些工具和模型，我的水平有所提升。

I just like I'm so fired up mostly because I've always been a I I'm an engineer, but a mediocre one, and and now I'm a a less mediocre one thanks to the the tools and the models.

所以我可能会梦想着在这方面有所建树。

So I'd probably be dreaming of building on something like that.

太棒了。

Awesome.

嘿，

Well, hey.

这次对话真是太有趣了，充满了深刻的见解和丰富的信息。

This has been so much fun, packed with insights, so much information.

非常感谢你抽出时间，我们十二个月后再聚，看看AI已经发展到什么程度了。

Thank you so much for your time, and let's come back in twelve months and see just how far AI has progressed.

不是十二个月

That's it's it's not twelve months.

是又一轮三百六十五天

It's another three sixty five.

对吧，亚历克？

Right, Alec?

就是这样

There you go.

对吧？

Right?

这就是我们的计划

That's our plan.

感谢邀请我

Thank you for having me.

谢谢提供这个平台，期待我们下一次的对话

Thanks for providing the platform, and I'm excited for our next conversation.

我们开始吧。

Let's do it.

感谢收听a16z播客。

Thanks for listening to the a 16 z podcast.

如果你喜欢这期节目，请到ratethispodcast.com/a16z留下评价让我们知道。

If you enjoyed the episode, let us know by leaving a review at ratethispodcast.com/a16z.

我们还有更多精彩对话即将呈现。

We've got more great conversations coming your way.

下次见。

See you next time.

本信息仅供教育用途，不构成对任何投资或金融产品的购买、持有或出售推荐。

This information is for educational purposes only and is not a recommendation to buy, hold, or sell any investment or financial product.

本播客由第三方制作，可能包含付费推广广告、其他公司提及以及与a16z无关联的个人内容。

This podcast has been produced by a third party and may include paid promotional advertisements, other company references, individuals and unaffiliated with A16Z.

此类广告、公司及个人均未获得Ah Capital Management LLC、a16z或其任何关联方的认可。

Such advertisements, companies, and individuals are not endorsed by Ah Capital Management LLC, A16Z, or any of its affiliates.