人工智能之母谈工作、机器人与为何世界模型是下一个前沿 | 费菲菲李博士

不能。

can No.

确实没有。

We're not.

事实上情况更糟。

Fact, it's even worse.

让我们给AI所有数据，包括牛顿时代没有的现代天体仪器数据，然后要求AI创建十七世纪第六套关于物体运动定律的方程组。

Let's give AI all the data, including modern instruments data of celestial bodies, which Newton did not have, and give it to them and just ask AI to create the sixth seventeenth century set of equations on the laws of bodily movements.

现在的AI还做不到这一点。

Today's AI cannot do that.

好吧。

All right.

我们还有很长的路要走，这就是我的

We're ways away is what I'm

看法。

hearing.

好的。

Okay.

那我们谈谈世界模型吧。

So let's talk about world models.

在我看来，这又是一个你领先于大众的绝佳例证。

This is to me, this is just another really amazing example of you being ahead of where people end up.

你早就提出过，我们只需要大量干净的数据供AI和神经网络学习。

So you were way ahead on, okay, we just need a lot of clean data for AI and neural networks to learn.

你谈论世界模型这个概念已经很久了。

You've been talking about this idea of world models for a long time.

你还为此创办了一家公司。

You started a company to build.

本质上，现在有的是语言模型。

Essentially, there's language models.

这是完全不同的东西。

This is a different thing.

这是一个世界模型。

This is a world model.

我们会讨论它是什么。

We'll talk about what that is.

就在我准备这个的时候，埃隆也在谈论世界模型。

And now, as I was preparing for this, Elon's like talking about world models.

黄仁勋也在谈论世界模型。

Jensen's talking about world models.

我知道谷歌正在研究这些东西。

I know Google's working on this stuff.

你们已经研究这个很久了，而且实际上刚刚发布了一些内容，我们正好要在这期播客播出前讨论。

You've been at this for a long time, and you actually just launched something that's going to we're going to talk about right before this podcast airs.

谈谈什么是世界模型？

Talk about what is a world model?

为什么它如此重要？

Why is it so important?

看到越来越多人像埃隆、黄仁勋那样讨论世界模型，我非常兴奋。

I'm very excited to see that more and more people are talking about world models like Elon, like Jensen.

我一生都在思考如何真正推动人工智能发展。

I have been thinking about really how to push AI forward all my life.

对吧？

Right?

过去几年从研究界诞生的大型语言模型，以及OpenAI等等，都极具启发性，即使对我这样的研究者来说也是如此。

And the large language models that came out of the research world and then OpenAI and and all this for the past few years were extremely inspiring, even for researcher like me.

我记得GPT-2发布的时候，大概是在2020年底。

I remembered when GPT two came out, and that was in, I think, late two thousand twenty.

我当时是联合主任。

I was co director.

现在依然是，但那时我是斯坦福以人为本人工智能研究所的全职联合主任。

I still am, but I was at that time full time co director of Stanford's Human Centered AI Institute.

我记得当时公众还不了解大语言模型的能力，但作为研究者，我们已经看到了。

And I I remember it was you know, the public was not aware of the power of the large language model yet, but as researchers, we were seeing it.

我们看到了未来。

We're seeing the future.

我和自然语言处理领域的同事如Percy Liang、Chris Batting进行了相当长时间的讨论。

And I had pretty long conversations with my natural language processing colleagues like Percy Liang and Chris Batting.

我们讨论这项技术将变得多么关键。

We were talking about how critical this technology is going to be.

斯坦福人工智能研究所（以人为本人工智能研究所HAI）是第一个建立基础模型研究中心的机构。

And the Stanford AI Institute, Human Centered AI Institute, HAI, was the first one to establish a full research center foundation model.

Percy Liang和许多研究人员领导发表了第一篇关于基础模型的学术论文。

We were Percy Liang and many researchers led the first academic paper foundation model.

这让我深受启发。

So it was just very inspiring for me.

当然，我来自视觉智能领域，当时就在思考我们可以在语言之外推进多少事情，因为人类运用空间智能和对世界的理解完成了许多超越语言范畴的事情。

So, of course, I come from the world of visual intelligence, and I was just thinking there's so much we can push forward beyond language because humans humans have used our sense of spatial intelligence, a world understanding to do so many things, and they are beyond language.

想象一个非常混乱的急救现场，无论是火灾、交通事故还是自然灾害。

Think about a very chaotic first responder scene, whether it's fire or some traffic accident or some natural disaster.

当你沉浸于场景中，思考人们如何组织救援、阻止灾难蔓延、扑灭火灾时，会发现大部分行动都源于对物体、环境和态势的本能理解与空间感知。

And if you immerse yourself in the scene and think about how people organize themselves to rescue people, to stop further disasters, to put down fires, to a lot of that is movements, spontaneous understanding of objects, worlds, situational awareness.

语言虽是其中一环，但在许多紧急情况下，语言无法帮你扑灭火焰。

Language is part of that, but a lot of those situations, language cannot get you to put on the fire.

那么这究竟意味着什么？

So that is what is that?

我思考了很多。

I was thinking a lot.

与此同时，我进行了大量机器人研究，并逐渐意识到：连接语言之外附加智能的关键，在于将具身AI（即机器人）、视觉智能与理解世界的空间智能相结合。

And in the meantime, was doing a lot of robotics research and it dawned on me that the linchpin of connecting the additional intelligence in addition to language and connecting embodied AI, which are robotics, connecting visual intelligence is the sense of spatial intelligence about understanding the world.

就在那时——我记得是2024年——我在TED演讲中探讨了空间智能与世界模型。

And that's when, I think I, it was 2024, I gave a TED talk about spatial intelligence and world models.

这个构想最早可追溯至2022年，基于我的机器人与计算机视觉研究逐渐成形。

And I start formulating this idea back in 2022 based on my robotics and computer vision research.

我明确意识到：必须集结最杰出的技术专家，以最快速度将这项技术变为现实。

And then one thing that was really clear to me is that I really want to work with the brightest technologists and move as fast as possible to bring this technology to life.

于是我们创立了名为World Labs的公司。

And that's when we founded this company called World Labs.

公司名称中的'World'一词正体现了我们对世界建模与空间智能的坚定信念。

And you can see the word world is in the title of our company because we believe so much in world modeling and spatial intelligence.

人们已习惯将聊天机器人简单等同于大语言模型。

People are so used to just chatbots, and that's a large language model.

理解世界模型的简易方式是：你描述一个场景，它就能生成无限可探索的虚拟世界。

A simple way to understand a world model is you basically describe a scene and it generates an infinitely explorable world.

我们会链接到你发布的内容，稍后会讨论，但这是否是一种简单的理解方式？

We'll link to the thing you launched, which we'll talk about, but just is that a simple way to understand it?

这是其中一部分，Lenny。

That's part of it, Lenny.

我认为理解世界模型的一种简单方式是：这个模型能让任何人通过提示（无论是图像还是句子）在脑海中创造任何世界，并能在这个世界中互动——无论是浏览行走、拾取物品还是改变事物，同时还能在这个世界中进行推理。

I think a simple way to understand a world model is that this model can allow anyone to create any worlds in their mind's eye by prompting, whether it's an image or a sentence, and also be able to interact in this world, whether you're browsing and walking or picking objects up or changing things, as well as to reason within this world.

例如，如果使用这个世界模型输出的代理是个机器人，它应该能够规划路径并帮助整理厨房等任务。

For example, if if the person consuming if the agent consuming this output of the world model is a robot, It should be able to plan its path and help to tidy the kitchen, for example.

因此世界模型是一个基础，你可以用它来推理、互动和创造世界。

So world model is a foundation that that you can use to reason, to interact, and to create worlds.

很棒。

Great.

是的。

Yeah.

所以机器人领域感觉可能是AI研究者的下一个重点方向，就像对世界的影响一样。

So robots feels like that's potentially the next big focus for AI researchers and just like the impact on the world.

你这里说的是，这是让机器人真正在现实世界中工作的关键缺失环节——理解世界如何运作。

What you're saying here is, this is a key missing piece of making robots actually work in the real world, understanding how the world works.

是的。

Yeah.

首先，我认为这不仅仅局限于机器人领域。

Well, first of all, I do think there's more than robots.

这很令人兴奋。

That's exciting.

不过，我完全同意你刚才说的所有观点。

So but I agree with everything you just said.

我认为世界建模和空间智能是具身人工智能缺失的关键部分。

I think world modeling and spatial intelligence is a key missing piece of embodied AI.

我还认为，我们不应低估人类作为具身主体的能力，以及人类如何能被人工智能的智能所增强。

I also think let's not underestimate that humans are embodied agents and humans can be augmented by AI's intelligence.

就像今天，人类虽然是语言动物，但在执行语言任务（包括软件工程）时，我们很大程度上被人工智能增强了能力。

Just like today, humans are language animals, but we're very much augmented by AI when helping us to, you know, do language tasks, including software engineering.

我认为我们不应低估——或者说我们往往忽略了——人类作为具身主体实际上能从世界模型和空间智能模型中获益良多，机器人同样可以。

I think that we shouldn't underestimate, or maybe we tend not to talk about how humans as an embodied agent can actually benefit so much from world models and spatial intelligent models, as well as robots can.

所以这里的关键突破在于机器人技术，这可是大事。

So the big unlocks here, robots, which a huge deal.

如果这能实现，想象一下我们每个人都能拥有机器人帮我们处理各种事务，比如在灾难中提供援助等等。

If this works out, imagine each of Us has robots doing a bunch of stuff for us, goes into, you know, they help us with disasters, things like that.

游戏显然是个很酷的例子，就像那些能无限畅玩、完全由你想象创造的游戏。

Games, obviously, is a really cool example, just like infinitely playable games that you just invent out of your head.

还有创意领域，感觉就像纯粹为了乐趣，发挥创意，构想各种奇妙的新世界和环境。

And then creativity feels like just like being fun, having fun, being creative, thinking of wild new worlds and environments.

当然还有设计领域。

And also design.

人类设计从机器到建筑到住宅，还包括科学发现，对吧？

Humans design from machines to buildings to homes and also scientific discovery, right?

可能性实在太多了。

There is so much.

我喜欢用DNA结构发现的例子来说明。

I like to use the example of the discovery of the structure of DNA.

如果你观察DNA发现史上最重要的部分之一，就是罗莎琳德·富兰克林拍摄的X射线衍射照片。

If you look at one of the most important piece in DNA's discovery history is the X-ray diffraction photo that was captured by Rosalind Franklin.

那是一张二维平面照片，结构看起来像一个带有衍射图案的十字架。

And it was a flat two d photo of a structure that looks like it looks like a cross with diffractions.

你可以谷歌搜索这些照片。

You can Google those photos.

但凭借那张二维平面照片，人类——特别是两位重要人物詹姆斯·沃森和弗朗西斯·克里克，结合其他信息，能够在三维空间中进行推理，推导出DNA高度立体的双螺旋结构。

But with that two d flat photo, humans, especially two important humans, James Watson and Francis Crick, in addition to their other information, was able to reason in three d space and deduce a highly three-dimensional double helix structure of the DNA.

而这种结构绝不可能是二维的。

And that structure cannot possibly be two d.

你无法通过二维思维推导出这种结构。

You cannot think in two d and deduce that structure.

你必须运用三维空间思维，利用人类的空间智能。

You have to think in three d spatial, use the human spatial intelligence.

因此我认为即使在科学发现中，空间智能或AI辅助的空间智能也至关重要。

So I think even in scientific discovery, spatial intelligence or AI assisted spatial intelligence is critical.

这让我想起克里斯·迪克森说过的话：下一件大事最初总会让人觉得像个玩具。

This is such an example of, I think it was Chris Dixon that had this line that the next big thing is gonna start off feeling like a toy.

当ChatGPT刚推出时，我记得萨尔·莫明发推说：'我们在玩一个很酷的东西，来看看'。

When ChatGPT just came out of like, I remember Sal Momin just tweeted, it's like, here's a cool thing we're playing with, check it out.

现在它已成为史上增长最快的产品，改变了世界。

Now it's the fastest growing product in all of history, changed the world.

是啊。

Yeah.

往往正是那些看起来‘嗯，这个挺酷’、玩起来有趣的东西，最终最能改变世界。

And it's oftentimes the things that just look like, okay, this is cool, that it's fun to play with and end up changing the world most.

没错。

Yeah.

本节目由客户通讯云平台Cinch赞助播出。

This episode is brought to you by Cinch, the customer communications cloud.

关于数字化客户通讯，有件事你需要知道。

Here's the thing about digital customer communications.

无论是发送营销活动、验证码还是账户提醒，你都需要确保信息能可靠触达用户。

Whether you're sending marketing campaigns, verification codes, or account alerts, you need them to reach users reliably.

这就是Cinch的用武之地。

That's where Cinch comes in.

全球超过15万家企业（包括十大科技巨头中的八家）使用Cinch的API，将消息、邮件和通话功能集成到产品中。

Over 150,000 businesses, including eight of the top 10 largest tech companies globally, use Cinch's API to build messaging, email, and calling into their products.

产品团队需要关注通讯领域的重要变革：富媒体通讯服务（RCS）。

And there's something big happening in messaging that product teams need to know about: rich communication services, or RCS.

你可以把RCS理解为短信2.0版本。

Think of RCS as SMS two point zero.

用户不再收到来自陌生号码的短信，而是直接看到经过认证的企业名称和标识——无需下载任何新应用。

Instead of getting texts from a random number, your users will see your verified company name and logo without needing to download anything new.

这种体验更安全且品牌感更强，还支持交互式轮播图和智能回复建议等功能。

It's a more secure and branded experience, plus you get features like interactive carousels and suggested replies.

这就是为什么这很重要。

And here's why this matters.

美国运营商开始采用RCS。

US carriers are starting to adopt RCS.

Cynch已在帮助各大品牌全球发送RCS消息，并让Lenny的播客听众在美国市场热潮前优先注册。

Cynch is already helping major brands send RCS messages around the world, and they're helping Lenny's podcast listeners get registered first before the rush hits The US market.

了解更多并开始使用请访问cinch.com/lenny。

Learn more and get started at cinch.com/lenny.

网址是sinch.com/lenny。

That's sinch.com/lenny.

我联系了本·霍洛维茨，他非常欣赏你的事业，是你的忠实粉丝。

I reached out to Ben Horowitz who loves what you're doing, a big fan of yours.

他们是投资人，我很信任他们。

They're investors, I believe in.

是的。

Yeah.

我们认识很多年了。

We we've known each other for for many years.

不过现在他们确实是Warlabs的投资人。

But, yes, right now, they are investors of Warlabs.

太棒了。

Amazing.

好的。

Okay.

于是我问他应该问你些什么，他建议我问你，为什么苦涩的教训单靠它自己不太可能对机器人奏效？

So I asked him what I should ask you about, and he suggested ask you, why is the bitter why is the bitter lesson alone not likely to work for robots?

首先，请解释一下AI历史上的苦涩教训是什么，然后说明为什么这无法让我们在机器人领域达到预期目标。

So first of all, just explain what the bitter lesson was in the history of AI, and then just why that won't get us to where we want to be with robots.

首先，其实存在很多苦涩的教训。

So, well, first of all, there are many bitter lessons.

但大家常提到的苦涩教训，是最近获得图灵奖的理查德·萨顿写的一篇论文。

But the bitter lessons everybody refers to is a paper written by Richard Sutton, who won the Turing Award recently.

他主要从事强化学习研究。

And he does a lot of reinforcement learning.

理查德说过，纵观历史，特别是AI算法发展历程，最终胜出的总是拥有海量数据的简单模型，而非数据量较少的复杂模型。

And Richard has said, right, if you look at the history, especially the algorithmic development of AI, it turns out simpler model with a ton of data always win at the end of the day instead of the more complex model with less data.

其实这篇论文是在ImageNet问世多年后才发表的。

I mean, that was actually This paper came years after ImageNet.

对我来说那并不苦涩。

That to me was not bitter.

那是个甜蜜的教训。

It was a sweet lesson.

正因如此我才创建了ImageNet，因为我坚信大数据能发挥这种作用。

That's why I built ImageNet, because I believe that big data plays that role.

那么为什么苦涩的教训无法单独在机器人领域见效？

So why can a bitter lesson work in robotics alone?

首先我认为需要肯定我们当前的成就。

Well, first of all, I think we need to give credit to where we are today.

机器人技术仍处于实验的早期阶段。

Robotics is very much in the early days of experimentation.

这项研究远不如语言模型那样成熟。

The research is not nearly as mature as, say, language models.

许多人仍在尝试不同的算法，其中部分算法由大数据驱动。

So many people are still experimenting with different algorithms, and some of those algorithms are driven by big data.

因此我认为大数据将继续在机器人技术中发挥作用。

So I do think big data will continue to play a role in robotics.

但机器人技术的难点是什么？

But what is hard for robotics?

主要有几个方面。

There are a couple of things.

一是数据获取更为困难。

One is that it's harder to get data.

数据获取要困难得多。

It's a lot harder to get data.

你可能会说，网上有现成的数据。

You can say, well, there's web data.

这正是当前机器人研究使用网络视频的原因。

This is where the latest robotics research is using web videos.

我认为网络视频确实发挥了作用。

And I think web videos do play a role.

但作为从事计算机视觉、空间智能和机器人技术的研究者，当我思考语言模型成功的原因时，实在很羡慕语言领域的同行——他们拥有完美的数据闭环：训练数据是文字（最终转化为词元），模型输出的也是文字。

But if you think about what made language model worth a very as someone who does computer vision and spatial intelligence and robotics, I'm very jealous of my colleagues language because they had this perfect setup where their training data are in words, eventually tokens, and then they produce a model that outputs words.

因此，你期望获得的结果（我们称之为目标函数）与训练数据之间存在着完美的匹配关系。

So you have this perfect alignment between what you hope to get, which we call objective function, and what your training data looks like.

但机器人技术则截然不同。

But robotics is different.

甚至空间智能也大相径庭。

Even spatial intelligence is different.

你希望机器人输出动作，但训练数据中缺乏三维世界里的动作样本。

You hope to get actions out of robots, but your training data lacks actions in three d worlds.

而这正是机器人必须实现的功能，对吧？

And that's what robots have to do, right?

在三维世界里执行动作。

Actions in three d worlds.

所以你必须另辟蹊径，用他们常说的'方枘圆凿'的方式解决问题。

So you have to find different ways to fit a what do they call a a a a square in a round hole.

我们拥有的是海量的网络视频资源。

What we have is tons of web videos.

因此我们开始讨论补充数据，比如遥操作数据或合成数据，让机器人通过大量数据来验证'苦涩教训'假说。

So then we have to start talking about adding, supplementing data such as teleoperation data or synthetic data so that the robots are trained with this hypothesis of bitter lesson, which is large amount of data.

我认为仍有希望，因为我们在世界建模方面的进展终将为机器人解锁这些信息。

I think there is still hope because even what we are doing in world modeling will really unlock a lot of this information for robots.

但我们必须保持谨慎，毕竟这项研究刚起步，'苦涩教训'尚未完全验证，因为我们还未彻底解决数据问题。

But I think we have to be careful because we're at the early days of this, and bitter lesson is still to be tested because we haven't fully figured out the data for.

关于机器人学'苦涩教训'的另一个现实是——相较于语言模型甚至空间模型，机器人本质上是物理系统。

Another part of the bitter lesson of robotics I think we should be so realistic about is, again, compared to language models or even spatial models, robots are physical systems.

因此，机器人更接近自动驾驶汽车，而非大型语言模型。

So robots are closer to self driving cars than a large language model.

认识到这一点非常重要。

And that's very important to recognize.

这意味着要让机器人工作，我们不仅需要大脑，还需要物理躯体。

That means that in order for robots to work, we not only need brains, we also need the physical body.

我们还需要应用场景。

We also need application scenarios.

如果你回顾自动驾驶汽车的历史，我的同事塞巴斯蒂安·特龙在2005或2006年用斯坦福的汽车赢得了第一届DARPA挑战赛。

If you look at the history of self driving car, my colleague, Sebastian Thrum, took Stanford's car to win the first DARPA challenge in 2006 or 2005.

从那个自动驾驶汽车原型至今已有二十年。

It's twenty years since that prototype of a self driving car.

从能在内华达沙漠行驶130英里，到今天Waymo出现在旧金山街头。

Being able to drive 130 miles in the Nevada Desert to today's Waymo and on the street of San Francisco.

而我们甚至还没有完成。

And we're not even done yet.

仍有很长的路要走。

There's still a lot.

所以这是一段二十年的旅程。

So that's a twenty year journey.

而自动驾驶汽车是简单得多的机器人。

And self driving cars are much simpler robots.

它们只是金属盒子在二维平面上移动，目标是不触碰任何东西。

They're just metal boxes running on two d surfaces, And the goal is not to touch anything.

机器人是在三维世界中运行的三维物体，其目标是接触物体。

Robot is three d things running in three d world, and the goal is to touch things.

因此这段旅程将会，你知道的，包含许多方面和元素。

So the journey is gonna be, you know, there's many aspects, elements.

当然可以说，自动驾驶汽车的早期算法属于深度学习前时代。

And of course, one could say, well, the self driving car early algorithm were pre deep learning era.

所以深度学习正在加速大脑的进化。

So deep learning is accelerating the brains.

我认为确实如此。

And I think that's true.

这就是我投身机器人领域的原因。

That's why I'm in robotics.

这就是我专注空间智能并为此感到兴奋的原因。

That's why I'm in spatial intelligence and I'm excited by it.

但与此同时，汽车工业非常成熟，产品化还涉及成熟用例、供应链和硬件。

But in the meantime, the car industry is very mature and productizing also involves the mature use cases, supply chains, the hardware.

所以我认为现在研究这些问题非常有趣。

So I think it's a very interesting time to work in these problems.

但本说得没错。

But it's true Ben is right.

我们可能仍需经历一些惨痛教训。

We might still be subject to a number of bitter lessons.

从事这项工作时，你是否会对大脑的运作方式及其为我们完成这一切的能力感到敬畏？

Doing this work, do you ever just feel awe for the way the brain works and is able to do all of this for us?

仅仅为了让机器能行走而不撞到东西或摔倒的这种复杂性，是否让你对我们已有的能力更加敬畏？

Just the complexity, to get a machine to just walk around and not hit things and fall, does just give you more respect for what we've already got?

完全同意。

Totally.

我们仅需约20瓦的功率运行。

We operate on about 20 watts.

这比我此刻房间里任何灯泡的亮度都要低。

That's dimmer than any light bulb in the room I'm in right now.

然而我们能做这么多事情。

And yet we can do so much.

所以实际上，我越从事AI工作，就越对人类产生敬意。

So I think actually the more I work in AI, the more I respect humans.

让我们聊聊你们刚推出的产品Marble，这个名字很可爱。

Let's talk about this product you just launched called Marble, a very cute name.

关于它是什么，为什么它很重要。

About what this is, why this is important.

我一直在试用它。

I've been playing with it.

它太不可思议了。

It's incredible.

我们会附上链接方便大家查看。

We'll link to it and for folks to check it out.

Marble是什么？

What is Marble?

是的。

Yeah.

我非常兴奋。

I'm very excited.

首先，Marble是Warlabs推出的首批产品之一。

So first of all, Marble is one of the first product that Warlabs has rolled out.

World Labs是一家前沿基础模型公司。

World Labs is a foundation frontier model company.

我们由四位技术背景深厚的联合创始人资助。

We are funded by four co founders who have deep technical history.

我的联合创始人Justin Johnson、Christoph Lassner和Ben Meldenhall都来自AI、计算机图形学和计算机视觉的研究领域。

My co founders, Justin Johnson, Christoph Lassner, and Ben Meldenhall, we all come from the research field of AI, computer graphics, computer vision.

我们相信空间智能和世界建模至少与语言模型同等重要，并且能与语言模型形成互补。

And we believe that spatial intelligence and world modeling is as important, if not more, to language models and complementary to language models.

因此我们希望抓住这个机会，建立一个能将前沿模型与产品连接起来的深度技术研究实验室。

So we wanted to seize this opportunity to create deep tech research lab that can connect the dots between frontier models with products.

Marble是一款基于我们前沿模型构建的应用。

So Marble is an app that's built upon our frontier models.

我们花了一年多时间构建世界上首个能输出真正三维世界的生成模型。

We've spent a year and plus building the world's first generative model that can output genuinely three d worlds.

这是个非常非常困难的问题。

That's a very, very hard problem.

而且这个过程非常艰难。

And it was a very hard process.

拥有一支非凡的团队，创始团队由来自顶尖团队的技术专家组成。

Have a team of incredible, founding team of incredible technologists from, you know, incredible teams.

就在一两个月前，我们首次见证仅需输入一句话和图像（或多张图像）就能创建出可自由探索的世界。

And then around just a month or two ago, we saw the first time that we we can just prompt with a sentence and the image and multiple images and create worlds that we can just navigate in.

如果使用我们的谷歌眼镜选项（我们提供该功能），你甚至能四处走动。

If you put it on goggle, which we have an option to let you do that, you can even walk around.

对吧？

Right?

尽管我们已开发很长时间，这依然令人震撼不已。

So it was even though we've been building this for quite a while, it was still just awe inspiring.

我们希望能让需要的人用上它。

And we wanted to get into the hands of people who need it.

我们知道众多创作者、设计师、考虑机器人模拟的人、探索可导航交互沉浸世界应用场景的人，以及游戏开发者都会发现它的价值。

And then we know that so many creators, designers, people who are thinking about robotic simulation, people who are thinking about different use cases of navigable, interactable, immersive worlds, game developers will find this useful.

因此我们开发了Marble作为第一步。

So we develop developed Marble as a first step.

虽然仍处于早期阶段，但这是全球首个实现该功能的模型，也是首个支持纯提示生成的产品。

It's it's, again, still very early, but it's the world's first model doing this, and it's the world's first product that allows people to just prompt.

我们称之为‘提示即世界’。

We call it prompt to worlds.

我一直在试用。

Well, I've been playing around.

这太疯狂了。

It is insane.

就像，你可以拥有一个更羞涩的世界，在那里你基本上可以无限漫步在中土世界，虽然那里还没有人，但这太疯狂了。

Like, you could just have a little shyer world where you just infinitely walk around Middle Earth, basically, and there's no there's no one there yet, but it's insane.

你可以去任何地方。

You just go anywhere.

这就像是个反乌托邦世界。

There's like dystopian world.

我只是在浏览所有这些例子。

I'm just looking at all these examples.

是的。

Yes.

其实我最喜欢的部分是，我不知道。

And my favorite part actually, I don't know.

我不确定是否存在功能缺陷。

I don't know if there's a feature bug.

你能在世界完全渲染出所有纹理前看到那些小点，我特别喜欢窥探这个模型的运作机制。

You can see like the dots of the world before it actually renders with all the textures and I just love to get a glimpse into what is going on with this model.

基本上，

Basically,

听起来真是太酷了。

That I is like so cool to hear.

因为作为研究者，这正是我学习的地方——引导你进入世界的那些点其实是特意设计的可视化功能。

Because this is where, as a researcher, I'm learning because the dots that lead you into the world was an intentional feature visualization.

它并不属于模型本身。

It is not part of the model.

这个模型实际上只是生成了世界。

The model actually just generates the world.

但我们试图找到一种方法引导人们进入这个世界。

But we were trying to find a way to guide people into the world.

许多工程师开发了不同版本，但我们最终聚焦于那个点。

And a number of engineers worked on different versions, but we converged on the dot.

很多人——你不是唯一一个——告诉我们这种体验多么令人愉悦。

And so many people, you're not the only one, told us how delightful that experience is.

听到这个刻意设计的可视化功能（不仅仅是硬核大模型）确实取悦了用户，我们感到非常满足。

And it was really satisfying for us to hear that this intentional visualization feature that's not just the big hardcore model actually has delighted our users.

哇。

Wow.

所以你们添加这个功能是为了让人类更容易理解发生了什么，获得更多乐趣。

So you add that to make it more to have humans understand what's going on more, getting more delightful.

哇，这太有趣了。

Wow, that is hilarious.

让我联想到语言模型——虽然不完全相同——但它们会谈论自己在想什么。是的。

Makes me think about LMs in the way they it's not the same thing, but they talk about what they're thinking and what they're Yes.

确实如此。

It is.

确实如此。

It is.

这也让我联想到《黑客帝国》。

It also makes me think about just the matrix.

这简直就是《黑客帝国》般的体验。

Like, it's exactly the matrix experience.

我不确定那是否是你们的灵感来源。

I don't know if that was your inspiration.

嗯，就像我说的，有很多工程师参与了这个项目。

Well, like I said, a number of engineers worked on that.

那可能是他们的灵感来源。

It could be their inspiration.

这已经刻在他们的潜意识里了。

It's in their subconscious.

是啊。

Yeah.

好的。

Okay.

那么对于想要尝试使用这个技术的人，目前有哪些实际应用可以立即上手？

So just for folks that may want to play around with this, maybe use it, what's like, what are some applications today that folks can start using today?

你们这次发布的主要目标是什么？

What's what's your goal with this launch?

没错。

Yeah.

我们确实认为世界建模技术具有广泛适用性，而且已经看到了一些非常激动人心的应用场景。

So we do believe that world modeling is very horizontal, but we're already seeing some really exciting use cases.

比如电影虚拟制作，因为他们需要能与摄像机同步的三维世界。

Virtual production for movies, because what they need are three d worlds that they can align with the camera.

所以当演员在上面表演时，他们可以调整摄像机位置，把各个片段拍得很好。

So when the actors are acting on it, they can, you know, they can position the camera and shoot the the segments really well.

我们已经看到了惊人的应用效果。

And we're already seeing incredible use.

其实我不知道你是否看过我们展示Marble的发布视频。

In fact, I don't know if you have seen our launch video showing Marble.

那是由一家虚拟制作公司制作的。

It was produced by a virtual production company.

我们与索尼合作，他们使用Marble场景来拍摄那些视频。

We we collaborated with Sony, and they use Marble scenes to shoot those videos.

我们当时与那些技术美术师和导演合作，他们说这使制作时间缩短了40倍。

So our we were collaborating with those technical artists and directors, and they were saying this has cut our production time by 40x.

实际上，必须

In fact, it has to

是40倍？

be 40x?

实际上必须如此，因为我们只有一个月时间做这个项目，而他们需要拍摄的内容太多了。

In fact, it has to because we only had one month to work on this project, and there were so many things they were trying to shoot.

使用Marble确实极大地加速了VFX和电影的虚拟制作流程。

So using Marble really, really significantly accelerated the production of virtual virtual production for VFX and movies.

这只是其中一个应用场景。

That's one use case.

我们已经看到用户将我们的Marble场景导出网格后用于游戏开发，无论是VR游戏还是他们开发的休闲游戏。

We are already seeing our users putting taking our marble scene and taking the mesh export and putting games, you know, whether it's games on VR or games, just fun games that they have developed.

我们刚才展示了一个机器人仿真实例，因为我现在仍是一名从事机器人训练的研究员。

We have had we were showing an example of robotic simulation because when I was, I mean, I still am a researcher doing robotic training.

最大的痛点之一就是为训练机器人创建合成数据。

One of the biggest pain point is to create synthetic data for training robots.

而这些合成数据需要具备高度多样性。

And this synthetic data needs to be very diverse.

它们需要来自不同环境，包含可供操作的不同物体。

They need to come from different environments with different objects to manipulate.

实现途径之一就是让计算机进行模拟。

And one path to it is to ask computers to simulate.

否则人类就得——你知道的——为机器人逐个构建所有素材。

Otherwise, humans have to, you know, build every single asset for robots.

那样会耗费更长时间。

That's just gonna take a lot longer.

已经有研究人员联系我们，希望使用Marble来创建这些合成环境。

So we already have researchers reaching out and wanting to use marble to create those synthetic environments.

我们还收到了关于如何使用Marble的意外用户反馈。

We also have unexpected user outreach in terms of how they want to use Marble.

例如，有个心理学团队联系我们想用Marble进行心理学研究。

For example, a psychologist team called us to use Marble to do psychology research.

原来他们研究的某些精神病患者，需要了解其大脑对不同特征沉浸场景的反应。

It turned out some of the psychiatric patients they study, they need to understand how their brain respond to different immersive scenes of different features.

比如混乱的场景、整洁的场景，或任何你能命名的场景。

For example, messy scenes or clean scenes or whatever you name it.

研究人员很难接触到这类沉浸式场景，而且他们需要花费太多时间和预算来创建。

And it's very hard for researchers to get their hands on these kind of immersive scenes, and it will take them too long and too much budget to create.

而Marble几乎能瞬间将大量这类实验环境交到他们手中。

And Marble is a really almost instantaneous way of getting so many of these experimental environments into their hands.

所以我们目前看到了多种应用场景，视觉特效师、游戏开发者、模拟开发者和设计师都非常兴奋。

So we're seeing multiple use cases at this point, but the VFX, the game developers, the simulation developers, as well as designers are very excited.

这正是AI领域的工作方式。

This is very much the way things work in AI.

我在播客中采访过其他AI领袖，他们总是说：尽早把产品放出去，这样才能发现最重要的应用场景在哪里。

I've had other AI leaders on the podcast, and it's always like, put things out there early as soon as you can to discover where the big use cases are.

ChatGPT的负责人告诉我，他们刚推出ChatGPT时，他整天刷TikTok看人们怎么使用它、讨论什么话题，这些观察帮助他们确定了重点发展方向。

The head of ChatGPT told me how when they first put out ChatGPT, he was just scanning TikTok to see how people were using it and all the things they were talking about, and that's what convinced them where to lean in and help them see how people actually want to use it.

我特别喜欢最后一个关于治疗的应用场景。

I love this last use case of like for therapy.

想象一下，比如恐高症患者面对高处，或是面对蛇、蜘蛛之类的场景——

Just imagining like, like heights, people seeing, dealing with heights or snakes or spiders that which

太神奇了。

It's amazing.

昨晚我朋友真的打电话给我，聊他的恐高症，还问我是否该用marble治疗。

A friend of mine last night literally called me and talked about his height scare and asked me if marble should be used.

你直接想到这个应用真是太棒了。

That's amazing you went straight there.

因为我在想象所有暴露疗法的场景，这个工具会非常适用。

That's, you know, because I'm imagining all the, like the exposure therapy stuff, like this could be so good for that.

这太酷了。

That is so cool.

好的，让我——我本该早点问这个的，但我觉得这里有个关键问题：它与VO3这类视频生成模型究竟有何不同？

Okay, so let me, I should have asked you this before, but I think there's a, there's gonna be a question of just how does this differ from things like VO3 and other video generation models?

对我来说区别很明显，不过解释一下它与现有视频AI工具的不同之处或许会有所帮助。

It's pretty clear to me, but I think it might be helpful just to explain how this is different from all the video AI tools people have seen.

WordLab的核心观点是：空间智能至关重要，且空间智能不仅关乎视频。

WordLab's thesis is that spatial intelligence is fundamentally very important, and spatial intelligence is not just about videos.

事实上，世界并非被动观看视频流逝的过程，对吧？

In fact, the world is not passively watching videos passing by, right?

柏拉图用洞穴寓言来描述视觉认知。

Love Plato has the allegory of the cave analogy to describe vision.

他说想象一个被绑在椅子上的囚犯——不太人道——在洞穴中观看面前的人生剧场。

He said that imagine a prisoner tied on his chair, not very humane, but in a cave watching a full life theater in front of him.

但真实的演员表演其实发生在他背后。

But the actual live theater that actors are acting is behind his back.

只是通过光线将动作投射在洞穴墙壁上。

It was just lit so that the projection of the action is on a wall of the cave.

这个囚犯的任务就是理解眼前发生的真相。

And then the task of this prisoner is to figure out what's going on.

这个极端例子生动展现了视觉的本质：从二维信息解读三维（或四维）世界。

It's a pretty extreme example, but it really shows, it describes what vision is about, is that to make sense of the three d world or four d world out of two d.

因此在我看来，空间智能远比创造扁平二维世界更为深刻。

So spatial intelligence to me is deeper than only creating that flat two d world.

对我来说，空间智能是创造、推理、互动、理解深度空间世界的能力，无论是二维、三维还是四维，包括动态变化等一切。

Spatial intelligence to me is the ability to create, reason, interact, make sense of deeply spatial world, whether it's two d or three d or four d, including dynamics and all that.

所以WorldLab正专注于这一领域。

So so WorldLab is focusing on that.

当然，单纯创作视频的能力本身也可能是其中的一部分。

And, of course, the ability to create videos per se could be part of this.

事实上，就在几周前，我们推出了全球首个可在单块H100 GPU上实时演示的视频生成技术。

And in fact, just a couple of weeks ago, we rolled out the world's first real time demoable real time video generation on a single h 100 GPU.

因此我们的技术体系包含了这一部分。

So we we we part of our technology includes that.

但我觉得Marble截然不同，因为我们真正希望创作者、设计师和开发者能掌握一个能生成具有三维结构世界的模型，用于他们的工作。

But I think Marble is very different because we really want creators, designers, developers to have in their hands a model that can give them worlds with three d structure so they can use it for for their work.

这正是Marble如此独特的原因。

And that's where that's why Marble is so different.

在我看来，这是一个充满无限可能性的平台。

The way I see it is it's a it's a platform for a ton of opportunity to do stuff.

正如你所说，视频就像是——这里有个一次性视频，很有趣很酷，然后就这样了。

As you describe, videos are just like, here's a one off video that's very fun and cool, and you could and that's it.

就这样。

That's it.

然后你就继续前进。

And then you move on.

顺便说，在Marble里我们允许用户以视频形式导出内容。

By the way, we could in Marble, we could allow people to export in video form.

就像你说的那样，你可以进入一个世界。

So you could actually like you said, you go into a world.

比如说这是一个霍比特人的洞穴。

So so let's say it's a hobbit cave.

特别是作为创作者，你会有一种非常具体的方式，按照导演脑海中的轨迹来移动摄像机。

You can actually especially as a creator, you have such a specific way of moving the camera in a trajectory in the director's mind.

对吧？

Right?

然后你可以将其从Marble导出为视频。

And then you can export that from Marble into video.

要创造出这样的东西需要什么条件？

What does it take to create something like this?

团队规模有多大？

Just like how big is the team?

你们使用多少GPU？

How GPUs you work in?

任何可以分享的信息都可以。

Anything you can share there.

我不知道这些信息有多少是保密的，但创造并推出你们这样的产品需要什么条件？

I don't know how much of this is private information, but just what does it take to create something like this that you've launched here?

这需要大量的脑力。

It takes a lot of brain power.

我们刚才说到每个大脑20瓦的功率。

We just talk about 20 watts per brain.

从这个角度看，数字虽小，但确实令人难以置信。

So from that point of view, it's a small number, but it's actually incredible.

要知道，这是五亿年进化才赋予我们的力量。

You know, it's half billion years of evolution to give us those power.

我们

We

拥有

have

目前团队约有30人，主要是研究员或研究工程师。

a team of 30 ish people now, and we are predominantly researchers or research engineers.

但我们也有设计师和产品人员。

And but we also have designers and and product.

我们真心希望创建一家以空间智能深度技术为核心的公司，同时也在认真打造产品。

We we actually really believe that we wanna create a company that's anchored in the deep tech of spatial intelligence, but we we we are actually building serious products.

因此我们实现了研发与产品化的融合。

So we have this integration of R and D and productization.

当然，我们使用了大量GPU。

And of course, we use, you know, a ton of GPUs.

啊这

A that's

技术层面是这样的。

the technical to hear.

恭喜产品发布。

Well, congrats on the launch.

我知道这是一个重要的里程碑。

I know this is a huge milestone.

知道这需要付出巨大的努力。

Know this took a ton of work.

谢谢苏

Thank So

我只想对你和你的团队表示祝贺。

I just want to say congrats to you and your team.

让我简单谈谈你的创始人历程。

Let me talk about your founder journey for a moment.

所以你是这家公司的创始人。

So you're a founder of this company.

你是多少年前创办的？

You started how many years ago?

几年前？

A couple years ago?

两、三年前？

Two, three years ago?

一年前。

A year ago.

一年前。

A year ago.

一年一年？

A year year?

好的。

Okay.

整整十八个月。

Very eighteen month.

是啊。

Yeah.

好的。

Okay.

在你开始之前，有什么是你希望自己早知道、能对十八个月前的自己耳语的事情吗？

What's something you wish you knew before you started this that you wish you could, like, whisper into the ear of eighteen months ago?

嗯，我一直希望能预知科技的未来。

Well, I continue to wish I know the future of technology.

我认为这实际上是我们创始优势之一——我们通常比大多数人更早看到未来。

I think actually that's one of our founding advantage is that we see the future earlier in general than most people.

但即便如此，伙计，未知和即将到来的事物还是如此令人兴奋和惊叹。

But still, man, this is so exciting and so amazing that what's unknown and what's coming.

但我知道你问这个问题的重点不在于科技的未来。

But I know the reason you're asking me this question is not about the future of technology.

你大概更想说的是——听着，我20岁时可没创办过这种规模的公司。

You're probably more, you know, look, I did not start a company of this scale at 20 year old.

所以我19岁时开了家干洗店，但那规模小得多。

So, you know, I started a dry cleaner when I was 19, but that's a little smaller scale.

我们得谈谈

We gotta talk

关于那件事。

about that.

后来我创立了谷歌云AI部门，又在斯坦福大学建立了一个研究所，但这两者性质截然不同。

And then I, you know, founded Google Cloud AI, and then I founded an institute at Stanford, but those are different beasts.

确实，相比那些20岁的创业者，作为经历过磨砺的创始人，我感觉自己准备得更充分些。

I did feel I was a little more prepared as a a founder of the grinding journey I compared to maybe the 20 year old founders.

但我仍时常感到震惊，甚至陷入焦虑——AI领域的技术竞争和人才争夺竟如此激烈。

But I still I'm surprised and it puts me into paranoia sometimes that how intensely competitive AI landscape is from the technology itself, as well as talents.

要知道当初我创业时，根本想象不到某些顶尖人才的薪酬会达到如此惊人的数字。

And, you know, when I founded the company, we did not have these incredible stories of how much certain talents would cost, you know?

这些情况不断刷新我的认知，我必须时刻保持警觉。

So these are things that continue to surprise me, and I have to be very alert about.

所以你提到的竞争，主要是人才争夺战，以及整个领域令人目眩的发展速度。

So the competition you're talking about is, yeah, the competition for talent, the speed at which just how things are moving.

确实。

Yeah.

没错。

Yeah.

我想回到你之前提到的观点：纵观你的职业生涯，你几乎参与了所有孕育当今重大突破的核心人类集群。

You mentioned this point that I want to come back to that you, if you just look over the course of your career, you were like at all of the major collections of humans that led to so many of the breakthroughs that are happening today.

我们讨论过ImageNet，还有斯坦福SAIL实验室——许多突破性工作诞生地，谷歌云也是重大突破的温床。

Obviously we talk about ImageNet, also just SAIL at Stanford is where a lot of the work happened, Google Cloud, which a lot of the breakthroughs happened.

是什么吸引你投身这些地方的？

What brought you to those places?

对于那些寻求职业发展、想要站在未来中心的人们来说，能否分享一下是什么力量推动你不断转换环境、加入那些可能对他人有启发的团队？

For people looking for how to advance in their career, be at the center of the future, just like, is there a through line there of just what pulled you from place to place and pulled you into those groups that might be helpful for people to hear?

这确实是个好问题，兰尼，因为我确实经常思考这一点。

This is actually a great question, Lanny, because I do think about it.

显然，我们之前讨论过是好奇心和热情将我引向了人工智能领域。

And obviously, we talked about its curiosity and passion that brought me to AI.

这更像是一颗科学北极星，对吧？

That is more a scientific north star, right?

我当时并不在乎AI是否真的能成气候。

I did not care if AI was a thing or not.

这是其中一部分原因。

So that was one part.

但关于我如何最终选择了包括创立World Labs在内的特定工作场所，我想我很感谢自己——或许该感谢父母的基因。

But how did I end up choosing in the particular places I work in, including starting World Labs, is I think I'm very grateful to myself or maybe to my parents' genes.

我在知识探索上是个非常无所畏惧的人。

I'm an intellectually very fearless person.

必须说，当我招聘年轻人时，我特别看重这种特质。因为想要有所作为，就必须接受你正在创造新事物或探索未知领域——这种品质至关重要。

And I have to say, when I hire young people, I look for that because I think that's a very important quality if one wants to make a difference, is that when you want to make a difference, you have to accept that you're creating something new or you're diving into something new.

前人都没做过这些事。

People haven't done that.

如果你具备这种自我认知，就几乎必须允许自己保持无畏和勇敢。

And if you have that self awareness, you almost have to allow yourself to be fearless and to be courageous.

比如当我来到斯坦福时，在学术圈里，我离那个叫做'终身教职'的东西很近——就是能在普林斯顿获得永久职位。

So when I, for example, came to Stanford, you know, in the world of academia, I was very close to this thing called tenure, which is, you know, have the job forever at Princeton.

但我选择来斯坦福是因为我热爱普林斯顿。

But I chose to come to Stanford because I love Princeton.

那是我的母校。

It's my alma mater.

就在那一刻，斯坦福有那么多杰出人才，硅谷的生态系统如此出色，让我愿意冒险重新开始我的终身教职计时。

It's just at that moment, there are people who are so amazing at Stanford and the Silicon Valley ecosystem was so amazing that I was okay to take a risk of restarting my tenure clock.

成为SAIL首位女性主任时，实际上相对而言我当时还是一名非常年轻的教员。

Going to becoming the first female director of SAIL, I was actually, relatively speaking, a very young faculty at that time.

我想这么做是因为我在乎那个群体。

And I wanted to do that because I care about that community.

我没有花太多时间考虑所有可能的失败情况。

I didn't spend too much time thinking about all the failure cases.

显然我很幸运得到了资深教员们的支持，但我只是想有所作为。

Obviously, I was very lucky that the more senior faculty supported me, but I just wanted to make a difference.

后来去谷歌也是类似的情况。

And then going to Google was similar.

我想与杰夫·迪恩、杰夫·辛顿这样的杰出学者，以及所有那些了不起的人共事。

I wanted to work with people like Jeff Dean, Jeff Hinton, and all these incredible academics, the incredible people.

你知道吗？

You know?

世界实验室的情况也是如此。

So so the same with World Labs.

我怀有这份热情，也相信志同道合的人能创造奇迹。

I I have this passion, and I also believe that people with the same mission can do incredible things.

这就是它如何指引我度过人生的。

So that's how it guided my through through life.

我不会过度思考所有可能出错的事情，因为实在太多了。

I don't overthink of all possible things that can go wrong because that's too many.

我觉得这是个重要因素。

I feel like that's an important element.

这不是在关注消极面，而是更多地关注人、使命和让你兴奋的事物，你觉得呢？

This is not focusing on the downside, focusing more on the people, the mission, what gets you excited, what do think?

我很好奇。

I curiosity.

是的。

Yeah.

我想对所有AI领域的年轻人才、工程师和研究人员说一句，因为你们中有些人申请了WorldLabs。

I do wanna say one thing to all the young talents in AI, the engineers, the researchers out there, because some of you apply to WorldLabs.

我感到非常荣幸你们考虑过WorldLabs。

I I feel very privileged you considered WorldLabs.

我发现现在很多年轻人在决定工作时会考虑方程式的每一个细节。

I do find many of the young people today think about every single aspect of an equation when they decide on jobs.

某种程度上，也许这就是他们想要的方式。

At some point, maybe, you know, maybe that's the way they want to do it.

但有时我确实想鼓励年轻人关注重要的事情，因为我发现自己与求职者交谈时总是不自觉地进入导师模式，不一定是在招聘或不招聘，而只是处于指导状态。

But sometimes I do want to encourage young people to focus on what's important because I find myself constantly in mentoring mode when I talk to job candidates, not necessarily recruiting or not recruiting, but just in mentoring mode.

当我看到一位才华横溢的年轻人过度纠结于工作考虑的每个微小维度时，或许最重要的问题是：你的热情在哪里？

When I see an incredible young talent who is over focusing on every minute dimension and aspect of considering a job, when maybe the most important thing is where's your passion?

你认同这个使命吗？

Do you align with the mission?

你相信并对这个团队有信心吗？

Do you believe and have faith in this team?

只需专注于你能产生的影响，以及你能参与的工作和合作的团队。

And just focus on the impact and you can make and the kind of work and team you can work with.

是啊，这很艰难。

Yeah, it's tough.

现在对AI领域的人来说很艰难。

It's tough for people in the AI space now.

他们面临太多压力，太多新闻，太多变化，太多错失恐惧症。

There's just so much at them, so much news, so much happening, so much FOMO.

确实如此。

That's true.

我能看出这种压力。

I could see the stress.

所以我认为这个建议非常重要，就像什么能真正让你对所做的事情感到满足，而不仅仅是哪家公司增长最快，谁会赢，我不知道。

So I think that advice is really important, just like what will actually make you feel fulfilled in what you're doing, not just where's the fastest growing company, where's the, who's going to win, I don't know.

我想确保我问到你今天在斯坦福大学HCI的工作。

I want to make sure I ask you about the work you're doing today at Stanford at the HCI.

我认为这是

I think it's the

HAI，以人为中心的人工智能研究所。

HAI, Human Centered AI Institute.

你在那里做什么？

What you what are you doing there?

我知道你现在还在做这件事。

I know this is a thing you do on the side still.

是的，HAI（以人为本的人工智能研究所）是由我和包括John H教授在内的一群学者共同创立的。

So, yes, I HAI, Human Centered AI Institute, was co founded by me and a group of faculty like Professor John H.

Mendy教授、James Landy教授、Chris Manning教授等人早在2018年就参与了。

Mendy, Professor James Landy, Professor Chris Manning back in 2018.

当时我其实刚结束在谷歌的学术休假。

I was actually finishing my sabbatical at Google.

这对我来说是个极其重要的决定——我本可以留在工业界，但在谷歌的经历让我明白：AI将成为改变人类文明的技术。

And it was a very, very important decision for me because I could have stayed in industry, but my time at Google taught me one thing is AI is gonna be a civilizational technology.

我逐渐意识到这对人类有多重要，甚至促使我在2018年为《纽约时报》撰文，探讨建立AI开发和应用的指导框架的必要性。

And they it's it dawned on me how important this is to humanity to the point that I actually wrote a piece in New York Times that year, 2018, to talk about the need for a guiding framework to develop and to to apply AI.

这个框架必须以人类福祉为根基，以人为核心。

And that framework has to be anchored in human benevolence, is human centeredness.

我认为斯坦福——这所位于硅谷心脏地带、孕育了从英伟达到谷歌等伟大企业的世界顶尖学府——应当成为创建这种以人为本AI框架的思想领袖，并将其贯彻到我们的研究、教育、政策及生态建设中。

And I felt that Stanford, one of the world's top university in the heart of Silicon Valley that gave birth to important companies from NVIDIA to Google, should be a thought leader to create this human centered AI framework and to actually embody that in our research, education, and policy and ecosystem work.

于是我创立了HAI。

So I founded HAI.

转眼六七年过去，它已成为全球最大的人工智能研究机构，开展以人为本的研究、教育、生态拓展和政策影响工作。

You know, after fast forward, after six, seven years, it has become the world's largest AI institute that does human centered research, education, ecosystem outreach, and policy impact.

它汇聚了斯坦福八大学院——从医学院、教育学院到可持续发展学院、商学院、工程学院、人文学院、法学院——数百名教职员工。

It involves hundreds of faculty across all eight schools at Stanford, from medicine to education, to sustainability, to business, to engineering, to humanities, to law.

我们支持研究人员，特别是在数字经济、法律研究、政治学、新药研发、新算法乃至超越Transformer模型等跨学科领域。

And we support researchers, especially at the interdisciplinary area from digital economy to legal studies, to political science, to discovery of new drugs, to to new algorithms, to that's beyond transformers.

我们还特别重视政策研究，因为当我们成立HAI时，我意识到硅谷与华盛顿、布鲁塞尔乃至世界其他地区之间缺乏对话。

We also actually put a very strong focus on on policy because when we started HAI, I realized that Silicon Valley did not talk to Washington, DC and or Brussels or other parts of the world.

鉴于这项技术的重要性，我们需要让各方都参与进来。

And given how important this technology is, we need to bring everybody on board.

因此我们创建了多项计划，从国会速成班到AI指数报告，再到政策简报会。

So we created multiple programs from congressional bootcamp to AI index report to policy briefing.

我们尤其参与了政策制定工作，包括推动在前特朗普政府时期通过的国家AI研究云法案，并参与州级AI监管讨论。

And we especially participated in policymaking, including advocating for a national AI research cloud bill that was passed in the first Trump administration and participate participating in state level regulatory AI discussions.

我们做了很多工作。

So there's a lot we did.

我继续担任领导者之一，尽管在运营层面参与较少，因为我不仅关心创造这项技术，更关注如何正确使用它。

And I continue to be one of the leaders, even though I'm much less involved operationally, because I care not only we create this technology, but we use it in the right way.

哇。

Wow.

我之前不知道你还做了这么多其他工作。

I was not aware of all that other work you were doing.

听你讲述时，我想起查理·芒格的名言：'选择一个简单的理念，并非常认真地对待它'。

As you're talking, was reminded Charlie Munger had this quote, Take a simple idea and take it very seriously.

感觉你在诸多领域都践行着这个理念并持之以恒，多年来产生的全方位影响令人难以置信。

I feel like you've done that in so many different ways and stayed with it, and it's unbelievable the impact that you've had in so many ways over the years.

我跳过快速问答环节，直接问你最后一个问题。

I'm gonna skip the lightning round, and I'm just gonna ask you one last question.

你还有什么想分享的吗？还有什么想留给听众的？

Is there anything else that you wanted to share, anything else you wanna leave listeners with?

我对AI Lenny感到非常兴奋。

I'm very excited by AI Lenny.

我想回答一个我在世界各地旅行时总被问到的问题：无论我是音乐家、中学教师、护士、会计、农民，我在AI中能扮演什么角色？AI是会接管我的生活和工作吗？

I wanna answer one question that when I travel around the world, everybody asks me is that if I'm a musician, if I'm a teacher, middle school teacher, if I'm a nurse, if I'm an accountant, if I'm a farmer, do I have a role in AI, or is AI just gonna take over my life or my work?

我认为这是关于AI最重要的问题。

And I think this is the most important question of AI.

我发现硅谷往往不会与人们——无论是像我们这样的还是不像我们这样的，而是指所有人——进行心与心的对话。

And I find that in Silicon Valley, we tend not to speak heart to heart with people, with people like us and and not like us in Silicon Valley, but like all of us.

我们总是随意抛出诸如无限生产力、无限休闲时间或无限力量之类的词汇。

We tend to just toss around words like infinite productivity or infinite leisure time or infinite power or whatever.

但归根结底，AI是关于人的。

But at the end of the day, AI is about people.

当人们问我这个问题时，我的回答是响亮的肯定。

And when people ask me that question, it's a resounding yes.

每个人在AI中都有自己的角色。

Everybody has a role in AI.

这取决于你的职业和追求，但任何技术都不应剥夺人类的尊严。

It depends on what what you do and what you want, but no technology should take away human dignity.

人类尊严和自主权应成为每项技术开发、部署及治理的核心。

And the human dignity and the agency should be at the heart of the development, the deployment, as well as the governance of every technology.

如果你是个年轻艺术家，热爱讲故事，就把AI当作工具来拥抱吧。

So if you are a young artist and your passion is storytelling, embrace AI as a tool.

事实上，拥抱漫威吧。

In fact, embrace Marvel.

我希望它能成为你的工具，因为你讲述故事的方式独一无二，这个世界依然需要它。

I hope it becomes a tool for you because the way you tell your story is unique and the world still needs it.

但你要如何讲述你的故事呢？

But how you tell your story?

如何用最不可思议的工具以最独特的方式讲述你的故事，这很重要。

How do you use the most incredible tool to tell your story in the most unique way is important.

而这个声音需要被听见。

And that voice needs to be heard.

即使你是临近退休的农民，AI依然重要，因为你也是公民。

If you're a farmer near retirement, AI still matters because you're a citizen.

你可以参与社区事务。

You can participate in your community.

你应该对AI的使用方式和应用领域有发言权。

You should have a voice in how AI is used, how AI is applied.

你可以与志同道合的人合作，鼓励大家运用AI让生活更轻松。

You you work with people that you can, you know, encourage all of all of you to use AI to make life easier for you.

如果你是护士，我希望你知道——至少在我的职业生涯中，我深耕医疗研究，因为我坚信AI技术应该大力辅助医护人员，无论是通过智能摄像头提供更多信息，还是机器人协助，毕竟我们的护士们工作过度、疲惫不堪。

If you're a nurse, I hope you know that at least in my career, I have worked so much in health care research because I feel our health care workers should be greatly augmented and helped by AI technology, whether it's smart cameras to feed more information or robotic assistance, because our nurses are overworked, overfatigued.

随着社会老龄化，我们需要更多帮助来照料人们。

And as our society ages, we need more help for people to be taken care of.

因此AI可以扮演这个角色。

So AI can play that role.

所以我想说的是，这非常重要，即使像我这样的技术专家也真诚地认为每个人在人工智能中都有自己的角色。

So I just wanna say that it's so important that even a technologist like me are sincere about that everybody has a role in AI.

多么完美的结束方式啊。

What a beautiful way to end it.

这与我们最初讨论的内容相呼应——人工智能在我们生活中的影响取决于我们每个人如何承担责任。

Such a tie back to where we started about how it's up to us and taking individual responsibility for what AI will do in our lives.

最后一个问题，大家在哪里可以找到Marble？

Final question, where can folks find Marble?

如果他们想加入WorldLabs，可以去哪里尝试？

Where can they go maybe try to join WorldLabs if they want to?

网址是什么？

What's the website?

人们应该去哪里？

Where do people go?

WorldLabs的网站是www.worldlabs.ai。

Well, WorldLabs website is www.worldlabs.ai.

你可以在那里找到我们的研究进展。

And you can find our research progress there.

我们有技术博客。

We have technical blogs.

你可以在那里找到Marble产品。

You can find Marble the product there.

你可以在那里注册登录。

You can sign in there.

你可以在那里找到我们的招聘链接。

You can find our job posts link there.

你知道的，我们在旧金山。

You can, you know, we're in San Francisco.

我们喜欢与世界顶尖人才共事。

We love to work with the world's best talents.

太棒了。

Amazing.

菲菲，非常感谢你能来。

Fei Fei, thank you so much for being here.

谢谢你，莱尼。

Thank you, Lenny.

再见，各位。

Bye, everyone.

非常感谢大家的收听。

Thank you so much for listening.