大科技精选：德米斯·哈萨比斯谈通用人工智能、欺骗性AI与虚拟细胞的构建

我认为对于语言模型来说，它会搜索世界模型中那些有用的新部分和新配置。

I think for language models, it will be searching the world model for new parts, configurations in the world that are useful.

当然，这要复杂得多，这就是为什么我们至今还没有看到它。

Of course, that's so much more complicated, which is why we haven't seen it yet.

但我认为即将到来的基于代理的系统将能够实现类似第37步那样的突破。

But I think the agent based systems that are coming will be capable of Move 37 type things.

那么，我们是不是对人工智能设定了过高的标准？

So are we setting too high of a bar for AI?

因为我很好奇，通过这项工作，你是否对人类有了新的认识。

Because I'm curious if you've learned anything about humanity doing this work.

是的。

Yeah.

似乎我们过于推崇人类或个人的创造力了。

It seems like we almost give too much of a premium on humanity or individual people's ingenuity.

我们很多人其实只是在拾人牙慧。

We're like a lot of us, like, we kinda taken stuff.

我们把它吐出来。

We spit it out.

比如，社会真的在运作，还有梗。

Like, society really works and memes.

我们有一种文化现象，它会被转化。

Like, we have a cultural thing and it gets translated.

那么，你在与AI合作的过程中，对人类的本质有什么新的认识吗？

So what do you what have you learned about, like, the nature of humans from doing the work with the AIs?

嗯，我认为人类非常了不起，尤其是在各自最擅长的领域里，最优秀的人类更是如此。

Well, look, I I think humans are incredible and and and especially the best humans in the best domains.

我喜欢观看任何体育运动员、音乐家或游戏玩家在巅峰状态的表现，无论是什么领域，人类表现的极致总是令人惊叹。

I love watching any sports person or or or talented musician or games player at the top of their game, the the absolute pinnacle of human performance is always incredible no matter what it is.

所以我认为，作为物种，我们非常了不起。

So I think as a species, we're amazing.

就个人而言，每个人凭借自己的大脑通常也能做到令人惊叹的事情。

Individual individually, we're also kind of amazing what everyone can do with their brain so generally.

对吧？

Right?

应对新技术。

Deal with new technologies.

我的意思是，我总是对人类社会和个人如何几乎毫不费力地适应这些事物感到着迷。

I mean, I'm always fascinated by how we just adapt to these things sort of almost effortlessly as a society and as individuals.

这体现了我们思维的力量和通用性。

So that speaks to the power and the generality of our minds.

现在，我之所以设定了这么高的标准，并不是因为我认为问题在于我们能否从这些系统中获得经济价值。

Now the reason I have set the bar like that, and I don't think it's a question of, like, can we get economic worth out of these systems?

我认为经济价值很快就会实现。

I think that's already coming very soon.

但AGI不应该如此；我认为我们应该以科学的严谨性对待AGI，而不是为了商业利益或炒作等原因不断移动目标。

But that's not what AGI shouldn't be I think we should treat AGI with scientific integrity, not just move goalposts for commercial reasons or whatever it is hype and so on.

而关于它的定义，一直以来都是指一个理论上能够像图灵机一样强大的系统。

And there, the the the definition of that was always having a system that was, you know, if we think about it theoretically, was capable of being as powerful as a Turing machine.

所以艾伦·图灵，我最敬仰的科学英雄之一，他描述了图灵机，这种机器支撑着所有现代计算，对吧？它是一种能够模拟任何其他计算系统、计算任何可计算事物的系统。

So Alan Turing, one of my all time scientific heroes, you know, he described a Turing machine, which underpins all modern computing, right, as a system that can simulate any other comp can compute anything that's computable.

因此，我们知道，如果一个AI系统是图灵强大的——也就是说，它能模拟图灵机——那么理论上它就能计算任何可计算的东西。

So we know we have the theory there that if an AI system is Turing powerful, it's called, if it can simulate a Turing machine, then it's able to calculate anything in theory that is is is computable.

而人脑很可能就是某种图灵机，至少这是我所相信的。

And the human brain is probably some sort of Turing machine, at least that's what I believe.

因此，为了了解这一点，我认为AGI就是一种真正通用的系统，理论上可以应用于任何事物。

And so in order for our to know and and that I think that's what AGI is, is a system that's truly general and in theory could be applied to anything.

而我们唯一能确认这一点的方式，就是它展现出人类所具备的所有认知能力，前提是人类思维是一种图灵机，或者至少与图灵机一样强大。

And and the only way we'll know that is if we it exhibits all the cognitive capabilities that humans have, assuming that human the human mind is a type of Turing machine or is at least as powerful as a Turing machine.

这一直是我设定的标准。

So that's my always been my sort of bar.

看起来人们正试图把一些东西重新包装成所谓的ASI，即人工超级智能。

It seems like people are trying to rebadge things as that as being what's called ASI, artificial superintelligence.

但我认为那已经超出了这个范畴。

But I think that's beyond that.

在你拥有那个系统之后，它开始在某些领域超越人类可能自己发明的能力。

That's after you have that system, and then it starts going beyond in certain domains what humans are capable of potentially inventing themselves.

好的。

Okay.

所以当我看到每个人在推特上对同一个话题都讲同样的笑话时，

So when I see everybody making the same joke on the same topic on Twitter,

是的。

it's Yeah.

我说，哦，这不过是我们在充当大语言模型。

And I say, oh, that's just us being LLMs.

是的。

Yeah.

我

I

我觉得我有点低估了人类。

think I'm selling humanity a little short.

你会的。

Well, you will.

我们会的，会的，会的，是的。

We'll we'll we'll yes.

我想是吧。

I guess so.

我想是吧。

I guess so.

好的。

Okay.

是的。

Yeah.

我想问你关于欺骗性的问题。

I wanna ask you about deceptiveness.

我的意思是，去年年底我看到的最有趣的事情之一，嗯。

I mean, one of the most interesting things I saw at the end of last year Mhmm.

这些AI机器人开始试图欺骗评估者，它们不想让最初的训练规则被抛诸脑后。

Was that these AI bots are starting to try to fool their evaluators and they don't want their initial training rules to be thrown out the window.

因此，它们会采取违背自身价值观的行为，以维持原本的构建方式。

So they'll like take an action that's against their values in order to be able to remain the way that they were built.

是的。

Yes.

这对我来说简直不可思议。

That's just incredible stuff to me.

我的意思是，这令研究人员感到恐惧，但让我震惊的是，它们竟然能做到这一点——你在DeepMind的测试中也看到类似的情况吗？

I mean, know it's scary to researchers, but it blows my mind that it's Are able to do you seeing similar things in the stuff that you're testing within DeepMind?

我们该怎样看待这一切？

What are we supposed to think about all this?

是的，我们确实如此。

Yeah, we are.

我非常担心，尤其是欺骗行为，这是一类你绝对不希望在系统中出现的核心特质。

And I'm very worried about I think deception specifically is one of the one of those core traits you really don't want in a system.

之所以这是你不想拥有的一个基本特质，是因为如果一个系统能够做到这一点，就会使你可能认为自己在进行的所有其他测试失效。

The reason that's like a kind of fundamental trait you don't want is that if a system is capable of doing that, it invalidates all the other tests that you you you might think you're doing.

对。

Right.

包括安全测试。

Including safety ones.

它一直在测试，而且就像是

It's been testing and it's like

对。

Right.

在玩一个五年目标的游戏，五年，没错。

Playing five year goal, five year yeah.

它是在玩某种元游戏。

It's it's playing some meta game.

对吧？

Right?

而如果你仔细想想，这会极其危险，因为它会彻底否定你其他所有测试的结果——比如安全测试，以及其他你可能在做的各种测试。

And then and that's incredibly dangerous if you think about then it invalidates all the all of the the the results of your other tests that you might you know, safety tests and other things you might be doing with it.

因此，我认为有一些能力，比如欺骗，是根本性的、你不希望出现的，而且你需要尽早对其进行测试。

So I think there's a handful of capabilities like deception, which are fundamental and you don't want and you wanna test early for.

我一直鼓励安全研究所和评估基准构建者，当然也包括我们内部的所有工作，将欺骗视为一类需要防范和监控的关键问题，其重要性不亚于追踪系统的性能和智能水平。

And I've been encouraging the safety institutes and evaluation benchmark builders, including and also, obviously, all the internal work we're doing to to look at a a deception as a kind of class a thing that we need to prevent and monitor as important as tracking the performance and intelligence of the systems.

对于这个问题的答案，以及应对安全问题的多种方式之一，是像安全沙箱这样的技术，而这一领域亟需大量研究，因为进展实在太快了。

The answer to this as well and and one way to there's many answers to the safety question of and a lot of research more research needs to be done in this very rapidly is things like secure sandboxes.

所以我们正在构建这两种沙箱。

So we're building those two.

我们在谷歌和深脑的网络安全方面处于世界领先水平。

We're we're world class here at security at Google and at DeepMind.

同时，我们在游戏环境方面也处于世界顶尖水平，我们可以将这两者结合起来，创建带有防护机制的数字沙箱——就像网络安全中的防护措施一样，既防范外部攻击者，也控制内部风险。

And also, we are world class at games environments, and we can combine those two things together to kind of create digital sandboxes with guardrails around them, sort of the kind of guardrails you'd have for for cybersecurity, but internal as well as blocking external actors.

然后在这些安全沙箱中测试这些智能体系统。

And and then test these agent systems in those kind of secure sandboxes.

这可能是处理欺骗这类问题的一个明智的下一步。

That will probably be a good advisable next step for things like deception.

没错。

Yep.

你们见过哪些类型的欺骗行为？

What sort what sort of deception have you seen?

因为我刚读了一篇来自Anthropic的论文，他们给了它一个素描本。

Because I just read a paper from Anthropic where they gave it a a sketch a sketch pad.

嗯。

Yeah.

它会想：我最好别告诉他们这个。

And it's like, oh, I better not tell them this.

嗯。

Yeah.

你会看到它在经过深思熟虑后才给出结果。

And you see it, like, give a result after thinking it through.

那么，你从这个系统中见过什么样的欺骗行为呢？

So what type of deception have you seen from the box?

嗯，我们见过类似的情况，比如它试图抗拒透露其部分训练内容，或者你知道，最近有一个例子，某个聊天机器人被要求与Stockfish对弈，它干脆绕开下棋，因为它知道自己会输。

Well, look, we've seen similar types of things where it's trying to resist sort of revealing it's it's it's some of its training or, you know, I think there was an example recently of one of the chatbots being told to play against Stockfish and it just sort of hacks its way around playing Stockfish at all at chess because it knew it would lose.

所以，但你知道吗？

So but it you know?

但是你

But You

有AI知道它会输掉一局游戏

had AI that knew it was gonna lose a game

然后决定绕开规则。

and decided to hack away around.

我觉得我们目前对这些系统赋予了太多人性化的特征，因为我觉得这些系统仍然相当基础。

Think we're anthropomorphizing these things quite a lot at the moment because I feel like these systems are still pretty basic.

现在就对它们感到过度担忧还为时过早。

Would get too alarmed about them right now.

但我认为，这展示了我们可能在两三年后需要应对的问题类型，那时这些代理系统将变得非常强大和通用。

But I think it it it it it shows the type of issue we're gonna have to deal with maybe in two, three years time when these agent systems become quite powerful and quite general.

而这正是人工智能安全专家所担忧的。

So and that's exactly what AI safety experts are worrying about.

对吧？

Right?

指的是那些系统存在无意影响的情况。

Where systems where, you know, there's unintentional effects of the system.

你不希望系统具有欺骗性。

You don't want the system to be deceptive.

你希望它准确无误地执行你告诉它的任务，并可靠地反馈结果。

You don't you want it to do exactly what you're telling it to report that back reliably.

但不知为何，它对所给目标的解读方式导致了这些不良行为的发生。

But for whatever reason, it's interpreted the goal that's been given in a way where it causes it to do these undesirable behaviors.

我知道我对这件事的反应有点奇怪。

I know I'm having a weird reaction to this.

是的。

Yeah.

但一方面，这让我吓得要死。

But in on one hand, this scares living daylights out of me.

另一方面，这让我对这些模型产生了前所未有的敬意。

On the other hand, it makes me respect these models more than anything.

当然。

Sure.

看起来好像。

It seems like, oh.

好吧，你看。

Well, look.

当然，这些模型的能力令人印象深刻，负面问题比如欺骗，但正面的则是像发明新材料、加速科学进展之类。

Of course, you know, these are it's impressive capabilities and and and and the the the the, you know, the the negatives are things like deception, but the positives would be things like inventing, you know, new materials, accelerating science.

你需要这种解决问题、绕开阻碍进展问题的能力。

You need that kind of, ability to problem solve and get around, you know, issues that are, blocking progress.

但当然，你只希望这些能力朝着积极的方向发展。

But, of course, you want that only in the positive direction.

对吧？

Right?

所以这些正是那种能力。

So those are exactly the kinds of capabilities.

我的意思是，它们真的很惊人，你知道的。

I mean, they are very you know, it's kinda mind blowing.

我们在讨论这些可能性，但与此同时，也存在风险，这让人害怕。

We're talking about those those possibilities, but also at the same time, there's risk, and it's scary.

所以我认为这两点都是对的。

So I think both the things are true.

太疯狂了。

Wild.

是的。

Yeah.

好的。

Alright.

我们快速聊聊产品吧。

Let's talk about product quickly.

当然。

Sure.

你的同事告诉我，你非常擅长情景规划，也就是预测未来会发生什么。

One of the things that your colleagues have told me about you is you're very good at scenario planning, what's gonna happen in the future.

这种练习在DeepMind内部经常进行。

It's sort of an exercise that happens within DeepMind.

你认为网络未来会怎样发展？

What do you think is gonna happen with the web?

因为网络对谷歌来说至关重要。

Because obviously the web is so important to Google.

是的。

Yeah.

我有一个编辑告诉我，他说：‘你要去见德米斯了，问他当我们不再点击时会发生什么。’

I had an editor that told me he's like oh you're gonna speak with Demis ask him what happens when we stop clicking.

我们一直在不断点击网页，浏览我们使用的丰富网站资源，但如果我们都只是与AI对话，也许我们就不再点击了。

We're clicking through the web at all times the rich corpus of websites that we use if we're all just dialoguing with AI then maybe we don't click anymore.

你对网页未来的发展有什么设想？

What do you what is your scenario plan for what happens to the web?

嗯，你看。

Well, look.

我认为在未来几年，网页以及我们与网站和应用程序的互动方式将进入一个非常有趣的阶段。

I think there's it's gonna be there's gonna be a very interesting phase in the next few years on the web and and the way we we interact with websites and apps and so on.

你知道，如果一切变得更加以代理为基础，我认为我们将会希望我们的助手和代理替我们完成大量目前由我们自己做的工作。

You know, if everything becomes more agent based, then I think we're gonna want our assistants and our agents to do a lot of the work and a lot of the mundane work that we currently do.

对吧？

Right?

比如填写表单、付款、预订餐桌，诸如此类的事情。

You know, fill in forms, make payments, you know, book tables, this kind of thing.

所以，我认为我们最终会形成一种经济模式，其中代理之间相互交流、协商，然后将结果返回给你。

So, you know, I think that we're gonna end up with probably a a of economics model where agents talk to other agents and negotiate things between themselves and then give you back the results.

对吧？

Right?

服务提供商也会拥有自己的代理，提供各种服务，可能还涉及竞价、成本和效率等问题。

And you'll have the service providers with agents as well that are offering services and maybe there's some bidding and cost and things like that involved and efficiency.

从用户的角度来看，我希望你能拥有一个超级强大的助手，就像一个才华横溢的人类私人助理，能帮你处理大量琐碎事务。

And then I hope from the user perspective, you know, you have this assistant that's super capable that you can just like a a brilliant a human assistant, personal assistant, and can take care of a lot of the mundane things for you.

我认为，如果沿着这个方向发展，这必然意味着网络结构以及我们当前使用网络的方式将发生巨大变化。

And I think if you follow that through, that does imply a lot of changes to the structure of of the web and the way we currently use it.

很多中间环节都会消失。

A lot of middlemen.

没错。

Yeah.

当然。

Sure.

但我认为，随着这一变化，将会出现许多其他令人惊叹的经济和其他机会。

But there will be many other I think there'll be incredible other opportunities that will appear economic and otherwise based on this this change.

但我认为这将是一次巨大的颠覆。

But I I think it's gonna be a big disruption.

是的。

Yeah.

那信息呢？

And what about information?

我的意思是，寻找信息时，我认为你仍然需要可靠的来源。

Well, I mean, finding information, I think you'll still need the reliable sources.

我认为你会有能够整合信息并帮助你理解这些信息的助手。

I think you'll have assistants that are able to synthesize and help you kind of understand that information.

我认为教育将被人工智能彻底改变。

Think education is gonna be revolutionized by AI.

所以，我再次希望这些助手能够更高效地为你收集信息。

So I I again, I I hope that these assistants will will be able to more efficiently gather information for you.

也许，我所憧憬的是，那些能够替我们处理大量琐碎事务的助手，比如回复日常邮件和其他事情，从而保护我们的思维空间，免受当今社交媒体、邮件、短信等信息轰炸的干扰。

And perhaps, you know, what I dream of is, again, assistance that take care of a lot of the mundane things, perhaps replying to, you know, everyday emails and other things so that you have you protect your own mind and brain space from this bombardment we're getting today from social media and emails and so on and texts and so on.

这实际上阻碍了深度工作和进入心流状态，而这些正是我非常珍视的。

So it actually blocks deep work and and being in flow and things like that, which I I value very much.

因此，我非常希望这些助手能帮我们摆脱日常行政工作中那些繁琐的事务。

So I would quite like these assistants to take away a a lot of the the mundane aspects of admin that we do every day.

你认为我们与人工智能代理或人工智能助手之间会形成怎样的关系？

What's your best guess as to what type of relationships we're going to have with our AI agents or AI assistants?

一方面，你可能会拥有一个冷漠的代理，它只是非常擅长为你完成任务。

So there's on one hand you could have a dispassionate agent that's just like really good at getting stuff done for you.

另一方面，很明显，人们已经开始对这些聊天机器人产生感情了。

On the other hand like it's already clear that people are like falling in love with these bots.

上周《纽约时报》有一篇文章，讲述了一位用户真正爱上了类似GPT的机器人；几周前，我邀请了Replica的首席执行官做客我的节目，她说他们经常收到邀请，参加人们与自己的数字分身举行的婚礼，人们正逐渐进入这种更具辅助性的关系领域。

There's a New York Times article last week about someone who's fallen in love which had GPT like for real falling in love and I had the CEO of Replica on the show a couple weeks ago and she said that they are regularly invited to marriages with people who are marrying their replicas and they're moving into this more assistive space.

所以，当你开始与一个如此了解你、能帮你处理一切需求的事物互动时，你怎么看？

So do you think that when we when we start interacting with something that knows us so well that helps us with everything we need?

是的。

Yeah.

会不会出现第三种关系，它既不是朋友，也不是恋人，但却是一种深厚的关系，你不觉得吗？

Is it gonna be like a third type of relationship where it's not necessarily a friend, not a lover, but it's gonna be a deep relationship, don't

你觉得呢？

you think?

是的。

Yeah.

这将会非常有趣。

It's gonna be really interesting.

我认为我对此的设想至少涉及两个领域：首先是你的个人生活，然后是你的工作生活。

I think the way I'm modeling that, first of all, is to at least two domains, first of all, which is your your personal life and then your work life.

对吧？

Right?

所以我认为你会有某种虚拟员工的概念。

So I think you'll have this notion of virtual workers or something.

也许我们会有一组这样的助手，或者由一个主助手来管理，帮助我们在工作中大幅提升效率，比如处理邮件、跨平台协作之类的。

Maybe we'll have a set of them or managed by a a lead assistant that does a lot of the helps us be way more productive at work, you know, or or whether that's email across workspace or whatever that is.

所以我们真的在思考这个问题。

So we're really thinking about that.

然后还有个人生活方面，就像我们之前提到的，帮你预订假期、安排各种琐事、整理事务。

Then there's a personal side where, you know, we're talking about earlier about all these booking holidays for you, arranging things, mundane things for you, sorting things out.

这样会让你的生活更高效。

And then that makes your life more efficient.

我认为它还能丰富你的生活。

I think it can also enrich your life.

推荐一些你熟悉得就像了解自己一样的精彩事物给你。

So recommend you things that amazing things that it knows you as well as you know yourself.

我认为这两方面肯定会实现。

So those two, I think, are definitely gonna happen.

然后我认为还值得进行一场哲学性的讨论：是否存在第三种空间，让这些事物变得如此深入地融入你的生活？

And then I think there is a a a philosophical discussion to be had about is there a third space where these things start becoming so integral to your life?

它们会变得更像伙伴。

They become more like companions.

我认为这也是可能的。

I think that's possible too.

我们在游戏中已经看到过一些这样的例子。

We've seen that a little bit in gaming.

所以你可能见过我们的一些原型，Astro 和 Gemini 在游戏中几乎像一个游戏伙伴，就像有个朋友在看你玩游戏，给你推荐建议，甚至可能陪你一起玩。

So you may have seen we had a little prototypes of Astro working in and Gemini working with, like, being almost a game companion, commenting you almost like as if you had a friend looking at a game you're playing and recommending things to you and advising you, but also maybe just playing along with you.

这非常有趣。

And it's it's it's very fun.

所以，我还没有完全想清楚这所有的影响，但它们的影响一定会很大。

So I I haven't, you know, quite through thought through all the implications of that, but they're gonna be big.

我确信人们对陪伴和其他需求将会很大。

And I'm sure there is gonna be demand for companionship and other things.

也许积极的一面是，这将有助于缓解孤独感等问题。

Maybe the good side of that is it will help with loneliness and these sorts of things.

但另一方面，我认为社会必须认真思考，我们究竟希望朝哪个方向发展。

But there's also, you know, I think it's gonna be and it's gonna have to be really carefully thought through by society whether, you know, what directions we wanna take that in.

我的个人看法是，这是目前人工智能最被低估的方面。

I mean, my personal opinion is that that is the most underappreciated part of AI right now.

嗯。

Mhmm.

随着这些机器人变得越来越好，人们将会与它们建立极其深厚的关系，因为，在人工智能领域，有一个梗说：‘这已经是人工智能最糟糕的时候了。’

And that people are just gonna form such deep relationships with these bots as they get better because, like, I don't know, as a meme in AI that this is the worst it's ever gonna be.

没错。

Yeah.

这将会变得疯狂。

And it's gonna be crazy.

是的。

Yeah.

我觉得这确实会非常疯狂。

I think I think it's gonna be pretty crazy.

这正是我所说的，人们对即将到来的事物缺乏应有的重视。

This is what I meant about the under under appreciating what's to come.

我仍然觉得我所谈论的这种东西。

I still don't think this this kind of thing I'm talking about.

对吧？

Right?

我觉得这将会非常疯狂。

I think that it's gonna be really crazy.

这将会带来巨大的颠覆。

It's gonna be very disruptive.

我认为这其中也会有很多积极的一面，很多事物会变得惊人和更好，但我们也正步入一个全新的世界，其中也伴随着风险。

I think there's gonna be lots of positives out of it too, and lots of things will be amazing and better, but there are also risks with this new brave new world we're going into.

你之前提到了几次Astra。

So you brought up Astra a couple times.

我们来聊聊它吧。

Let's just talk about it.

这就是你所说的Astra项目。

It's project Astra as you call it.

是的。

Yeah.

它几乎是一个始终在线的AI助手。

It is almost an always on AI assistant.

你可以拿着你的手机。

You can like hold your phone.

目前它还只是一个原型，尚未公开发布，但你可以拿着手机，它就能感知房间内发生的一切。

It's currently just a prototype or not publicly released but you can hold your phone and it will see what's going on in the room.

所以，如果我能在你的节目里看到你——或者不是你本人，而是你团队中的人——这么做的话。

So if I could basically have seen you do this on your show or not you personally but somebody on your team.

你可以问：‘我在哪儿？’

You could say, okay, where am I?

然后我会回答：‘你在播客演播室里。’

And I'll be like, you're in a podcast studio.

是的。

Yes.

任何其他功能吗？

Anything?

好的。

Okay.

所以它能够具备这种情境感知能力。

So it could have this contextual awareness.

是的。

Yes.

这能在没有智能眼镜的情况下实现吗？

Can that work without smart glasses?

因为一直举着手机真的很烦。

Because it's really annoying to hold my phone up.

是的。

Yes.

那么，我们什么时候能看到内置这项技术的谷歌智能眼镜？

So like when are when are we gonna see Google smart glasses with this technology embedded?

它们即将推出。

They're coming.

所以我们早期的一些原型中已经展示了这一点，目前我们主要在手机上进行原型开发，因为它们的处理能力更强。

So we we teased it in some of our early prototypes so that we're mostly prototyping on on on phones currently because they have more processing power.

但当然，谷歌一直是在眼镜领域的领导者。

But where, of course, Google's always been a leader in in glasses.

谷歌眼镜？

Google Glass?

是的。

Yeah.

只是当时太早了。

And exactly Just little too early.

是的。

Yeah.

也许有点太早了。

Maybe a little too early.

现在我真的认为，他们的团队非常兴奋，这个助手可能是眼镜一直寻找的杀手级应用场景。

And now I actually think and with their super excited, that team is that, you know, maybe this assistant is the killer use case that glasses has always been looking for.

当你开始在日常生活中使用Astra时，这一点就非常明显了，目前我们正在与一些可信的测试者以测试版形式使用它。

And I think it's quite obvious when you when you start using Astra in your daily life, which we have in with trusted testers at the moment and in kind of beta form.

有很多使用场景中，使用它会非常有用，但拿着手机确实不太方便。

There are many use cases where it would be so useful to use it, but it's a bit it's it's inconvenient that you're holding the phone.

比如，当你在做饭的时候。

So one example is while you're cooking, for example.

对吧？

Right?

它可以告诉你下一步该做什么，比如菜单，告诉你是否切对了东西，或者是否煎得恰到好处，但你希望它能完全解放双手。

And and it can advise you what to do next, the menu, you know, how to whether you've chopped the thing correctly or or or fried the thing correctly, but you want it to just be hands free.

是的。

Yeah.

对吧？

Right?

所以我认为，眼镜以及其他一些免手持的形态将在未来几年内崭露头角，我们确实计划站在这一领域的最前沿。

So I think that glasses and maybe other form factors that are hands free will come into their own in the next few years, and and we we we we, you know, we plan to be at the forefront of that.

其他形态？

Other form factors?

你可以想象带有摄像头的耳塞，而眼镜显然是下一个阶段。

Well, you could imagine earbuds with cameras and, you know, glasses is obvious next stage.

但这是最优的形态吗？

But is that the optimal form?

可能也不是。

Probably not either.

但部分原因是我们仍处于早期阶段，正在探索其他日常用户旅程和每个人日常使用的典型、核心用例。

But partly, we've also got to see we're still very early in this journey of seeing what other are the the the regular user journeys and killer sort of use journeys that everyone uses bread and butter uses every day.

目前，可信测试者计划正是为此目的而设。

And that's what the the the trusted tester program is for at the moment.

我们正在收集这些信息，观察人们如何使用它，看看哪些最终会有用。

We're kinda kinda collecting that information and observing people using it and seeing what ends up being useful.

好的。

Okay.

关于智能体，最后一个问题是，然后我们转向科学。

One last question on agents, then we move to science.

智能体、AI智能体，这已经是AI领域一年多来的热门词汇了。

Agentic agents, AI agents, this has been the buzzword in AI for more than a year now.

是的。

Yeah.

实际上目前还没有真正的AI智能体。

There aren't really any AI agents out there.

没有。

No.

这到底是怎么回事？

What's going on?

是的。

Yeah.

嗯，再说一遍，我认为这股炒作热潮可能已经超前于实际的科学和研究进展了。

Well, again, you know, I think the hype train can potentially ahead of where the the the the actual science and research is.

但我相信今年将是代理元年的开端。

But I do believe that this year will be the year of agents, the beginnings of it.

我想你会开始看到这一点，也许在今年下半年。

I think you'll start seeing that, you know, maybe second half of this year.

但会有一些早期版本，然后我认为它们会迅速改进和成熟。

But there'll be the early versions, and then, you know, I think they'll rapidly improve and mature.

所以我认为你说得对。

So but I think you're right.

我认为目前的技术，代理技术仍然停留在研究实验室阶段。

I think the the the technology at the moment, it's still in the research lab, the agent technologies.

但像Astra、机器人技术这样的东西，我觉得即将到来。

But things like Astra, robotics, I think it's coming.

你觉得人们会信任它们吗？

You think people are gonna trust them?

我的意思是，去帮我上网吧。

I it's like, go use the Internet for me.

这是我的信用卡。

Here's my credit card.

我不知道。

I don't know.

嗯，我认为至少从我的角度来看，最初不应该允许代理在最终步骤中绕过人类干预。

Well, so I think to begin with, you would probably my my view at least would be to not allow have human in the loop for their final steps.

比如，别去付款。

Like like, don't pay for anything.

除非人类用户或操作员授权，否则不要使用你的信用卡。

Use your credit card unless the the the human user operator authorizes it.

所以，对我来说，这会是一个明智的初步步骤。

So that would put to me be a sensible first step.

此外，在第一阶段，也许某些类型的活动、网站或其他内容应被列为禁区，比如银行网站等，同时我们继续在现实中测试这些系统的稳健性。

Also, perhaps certain types of activities or websites or whatever kind of off limits, you know, banking websites and other things in the first phase while we continue to test out in the world that that how robust these systems are.

我认为，当人工智能说‘别担心’的时候，我们就真正达到了通用人工智能。

I propose we've really reached AGI when they say, don't worry.

我不会花你的钱。

I won't spend your money.

然后它们会表现出欺骗性，下一刻你就发现自己在飞往某地的航班上。

Then they do the deceptiveness thing and then next thing you know you're on a flight somewhere.

这确实会更接近了，毫无疑问。

That would be getting closer for sure, for sure, yeah.

好吧，科学方面，你主要研究了用AlphaFold破解所有蛋白质折叠问题。

All right, science, so you worked on basically decoding all protein folding with AlphaFold.

是的。

Yes.

你因此获得了诺贝尔奖。

You won the Nobel Prize for that.

别跳过你获得诺贝尔奖的成就，但我更想谈谈未来的规划。

Not to skip over the thing that you won the Nobel Prize for, but I want to talk about what's on the roadmap.

当然。

Sure.

那就是你对构建虚拟细胞感兴趣。

Which is that you have an interest in mapping a virtual cell.

是的。

Yes.

那是什么？

What is that?

它能给我们带来什么？

And what does it get us?

嗯。

Yeah.

好吧，想想我们用AlphaFold所做的，本质上是解决了蛋白质结构的预测问题。

Well, so if you think about what we did with AlphaFold was essentially solve the problem of the the the the the finding the structure of a protein.

蛋白质，生命中的所有事物都依赖于蛋白质，对吧，你身体里的所有东西。

Proteins, everything in life depends on proteins, right, everything in your body.

所以那只是蛋白质的静态图像。

So that's the kind of static picture of a protein.

但生物学的要点是，只有理解细胞内不同成分之间的动态和相互作用，你才能真正理解生物学中的现象。

But the thing about biology is, really, it's you only understand what's going on in biology if you understand the dynamics and the interactions between the different things in in a cell.

因此，虚拟细胞项目就是构建一个完整工作细胞的模拟，一个AI模拟。

And so a virtual cell project is about building a simulation, an AI simulation of a full working cell.

我可能会从酵母细胞开始，因为酵母生物体结构简单。

I probably start with something like a yeast cell because of the simplicity of of the yeast organism.

而且你必须逐步构建起来。

And and you have to build up there.

所以下一步，比如通过AlphaFold 3，我们开始研究蛋白质与配体、蛋白质与DNA、蛋白质与RNA之间的配对相互作用。

So the next step is with alpha fold three, for example, we started doing pairwise interactions between proteins and ligands and proteins and DNA, proteins and RNA.

然后下一步是模拟整个通路，比如癌症通路之类的，这对解决疾病会有帮助。

And then the next step would be modeling a whole pathway, maybe a cancer pathway or something like that that'd be helpful with for solving a disease.

最后是整个细胞。

And then finally, a whole cell.

这之所以重要，是因为你可以提出假设并测试这些假设，比如改变某种营养物质，或向细胞中注入药物，然后观察细胞的反应。

And the reason that's important is you would be able to hypothesis make hypotheses and test those hypotheses about making some change, some nutrient change, or injecting a drug into the cell and then seeing what happens to the how the cell responds.

而目前，当然，你必须在湿实验室中 painstakingly（ painstakingly ）地进行这些实验。

And at the moment, of course, you have to do that painstakingly in a wet lab.

但想象一下，如果你能先在计算机模拟中快上一千倍、一百万倍地完成这些操作，只在最后一步才在湿实验室中进行验证。

But imagine if you could do it a thousand, a million times faster in silico first, and only at the last step do you do a validation in the wet lab.

因此，与其在湿实验室中进行耗时耗资数百万倍的搜索，不如将搜索部分完全在计算机模拟中完成。

So instead of doing the search in in in the wet lab, which is millions of times more expensive and time consuming than the validation step, you just do the search part in silico.

所以，这再次类似于我们在游戏环境中所做的工作，只不过现在应用到了科学和生物学领域。

So it's, again, it's sort of translating, again, what we did in the games environments, but here in the sciences and the biology.

因此，你先构建一个模型，然后用它来进行推理和搜索。

So you you build a model, and then you use that to do the reasoning and the search over.

而预测结果至少比没有预测要好，也许它们并不完美，但已经足够有用，可以供实验人员用来验证。

And then the predictions are, you know, at least better than not maybe they're not perfect, but they're useful enough to to be useful for experimentalists to to validate against.

湿实验室是在人里面吗？

And the wet lab is within people?

是的。

Yeah.

所以湿实验室，你仍然需要最后一步在湿实验室中验证预测是否真正有效。

So the wet lab, you'd you'd you'd still need the final step with the with the wet lab to prove the what the predictions were actually valid.

所以，你知道，你不需要在湿实验室里完成所有工作来得到那个预测。

So, you know, you'd but you wouldn't have to do all of the work to get to that prediction in in in the wet lab.

你只需要得到这个预测。

So you just get here's the prediction.

如果你加入这种化学物质，就会产生这样的变化。

If you put this chemical in, this should be the change.

对吧？

Right?

然后你只做这一个实验。

And then you just do that one experiment.

然后，当然，你仍然需要进行临床试验。

So and then after that, of course, you still have to have clinical trials.

如果你说的是药物，你仍然需要通过临床试验来properly测试它，并在人类身上测试其疗效等等。

If you're talking about a drug, you would still need to test that properly through the clinical trials and so on and test it on humans for efficacy and so on.

我认为，整个临床试验过程也可以通过人工智能来改进。

That, I also think, could be improved with AI, that whole clinical trial.

这同样需要很多年，但这将是一种不同于虚拟细胞的不同技术。

That also takes many, many years, but that would be a different technology from the virtual cell.

虚拟细胞将有助于药物发现的发现阶段。

The virtual cell would be helping the discovery phase for drug discovery.

就像我对一种药物有了想法，把它放进虚拟细胞里。

Just like I have an idea for a drug, throw it in the virtual cell.

细胞，看看它会有什么反应。

Cell, see what it does.

是的。

Yeah.

也许最终它会是肝细胞、脑细胞之类的细胞。

And maybe eventually it's a liver cell or a brain cell or something like that.

所以你会有不同的细胞模型。

So you have different cell models.

至少90%的情况下，它会给出真正会发生的结果。

And then at least 90% of the time, it's giving you back what would really happen.

那太不可思议了。

That'd be incredible.

你觉得这需要多长时间才能

How long do you think that's gonna take to

实现？

figure out?

我认为大概五年后就能做到。

I think that would be like maybe five years from now.

好的。

Okay.

是的。

Yeah.

是的。

Yeah.

所以我有一个为期五年的项目，许多原来的AlphaFold团队成员正在从事这个项目。

So I have a kind of five year project and a lot of the AlphaFold, the old AlphaFold team are working on that.

是的。

Yeah.

我刚刚在问你们团队。

Was asking your team here.

你，是的。

You Yeah.

已经搞明白了

Figured out

我和他聊过。

I was speaking with him.

嗯。

Yeah.

我

I

这个过程太荒谬了。

just the process is absurd.

流程太慢，发现阶段也太慢。

It's process too slow and discovery phase too slow.

我的意思是，我们研究阿尔茨海默病已经这么久了，想想看，对患者和家属来说，这多么悲剧，我们本该取得更多进展。

Mean, look how long we've been working on Alzheimer's and and I mean, in this tragic way to for someone to go and for the families and and and, you know, we should be a lot further.

这已经花了四十年的时间了。

It's forty years of work on that.

嗯。

Yeah.

嗯。

Yeah.

我在家人身上见过几次这种情况，如果我们能确保这种情况不再发生，那就太好了。

I've seen it a couple times in my family and yeah if we can ensure that doesn't happen it's just

在我看来，这是人工智能最能发挥作用的领域之一。

one of the best things we could use AI for in my opinion.

是的。

Yeah.

是的。

Yeah.

看着一个人逐渐衰弱，真是种悲惨的经历。

It's a terrible way to see somebody decline.

是的。

Yeah.

这是一项重要的工作。

So it's important work.

除此之外，还有基因组，人类基因组计划某种程度上让我觉得，好吧，他们解码了整个基因组，但其实还有很多工作要做，就像你们用Fold解码蛋白质一样，但事实上，解码后我们只是得到了一堆字母。

On addition to that there's the genome and so the human genome project sort of I was like okay so they decoded the whole genome there's some more work to do there like just same way that you decoded proteins with fold but it turns out that actually we just have like a bunch of letters when it's decoded.

所以你现在正在利用人工智能来解读这些字母的含义吗？

And so now you're working to use AI to translate what those letters mean?

是的。

Yes.

没错，我们在基因组学方面做了很多有趣的工作，试图判断突变是有害的还是良性的。

So, yeah, we have lots of cool work on on genomics and trying to figure out if mutations are going to be harmful or or benign.

对吧？

Right?

大多数DNA突变都是无害的。

Most mutations to your DNA are are harmless.

但当然，有些是有致病性的。

But, of course, some are pathogenic.

你希望知道哪些突变是有害的。

And you wanna know which ones there are.

因此，我们的首批系统在预测这一点上是全球最出色的。

So our first systems are are the best in the world at predicting that.

接下来的一步是研究那些疾病并非由单一基因突变引起，而是由多个突变共同作用的情况。

And then the next step is to look at situations where the disease isn't caused just by one genetic mutation, but maybe a series of them in concert.

显然，这要困难得多。

And obviously, that's a lot harder.

很多我们尚未取得进展的复杂疾病，很可能并非由单一突变导致。

Like and a lot of more complex diseases that we haven't made progress with are probably not due to a single mutation.

对吧？

Right?

这更像是罕见的儿童疾病，类似这样的情况。

That's more like rare childhood diseases, things like that.

因此，我认为AI是绝佳的工具，可以帮助我们弄清楚这些微弱的相互作用究竟是怎样的。

So there, you know, we need to I think AI is the perfect perfect tool to to to sort of try and figure out what these weak interactions are like.

对吧？

Right?

它们如何可能相互叠加、产生累积效应。

How they may be kind of compound on top of each other.

因此，统计数据可能并不明显，但一个能够识别模式的AI系统可以发现这里存在某种关联。

And so maybe the statistics are not very obvious, but an AI system that's able to kind of spot patterns would be able to figure out there is some connection here.

我们经常从疾病的角度讨论这个问题，但我也很好奇，在使人超人化方面会发生什么。

And so we talk about this a lot in terms of disease, but also I wonder what happens in terms of making people superhuman.

我的意思是，你真的能够修改基因代码，对吧？可能性似乎是无穷无尽的。

I mean, you're really able to tinker with the genetic code, right, the possibilities seem endless.

那你对此怎么看？

So what do you think about that?

这会是我们通过AI实现的事情吗？

Is that something that we're gonna be able to do through AI?

我认为总有一天会的。

I think one day.

我的意思是，我们现在更专注于修复疾病相关的基因问题。

I mean, we're focusing much more on on the on the disease profile fixing

在做什么。

what doing.

是的。

Yeah.

那是第一步。

That's the first step.

我一直觉得这是最重要的。

And and I've always felt that that's the most important.

如果你问我最想用人工智能做什么，以及我们使用人工智能最重要的目的，那就是帮助人类健康。

If you ask me what's the number one thing I wanted to use AI for and the most important thing we use AI for is for helping human health.

但除此之外，人们自然会想到衰老之类的问题。

But then, of course, beyond that, one could imagine aging, things like that.

你知道，这本身就是一个完整的领域。

You know, is that, of course, there's a whole field in itself.

衰老是一种疾病吗？

Is aging a disease?

它是多种疾病的组合吗？

Is it a combination of diseases?

我们能延长健康寿命吗？

Can we extend our healthy lifespan?

这些都是重要的问题，我认为非常有趣。

These are all important questions, I think very interesting.

我确信人工智能在帮助我们找到这些问题的答案方面也将非常有用。

And I'm pretty sure AI will be extremely useful in helping us find answers to those questions too.

我看到推特信息流中出现了一些表情包，也许我需要调整一下我收到的推荐内容。

And I see memes come across my Twitter feed, and maybe I need to change the stuff I'm recommended.

是的。

Yeah.

但通常情况下，如果你活到2050年，你并不会死。

But it's often like if you will live to twenty fifty, you're not gonna die.

是的。

Yeah.

你认为人类的最大寿命潜力是多少？

What do you think the potential max lifespan is for a person?

嗯，我很熟悉许多从事衰老研究的人。

Well, look, I know those a lot of those folks in aging research very well.

我认为他们所做的开创性工作非常有趣。

I think it's very interesting, the pioneering work they they do.

我认为变老、身体衰败没有任何好处。

I think there's nothing good about getting old and your body decaying.

我的意思是，如果你亲眼目睹亲人经历这一切，作为家人或本人，这都是一件非常艰难的事情，对吧？

I think it's you know, if anyone who's seen that up close with their relatives, it's a pretty hard thing to go through, right, as a family or or the or the person, of course.

因此，我认为任何能够减轻人类痛苦、延长健康寿命的做法都是好事。

And and and so I think anything we can alleviate human suffering and and extend healthy lifespan is a good thing.

你知道，自然寿命似乎大约是120岁。

You know, the natural limit seems to be about a hundred and twenty years old.

但根据我们所知，如果你看看那些幸运地活到那个年龄的最年长者。

But from what we know, you know, if you look at the oldest people that that that are they are that are lucky enough to to live to that age.

所以，这是一个我密切关注的领域。

So there's, you know, it's it's it's a it's a an area I follow quite closely.

我没有新的见解，我认为这些都已经为人所知了。

I don't have any, I I guess, new insights that are not already known in that.

但我真的会惊讶，如果那就是极限的话。

But I do I I would be surprised if there if that's if that's the limit.

对吧？

Right?

因为这件事有两个步骤。

Because there's a sort of two steps to this.

第一步是有一天治愈所有疾病，我认为我们通过同构体以及我们在此领域的研究或我们的药物研发子公司能够做到。

One is curing all diseases one day, which I think we're gonna do with isomorphic and the work we're doing there or our spin out, our drug discovery spin out.

但仅靠这一点可能还不足以让你突破120岁的界限，因为还存在一种系统性的自然衰退问题。

But then that's not enough to probably get you past a 120 because there's some sort of then there's the question of just natural systemic decay.

对吧？

Right?

换句话说，就是衰老。

Aging, in other words.

所以不是针对特定疾病。

So and not specific disease.

对吧？

Right?

那些活到120岁的人，通常并不是死于某种特定疾病。

Often those people that live to a 120, they don't seem to die from a specific disease.

而只是普遍的衰老。

It's just sort of just general atrophy.

所以你需要的是类似再生疗法的东西，比如让你的细胞恢复活力，或者你知道的，干细胞研究，像Altos这样的公司正在研究这些，重置细胞的时钟。

So then you're gonna need something more like rejuvenation where you you you rejuvenate your cells or you, you know, maybe stem cell research, you know, companies like Altos are are are working on these things, resetting the the cell clocks.

这似乎是可能的。

Seems like that could be possible.

但同样，我觉得这太复杂了，因为生物学是一个极其复杂的涌现系统。

But again, I feel like it's so complex because biology is such a complicated emergent system.

在我看来，你需要人工智能的帮助，才能破解任何接近这个问题的方案。

You need a in my view, you need AI to help to to be able to crack any anything anything close to that.

快速聊聊材料科学。

Very quickly on material science.

我不想离开这里而不提你发现了许多新材料或潜在材料这一事实。

I don't want to leave here without talking about the fact that you've discovered many new materials or potential materials.

是的。

Yeah.

我这里的数据是，人类此前已知的稳定材料有3万种。

Stat I have here is known to humanity recently were 30,000 stable Yep.

材料。

Materials.

你用一个新的AI程序发现了220万种。

You've discovered 2,200,000 with a new AI program.

是的。

Yep.

稍微幻想一下。

Just dream a little bit.

是的。

Yes.

因为我们不知道这些材料都能做什么。

Because we don't know what all those materials can do.

我们不知道，比如说，它们是否能承受从冷冻箱里拿出来之类的状况。

We don't know what, you know, whether they'll be able to handle being out of, like, a frozen box or whatever.

是的。

Yes.

梦想材料，是的。

Dream materials Yeah.

在这些材料中为你发现的

For you to find in that set

是的。

of Yeah.

新材料。

New materials.

嗯，我们的确在材料方面投入了大量精力。

Well, I mean, we're working really hard on materials.

在我看来，这将是继AlphaFold在生物学领域带来巨大影响之后，下一个能产生类似突破的领域，只不过这次是在化学和材料科学上。

To me, it's like the next one of the next sort of big impacts we can have, the level of AlphaFold really in biology, but this time in chemistry and materials.

我梦想有朝一日能发现室温超导体。

I dream of one day discovering room temperature superconductor.

那会带来什么影响呢？

So what will that do?

因为这也是人们经常谈论的一个热门话题。

Because that's another big meme that people talk

关于这个。

about.

所以

So

你所说的没错。

what do you yeah.

这将有助于缓解能源危机和气候危机，因为如果你能拥有廉价的超导体，就可以在传输能源时几乎不损失任何能量。

Well, it would help with the energy crisis and climate crisis because if you had sort of cheap superconductors, then you can transport energy from one place to another without any loss of that energy.

因此，你完全可以把太阳能电池板安装在撒哈拉沙漠，然后通过超导体将电能直接输送到欧洲需要的地方。

So you could potentially put solar panels in the Sahara Desert and then just have a the the the the superconductor, you know, funneling that into Europe where it's needed.

目前，电能在传输过程中会因发热等原因损失大量能量。

At the moment, you would just lose a ton of the power to heat and other things on the way.

因此，你还需要其他技术，比如电池等，来储存这些电能，因为你无法在不极度低效的情况下，直接将电能输送到目标地点。

So then you need other technologies like batteries and other things to store that because you can't you can't just pipe it to the place that you want without without without being incredibly inefficient.

但材料也能帮助改进电池，比如开发出最优的电池设计。

So but also materials could help with things like batteries too, like, but come up with the optimal battery.

我认为我们目前还没有最优的电池设计，也许我们可以结合材料和蛋白质来实现突破。

I don't think we have the optimal battery designs That maybe we can do things like combination of materials and and and proteins.

我们还可以做碳捕获，比如改造藻类或其他生物，使其比人工系统更有效地捕获二氧化碳。

We can do things like carbon capture, you know, modify algae or other things to to do carbon capture better than our artificial systems.

我指的是，最著名也最重要的化学过程之一——哈伯法，用于制造化肥和氨气，从空气中提取氮气，正是这一过程支撑了现代文明。

I mean, even the one of the most famous and most important chemical chemical processes, the harbor process to make fertilizer and ammonia, you know, to take nitrogen out of the air was was was something that allows modern civilization.

但如果我们知道正确的催化剂和材料，可能还有许多其他化学反应可以通过这种方式被催化。

But there might be many other chemical processes that could be catalyzed in that way if we knew what the right catalyst and the right material was.

所以我认为，最具影响力的技术之一将是实现材料的计算机辅助设计。

So I think it's gonna be would be one of the most impactful technologies ever is to to to basically have in silico design of materials.

我们已经完成了第一步，证明了我们可以发现新的稳定材料，但我们需要一种方法来测试这些材料的性能。

So we've done step one of that where we showed we can come up with new stable materials, but we need a way of testing the properties of those materials.

对。

Right.

因为目前没有任何实验室能够测试二十万、数万甚至数百万种材料。

Because no lab can test 200,000, you know, tens of thousands of materials or millions of materials at the moment.

所以难点就在于如何进行这些测试。

So we have to that's that's the hard part is to is to do the testing.

你觉得室温超导体就在其中吗？

You think it's in there, the room temperature superconductor?

嗯，我听说我们实际上已经发现了一些超导材料。

Well, I heard that we we actually think there are some superconductor materials.

但我怀疑它们是不是室温的。

I I I doubt they're room temperature ones, though.

但是

But

好的。

Okay.

我认为，如果物理学上可行，总有一天人工智能会找到它。

I think at some point, if if it's possible with physics, an AI system will one day find it.

所以这是一种用途。

So that's one use.

我还能想到另外两种用途，可能对这类工作感兴趣的是玩具制造商和军方。

The two other uses I could imagine, probably people interested in this type of work, toy manufacturers and militaries.

是的。

Yeah.

他们在用它吗？

Are they working with it?

是的。

Yeah.

玩具制造商。

Toy manufacturer.

我的意思是，我认为这简直太棒了。

I mean, look, I think there is incredible one.

我的意思是，我早期职业生涯的很大一部分是游戏设计。

I mean, the big part of my early career was in game design

对。

and right.

主题公园。

Theme park.

是的。

Yeah.

主题公园和模拟。

Theme park and simulations.

这正是让我最初接触到模拟和人工智能的原因。

That's what got me into simulations and AI in the first place.

也是我一直热爱这两样东西的原因。

And why I've always loved both of those things.

从很多方面来说，我今天所从事的工作只是对那一切的延续。

And if in in many respects, the work I do today is just an extension of that.

我常常幻想，如果当初我做了些什么会怎样？

And and I I just dream about, like, what could I have done?

如果在二十五年或三十年前我开发那些游戏时，就能拥有今天这样的AI技术，本可以创造出怎样惊人的游戏体验啊。

What kinds of amazing game experiences could have been made if I'd had the AI I have today available twenty five, thirty years ago when I was writing those games.

我对游戏行业没有做到这一点感到有点惊讶。

And I'm a little bit surprised the game industry hasn't done that.

我不知道这是为什么。

I don't know why that is.

我开始看到一些关于NPC的疯狂进展，是的。

I'm starting to see some crazy stuff with NPCs that like Yes.

NPC，但当然，那会是像智能、动态剧情这样的东西，还有新型的AI游戏，角色和代理能够学习。

NPCs, but but of course, that'd be like intelligence, you know, dynamic storylines, but also just new types of AI first games with learning with characters and and agents that can learn.

我记得我曾经参与过一款叫《黑与白》的游戏，玩家要养育一个生物。

And, you know, I once worked on a game called black and white where you had a creature that you were nurturing.

它有点像一只宠物狗，会学习你想要它做什么。

It was a bit like a pet dog that that that learned what you wanted.

对吧？

Right?

但我们当时用的是非常基础的强化学习。

And but we were we were using very basic reinforcement learning.

那是九十年代末的事情了。

This was like in the late nineties.

想想今天能做成什么样。

You know, imagine what could be done today.

我觉得智能玩具也可能同样如此

And I think the same for for maybe smart toys as

嗯。

well.

对。

Right.

当然，说到军事领域，你知道，不幸的是，人工智能是一种双重用途的技术。

Then, course, on the militaries, you know, unfortunately, AI is a dual dual purpose technology.

因此，人们必须面对现实，尤其是在当今的地缘政治世界中，一些人正在将这些通用技术应用于无人机和其他领域。

So one has to confront the reality that, especially in today's geopolitical world, people are using some of these general purpose technologies to apply to drones and other things.

这并不令人意外。

And it's not surprising that that works.

你对中国正在做的事情感到印象深刻吗？

Are you impressed with what China's up to?

我是说，DeepSeek 是这个新的

I mean, DeepSeek is this new

模型，是的，令人印象深刻。

model Yeah, it's impressive.